Adds unsigned 32-bit integers a and b with unsigned 8-bit carry-in c_in
(carry or overflow flag), and store the unsigned 32-bit result in out, and the carry-out
is returned (carry or overflow flag).
Adds unsigned 64-bit integers a and b with unsigned 8-bit carry-in c_in
(carry or overflow flag), and store the unsigned 64-bit result in out, and the carry-out
is returned (carry or overflow flag).
Adds unsigned 32-bit integers a and b with unsigned 8-bit carry-in c_in
(carry or overflow flag), and store the unsigned 32-bit result in out, and
the carry-out is returned (carry or overflow flag).
Adds unsigned 64-bit integers a and b with unsigned 8-bit carry-in c_in
(carry or overflow flag), and store the unsigned 64-bit result in out, and
the carry-out is returned (carry or overflow flag).
Compute the bitwise OR of 8-bit masks a and b. If the result is all zeros, store 1 in dst, otherwise
store 0 in dst. If the result is all ones, store 1 in all_ones, otherwise store 0 in all_ones.
Compute the bitwise OR of 16-bit masks a and b. If the result is all zeros, store 1 in dst, otherwise
store 0 in dst. If the result is all ones, store 1 in all_ones, otherwise store 0 in all_ones.
Compute the bitwise OR of 32-bit masks a and b. If the result is all zeros, store 1 in dst, otherwise
store 0 in dst. If the result is all ones, store 1 in all_ones, otherwise store 0 in all_ones.
Compute the bitwise OR of 64-bit masks a and b. If the result is all zeros, store 1 in dst, otherwise
store 0 in dst. If the result is all ones, store 1 in all_ones, otherwise store 0 in all_ones.
Compute the bitwise AND of 8-bit masks a and b, and if the result is all zeros, store 1 in dst,
otherwise store 0 in dst. Compute the bitwise NOT of a and then AND with b, if the result is all
zeros, store 1 in and_not, otherwise store 0 in and_not.
Compute the bitwise AND of 16-bit masks a and b, and if the result is all zeros, store 1 in dst,
otherwise store 0 in dst. Compute the bitwise NOT of a and then AND with b, if the result is all
zeros, store 1 in and_not, otherwise store 0 in and_not.
Compute the bitwise AND of 32-bit masks a and b, and if the result is all zeros, store 1 in dst,
otherwise store 0 in dst. Compute the bitwise NOT of a and then AND with b, if the result is all
zeros, store 1 in and_not, otherwise store 0 in and_not.
Compute the bitwise AND of 64-bit masks a and b, and if the result is all zeros, store 1 in dst,
otherwise store 0 in dst. Compute the bitwise NOT of a and then AND with b, if the result is all
zeros, store 1 in and_not, otherwise store 0 in and_not.
Considers the input b as packed 64-bit integers and c as packed 8-bit integers.
Then groups 8 8-bit values from cas indices into the bits of the corresponding 64-bit integer.
It then selects these bits and packs them into the output.
Compare packed double-precision (64-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k.
Compare packed single-precision (32-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k.
Test each 32-bit element of a for equality with all other elements in a closer to the least significant bit. Each element’s comparison forms a zero extended bit vector in dst.
Test each 64-bit element of a for equality with all other elements in a closer to the least significant bit. Each element’s comparison forms a zero extended bit vector in dst.
Convert packed single-precision (32-bit) floating-point elements in two 256-bit vectors
a and b to packed BF16 (16-bit) floating-point elements, and store the results in a
256-bit wide vector.
Intel’s documentation
Convert packed BF16 (16-bit) floating-point even-indexed elements stored at memory locations starting at
location a to single precision (32-bit) floating-point elements, and store the results in dst.
Convert packed BF16 (16-bit) floating-point odd-indexed elements stored at memory locations starting at
location a to single precision (32-bit) floating-point elements, and store the results in dst.
Convert packed single-precision (32-bit) floating-point elements in a to packed BF16 (16-bit)
floating-point elements, and store the results in dst.
Intel’s documentation
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst.
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst.
Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst. Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets.
Conditionally multiplies the packed single-precision (32-bit) floating-point
elements in a and b using the high 4 bits in imm8,
sum the four products, and conditionally return the sum
using the low 4 bits of imm8.
Compute dot-product of BF16 (16-bit) floating-point pairs in a and b,
accumulating the intermediate single-precision (32-bit) floating-point elements
with elements in src, and store the results in dst.
Intel’s documentation
Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding signed 8-bit
integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding
32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding signed 8-bit
integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding
32-bit integer in src with signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding unsigned 8-bit
integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding
32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding unsigned 8-bit
integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding
32-bit integer in src with signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding unsigned 8-bit
integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding
32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding unsigned 8-bit
integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding
32-bit integer in src with signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding unsigned 16-bit
integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding
32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding unsigned 16-bit
integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding
32-bit integer in src with signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding signed 16-bit
integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding
32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding signed 16-bit
integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding
32-bit integer in src with signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding unsigned 16-bit
integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding
32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding unsigned 16-bit
integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding
32-bit integer in src with signed saturation, and store the packed 32-bit results in dst.
Fix up packed double-precision (64-bit) floating-point elements in a and b using packed 64-bit integers in c, and store the results in dst. imm8 is used to set the required flags reporting.
Fix up packed single-precision (32-bit) floating-point elements in a and b using packed 32-bit integers in c, and store the results in dst. imm8 is used to set the required flags reporting.
Multiplies packed double-precision (64-bit) floating-point elements in a
and b, and alternatively add and subtract packed elements in c to/from
the intermediate result.
Multiplies packed single-precision (32-bit) floating-point elements in a
and b, and alternatively add and subtract packed elements in c to/from
the intermediate result.
Multiplies packed double-precision (64-bit) floating-point elements in a
and b, and alternatively subtract and add packed elements in c from/to
the intermediate result.
Multiplies packed single-precision (32-bit) floating-point elements in a
and b, and alternatively subtract and add packed elements in c from/to
the intermediate result.
Multiplies packed double-precision (64-bit) floating-point elements in a
and b, and subtract packed elements in c from the negated intermediate
result.
Multiplies packed single-precision (32-bit) floating-point elements in a
and b, and subtract packed elements in c from the negated intermediate
result.
Test packed double-precision (64-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k.
imm can be a combination of:
Test packed single-precision (32-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k.
imm can be a combination of:
Convert the exponent of each packed double-precision (64-bit) floating-point element in a to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in dst. This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of each packed single-precision (32-bit) floating-point element in a to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in dst. This intrinsic essentially calculates floor(log2(x)) for each element.
Normalize the mantissas of packed double-precision (64-bit) floating-point elements in a, and store the results in dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of packed single-precision (32-bit) floating-point elements in a, and store the results in dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Performs an affine transformation on the packed bytes in x.
That is computes a*x+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs an affine transformation on the inverted packed bytes in x.
That is computes a*inv(x)+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
The inverse of a byte is defined with respect to the reduction polynomial x^8+x^4+x^3+x+1.
The inverse of 0 is 0.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs a multiplication in GF(2^8) on the packed bytes.
The field is in polynomial representation with the reduction polynomial
x^8 + x^4 + x^3 + x + 1.
Horizontal addition of adjacent pairs in the two packed vectors
of 4 64-bit floating points a and b.
In the result, sums of elements from a are returned in even locations,
while sums of elements from b are returned in odd locations.
Horizontal addition of adjacent pairs in the two packed vectors
of 8 32-bit floating points a and b.
In the result, sums of elements from a are returned in locations of
indices 0, 1, 4, 5; while sums of elements from b are locations
2, 3, 6, 7.
Horizontal subtraction of adjacent pairs in the two packed vectors
of 4 64-bit floating points a and b.
In the result, sums of elements from a are returned in even locations,
while sums of elements from b are returned in odd locations.
Horizontal subtraction of adjacent pairs in the two packed vectors
of 8 32-bit floating points a and b.
In the result, sums of elements from a are returned in locations of
indices 0, 1, 4, 5; while sums of elements from b are locations
2, 3, 6, 7.
Scatter 64-bit integers from a into memory using 32-bit indices. 64-bit elements are stored at addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale). scale should be 1, 2, 4 or 8.
Stores 4 double-precision (64-bit) floating-point elements from a to memory starting at location base_addr
at packed 32-bit integer indices stored in vindex scaled by scale
Stores 8 single-precision (32-bit) floating-point elements from a to memory starting at location base_addr
at packed 32-bit integer indices stored in vindex scaled by scale
Stores 4 double-precision (64-bit) floating-point elements from a to memory starting at location base_addr
at packed 64-bit integer indices stored in vindex scaled by scale
Stores 4 single-precision (32-bit) floating-point elements from a to memory starting at location base_addr
at packed 64-bit integer indices stored in vindex scaled by scale
Copy a to dst, then insert 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from b into dst at the location specified by imm8.
Copy a to dst, then insert 128 bits (composed of 2 packed double-precision (64-bit) floating-point
elements) from b into dst at the location specified by IMM8.
Copies a to result, then inserts 128 bits (composed of 2 packed
double-precision (64-bit) floating-point elements) from b into result
at the location specified by imm8.
Copies a to result, then inserts 128 bits (composed of 4 packed
single-precision (32-bit) floating-point elements) from b into result
at the location specified by imm8.
Loads 256-bits of integer data from unaligned memory into result.
This intrinsic may perform better than _mm256_loadu_si256 when the
data crosses a cache line boundary.
Load 256-bits (composed of 8 packed 32-bit integers) from memory into dst. mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Load 256-bits (composed of 4 packed 64-bit integers) from memory into dst. mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Loads 256-bits (composed of 4 packed double-precision (64-bit)
floating-point elements) from memory into result.
mem_addr must be aligned on a 32-byte boundary or a
general-protection exception may be generated.
Loads 256-bits (composed of 8 packed single-precision (32-bit)
floating-point elements) from memory into result.
mem_addr must be aligned on a 32-byte boundary or a
general-protection exception may be generated.
Loads 256-bits of integer data from memory into result.
mem_addr must be aligned on a 32-byte boundary or a
general-protection exception may be generated.
Loads two 128-bit values (composed of 4 packed single-precision (32-bit)
floating-point elements) from memory, and combine them into a 256-bit
value.
hiaddr and loaddr do not need to be aligned on any particular boundary.
Loads two 128-bit values (composed of 2 packed double-precision (64-bit)
floating-point elements) from memory, and combine them into a 256-bit
value.
hiaddr and loaddr do not need to be aligned on any particular boundary.
Loads two 128-bit values (composed of integer data) from memory, and combine
them into a 256-bit value.
hiaddr and loaddr do not need to be aligned on any particular boundary.
Loads 256-bits (composed of 4 packed double-precision (64-bit)
floating-point elements) from memory into result.
mem_addr does not need to be aligned on any particular boundary.
Loads 256-bits (composed of 8 packed single-precision (32-bit)
floating-point elements) from memory into result.
mem_addr does not need to be aligned on any particular boundary.
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the high 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst.
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the high 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst.
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the low 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst.
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the low 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst.
Multiplies packed signed 16-bit integers in a and b, producing
intermediate signed 32-bit integers. Horizontally add adjacent pairs
of intermediate 32-bit integers.
Vertically multiplies each unsigned 8-bit integer from a with the
corresponding signed 8-bit integer from b, producing intermediate
signed 16-bit integers. Horizontally add adjacent pairs of intermediate
signed 16-bit integers
Shuffle 8-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set).
Shuffle 32-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set).
Shuffle 64-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set)
Shuffle single-precision (32-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Compute the absolute value of packed signed 8-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the absolute value of packed signed 16-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the absolute value of packed signed 32-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the absolute value of packed signed 64-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed 8-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed 16-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed 32-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed 64-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed signed 8-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed signed 16-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed unsigned 8-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed unsigned 16-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate pairs of 16-byte blocks in a and b into a 32-byte temporary result, shift the result right by imm8 bytes, and store the low 16 bytes in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate a and b into a 64-byte immediate result, shift the result right by imm8 32-bit elements, and store the low 32 bytes (8 elements) in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate a and b into a 64-byte immediate result, shift the result right by imm8 64-bit elements, and store the low 32 bytes (4 elements) in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Performs element-by-element bitwise AND between packed 32-bit integer elements of a and b, storing the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise AND of packed 64-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise AND of packed double-precision (64-bit) floating point numbers in a and b
and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Compute the bitwise AND of packed single-precision (32-bit) floating point numbers in a and b
and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Compute the bitwise NOT of packed 32-bit integers in a and then AND with b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise NOT of packed 64-bit integers in a and then AND with b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise NOT of packed double-precision (64-bit) floating point numbers in a and then
bitwise AND with b and store the results in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Compute the bitwise NOT of packed single-precision (32-bit) floating point numbers in a and then
bitwise AND with b and store the results in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Average packed unsigned 8-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Average packed unsigned 16-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Considers the input b as packed 64-bit integers and c as packed 8-bit integers.
Then groups 8 8-bit values from cas indices into the bits of the corresponding 64-bit integer.
It then selects these bits and packs them into the output.
Broadcasts the lower 2 packed single-precision (32-bit) floating-point elements from a to all
elements of dst using writemask k (elements are copied from src if the corresponding bit is not set).
Broadcast the 4 packed single-precision (32-bit) floating-point elements from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcasts the 2 packed double-precision (64-bit) floating-point elements from a to all
elements of dst using writemask k (elements are copied from src if the corresponding bit is not set).
Broadcasts the lower 2 packed 32-bit integers from a to all elements of dst using writemask k
(elements are copied from src if the corresponding bit is not set).
Broadcast the 4 packed 32-bit integers from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcasts the 2 packed 64-bit integers from a to all elements of dst using writemask k
(elements are copied from src if the corresponding bit is not set).
Broadcast the low packed 8-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low packed 32-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low packed 64-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low double-precision (64-bit) floating-point element from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low single-precision (32-bit) floating-point element from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low packed 16-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed 32-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed 64-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed 32-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Contiguously store the active 8-bit integers in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active 16-bit integers in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active 32-bit integers in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active 64-bit integers in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active double-precision (64-bit) floating-point elements in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active single-precision (32-bit) floating-point elements in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active double-precision (64-bit) floating-point elements in a (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Contiguously store the active single-precision (32-bit) floating-point elements in a (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Test each 32-bit element of a for equality with all other elements in a closer to the least significant bit using writemask k (elements are copied from src when the corresponding mask bit is not set). Each element’s comparison forms a zero extended bit vector in dst.
Test each 64-bit element of a for equality with all other elements in a closer to the least significant bit using writemask k (elements are copied from src when the corresponding mask bit is not set). Each element’s comparison forms a zero extended bit vector in dst.
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:
Sign extend packed 8-bit integers in a to packed 16-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 8-bit integers in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 8-bit integers in the low 4 bytes of a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 16-bit integers in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 16-bit integers in a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed 32-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 32-bit integers in a to packed 16-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 32-bit integers in a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 32-bit integers in a to packed 8-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed 32-bit integers in a to packed 16-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed 64-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 16-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed signed 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed 64-bit integers in a to packed 8-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed 64-bit integers in a to packed 16-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed 64-bit integers in a to packed 32-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Zero extend packed unsigned 8-bit integers in a to packed 16-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 8-bit integers in the low 8 bytes of a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 8-bit integers in the low 4 bytes of a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 16-bit integers in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 16-bit integers in the low 8 bytes of a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 32-bit integers in a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed unsigned 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed single-precision (32-bit) floating-point elements in two vectors a and b
to packed BF16 (16-bit) floating-point elements and store the results in single vector
dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Intel’s documentation
Convert packed single-precision (32-bit) floating-point elements in a to packed BF16 (16-bit)
floating-point elements, and store the results in dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
Intel’s documentation
Converts packed BF16 (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 32-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed 16-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed 8-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 32-bit integers in a to packed 16-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 64-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 16-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 32-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 8-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 64-bit integers in a to packed 16-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 64-bit integers in a to packed 32-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Convert packed double-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 32-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed unsigned 16-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed 8-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 32-bit integers in a to packed unsigned 16-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 64-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 16-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 32-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed 8-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 64-bit integers in a to packed 16-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 64-bit integers in a to packed 32-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets.
Divide packed double-precision (64-bit) floating-point elements in a by packed elements in b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Divide packed single-precision (32-bit) floating-point elements in a by packed elements in b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute dot-product of BF16 (16-bit) floating-point pairs in a and b,
accumulating the intermediate single-precision (32-bit) floating-point elements
with elements in src, and store the results in dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
Intel’s documentation
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 8-bit integers from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 16-bit integers from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 32-bit integers from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 64-bit integers from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active double-precision (64-bit) floating-point elements from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active single-precision (32-bit) floating-point elements from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 8-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 16-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 32-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 64-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active double-precision (64-bit) floating-point elements from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active single-precision (32-bit) floating-point elements from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Extract 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from a, selected with imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Extracts 128 bits (composed of 2 packed double-precision (64-bit) floating-point elements) from a,
selected with IMM8, and stores the result in dst using writemask k (elements are copied from src
if the corresponding bit is not set).
Extract 128 bits (composed of 4 packed 32-bit integers) from a, selected with IMM1, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Extracts 128 bits (composed of 2 packed 64-bit integers) from a, selected with IMM8, and stores
the result in dst using writemask k (elements are copied from src if the corresponding bit is not set).
Fix up packed double-precision (64-bit) floating-point elements in a and b using packed 64-bit integers in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.
Fix up packed single-precision (32-bit) floating-point elements in a and b using packed 32-bit integers in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Test packed double-precision (64-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the
corresponding mask bit is not set).
imm can be a combination of:
Test packed single-precision (32-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the
corresponding mask bit is not set).
imm can be a combination of:
Convert the exponent of each packed double-precision (64-bit) floating-point element in a to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of each packed single-precision (32-bit) floating-point element in a to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Normalize the mantissas of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Performs an affine transformation on the packed bytes in x.
That is computes a*x+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs an affine transformation on the inverted packed bytes in x.
That is computes a*inv(x)+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
The inverse of a byte is defined with respect to the reduction polynomial x^8+x^4+x^3+x+1.
The inverse of 0 is 0.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs a multiplication in GF(2^8) on the packed bytes.
The field is in polynomial representation with the reduction polynomial
x^8 + x^4 + x^3 + x + 1.
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Stores 8 32-bit integer elements from a to memory starting at location base_addr at packed 32-bit integer
indices stored in vindex scaled by scale using writemask k (elements whose corresponding mask bit is not set
are not written to memory).
Stores 4 64-bit integer elements from a to memory starting at location base_addr at packed 32-bit integer
indices stored in vindex scaled by scale using writemask k (elements whose corresponding mask bit is not set
are not written to memory).
Stores 4 double-precision (64-bit) floating-point elements from a to memory starting at location base_addr
at packed 32-bit integer indices stored in vindex scaled by scale using writemask k (elements whose corresponding
mask bit is not set are not written to memory).
Stores 8 single-precision (32-bit) floating-point elements from a to memory starting at location base_addr
at packed 32-bit integer indices stored in vindex scaled by scale using writemask k (elements whose corresponding
mask bit is not set are not written to memory).
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Stores 4 32-bit integer elements from a to memory starting at location base_addr at packed 64-bit integer
indices stored in vindex scaled by scale using writemask k (elements whose corresponding mask bit is not set
are not written to memory).
Stores 4 64-bit integer elements from a to memory starting at location base_addr at packed 64-bit integer
indices stored in vindex scaled by scale using writemask k (elements whose corresponding mask bit is not set
are not written to memory).
Stores 4 double-precision (64-bit) floating-point elements from a to memory starting at location base_addr
at packed 64-bit integer indices stored in vindex scaled by scale using writemask k (elements whose corresponding
mask bit is not set are not written to memory).
Stores 4 single-precision (32-bit) floating-point elements from a to memory starting at location base_addr
at packed 64-bit integer indices stored in vindex scaled by scale using writemask k (elements whose corresponding
mask bit is not set are not written to memory).
Copy a to tmp, then insert 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from b into tmp at the location specified by imm8. Store tmp to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Copy a to tmp, then insert 128 bits (composed of 2 packed double-precision (64-bit) floating-point
elements) from b into tmp at the location specified by IMM8, and copy tmp to dst using writemask k
(elements are copied from src if the corresponding bit is not set).
Copy a to tmp, then insert 128 bits (composed of 4 packed 32-bit integers) from b into tmp at the location specified by imm8. Store tmp to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Copy a to tmp, then insert 128 bits (composed of 2 packed 64-bit integers) from b into tmp at the
location specified by IMM8, and copy tmp to dst using writemask k (elements are copied from src if
the corresponding bit is not set).
Load packed 32-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Load packed 64-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Load packed double-precision (64-bit) floating-point elements from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Load packed single-precision (32-bit) floating-point elements from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Load packed 8-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 16-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 32-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 64-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed double-precision (64-bit) floating-point elements from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed single-precision (32-bit) floating-point elements from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Counts the number of leading zero bits in each packed 32-bit integer in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Counts the number of leading zero bits in each packed 64-bit integer in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the high 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst using writemask k (elements are copied
from k when the corresponding mask bit is not set).
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the low 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst using writemask k (elements are copied
from k when the corresponding mask bit is not set).
Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed unsigned 8-bit integers in a by packed signed 8-bit integers in b, producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Move packed double-precision (64-bit) floating-point elements from a to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Move packed single-precision (32-bit) floating-point elements from a to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Duplicate even-indexed double-precision (64-bit) floating-point elements from a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Duplicate odd-indexed single-precision (32-bit) floating-point elements from a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Duplicate even-indexed single-precision (32-bit) floating-point elements from a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the low signed 32-bit integers from each packed 64-bit element in a and b, and store the signed 64-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the low unsigned 32-bit integers from each packed 64-bit element in a and b, and store the unsigned 64-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the packed signed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the packed unsigned 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Truncate each intermediate integer to the 18 most significant bits, round by adding 1, and store bits [16:1] to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the packed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the low 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the packed 32-bit integers in a and b, producing intermediate 64-bit integers, and store the low 32 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed 64-bit integers in a and b, producing intermediate 128-bit integers, and store
the low 64 bits of the intermediate integers in dst using writemask k (elements are copied from
src if the corresponding bit is not set).
For each 64-bit element in b, select 8 unaligned bytes using a byte-granular shift control within the corresponding 64-bit element of a, and store the 8 assembled bytes to the corresponding 64-bit element of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise OR of packed 32-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise OR of packed 64-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise OR of packed double-precision (64-bit) floating point numbers in a and b
and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Compute the bitwise OR of packed single-precision (32-bit) floating point numbers in a and b
and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Convert packed signed 16-bit integers from a and b to packed 8-bit integers using signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers from a and b to packed 16-bit integers using signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 16-bit integers from a and b to packed 8-bit integers using unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers from a and b to packed 16-bit integers using unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 128-bit lanes using the control in b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 8-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 32-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 64-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 64-bit integers in a within 256-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 256-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 8-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 32-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 64-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
double-precision (64-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src to dst if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
single-precision (32-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src to dst if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Extract the reduced argument of packed double-precision (64-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using writemask k (elements are
copied from src to dst if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed single-precision (32-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using writemask k (elements are
copied from src to dst if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Round packed double-precision (64-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed single-precision (32-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Scale the packed double-precision (64-bit) floating-point elements in a using values from b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by imm8 bits, and store the upper 16-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by imm8 bits, and store the upper 32-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by imm8 bits, and store the upper 64-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 16-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 32-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 64-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by imm8 bits, and store the lower 16-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by imm8 bits, and store the lower 32-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by imm8 bits, and store the lower 64-bits in dst using writemask k (elements are copied from src“ when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 16-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 32-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 64-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 8-bit integers in a within 128-bit lanes using the control in the corresponding 8-bit element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 32-bit integers in a within 128-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 4 single-precision (32-bit) floating-point elements) selected by imm8 from a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 2 double-precision (64-bit) floating-point elements) selected by imm8 from a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 4 32-bit integers) selected by imm8 from a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 2 64-bit integers) selected by imm8 from a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements within 128-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 16-bit integers in the high 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the high 64 bits of 128-bit lanes of dst, with the low 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 16-bit integers in the low 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the low 64 bits of 128-bit lanes of dst, with the high 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Store packed 32-bit integers from a into memory using writemask k.
mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Store packed 64-bit integers from a into memory using writemask k.
mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Store packed double-precision (64-bit) floating-point elements from a into memory using writemask k.
mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Store packed single-precision (32-bit) floating-point elements from a into memory using writemask k.
mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Store packed double-precision (64-bit) floating-point elements from a into memory using writemask k.
mem_addr does not need to be aligned on any particular boundary.
Store packed single-precision (32-bit) floating-point elements from a into memory using writemask k.
mem_addr does not need to be aligned on any particular boundary.
Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed 16-bit integers in b from packed 16-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed 32-bit integers in b from packed 32-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed 64-bit integers in b from packed 64-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed double-precision (64-bit) floating-point elements in b from packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed single-precision (32-bit) floating-point elements in b from packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed signed 8-bit integers in b from packed 8-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed signed 16-bit integers in b from packed 16-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 32-bit integer, the corresponding bit from src, a, and b are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst using writemask k at 32-bit granularity (32-bit elements are copied from src when the corresponding mask bit is not set).
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 64-bit integer, the corresponding bit from src, a, and b are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst using writemask k at 64-bit granularity (64-bit elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise AND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero.
Compute the bitwise AND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero.
Compute the bitwise AND of packed 32-bit integers in a and b, producing intermediate 32-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero.
Compute the bitwise AND of packed 64-bit integers in a and b, producing intermediate 64-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero.
Compute the bitwise NAND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero.
Compute the bitwise NAND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero.
Compute the bitwise NAND of packed 32-bit integers in a and b, producing intermediate 32-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero.
Compute the bitwise NAND of packed 64-bit integers in a and b, producing intermediate 64-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero.
Unpack and interleave 8-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 16-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 32-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 64-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave double-precision (64-bit) floating-point elements from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave single-precision (32-bit) floating-point elements from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 8-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 16-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 32-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 64-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave double-precision (64-bit) floating-point elements from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave single-precision (32-bit) floating-point elements from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise XOR of packed 32-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise XOR of packed 64-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise XOR of packed double-precision (64-bit) floating point numbers in a and b
and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Compute the bitwise XOR of packed single-precision (32-bit) floating point numbers in a and b
and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Loads packed 32-bit integers from memory pointed by mem_addr using mask
(elements are zeroed out when the highest bit is not set in the
corresponding element).
Loads packed 64-bit integers from memory pointed by mem_addr using mask
(elements are zeroed out when the highest bit is not set in the
corresponding element).
Loads packed double-precision (64-bit) floating-point elements from memory
into result using mask (elements are zeroed out when the high bit of the
corresponding element is not set).
Loads packed single-precision (32-bit) floating-point elements from memory
into result using mask (elements are zeroed out when the high bit of the
corresponding element is not set).
Stores packed 32-bit integers from a into memory pointed by mem_addr
using mask (elements are not stored when the highest bit is not set
in the corresponding element).
Stores packed 64-bit integers from a into memory pointed by mem_addr
using mask (elements are not stored when the highest bit is not set
in the corresponding element).
Compute the absolute value of packed signed 8-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the absolute value of packed signed 16-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the absolute value of packed signed 32-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the absolute value of packed signed 64-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed 8-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed 16-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed 32-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed 64-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed signed 8-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed signed 16-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed unsigned 8-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed unsigned 16-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate pairs of 16-byte blocks in a and b into a 32-byte temporary result, shift the result right by imm8 bytes, and store the low 16 bytes in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate a and b into a 64-byte immediate result, shift the result right by imm8 32-bit elements, and store the low 32 bytes (8 elements) in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate a and b into a 64-byte immediate result, shift the result right by imm8 64-bit elements, and store the low 32 bytes (4 elements) in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise AND of packed 32-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise AND of packed 64-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise AND of packed double-precision (64-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Compute the bitwise AND of packed single-precision (32-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Compute the bitwise NOT of packed 32-bit integers in a and then AND with b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise NOT of packed 64-bit integers in a and then AND with b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise NOT of packed double-precision (64-bit) floating point numbers in a and then
bitwise AND with b and store the results in dst using zeromask k (elements are zeroed out if the
corresponding bit is not set).
Compute the bitwise NOT of packed single-precision (32-bit) floating point numbers in a and then
bitwise AND with b and store the results in dst using zeromask k (elements are zeroed out if the
corresponding bit is not set).
Average packed unsigned 8-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Average packed unsigned 16-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcasts the lower 2 packed single-precision (32-bit) floating-point elements from a to all
elements of dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Broadcast the 4 packed single-precision (32-bit) floating-point elements from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcasts the 2 packed double-precision (64-bit) floating-point elements from a to all
elements of dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Broadcasts the lower 2 packed 32-bit integers from a to all elements of dst using zeromask k
(elements are zeroed out if the corresponding bit is not set).
Broadcast the 4 packed 32-bit integers from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low packed 8-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low packed 32-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low packed 64-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low double-precision (64-bit) floating-point element from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low single-precision (32-bit) floating-point element from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low packed 16-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Contiguously store the active double-precision (64-bit) floating-point elements in a (those with their respective bit set in zeromask k) to dst, and set the remaining elements to zero.
Contiguously store the active single-precision (32-bit) floating-point elements in a (those with their respective bit set in zeromask k) to dst, and set the remaining elements to zero.
Test each 32-bit element of a for equality with all other elements in a closer to the least significant bit using zeromask k (elements are zeroed out when the corresponding mask bit is not set). Each element’s comparison forms a zero extended bit vector in dst.
Test each 64-bit element of a for equality with all other elements in a closer to the least significant bit using zeromask k (elements are zeroed out when the corresponding mask bit is not set). Each element’s comparison forms a zero extended bit vector in dst.
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Sign extend packed 8-bit integers in a to packed 16-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 8-bit integers in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 8-bit integers in the low 4 bytes of a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 16-bit integers in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 16-bit integers in a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 32-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 32-bit integers in a to packed 16-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 32-bit integers in a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 16-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed signed 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Zero extend packed unsigned 8-bit integers in a to packed 16-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 8-bit integers in the low 8 bytes of a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 8-bit integers in the low 4 bytes of a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 16-bit integers in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 16-bit integers in the low 8 bytes of a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 32-bit integers in a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed unsigned 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed single-precision (32-bit) floating-point elements in two vectors a and b
to packed BF16 (16-bit) floating-point elements, and store the results in single vector
dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Intel’s documentation
Convert packed single-precision (32-bit) floating-point elements in a to packed BF16 (16-bit)
floating-point elements, and store the results in dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
Intel’s documentation
Converts packed BF16 (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out
when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 16-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 32-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the results in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set).
Convert packed double-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set).
Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed unsigned 16-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 16-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 32-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets.
Divide packed double-precision (64-bit) floating-point elements in a by packed elements in b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Divide packed single-precision (32-bit) floating-point elements in a by packed elements in b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute dot-product of BF16 (16-bit) floating-point pairs in a and b,
accumulating the intermediate single-precision (32-bit) floating-point elements
with elements in src, and store the results in dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
Intel’s documentation
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 8-bit integers from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 16-bit integers from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 32-bit integers from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 64-bit integers from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active double-precision (64-bit) floating-point elements from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active single-precision (32-bit) floating-point elements from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 8-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 16-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 32-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 64-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active double-precision (64-bit) floating-point elements from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active single-precision (32-bit) floating-point elements from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Extract 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from a, selected with imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Extracts 128 bits (composed of 2 packed double-precision (64-bit) floating-point elements) from a,
selected with IMM8, and stores the result in dst using zeromask k (elements are zeroed out if the
corresponding bit is not set).
Extract 128 bits (composed of 4 packed 32-bit integers) from a, selected with IMM1, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Extracts 128 bits (composed of 2 packed 64-bit integers) from a, selected with IMM8, and stores
the result in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Fix up packed double-precision (64-bit) floating-point elements in a and b using packed 64-bit integers in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.
Fix up packed single-precision (32-bit) floating-point elements in a and b using packed 32-bit integers in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert the exponent of each packed double-precision (64-bit) floating-point element in a to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of each packed single-precision (32-bit) floating-point element in a to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Normalize the mantissas of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Performs an affine transformation on the packed bytes in x.
That is computes a*x+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs an affine transformation on the inverted packed bytes in x.
That is computes a*inv(x)+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
The inverse of a byte is defined with respect to the reduction polynomial x^8+x^4+x^3+x+1.
The inverse of 0 is 0.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs a multiplication in GF(2^8) on the packed bytes.
The field is in polynomial representation with the reduction polynomial
x^8 + x^4 + x^3 + x + 1.
Copy a to tmp, then insert 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from b into tmp at the location specified by imm8. Store tmp to dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Copy a to tmp, then insert 128 bits (composed of 2 packed double-precision (64-bit) floating-point
elements) from b into tmp at the location specified by IMM8, and copy tmp to dst using zeromask k
(elements are zeroed out if the corresponding bit is not set).
Copy a to tmp, then insert 128 bits (composed of 4 packed 32-bit integers) from b into tmp at the location specified by imm8. Store tmp to dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Copy a to tmp, then insert 128 bits (composed of 2 packed 64-bit integers) from b into tmp at the
location specified by IMM8, and copy tmp to dst using zeromask k (elements are zeroed out if the
corresponding bit is not set).
Load packed 32-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Load packed 64-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Load packed double-precision (64-bit) floating-point elements from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Load packed single-precision (32-bit) floating-point elements from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Load packed 8-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 16-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 32-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 64-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed double-precision (64-bit) floating-point elements from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed single-precision (32-bit) floating-point elements from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Counts the number of leading zero bits in each packed 32-bit integer in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Counts the number of leading zero bits in each packed 64-bit integer in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the high 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst using writemask k (elements are zeroed
out when the corresponding mask bit is not set).
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the low 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst using writemask k (elements are zeroed
out when the corresponding mask bit is not set).
Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed unsigned 8-bit integers in a by packed signed 8-bit integers in b, producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Move packed double-precision (64-bit) floating-point elements from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Move packed single-precision (32-bit) floating-point elements from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Duplicate even-indexed double-precision (64-bit) floating-point elements from a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Duplicate odd-indexed single-precision (32-bit) floating-point elements from a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Duplicate even-indexed single-precision (32-bit) floating-point elements from a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the low signed 32-bit integers from each packed 64-bit element in a and b, and store the signed 64-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the low unsigned 32-bit integers from each packed 64-bit element in a and b, and store the unsigned 64-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the packed signed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the packed unsigned 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Truncate each intermediate integer to the 18 most significant bits, round by adding 1, and store bits [16:1] to dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the packed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the low 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the packed 32-bit integers in a and b, producing intermediate 64-bit integers, and store the low 32 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed 64-bit integers in a and b, producing intermediate 128-bit integers, and store
the low 64 bits of the intermediate integers in dst using zeromask k (elements are zeroed out if
the corresponding bit is not set).
For each 64-bit element in b, select 8 unaligned bytes using a byte-granular shift control within the corresponding 64-bit element of a, and store the 8 assembled bytes to the corresponding 64-bit element of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise OR of packed 32-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise OR of packed 64-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise OR of packed double-precision (64-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Compute the bitwise OR of packed single-precision (32-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed signed 16-bit integers from a and b to packed 8-bit integers using signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers from a and b to packed 16-bit integers using signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 16-bit integers from a and b to packed 8-bit integers using unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers from a and b to packed 16-bit integers using unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 128-bit lanes using the control in b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 8-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 32-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 64-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 64-bit integers in a within 256-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 256-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 8-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 32-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 64-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
double-precision (64-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
single-precision (32-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Extract the reduced argument of packed double-precision (64-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using zeromask k (elements are
zeroed out if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed single-precision (32-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using zeromask k (elements are
zeroed out if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Round packed double-precision (64-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed single-precision (32-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Scale the packed double-precision (64-bit) floating-point elements in a using values from b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low packed 16-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by imm8 bits, and store the upper 16-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by imm8 bits, and store the upper 32-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by imm8 bits, and store the upper 64-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 16-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 32-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 64-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by imm8 bits, and store the lower 16-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by imm8 bits, and store the lower 32-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by imm8 bits, and store the lower 64-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 16-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 32-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 64-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle packed 8-bit integers in a according to shuffle control mask in the corresponding 8-bit element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 32-bit integers in a within 128-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 4 single-precision (32-bit) floating-point elements) selected by imm8 from a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 2 double-precision (64-bit) floating-point elements) selected by imm8 from a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 4 32-bit integers) selected by imm8 from a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 2 64-bit integers) selected by imm8 from a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements within 128-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 16-bit integers in the high 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the high 64 bits of 128-bit lanes of dst, with the low 64 bits of 128-bit lanes being copied from a to dst, using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 16-bit integers in the low 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the low 64 bits of 128-bit lanes of dst, with the high 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed 16-bit integers in b from packed 16-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed 32-bit integers in b from packed 32-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed 64-bit integers in b from packed 64-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed double-precision (64-bit) floating-point elements in b from packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed single-precision (32-bit) floating-point elements in b from packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed signed 8-bit integers in b from packed 8-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed signed 16-bit integers in b from packed 16-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 32-bit integer, the corresponding bit from a, b, and c are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst using zeromask k at 32-bit granularity (32-bit elements are zeroed out when the corresponding mask bit is not set).
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 64-bit integer, the corresponding bit from a, b, and c are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst using zeromask k at 64-bit granularity (64-bit elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 8-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 16-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 32-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 64-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave double-precision (64-bit) floating-point elements from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave single-precision (32-bit) floating-point elements from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 8-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 16-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 32-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 64-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave double-precision (64-bit) floating-point elements from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave single-precision (32-bit) floating-point elements from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise XOR of packed 32-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise XOR of packed 64-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise XOR of packed double-precision (64-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Compute the bitwise XOR of packed single-precision (32-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Loads 8 32-bit integer elements from memory starting at location base_addr at packed 32-bit integer
indices stored in vindex scaled by scale using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Loads 4 64-bit integer elements from memory starting at location base_addr at packed 32-bit integer
indices stored in vindex scaled by scale using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Loads 4 double-precision (64-bit) floating-point elements from memory starting at location base_addr
at packed 32-bit integer indices stored in vindex scaled by scale using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Loads 8 single-precision (32-bit) floating-point elements from memory starting at location base_addr
at packed 32-bit integer indices stored in vindex scaled by scale using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Loads 4 32-bit integer elements from memory starting at location base_addr at packed 64-bit integer
indices stored in vindex scaled by scale using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Loads 4 64-bit integer elements from memory starting at location base_addr at packed 32-bit integer
indices stored in vindex scaled by scale using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Loads 4 double-precision (64-bit) floating-point elements from memory starting at location base_addr
at packed 32-bit integer indices stored in vindex scaled by scale using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Loads 4 single-precision (32-bit) floating-point elements from memory starting at location base_addr
at packed 32-bit integer indices stored in vindex scaled by scale using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Sets each bit of the returned mask based on the most significant bit of the
corresponding packed double-precision (64-bit) floating-point element in
a.
Sets each bit of the returned mask based on the most significant bit of the
corresponding packed single-precision (32-bit) floating-point element in
a.
Computes the sum of absolute differences (SADs) of quadruplets of unsigned
8-bit integers in a compared to those in b, and stores the 16-bit
results in dst. Eight SADs are performed for each 128-bit lane using one
quadruplet from b and eight quadruplets from a. One quadruplet is
selected from b starting at on the offset specified in imm8. Eight
quadruplets are formed from sequential 8-bit integers selected from a
starting at the offset specified in imm8.
Multiplies the packed unsigned 16-bit integers in a and b, producing
intermediate 32-bit integers and returning the high 16 bits of the
intermediate integers.
Multiplies packed 16-bit integers in a and b, producing
intermediate signed 32-bit integers. Truncate each intermediate
integer to the 18 most significant bits, round by adding 1, and
return bits [16:1].
For each 64-bit element in b, select 8 unaligned bytes using a byte-granular shift control within the corresponding 64-bit element of a, and store the 8 assembled bytes to the corresponding 64-bit element of dst.
Shuffle double-precision (64-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst.
Shuffle single-precision (32-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
double-precision (64-bit) floating-point elements in a and b, and store the results in dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
single-precision (32-bit) floating-point elements in a and b, and store the results in dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in a, and store the results in dst. The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in a, and store the results in dst. The maximum relative error for this approximation is less than 2^-14.
Computes the approximate reciprocal of packed single-precision (32-bit)
floating-point elements in a, and returns the results. The maximum
relative error for this approximation is less than 1.5*2^-12.
Extract the reduced argument of packed double-precision (64-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed single-precision (32-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8 parameter, which can be one of:
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst.
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst.
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst.
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst.
Round packed double-precision (64-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed single-precision (32-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst. The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst. The maximum relative error for this approximation is less than 2^-14.
Computes the approximate reciprocal square root of packed single-precision
(32-bit) floating-point elements in a, and returns the results.
The maximum relative error for this approximation is less than 1.5*2^-12.
Computes the absolute differences of packed unsigned 8-bit integers in a
and b, then horizontally sum each consecutive 8 differences to
produce four unsigned 16-bit integers, and pack these unsigned 16-bit
integers in the low 16 bits of the 64-bit return value
This intrinsic is one of the two SHA512 message scheduling instructions.
The intrinsic performs an intermediate calculation for the next four SHA512
message qwords. The calculated results are stored in dst.
This intrinsic is one of the two SHA512 message scheduling instructions.
The intrinsic performs the final calculation for the next four SHA512 message
qwords. The calculated results are stored in dst.
This intrinsic performs two rounds of SHA512 operation using initial SHA512 state
(C,D,G,H) from a, an initial SHA512 state (A,B,E,F) from b, and a
pre-computed sum of the next two round message qwords and the corresponding
round constants from c (only the two lower qwords of the third operand). The
updated SHA512 state (A,B,E,F) is written to dst, and dst can be used as the
updated state (C,D,G,H) in later rounds.
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by imm8 bits, and store the upper 16-bits in dst).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by imm8 bits, and store the upper 32-bits in dst.
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by imm8 bits, and store the upper 64-bits in dst).
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 16-bits in dst.
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 32-bits in dst.
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 64-bits in dst.
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by imm8 bits, and store the lower 16-bits in dst.
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by imm8 bits, and store the lower 32-bits in dst.
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by imm8 bits, and store the lower 64-bits in dst.
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 16-bits in dst.
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 32-bits in dst.
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 64-bits in dst.
Shuffles 16-bit integers in the high 64 bits of 128-bit lanes of a using
the control in imm8. The low 64 bits of 128-bit lanes of a are copied
to the output.
Shuffles 16-bit integers in the low 64 bits of 128-bit lanes of a using
the control in imm8. The high 64 bits of 128-bit lanes of a are copied
to the output.
Negates packed 8-bit integers in a when the corresponding signed
8-bit integer in b is negative, and returns the results.
Results are zeroed out when the corresponding element in b is zero.
Negates packed 16-bit integers in a when the corresponding signed
16-bit integer in b is negative, and returns the results.
Results are zeroed out when the corresponding element in b is zero.
Negates packed 32-bit integers in a when the corresponding signed
32-bit integer in b is negative, and returns the results.
Results are zeroed out when the corresponding element in b is zero.
Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst.
This intrinsic performs four rounds of SM4 key expansion. The intrinsic operates on independent
128-bit lanes. The calculated results are stored in dst.
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst.
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst.
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst.
Store 256-bits (composed of 8 packed 32-bit integers) from a into memory. mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Store 256-bits (composed of 4 packed 64-bit integers) from a into memory. mem_addr must be aligned on a 32-byte boundary or a general-protection exception may be generated.
Stores 256-bits (composed of 4 packed double-precision (64-bit)
floating-point elements) from a into memory.
mem_addr must be aligned on a 32-byte boundary or a
general-protection exception may be generated.
Stores 256-bits (composed of 8 packed single-precision (32-bit)
floating-point elements) from a into memory.
mem_addr must be aligned on a 32-byte boundary or a
general-protection exception may be generated.
Stores the high and low 128-bit halves (each composed of 4 packed
single-precision (32-bit) floating-point elements) from a into memory two
different 128-bit locations.
hiaddr and loaddr do not need to be aligned on any particular boundary.
Stores the high and low 128-bit halves (each composed of 2 packed
double-precision (64-bit) floating-point elements) from a into memory two
different 128-bit locations.
hiaddr and loaddr do not need to be aligned on any particular boundary.
Stores the high and low 128-bit halves (each composed of integer data) from
a into memory two different 128-bit locations.
hiaddr and loaddr do not need to be aligned on any particular boundary.
Stores 256-bits (composed of 4 packed double-precision (64-bit)
floating-point elements) from a into memory.
mem_addr does not need to be aligned on any particular boundary.
Stores 256-bits (composed of 8 packed single-precision (32-bit)
floating-point elements) from a into memory.
mem_addr does not need to be aligned on any particular boundary.
Load 256-bits of integer data from memory into dst using a non-temporal memory hint. mem_addr
must be aligned on a 32-byte boundary or a general-protection exception may be generated. To
minimize caching, the data is flagged as non-temporal (unlikely to be used again soon)
Moves double-precision values from a 256-bit vector of [4 x double]
to a 32-byte aligned memory location. To minimize caching, the data is
flagged as non-temporal (unlikely to be used again soon).
Moves single-precision floating point values from a 256-bit vector
of [8 x float] to a 32-byte aligned memory location. To minimize
caching, the data is flagged as non-temporal (unlikely to be used again
soon).
Moves integer data from a 256-bit integer vector to a 32-byte
aligned memory location. To minimize caching, the data is flagged as
non-temporal (unlikely to be used again soon)
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 32-bit integer, the corresponding bit from a, b, and c are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst.
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 64-bit integer, the corresponding bit from a, b, and c are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst.
Compute the bitwise AND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero.
Compute the bitwise AND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero.
Compute the bitwise AND of packed 32-bit integers in a and b, producing intermediate 32-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero.
Compute the bitwise AND of packed 64-bit integers in a and b, producing intermediate 64-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero.
Computes the bitwise AND of 256 bits (representing double-precision (64-bit)
floating-point elements) in a and b, producing an intermediate 256-bit
value, and set ZF to 1 if the sign bit of each 64-bit element in the
intermediate value is zero, otherwise set ZF to 0. Compute the bitwise
NOT of a and then AND with b, producing an intermediate value, and set
CF to 1 if the sign bit of each 64-bit element in the intermediate value
is zero, otherwise set CF to 0. Return the CF value.
Computes the bitwise AND of 256 bits (representing single-precision (32-bit)
floating-point elements) in a and b, producing an intermediate 256-bit
value, and set ZF to 1 if the sign bit of each 32-bit element in the
intermediate value is zero, otherwise set ZF to 0. Compute the bitwise
NOT of a and then AND with b, producing an intermediate value, and set
CF to 1 if the sign bit of each 32-bit element in the intermediate value
is zero, otherwise set CF to 0. Return the CF value.
Computes the bitwise AND of 256 bits (representing integer data) in a and
b, and set ZF to 1 if the result is zero, otherwise set ZF to 0.
Computes the bitwise NOT of a and then AND with b, and set CF to 1 if
the result is zero, otherwise set CF to 0. Return the CF value.
Compute the bitwise NAND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k if the intermediate value is zero.
Compute the bitwise NAND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k if the intermediate value is zero.
Compute the bitwise NAND of packed 32-bit integers in a and b, producing intermediate 32-bit values, and set the corresponding bit in result mask k if the intermediate value is zero.
Compute the bitwise NAND of packed 64-bit integers in a and b, producing intermediate 64-bit values, and set the corresponding bit in result mask k if the intermediate value is zero.
Computes the bitwise AND of 256 bits (representing double-precision (64-bit)
floating-point elements) in a and b, producing an intermediate 256-bit
value, and set ZF to 1 if the sign bit of each 64-bit element in the
intermediate value is zero, otherwise set ZF to 0. Compute the bitwise
NOT of a and then AND with b, producing an intermediate value, and set
CF to 1 if the sign bit of each 64-bit element in the intermediate value
is zero, otherwise set CF to 0. Return 1 if both the ZF and CF values
are zero, otherwise return 0.
Computes the bitwise AND of 256 bits (representing single-precision (32-bit)
floating-point elements) in a and b, producing an intermediate 256-bit
value, and set ZF to 1 if the sign bit of each 32-bit element in the
intermediate value is zero, otherwise set ZF to 0. Compute the bitwise
NOT of a and then AND with b, producing an intermediate value, and set
CF to 1 if the sign bit of each 32-bit element in the intermediate value
is zero, otherwise set CF to 0. Return 1 if both the ZF and CF values
are zero, otherwise return 0.
Computes the bitwise AND of 256 bits (representing integer data) in a and
b, and set ZF to 1 if the result is zero, otherwise set ZF to 0.
Computes the bitwise NOT of a and then AND with b, and set CF to 1 if
the result is zero, otherwise set CF to 0. Return 1 if both the ZF and
CF values are zero, otherwise return 0.
Computes the bitwise AND of 256 bits (representing double-precision (64-bit)
floating-point elements) in a and b, producing an intermediate 256-bit
value, and set ZF to 1 if the sign bit of each 64-bit element in the
intermediate value is zero, otherwise set ZF to 0. Compute the bitwise
NOT of a and then AND with b, producing an intermediate value, and set
CF to 1 if the sign bit of each 64-bit element in the intermediate value
is zero, otherwise set CF to 0. Return the ZF value.
Computes the bitwise AND of 256 bits (representing single-precision (32-bit)
floating-point elements) in a and b, producing an intermediate 256-bit
value, and set ZF to 1 if the sign bit of each 32-bit element in the
intermediate value is zero, otherwise set ZF to 0. Compute the bitwise
NOT of a and then AND with b, producing an intermediate value, and set
CF to 1 if the sign bit of each 32-bit element in the intermediate value
is zero, otherwise set CF to 0. Return the ZF value.
Computes the bitwise AND of 256 bits (representing integer data) in a and
b, and set ZF to 1 if the result is zero, otherwise set ZF to 0.
Computes the bitwise NOT of a and then AND with b, and set CF to 1 if
the result is zero, otherwise set CF to 0. Return the ZF value.
Returns vector of type __m256d with indeterminate elements.
Despite using the word “undefined” (following Intel’s naming scheme), this non-deterministically
picks some valid value and is not equivalent to mem::MaybeUninit.
In practice, this is typically equivalent to mem::zeroed.
Returns vector of type __m256 with indeterminate elements.
Despite using the word “undefined” (following Intel’s naming scheme), this non-deterministically
picks some valid value and is not equivalent to mem::MaybeUninit.
In practice, this is typically equivalent to mem::zeroed.
Returns vector of type __m256i with with indeterminate elements.
Despite using the word “undefined” (following Intel’s naming scheme), this non-deterministically
picks some valid value and is not equivalent to mem::MaybeUninit.
In practice, this is typically equivalent to mem::zeroed.
Constructs a 256-bit floating-point vector of [4 x double] from a
128-bit floating-point vector of [2 x double]. The lower 128 bits
contain the value of the source vector. The upper 128 bits are set
to zero.
Constructs a 256-bit floating-point vector of [8 x float] from a
128-bit floating-point vector of [4 x float]. The lower 128 bits contain
the value of the source vector. The upper 128 bits are set to zero.
Constructs a 256-bit integer vector from a 128-bit integer vector.
The lower 128 bits contain the value of the source vector. The upper
128 bits are set to zero.
Concatenate pairs of 16-byte blocks in a and b into a 32-byte temporary result, shift the result right by imm8 bytes, and store the low 16 bytes in dst.
Unlike _mm_alignr_epi8, _mm256_alignr_epi8 functions, where the entire input vectors are concatenated to the temporary result,
this concatenation happens in 4 steps, where each step builds 32-byte temporary result.
Considers the input b as packed 64-bit integers and c as packed 8-bit integers.
Then groups 8 8-bit values from cas indices into the bits of the corresponding 64-bit integer.
It then selects these bits and packs them into the output.
Cast vector of type __m128d to type __m512d; the upper 384 bits of the result are undefined. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m256d to type __m512d; the upper 256 bits of the result are undefined. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m512d to type __m128d. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m512d to type __m256d. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m512d to type __m512. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m512d to type __m512i. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m128 to type __m512; the upper 384 bits of the result are undefined. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m256 to type __m512; the upper 256 bits of the result are undefined. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m512 to type __m128. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m512 to type __m256. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m512 to type __m512d. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m512 to type __m512i. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m128i to type __m512i; the upper 384 bits of the result are undefined. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m256i to type __m512i; the upper 256 bits of the result are undefined. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m512i to type __m512d. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m512i to type __m512. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m512i to type __m128i. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m512i to type __m256i. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Compare packed double-precision (64-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k.
Compare packed single-precision (32-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k.
Compare packed double-precision (64-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare packed single-precision (32-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Test each 32-bit element of a for equality with all other elements in a closer to the least significant bit. Each element’s comparison forms a zero extended bit vector in dst.
Test each 64-bit element of a for equality with all other elements in a closer to the least significant bit. Each element’s comparison forms a zero extended bit vector in dst.
Convert packed signed 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst. Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed signed 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst. Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed unsigned 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst. Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed unsigned 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst. Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst. Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst. Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed double-precision (64-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst.\
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst. Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst. Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed single-precision (32-bit) floating-point elements in a to packed double-precision (64-bit) floating-point elements, and store the results in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit) floating-point elements, and store the results in dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:
Performs element-by-element conversion of the lower half of packed 32-bit integer elements in v2 to packed double-precision (64-bit) floating-point elements, storing the results in dst.
Performs element-by-element conversion of the lower half of packed 32-bit unsigned integer elements in v2 to packed double-precision (64-bit) floating-point elements, storing the results in dst.
Convert packed single-precision (32-bit) floating-point elements in two 512-bit vectors
a and b to packed BF16 (16-bit) floating-point elements, and store the results in a
512-bit wide vector.
Intel’s documentation
Convert packed single-precision (32-bit) floating-point elements in a to packed BF16 (16-bit)
floating-point elements, and store the results in dst.
Intel’s documentation
Convert packed double-precision (64-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst.
Performs an element-by-element conversion of packed double-precision (64-bit) floating-point elements in v2 to single-precision (32-bit) floating-point elements and stores them in dst. The elements are stored in the lower half of the results vector, while the remaining upper half locations are set to 0.
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst.
Convert packed single-precision (32-bit) floating-point elements in a to packed double-precision (64-bit) floating-point elements, and store the results in dst.
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit) floating-point elements, and store the results in dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:
Performs element-by-element conversion of the lower half of packed single-precision (32-bit) floating-point elements in v2 to packed double-precision (64-bit) floating-point elements, storing the results in dst.
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC
to the sae parameter.
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC
to the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC
to the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC
to the sae parameter.
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst.
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst.
Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst. Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets.
Compute dot-product of BF16 (16-bit) floating-point pairs in a and b,
accumulating the intermediate single-precision (32-bit) floating-point elements
with elements in src, and store the results in dst.Compute dot-product of BF16 (16-bit)
floating-point pairs in a and b, accumulating the intermediate single-precision (32-bit)
floating-point elements with elements in src, and store the results in dst.
Intel’s documentation
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst.
Fix up packed double-precision (64-bit) floating-point elements in a and b using packed 64-bit integers in c, and store the results in dst. imm8 is used to set the required flags reporting.
Fix up packed single-precision (32-bit) floating-point elements in a and b using packed 32-bit integers in c, and store the results in dst. imm8 is used to set the required flags reporting.
Fix up packed double-precision (64-bit) floating-point elements in a and b using packed 64-bit integers in c, and store the results in dst. imm8 is used to set the required flags reporting.\
Fix up packed single-precision (32-bit) floating-point elements in a and b using packed 32-bit integers in c, and store the results in dst. imm8 is used to set the required flags reporting.\
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst.
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst.
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst.\
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst.\
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst.
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst.
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst.\
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst.\
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst.
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst.
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst.\
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst.\
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst.
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst.
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst.\
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst.\
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst.
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst.
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst.\
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst.\
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst.
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst.
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst.\
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst.\
Test packed double-precision (64-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k.
imm can be a combination of:
Test packed single-precision (32-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k.
imm can be a combination of:
Convert the exponent of each packed double-precision (64-bit) floating-point element in a to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in dst. This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of each packed single-precision (32-bit) floating-point element in a to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in dst. This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of each packed double-precision (64-bit) floating-point element in a to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in dst. This intrinsic essentially calculates floor(log2(x)) for each element.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the exponent of each packed single-precision (32-bit) floating-point element in a to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in dst. This intrinsic essentially calculates floor(log2(x)) for each element.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of packed double-precision (64-bit) floating-point elements in a, and store the results in dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of packed single-precision (32-bit) floating-point elements in a, and store the results in dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of packed double-precision (64-bit) floating-point elements in a, and store the results in dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of packed single-precision (32-bit) floating-point elements in a, and store the results in dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Performs an affine transformation on the packed bytes in x.
That is computes a*x+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs an affine transformation on the inverted packed bytes in x.
That is computes a*inv(x)+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
The inverse of a byte is defined with respect to the reduction polynomial x^8+x^4+x^3+x+1.
The inverse of 0 is 0.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs a multiplication in GF(2^8) on the packed bytes.
The field is in polynomial representation with the reduction polynomial
x^8 + x^4 + x^3 + x + 1.
Gather 32-bit integers from memory using 32-bit indices. 32-bit elements are loaded from addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst. scale should be 1, 2, 4 or 8.
Gather 64-bit integers from memory using 32-bit indices. 64-bit elements are loaded from addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst. scale should be 1, 2, 4 or 8.
Gather double-precision (64-bit) floating-point elements from memory using 32-bit indices. 64-bit elements are loaded from addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst. scale should be 1, 2, 4 or 8.
Gather single-precision (32-bit) floating-point elements from memory using 32-bit indices. 32-bit elements are loaded from addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst. scale should be 1, 2, 4 or 8.
Loads 8 64-bit integer elements from memory starting at location base_addr at packed 32-bit integer
indices stored in the lower half of vindex scaled by scale and stores them in dst.
Loads 8 double-precision (64-bit) floating-point elements from memory starting at location base_addr
at packed 32-bit integer indices stored in the lower half of vindex scaled by scale and stores them in dst.
Stores 8 64-bit integer elements from a to memory starting at location base_addr at packed 32-bit integer
indices stored in the lower half of vindex scaled by scale.
Stores 8 double-precision (64-bit) floating-point elements from a to memory starting at location base_addr
at packed 32-bit integer indices stored in the lower half of vindex scaled by scale.
Scatter 32-bit integers from a into memory using 32-bit indices. 32-bit elements are stored at addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale). scale should be 1, 2, 4 or 8.
Scatter 64-bit integers from a into memory using 32-bit indices. 64-bit elements are stored at addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale). scale should be 1, 2, 4 or 8.
Scatter double-precision (64-bit) floating-point elements from a into memory using 32-bit indices. 64-bit elements are stored at addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale). scale should be 1, 2, 4 or 8.
Scatter single-precision (32-bit) floating-point elements from a into memory using 32-bit indices. 32-bit elements are stored at addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale). scale should be 1, 2, 4 or 8.
Gather 32-bit integers from memory using 64-bit indices. 32-bit elements are loaded from addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst. scale should be 1, 2, 4 or 8.
Gather 64-bit integers from memory using 64-bit indices. 64-bit elements are loaded from addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst. scale should be 1, 2, 4 or 8.
Gather double-precision (64-bit) floating-point elements from memory using 64-bit indices. 64-bit elements are loaded from addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst. scale should be 1, 2, 4 or 8.
Gather single-precision (32-bit) floating-point elements from memory using 64-bit indices. 32-bit elements are loaded from addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst. scale should be 1, 2, 4 or 8.
Scatter 32-bit integers from a into memory using 64-bit indices. 32-bit elements are stored at addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale). scale should be 1, 2, 4 or 8.
Scatter 64-bit integers from a into memory using 64-bit indices. 64-bit elements are stored at addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale). scale should be 1, 2, 4 or 8.
Scatter double-precision (64-bit) floating-point elements from a into memory using 64-bit indices. 64-bit elements are stored at addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale). scale should be 1, 2, 4 or 8.
Scatter single-precision (32-bit) floating-point elements from a into memory using 64-bit indices. 32-bit elements are stored at addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale) subject to mask k (elements are not stored when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Copy a to dst, then insert 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from b into dst at the location specified by imm8.
Copy a to dst, then insert 256 bits (composed of 8 packed single-precision (32-bit) floating-point
elements) from b into dst at the location specified by IMM8.
Copy a to dst, then insert 128 bits (composed of 2 packed double-precision (64-bit) floating-point
elements) from b into dst at the location specified by IMM8.
Copy a to dst, then insert 256 bits (composed of 4 packed double-precision (64-bit) floating-point elements) from b into dst at the location specified by imm8.
Load 512-bits (composed of 16 packed 32-bit integers) from memory into dst. mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Load 512-bits (composed of 8 packed 64-bit integers) from memory into dst. mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Load 512-bits (composed of 8 packed double-precision (64-bit) floating-point elements) from memory into dst. mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Load 512-bits (composed of 16 packed single-precision (32-bit) floating-point elements) from memory into dst. mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Loads 512-bits (composed of 8 packed double-precision (64-bit)
floating-point elements) from memory into result.
mem_addr does not need to be aligned on any particular boundary.
Loads 512-bits (composed of 16 packed single-precision (32-bit)
floating-point elements) from memory into result.
mem_addr does not need to be aligned on any particular boundary.
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the high 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst.
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the low 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst.
Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst.
Vertically multiply each unsigned 8-bit integer from a with the corresponding signed 8-bit integer from b, producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in dst.
Shuffle 8-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set).
Shuffle 32-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set).
Shuffle 64-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set)
Shuffle single-precision (32-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).\
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).\
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).\
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).\
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).\
Compute the absolute value of packed signed 8-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the absolute value of packed signed 16-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Computes the absolute value of packed 32-bit integers in a, and store the
unsigned results in dst using writemask k (elements are copied from
src when the corresponding mask bit is not set).
Compute the absolute value of packed signed 64-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Finds the absolute value of each packed double-precision (64-bit) floating-point element in v2, storing the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Finds the absolute value of each packed single-precision (32-bit) floating-point element in v2, storing the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed 8-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed 16-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed 32-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed 64-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Add packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Add packed signed 8-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed signed 16-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed unsigned 8-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed unsigned 16-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate pairs of 16-byte blocks in a and b into a 32-byte temporary result, shift the result right by imm8 bytes, and store the low 16 bytes in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate a and b into a 128-byte immediate result, shift the result right by imm8 32-bit elements, and store the low 64 bytes (16 elements) in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate a and b into a 128-byte immediate result, shift the result right by imm8 64-bit elements, and store the low 64 bytes (8 elements) in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Performs element-by-element bitwise AND between packed 32-bit integer elements of a and b, storing the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise AND of packed 64-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise AND of packed double-precision (64-bit) floating point numbers in a and b
and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Compute the bitwise AND of packed single-precision (32-bit) floating point numbers in a and b
and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Compute the bitwise NOT of packed 32-bit integers in a and then AND with b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise NOT of packed 64-bit integers in a and then AND with b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise NOT of packed double-precision (64-bit) floating point numbers in a and then
bitwise AND with b and store the results in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Compute the bitwise NOT of packed single-precision (32-bit) floating point numbers in a and then
bitwise AND with b and store the results in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Average packed unsigned 8-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Average packed unsigned 16-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Considers the input b as packed 64-bit integers and c as packed 8-bit integers.
Then groups 8 8-bit values from cas indices into the bits of the corresponding 64-bit integer.
It then selects these bits and packs them into the output.
Broadcasts the lower 2 packed single-precision (32-bit) floating-point elements from a to all
elements of dst using writemask k (elements are copied from src if the corresponding bit is not set).
Broadcast the 4 packed single-precision (32-bit) floating-point elements from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcasts the 8 packed single-precision (32-bit) floating-point elements from a to all
elements of dst using writemask k (elements are copied from src if the corresponding bit is not set).
Broadcasts the 2 packed double-precision (64-bit) floating-point elements from a to all
elements of dst using writemask k (elements are copied from src if the corresponding bit is not set).
Broadcast the 4 packed double-precision (64-bit) floating-point elements from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcasts the lower 2 packed 32-bit integers from a to all elements of dst using writemask k
(elements are copied from src if the corresponding bit is not set).
Broadcast the 4 packed 32-bit integers from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcasts the 8 packed 32-bit integers from a to all elements of dst using writemask k
(elements are copied from src if the corresponding bit is not set).
Broadcasts the 2 packed 64-bit integers from a to all elements of dst using writemask k
(elements are copied from src if the corresponding bit is not set).
Broadcast the 4 packed 64-bit integers from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low packed 8-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low packed 32-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low packed 64-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low double-precision (64-bit) floating-point element from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low single-precision (32-bit) floating-point element from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low packed 16-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare packed single-precision (32-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare packed signed 8-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed 32-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed 64-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed 32-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b for not-less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b for not-less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b for not-less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b for not-less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b to see if neither is NaN, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b to see if neither is NaN, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b to see if either is NaN, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b to see if either is NaN, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Contiguously store the active 8-bit integers in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active 16-bit integers in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active 32-bit integers in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active 64-bit integers in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active double-precision (64-bit) floating-point elements in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active single-precision (32-bit) floating-point elements in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active double-precision (64-bit) floating-point elements in a (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Contiguously store the active single-precision (32-bit) floating-point elements in a (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Test each 32-bit element of a for equality with all other elements in a closer to the least significant bit using writemask k (elements are copied from src when the corresponding mask bit is not set). Each element’s comparison forms a zero extended bit vector in dst.
Test each 64-bit element of a for equality with all other elements in a closer to the least significant bit using writemask k (elements are copied from src when the corresponding mask bit is not set). Each element’s comparison forms a zero extended bit vector in dst.
Convert packed signed 32-bit integers in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Convert packed signed 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set). Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed signed 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set). Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed unsigned 32-bit integers in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Convert packed unsigned 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set). Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed unsigned 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set). Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set). Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set). Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed double-precision (64-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set). Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set). Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed single-precision (32-bit) floating-point elements in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the rounding[3:0] parameter, which can be one of:
Sign extend packed 8-bit integers in a to packed 16-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 8-bit integers in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 8-bit integers in the low 8 bytes of a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 16-bit integers in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 16-bit integers in a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed 32-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 32-bit integers in a to packed 16-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 32-bit integers in a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 32-bit integers in a to packed 8-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed 32-bit integers in a to packed 16-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Performs element-by-element conversion of the lower half of packed 32-bit integer elements in v2 to packed double-precision (64-bit) floating-point elements, storing the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 16-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed signed 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed 64-bit integers in a to packed 8-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed 64-bit integers in a to packed 16-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed 64-bit integers in a to packed 32-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Zero extend packed unsigned 8-bit integers in a to packed 16-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 8-bit integers in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 8-bit integers in the low 8 bytes of a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 16-bit integers in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 16-bit integers in a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 32-bit integers in a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Performs element-by-element conversion of the lower half of 32-bit unsigned integer elements in v2 to packed double-precision (64-bit) floating-point elements, storing the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed unsigned 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed single-precision (32-bit) floating-point elements in two vectors
a and b to packed BF16 (16-bit) floating-point elements, and store the results
in single vector dst using writemask k (elements are copied from src when the
corresponding mask bit is not set).
Intel’s documentation
Convert packed single-precision (32-bit) floating-point elements in a to packed BF16 (16-bit)
floating-point elements, and store the results in dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
Intel’s documentation
Converts packed BF16 (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Performs an element-by-element conversion of packed double-precision (64-bit) floating-point elements in v2 to single-precision (32-bit) floating-point elements and stores them in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). The elements are stored in the lower half of the results vector, while the remaining upper half locations are set to 0.
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the rounding[3:0] parameter, which can be one of:
Performs element-by-element conversion of the lower half of packed single-precision (32-bit) floating-point elements in v2 to packed double-precision (64-bit) floating-point elements, storing the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 32-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed 16-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed 8-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 32-bit integers in a to packed 16-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 64-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 16-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 32-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 8-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 64-bit integers in a to packed 16-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 64-bit integers in a to packed 32-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the
corresponding bit is not set). Exceptions can be suppressed by passing _MM_FROUND_NO_EXC to the sae parameter.
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the
corresponding bit is not set). Exceptions can be suppressed by passing _MM_FROUND_NO_EXC to the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the
corresponding bit is not set). Exceptions can be suppressed by passing _MM_FROUND_NO_EXC to the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the
corresponding bit is not set). Exceptions can be suppressed by passing _MM_FROUND_NO_EXC to the sae parameter.
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Convert packed double-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 32-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed unsigned 16-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed 8-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 32-bit integers in a to packed 16-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 64-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 16-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 32-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed 8-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 64-bit integers in a to packed 16-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 64-bit integers in a to packed 32-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets.
Divide packed double-precision (64-bit) floating-point elements in a by packed elements in b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Divide packed single-precision (32-bit) floating-point elements in a by packed elements in b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Divide packed double-precision (64-bit) floating-point elements in a by packed elements in b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Divide packed single-precision (32-bit) floating-point elements in a by packed elements in b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Compute dot-product of BF16 (16-bit) floating-point pairs in a and b,
accumulating the intermediate single-precision (32-bit) floating-point elements
with elements in src, and store the results in dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
Intel’s documentation
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 8-bit integers from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 16-bit integers from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 32-bit integers from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 64-bit integers from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active double-precision (64-bit) floating-point elements from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active single-precision (32-bit) floating-point elements from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 8-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 16-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 32-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 64-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active double-precision (64-bit) floating-point elements from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active single-precision (32-bit) floating-point elements from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Extract 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from a, selected with imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Extracts 256 bits (composed of 8 packed single-precision (32-bit) floating-point elements) from a,
selected with IMM8, and stores the result in dst using writemask k (elements are copied from src
if the corresponding bit is not set).
Extracts 128 bits (composed of 2 packed double-precision (64-bit) floating-point elements) from a,
selected with IMM8, and stores the result in dst using writemask k (elements are copied from src
if the corresponding bit is not set).
Extract 256 bits (composed of 4 packed double-precision (64-bit) floating-point elements) from a, selected with imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Extract 128 bits (composed of 4 packed 32-bit integers) from a, selected with IMM2, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Extracts 256 bits (composed of 8 packed 32-bit integers) from a, selected with IMM8, and stores
the result in dst using writemask k (elements are copied from src if the corresponding bit is not set).
Extracts 128 bits (composed of 2 packed 64-bit integers) from a, selected with IMM8, and stores
the result in dst using writemask k (elements are copied from src if the corresponding bit is not set).
Extract 256 bits (composed of 4 packed 64-bit integers) from a, selected with IMM1, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Fix up packed double-precision (64-bit) floating-point elements in a and b using packed 64-bit integers in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.
Fix up packed single-precision (32-bit) floating-point elements in a and b using packed 32-bit integers in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.
Fix up packed double-precision (64-bit) floating-point elements in a and b using packed 64-bit integers in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.\
Fix up packed single-precision (32-bit) floating-point elements in a and b using packed 32-bit integers in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.\
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).\
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).\
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).\
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).\
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).\
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).\
Test packed double-precision (64-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the
corresponding mask bit is not set).
imm can be a combination of:
Test packed single-precision (32-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the
corresponding mask bit is not set).
imm can be a combination of:
Convert the exponent of each packed double-precision (64-bit) floating-point element in a to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of each packed single-precision (32-bit) floating-point element in a to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of each packed double-precision (64-bit) floating-point element in a to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the exponent of each packed single-precision (32-bit) floating-point element in a to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Performs an affine transformation on the packed bytes in x.
That is computes a*x+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs an affine transformation on the inverted packed bytes in x.
That is computes a*inv(x)+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
The inverse of a byte is defined with respect to the reduction polynomial x^8+x^4+x^3+x+1.
The inverse of 0 is 0.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs a multiplication in GF(2^8) on the packed bytes.
The field is in polynomial representation with the reduction polynomial
x^8 + x^4 + x^3 + x + 1.
Gather 32-bit integers from memory using 32-bit indices. 32-bit elements are loaded from addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst using writemask k (elements are copied from src when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Gather 64-bit integers from memory using 32-bit indices. 64-bit elements are loaded from addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst using writemask k (elements are copied from src when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Gather double-precision (64-bit) floating-point elements from memory using 32-bit indices. 64-bit elements are loaded from addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst using writemask k (elements are copied from src when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Gather single-precision (32-bit) floating-point elements from memory using 32-bit indices. 32-bit elements are loaded from addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst using writemask k (elements are copied from src when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Loads 8 64-bit integer elements from memory starting at location base_addr at packed 32-bit integer
indices stored in the lower half of vindex scaled by scale and stores them in dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
Loads 8 double-precision (64-bit) floating-point elements from memory starting at location base_addr
at packed 32-bit integer indices stored in the lower half of vindex scaled by scale and stores them in dst
using writemask k (elements are copied from src when the corresponding mask bit is not set).
Stores 8 64-bit integer elements from a to memory starting at location base_addr at packed 32-bit integer
indices stored in the lower half of vindex scaled by scale using writemask k (elements whose corresponding
mask bit is not set are not written to memory).
Stores 8 double-precision (64-bit) floating-point elements from a to memory starting at location base_addr
at packed 32-bit integer indices stored in the lower half of vindex scaled by scale using writemask k
(elements whose corresponding mask bit is not set are not written to memory).
Scatter 32-bit integers from a into memory using 32-bit indices. 32-bit elements are stored at addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale) subject to mask k (elements are not stored when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Scatter 64-bit integers from a into memory using 32-bit indices. 64-bit elements are stored at addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale) subject to mask k (elements are not stored when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Scatter double-precision (64-bit) floating-point elements from a into memory using 32-bit indices. 64-bit elements are stored at addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale) subject to mask k (elements are not stored when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Scatter single-precision (32-bit) floating-point elements from a into memory using 32-bit indices. 32-bit elements are stored at addresses starting at base_addr and offset by each 32-bit element in vindex (each index is scaled by the factor in scale) subject to mask k (elements are not stored when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Gather 32-bit integers from memory using 64-bit indices. 32-bit elements are loaded from addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst using writemask k (elements are copied from src when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Gather 64-bit integers from memory using 64-bit indices. 64-bit elements are loaded from addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst using writemask k (elements are copied from src when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Gather double-precision (64-bit) floating-point elements from memory using 64-bit indices. 64-bit elements are loaded from addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst using writemask k (elements are copied from src when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Gather single-precision (32-bit) floating-point elements from memory using 64-bit indices. 32-bit elements are loaded from addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale). Gathered elements are merged into dst using writemask k (elements are copied from src when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Scatter 32-bit integers from a into memory using 64-bit indices. 32-bit elements are stored at addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale) subject to mask k (elements are not stored when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Scatter 64-bit integers from a into memory using 64-bit indices. 64-bit elements are stored at addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale) subject to mask k (elements are not stored when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Scatter double-precision (64-bit) floating-point elements from a into memory using 64-bit indices. 64-bit elements are stored at addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale) subject to mask k (elements are not stored when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Scatter single-precision (32-bit) floating-point elements from a into memory using 64-bit indices. 32-bit elements are stored at addresses starting at base_addr and offset by each 64-bit element in vindex (each index is scaled by the factor in scale) subject to mask k (elements are not stored when the corresponding mask bit is not set). scale should be 1, 2, 4 or 8.
Copy a to tmp, then insert 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from b into tmp at the location specified by imm8. Store tmp to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Copy a to tmp, then insert 256 bits (composed of 8 packed single-precision (32-bit) floating-point
elements) from b into tmp at the location specified by IMM8, and copy tmp to dst using writemask k
(elements are copied from src if the corresponding bit is not set).
Copy a to tmp, then insert 128 bits (composed of 2 packed double-precision (64-bit) floating-point
elements) from b into tmp at the location specified by IMM8, and copy tmp to dst using writemask k
(elements are copied from src if the corresponding bit is not set).
Copy a to tmp, then insert 256 bits (composed of 4 packed double-precision (64-bit) floating-point elements) from b into tmp at the location specified by imm8. Store tmp to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Copy a to tmp, then insert 128 bits (composed of 4 packed 32-bit integers) from b into tmp at the location specified by imm8. Store tmp to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Copy a to tmp, then insert 256 bits (composed of 8 packed 32-bit integers) from b into tmp at the
location specified by IMM8, and copy tmp to dst using writemask k (elements are copied from src if
the corresponding bit is not set).
Copy a to tmp, then insert 128 bits (composed of 2 packed 64-bit integers) from b into tmp at the
location specified by IMM8, and copy tmp to dst using writemask k (elements are copied from src if
the corresponding bit is not set).
Copy a to tmp, then insert 256 bits (composed of 4 packed 64-bit integers) from b into tmp at the location specified by imm8. Store tmp to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load packed 32-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Load packed 64-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Load packed double-precision (64-bit) floating-point elements from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Load packed single-precision (32-bit) floating-point elements from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Load packed 8-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 16-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 32-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 64-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed double-precision (64-bit) floating-point elements from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed single-precision (32-bit) floating-point elements from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Counts the number of leading zero bits in each packed 32-bit integer in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Counts the number of leading zero bits in each packed 64-bit integer in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the high 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst using writemask k (elements are copied
from k when the corresponding mask bit is not set).
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the low 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst using writemask k (elements are copied
from k when the corresponding mask bit is not set).
Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed unsigned 8-bit integers in a by packed signed 8-bit integers in b, producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare packed signed 8-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Move packed double-precision (64-bit) floating-point elements from a to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Move packed single-precision (32-bit) floating-point elements from a to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Duplicate even-indexed double-precision (64-bit) floating-point elements from a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Duplicate odd-indexed single-precision (32-bit) floating-point elements from a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Duplicate even-indexed single-precision (32-bit) floating-point elements from a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the low signed 32-bit integers from each packed 64-bit element in a and b, and store the signed 64-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the low unsigned 32-bit integers from each packed 64-bit element in a and b, and store the unsigned 64-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Multiply the packed signed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the packed unsigned 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Truncate each intermediate integer to the 18 most significant bits, round by adding 1, and store bits [16:1] to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the packed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the low 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the packed 32-bit integers in a and b, producing intermediate 64-bit integers, and store the low 32 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed 64-bit integers in a and b, producing intermediate 128-bit integers, and store
the low 64 bits of the intermediate integers in dst using writemask k (elements are copied from
src if the corresponding bit is not set).
Multiplies elements in packed 64-bit integer vectors a and b together, storing the lower 64 bits of the result in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
For each 64-bit element in b, select 8 unaligned bytes using a byte-granular shift control within the corresponding 64-bit element of a, and store the 8 assembled bytes to the corresponding 64-bit element of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise OR of packed 32-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise OR of packed 64-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise OR of packed double-precision (64-bit) floating point numbers in a and b and
store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Compute the bitwise OR of packed single-precision (32-bit) floating point numbers in a and b and
store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Convert packed signed 16-bit integers from a and b to packed 8-bit integers using signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers from a and b to packed 16-bit integers using signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 16-bit integers from a and b to packed 8-bit integers using unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers from a and b to packed 16-bit integers using unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 32-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). Note that this intrinsic shuffles across 128-bit lanes, unlike past intrinsics that use the permutevar name. This intrinsic is identical to _mm512_mask_permutexvar_epi32, and it is recommended that you use that intrinsic name.
Shuffle double-precision (64-bit) floating-point elements in a within 128-bit lanes using the control in b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 8-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 32-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 64-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 64-bit integers in a within 256-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 256-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 8-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 32-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 64-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
double-precision (64-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src to dst if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
single-precision (32-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src to dst if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
double-precision (64-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src to dst if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
single-precision (32-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src to dst if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Reduce the packed double-precision (64-bit) floating-point elements in a by multiplication using mask k. Returns the product of all active elements in a.
Reduce the packed single-precision (32-bit) floating-point elements in a by multiplication using mask k. Returns the product of all active elements in a.
Extract the reduced argument of packed double-precision (64-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using writemask k (elements are
copied from src to dst if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed single-precision (32-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using writemask k (elements are
copied from src to dst if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed double-precision (64-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using writemask k (elements are
copied from src to dst if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed single-precision (32-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using writemask k (elements are
copied from src to dst if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Round packed double-precision (64-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed single-precision (32-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed double-precision (64-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed single-precision (32-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Scale the packed double-precision (64-bit) floating-point elements in a using values from b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Scale the packed double-precision (64-bit) floating-point elements in a using values from b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by imm8 bits, and store the upper 16-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by imm8 bits, and store the upper 32-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by imm8 bits, and store the upper 64-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 16-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 32-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 64-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by imm8 bits, and store the lower 16-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by imm8 bits, and store the lower 32-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by imm8 bits, and store the lower 64-bits in dst using writemask k (elements are copied from src“ when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 16-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 32-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 64-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 8-bit integers in a within 128-bit lanes using the control in the corresponding 8-bit element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 32-bit integers in a within 128-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 4 single-precision (32-bit) floating-point elements) selected by imm8 from a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 2 double-precision (64-bit) floating-point elements) selected by imm8 from a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 4 32-bit integers) selected by imm8 from a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 2 64-bit integers) selected by imm8 from a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements within 128-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 16-bit integers in the high 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the high 64 bits of 128-bit lanes of dst, with the low 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 16-bit integers in the low 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the low 64 bits of 128-bit lanes of dst, with the high 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Compute the square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Shift packed 16-bit integers in a right by count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Store packed 32-bit integers from a into memory using writemask k.
mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Store packed 64-bit integers from a into memory using writemask k.
mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Store packed double-precision (64-bit) floating-point elements from a into memory using writemask k.
mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Store packed single-precision (32-bit) floating-point elements from a into memory using writemask k.
mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Store packed double-precision (64-bit) floating-point elements from a into memory using writemask k.
mem_addr does not need to be aligned on any particular boundary.
Store packed single-precision (32-bit) floating-point elements from a into memory using writemask k.
mem_addr does not need to be aligned on any particular boundary.
Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed 16-bit integers in b from packed 16-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed 32-bit integers in b from packed 32-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed 64-bit integers in b from packed 64-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed double-precision (64-bit) floating-point elements in b from packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed single-precision (32-bit) floating-point elements in b from packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed double-precision (64-bit) floating-point elements in b from packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Subtract packed single-precision (32-bit) floating-point elements in b from packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).\
Subtract packed signed 8-bit integers in b from packed 8-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed signed 16-bit integers in b from packed 16-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 32-bit integer, the corresponding bit from src, a, and b are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst using writemask k at 32-bit granularity (32-bit elements are copied from src when the corresponding mask bit is not set).
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 64-bit integer, the corresponding bit from src, a, and b are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst using writemask k at 64-bit granularity (64-bit elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise AND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero.
Compute the bitwise AND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero.
Compute the bitwise AND of packed 32-bit integers in a and b, producing intermediate 32-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero.
Compute the bitwise AND of packed 64-bit integers in a and b, producing intermediate 64-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero.
Compute the bitwise NAND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero.
Compute the bitwise NAND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero.
Compute the bitwise NAND of packed 32-bit integers in a and b, producing intermediate 32-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero.
Compute the bitwise NAND of packed 64-bit integers in a and b, producing intermediate 64-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero.
Unpack and interleave 8-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 16-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 32-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 64-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave double-precision (64-bit) floating-point elements from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave single-precision (32-bit) floating-point elements from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 8-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 16-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 32-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 64-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave double-precision (64-bit) floating-point elements from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave single-precision (32-bit) floating-point elements from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise XOR of packed 32-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise XOR of packed 64-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise XOR of packed double-precision (64-bit) floating point numbers in a and b and
store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Compute the bitwise XOR of packed single-precision (32-bit) floating point numbers in a and b and
store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Compute the absolute value of packed signed 8-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the absolute value of packed signed 16-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Computes the absolute value of packed 32-bit integers in a, and store the
unsigned results in dst using zeromask k (elements are zeroed out when
the corresponding mask bit is not set).
Compute the absolute value of packed signed 64-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed 8-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed 16-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed 32-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed 64-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Add packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Add packed signed 8-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed signed 16-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed unsigned 8-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed unsigned 16-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate pairs of 16-byte blocks in a and b into a 32-byte temporary result, shift the result right by imm8 bytes, and store the low 16 bytes in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate a and b into a 128-byte immediate result, shift the result right by imm8 32-bit elements, and stores the low 64 bytes (16 elements) in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate a and b into a 128-byte immediate result, shift the result right by imm8 64-bit elements, and stores the low 64 bytes (8 elements) in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise AND of packed 32-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise AND of packed 64-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise AND of packed double-precision (64-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Compute the bitwise AND of packed single-precision (32-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Compute the bitwise NOT of packed 32-bit integers in a and then AND with b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise NOT of packed 64-bit integers in a and then AND with b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise NOT of packed double-precision (64-bit) floating point numbers in a and then
bitwise AND with b and store the results in dst using zeromask k (elements are zeroed out if the
corresponding bit is not set).
Compute the bitwise NOT of packed single-precision (32-bit) floating point numbers in a and then
bitwise AND with b and store the results in dst using zeromask k (elements are zeroed out if the
corresponding bit is not set).
Average packed unsigned 8-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Average packed unsigned 16-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcasts the lower 2 packed single-precision (32-bit) floating-point elements from a to all
elements of dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Broadcast the 4 packed single-precision (32-bit) floating-point elements from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcasts the 8 packed single-precision (32-bit) floating-point elements from a to all
elements of dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Broadcasts the 2 packed double-precision (64-bit) floating-point elements from a to all
elements of dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Broadcast the 4 packed double-precision (64-bit) floating-point elements from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcasts the lower 2 packed 32-bit integers from a to all elements of dst using zeromask k
(elements are zeroed out if the corresponding bit is not set).
Broadcast the 4 packed 32-bit integers from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the 4 packed 64-bit integers from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low packed 8-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low packed 32-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low packed 64-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low double-precision (64-bit) floating-point element from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low single-precision (32-bit) floating-point element from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low packed 16-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Contiguously store the active double-precision (64-bit) floating-point elements in a (those with their respective bit set in zeromask k) to dst, and set the remaining elements to zero.
Contiguously store the active single-precision (32-bit) floating-point elements in a (those with their respective bit set in zeromask k) to dst, and set the remaining elements to zero.
Test each 32-bit element of a for equality with all other elements in a closer to the least significant bit using zeromask k (elements are zeroed out when the corresponding mask bit is not set). Each element’s comparison forms a zero extended bit vector in dst.
Test each 64-bit element of a for equality with all other elements in a closer to the least significant bit using zeromask k (elements are zeroed out when the corresponding mask bit is not set). Each element’s comparison forms a zero extended bit vector in dst.
Convert packed signed 32-bit integers in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Convert packed signed 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed signed 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed unsigned 32-bit integers in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Convert packed unsigned 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed unsigned 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed double-precision (64-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Rounding is done according to the ROUNDING parameter, which can be one of:
Convert packed single-precision (32-bit) floating-point elements in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the rounding[3:0] parameter, which can be one of:
Sign extend packed 8-bit integers in a to packed 16-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 8-bit integers in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 8-bit integers in the low 8 bytes of a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 16-bit integers in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 16-bit integers in a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 32-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 32-bit integers in a to packed 16-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 32-bit integers in a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 16-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed signed 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Zero extend packed unsigned 8-bit integers in a to packed 16-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 8-bit integers in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 8-bit integers in the low 8 bytes of a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 16-bit integers in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 16-bit integers in a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 32-bit integers in a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed unsigned 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed single-precision (32-bit) floating-point elements in two vectors
a and b to packed BF16 (16-bit) floating-point elements, and store the results
in single vector dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set).
Intel’s documentation
Convert packed single-precision (32-bit) floating-point elements in a to packed BF16 (16-bit)
floating-point elements, and store the results in dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
Intel’s documentation
Converts packed BF16 (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out
when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the rounding[3:0] parameter, which can be one of:
Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 16-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 32-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set). Exceptions can be suppressed by passing _MM_FROUND_NO_EXC to the sae parameter.
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set). Exceptions can be suppressed by passing _MM_FROUND_NO_EXC to the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set). Exceptions can be suppressed by passing _MM_FROUND_NO_EXC to the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set). Exceptions can be suppressed by passing _MM_FROUND_NO_EXC to the sae parameter.
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set).
Convert packed double-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set).
Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed unsigned 16-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 16-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 32-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets.
Divide packed double-precision (64-bit) floating-point elements in a by packed elements in b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Divide packed single-precision (32-bit) floating-point elements in a by packed elements in b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Divide packed double-precision (64-bit) floating-point elements in a by packed elements in b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Divide packed single-precision (32-bit) floating-point elements in a by packed elements in b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Compute dot-product of BF16 (16-bit) floating-point pairs in a and b,
accumulating the intermediate single-precision (32-bit) floating-point elements
with elements in src, and store the results in dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
Intel’s documentation
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 8-bit integers from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 16-bit integers from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 32-bit integers from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 64-bit integers from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active double-precision (64-bit) floating-point elements from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active single-precision (32-bit) floating-point elements from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 8-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 16-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 32-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 64-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active double-precision (64-bit) floating-point elements from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active single-precision (32-bit) floating-point elements from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Extract 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from a, selected with imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Extracts 256 bits (composed of 8 packed single-precision (32-bit) floating-point elements) from a,
selected with IMM8, and stores the result in dst using zeromask k (elements are zeroed out if the
corresponding bit is not set).
Extracts 128 bits (composed of 2 packed double-precision (64-bit) floating-point elements) from a,
selected with IMM8, and stores the result in dst using zeromask k (elements are zeroed out if the
corresponding bit is not set).
Extract 256 bits (composed of 4 packed double-precision (64-bit) floating-point elements) from a, selected with imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Extract 128 bits (composed of 4 packed 32-bit integers) from a, selected with IMM2, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Extracts 256 bits (composed of 8 packed 32-bit integers) from a, selected with IMM8, and stores
the result in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Extracts 128 bits (composed of 2 packed 64-bit integers) from a, selected with IMM8, and stores
the result in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Extract 256 bits (composed of 4 packed 64-bit integers) from a, selected with IMM1, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Fix up packed double-precision (64-bit) floating-point elements in a and b using packed 64-bit integers in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.
Fix up packed single-precision (32-bit) floating-point elements in a and b using packed 32-bit integers in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.
Fix up packed double-precision (64-bit) floating-point elements in a and b using packed 64-bit integers in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.\
Fix up packed single-precision (32-bit) floating-point elements in a and b using packed 32-bit integers in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.\
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in a using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Convert the exponent of each packed double-precision (64-bit) floating-point element in a to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of each packed single-precision (32-bit) floating-point element in a to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of each packed double-precision (64-bit) floating-point element in a to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the exponent of each packed single-precision (32-bit) floating-point element in a to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Performs an affine transformation on the packed bytes in x.
That is computes a*x+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs an affine transformation on the inverted packed bytes in x.
That is computes a*inv(x)+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
The inverse of a byte is defined with respect to the reduction polynomial x^8+x^4+x^3+x+1.
The inverse of 0 is 0.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs a multiplication in GF(2^8) on the packed bytes.
The field is in polynomial representation with the reduction polynomial
x^8 + x^4 + x^3 + x + 1.
Copy a to tmp, then insert 128 bits (composed of 4 packed single-precision (32-bit) floating-point elements) from b into tmp at the location specified by imm8. Store tmp to dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Copy a to tmp, then insert 256 bits (composed of 8 packed single-precision (32-bit) floating-point
elements) from b into tmp at the location specified by IMM8, and copy tmp to dst using zeromask k
(elements are zeroed out if the corresponding bit is not set).
Copy a to tmp, then insert 128 bits (composed of 2 packed double-precision (64-bit) floating-point
elements) from b into tmp at the location specified by IMM8, and copy tmp to dst using zeromask k
(elements are zeroed out if the corresponding bit is not set).
Copy a to tmp, then insert 256 bits (composed of 4 packed double-precision (64-bit) floating-point elements) from b into tmp at the location specified by imm8. Store tmp to dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Copy a to tmp, then insert 128 bits (composed of 4 packed 32-bit integers) from b into tmp at the location specified by imm8. Store tmp to dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Copy a to tmp, then insert 256 bits (composed of 8 packed 32-bit integers) from b into tmp at the
location specified by IMM8, and copy tmp to dst using zeromask k (elements are zeroed out if the
corresponding bit is not set).
Copy a to tmp, then insert 128 bits (composed of 2 packed 64-bit integers) from b into tmp at the
location specified by IMM8, and copy tmp to dst using zeromask k (elements are zeroed out if the
corresponding bit is not set).
Copy a to tmp, then insert 256 bits (composed of 4 packed 64-bit integers) from b into tmp at the location specified by imm8. Store tmp to dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load packed 32-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Load packed 64-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Load packed double-precision (64-bit) floating-point elements from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Load packed single-precision (32-bit) floating-point elements from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Load packed 8-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 16-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 32-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 64-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed double-precision (64-bit) floating-point elements from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed single-precision (32-bit) floating-point elements from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Counts the number of leading zero bits in each packed 32-bit integer in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Counts the number of leading zero bits in each packed 64-bit integer in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the high 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst using writemask k (elements are zeroed
out when the corresponding mask bit is not set).
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the low 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst using writemask k (elements are zeroed
out when the corresponding mask bit is not set).
Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed unsigned 8-bit integers in a by packed signed 8-bit integers in b, producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare packed signed 8-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Move packed double-precision (64-bit) floating-point elements from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Move packed single-precision (32-bit) floating-point elements from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Duplicate even-indexed double-precision (64-bit) floating-point elements from a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Duplicate odd-indexed single-precision (32-bit) floating-point elements from a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Duplicate even-indexed single-precision (32-bit) floating-point elements from a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the low signed 32-bit integers from each packed 64-bit element in a and b, and store the signed 64-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the low unsigned 32-bit integers from each packed 64-bit element in a and b, and store the unsigned 64-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Multiply packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Multiply the packed signed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the packed unsigned 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Truncate each intermediate integer to the 18 most significant bits, round by adding 1, and store bits [16:1] to dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the packed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the low 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the packed 32-bit integers in a and b, producing intermediate 64-bit integers, and store the low 32 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed 64-bit integers in a and b, producing intermediate 128-bit integers, and store
the low 64 bits of the intermediate integers in dst using zeromask k (elements are zeroed out if
the corresponding bit is not set).
For each 64-bit element in b, select 8 unaligned bytes using a byte-granular shift control within the corresponding 64-bit element of a, and store the 8 assembled bytes to the corresponding 64-bit element of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise OR of packed 32-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise OR of packed 64-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise OR of packed double-precision (64-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Compute the bitwise OR of packed single-precision (32-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed signed 16-bit integers from a and b to packed 8-bit integers using signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers from a and b to packed 16-bit integers using signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 16-bit integers from a and b to packed 8-bit integers using unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers from a and b to packed 16-bit integers using unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 128-bit lanes using the control in b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 8-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 32-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 64-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 64-bit integers in a within 256-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 256-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 8-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 32-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 64-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
double-precision (64-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
single-precision (32-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
double-precision (64-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
single-precision (32-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Extract the reduced argument of packed double-precision (64-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using zeromask k (elements are
zeroed out if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed single-precision (32-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using zeromask k (elements are
zeroed out if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed double-precision (64-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using zeromask k (elements are
zeroed out if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed single-precision (32-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using zeromask k (elements are
zeroed out if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Round packed double-precision (64-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed single-precision (32-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed double-precision (64-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed single-precision (32-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Scale the packed double-precision (64-bit) floating-point elements in a using values from b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Scale the packed double-precision (64-bit) floating-point elements in a using values from b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Broadcast the low packed 16-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by imm8 bits, and store the upper 16-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by imm8 bits, and store the upper 32-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by imm8 bits, and store the upper 64-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 16-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 32-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 64-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by imm8 bits, and store the lower 16-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by imm8 bits, and store the lower 32-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by imm8 bits, and store the lower 64-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 16-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 32-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 64-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle packed 8-bit integers in a according to shuffle control mask in the corresponding 8-bit element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 32-bit integers in a within 128-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 4 single-precision (32-bit) floating-point elements) selected by imm8 from a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 2 double-precision (64-bit) floating-point elements) selected by imm8 from a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 4 32-bit integers) selected by imm8 from a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 128-bits (composed of 2 64-bit integers) selected by imm8 from a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements within 128-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 16-bit integers in the high 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the high 64 bits of 128-bit lanes of dst, with the low 64 bits of 128-bit lanes being copied from a to dst, using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 16-bit integers in the low 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the low 64 bits of 128-bit lanes of dst, with the high 64 bits of 128-bit lanes being copied from a to dst, using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Compute the square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Shift packed 16-bit integers in a right by count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed 16-bit integers in b from packed 16-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed 32-bit integers in b from packed 32-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed 64-bit integers in b from packed 64-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed double-precision (64-bit) floating-point elements in b from packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed single-precision (32-bit) floating-point elements in b from packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed double-precision (64-bit) floating-point elements in b from packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Subtract packed single-precision (32-bit) floating-point elements in b from packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).\
Subtract packed signed 8-bit integers in b from packed 8-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed signed 16-bit integers in b from packed 16-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 32-bit integer, the corresponding bit from a, b, and c are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst using zeromask k at 32-bit granularity (32-bit elements are zeroed out when the corresponding mask bit is not set).
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 64-bit integer, the corresponding bit from a, b, and c are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst using zeromask k at 64-bit granularity (64-bit elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 8-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 16-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 32-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 64-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave double-precision (64-bit) floating-point elements from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave single-precision (32-bit) floating-point elements from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 8-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 16-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 32-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 64-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave double-precision (64-bit) floating-point elements from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave single-precision (32-bit) floating-point elements from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise XOR of packed 32-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise XOR of packed 64-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise XOR of packed double-precision (64-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Compute the bitwise XOR of packed single-precision (32-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed maximum values in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed maximum values in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed minimum values in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed minimum values in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Multiply the packed signed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst.
Multiply the packed unsigned 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst.
Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Truncate each intermediate integer to the 18 most significant bits, round by adding 1, and store bits [16:1] to dst.
For each 64-bit element in b, select 8 unaligned bytes using a byte-granular shift control within the corresponding 64-bit element of a, and store the 8 assembled bytes to the corresponding 64-bit element of dst.
Shuffle 32-bit integers in a across lanes using the corresponding index in idx, and store the results in dst. Note that this intrinsic shuffles across 128-bit lanes, unlike past intrinsics that use the permutevar name. This intrinsic is identical to _mm512_permutexvar_epi32, and it is recommended that you use that intrinsic name.
Shuffle double-precision (64-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst.
Shuffle single-precision (32-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
double-precision (64-bit) floating-point elements in a and b, and store the results in dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
single-precision (32-bit) floating-point elements in a and b, and store the results in dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
double-precision (64-bit) floating-point elements in a and b, and store the results in dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
single-precision (32-bit) floating-point elements in a and b, and store the results in dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in a, and store the results in dst. The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in a, and store the results in dst. The maximum relative error for this approximation is less than 2^-14.
Extract the reduced argument of packed double-precision (64-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed single-precision (32-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed double-precision (64-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed single-precision (32-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8 parameter, which can be one of:
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst.
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst.
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst.
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst.
Round packed double-precision (64-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed single-precision (32-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed double-precision (64-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed single-precision (32-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst. The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst. The maximum relative error for this approximation is less than 2^-14.
Compute the absolute differences of packed unsigned 8-bit integers in a and b, then horizontally sum each consecutive 8 differences to produce eight unsigned 16-bit integers, and pack these unsigned 16-bit integers in the low 16 bits of 64-bit elements in dst.
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by imm8 bits, and store the upper 16-bits in dst).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by imm8 bits, and store the upper 32-bits in dst.
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by imm8 bits, and store the upper 64-bits in dst).
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 16-bits in dst.
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 32-bits in dst.
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 64-bits in dst.
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by imm8 bits, and store the lower 16-bits in dst.
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by imm8 bits, and store the lower 32-bits in dst.
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by imm8 bits, and store the lower 64-bits in dst.
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 16-bits in dst.
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 32-bits in dst.
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 64-bits in dst.
Shuffle 16-bit integers in the high 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the high 64 bits of 128-bit lanes of dst, with the low 64 bits of 128-bit lanes being copied from a to dst.
Shuffle 16-bit integers in the low 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the low 64 bits of 128-bit lanes of dst, with the high 64 bits of 128-bit lanes being copied from a to dst.
Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst.
Shift packed 32-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst.
Shift packed 64-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst.
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst.
Shift packed 32-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst.
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst.
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst.
Shift packed 32-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst.
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst.
Store 512-bits (composed of 16 packed 32-bit integers) from a into memory. mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Store 512-bits (composed of 8 packed 64-bit integers) from a into memory. mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Store 512-bits (composed of 8 packed double-precision (64-bit) floating-point elements) from a into memory. mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Stores 512-bits (composed of 8 packed double-precision (64-bit)
floating-point elements) from a into memory.
mem_addr does not need to be aligned on any particular boundary.
Stores 512-bits (composed of 16 packed single-precision (32-bit)
floating-point elements) from a into memory.
mem_addr does not need to be aligned on any particular boundary.
Load 512-bits of integer data from memory into dst using a non-temporal memory hint. mem_addr
must be aligned on a 64-byte boundary or a general-protection exception may be generated. To
minimize caching, the data is flagged as non-temporal (unlikely to be used again soon)
Store 512-bits (composed of 8 packed double-precision (64-bit) floating-point elements) from a into memory using a non-temporal memory hint. mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Store 512-bits (composed of 16 packed single-precision (32-bit) floating-point elements) from a into memory using a non-temporal memory hint. mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Store 512-bits of integer data from a into memory using a non-temporal memory hint. mem_addr must be aligned on a 64-byte boundary or a general-protection exception may be generated.
Subtract packed double-precision (64-bit) floating-point elements in b from packed double-precision (64-bit) floating-point elements in a, and store the results in dst.
Subtract packed single-precision (32-bit) floating-point elements in b from packed single-precision (32-bit) floating-point elements in a, and store the results in dst.
Subtract packed double-precision (64-bit) floating-point elements in b from packed double-precision (64-bit) floating-point elements in a, and store the results in dst.\
Subtract packed single-precision (32-bit) floating-point elements in b from packed single-precision (32-bit) floating-point elements in a, and store the results in dst.\
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 32-bit integer, the corresponding bit from a, b, and c are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst.
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 64-bit integer, the corresponding bit from a, b, and c are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst.
Compute the bitwise AND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero.
Compute the bitwise AND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero.
Compute the bitwise AND of packed 32-bit integers in a and b, producing intermediate 32-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero.
Compute the bitwise AND of packed 64-bit integers in a and b, producing intermediate 64-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero.
Compute the bitwise NAND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k if the intermediate value is zero.
Compute the bitwise NAND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k if the intermediate value is zero.
Compute the bitwise NAND of packed 32-bit integers in a and b, producing intermediate 32-bit values, and set the corresponding bit in result mask k if the intermediate value is zero.
Compute the bitwise NAND of packed 64-bit integers in a and b, producing intermediate 64-bit values, and set the corresponding bit in result mask k if the intermediate value is zero.
Return vector of type __m512 with indeterminate elements.
Despite using the word “undefined” (following Intel’s naming scheme), this non-deterministically
picks some valid value and is not equivalent to mem::MaybeUninit.
In practice, this is typically equivalent to mem::zeroed.
Return vector of type __m512i with indeterminate elements.
Despite using the word “undefined” (following Intel’s naming scheme), this non-deterministically
picks some valid value and is not equivalent to mem::MaybeUninit.
In practice, this is typically equivalent to mem::zeroed.
Returns vector of type __m512d with indeterminate elements.
Despite using the word “undefined” (following Intel’s naming scheme), this non-deterministically
picks some valid value and is not equivalent to mem::MaybeUninit.
In practice, this is typically equivalent to mem::zeroed.
Returns vector of type __m512 with indeterminate elements.
Despite using the word “undefined” (following Intel’s naming scheme), this non-deterministically
picks some valid value and is not equivalent to mem::MaybeUninit.
In practice, this is typically equivalent to mem::zeroed.
Unpack and interleave double-precision (64-bit) floating-point elements from the high half of each 128-bit lane in a and b, and store the results in dst.
Unpack and interleave single-precision (32-bit) floating-point elements from the high half of each 128-bit lane in a and b, and store the results in dst.
Unpack and interleave double-precision (64-bit) floating-point elements from the low half of each 128-bit lane in a and b, and store the results in dst.
Unpack and interleave single-precision (32-bit) floating-point elements from the low half of each 128-bit lane in a and b, and store the results in dst.
Cast vector of type __m128d to type __m512d; the upper 384 bits of the result are zeroed. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m256d to type __m512d; the upper 256 bits of the result are zeroed. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m128 to type __m512; the upper 384 bits of the result are zeroed. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m256 to type __m512; the upper 256 bits of the result are zeroed. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m128i to type __m512i; the upper 384 bits of the result are zeroed. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Cast vector of type __m256i to type __m512i; the upper 256 bits of the result are zeroed. This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.
Add the lower double-precision (64-bit) floating-point element in a and b, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.\
Add the lower single-precision (32-bit) floating-point element in a and b, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.\
Decrypt 10 rounds of unsigned 8-bit integers in input using 128-bit AES key specified in the
384-bit key handle handle. Store the resulting unsigned 8-bit integers into the corresponding
elements of output. Returns 0 if the operation was successful, and 1 if the operation failed
due to a handle violation.
Decrypt 14 rounds of unsigned 8-bit integers in input using 256-bit AES key specified in the
512-bit key handle handle. Store the resulting unsigned 8-bit integers into the corresponding
elements of output. Returns 0 if the operation was successful, and 1 if the operation failed
due to a handle violation.
Decrypt 10 rounds of 8 groups of unsigned 8-bit integers in input using 128-bit AES key specified
in the 384-bit key handle handle. Store the resulting unsigned 8-bit integers into the corresponding
elements of output. Returns 0 if the operation was successful, and 1 if the operation failed
due to a handle violation.
Decrypt 14 rounds of 8 groups of unsigned 8-bit integers in input using 256-bit AES key specified
in the 512-bit key handle handle. Store the resulting unsigned 8-bit integers into the corresponding
elements of output. Returns 0 if the operation was successful, and 1 if the operation failed
due to a handle violation.
Encrypt 10 rounds of unsigned 8-bit integers in input using 128-bit AES key specified in the
384-bit key handle handle. Store the resulting unsigned 8-bit integers into the corresponding
elements of output. Returns 0 if the operation was successful, and 1 if the operation failed
due to a handle violation.
Encrypt 14 rounds of unsigned 8-bit integers in input using 256-bit AES key specified in the
512-bit key handle handle. Store the resulting unsigned 8-bit integers into the corresponding
elements of output. Returns 0 if the operation was successful, and 1 if the operation failed
due to a handle violation.
Encrypt 10 rounds of 8 groups of unsigned 8-bit integers in input using 128-bit AES key specified
in the 384-bit key handle handle. Store the resulting unsigned 8-bit integers into the corresponding
elements of output. Returns 0 if the operation was successful, and 1 if the operation failed
due to a handle violation.
Encrypt 14 rounds of 8 groups of unsigned 8-bit integers in input using 256-bit AES key specified
in the 512-bit key handle handle. Store the resulting unsigned 8-bit integers into the corresponding
elements of output. Returns 0 if the operation was successful, and 1 if the operation failed
due to a handle violation.
Considers the input b as packed 64-bit integers and c as packed 8-bit integers.
Then groups 8 8-bit values from cas indices into the bits of the corresponding 64-bit integer.
It then selects these bits and packs them into the output.
Round the packed double-precision (64-bit) floating-point elements in a
up to an integer value, and stores the results as packed double-precision
floating-point elements.
Round the packed single-precision (32-bit) floating-point elements in a
up to an integer value, and stores the results as packed single-precision
floating-point elements.
Round the lower double-precision (64-bit) floating-point element in b
up to an integer value, store the result as a double-precision
floating-point element in the lower element of the intrinsic result,
and copies the upper element from a to the upper element
of the intrinsic result.
Round the lower single-precision (32-bit) floating-point element in b
up to an integer value, store the result as a single-precision
floating-point element in the lower element of the intrinsic result,
and copies the upper 3 packed elements from a to the upper elements
of the intrinsic result.
Compare packed double-precision (64-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k.
Compare packed single-precision (32-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k.
Compare the lower double-precision (64-bit) floating-point element in a and b based on the comparison operand specified by imm8, and store the result in mask vector k.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower single-precision (32-bit) floating-point element in a and b based on the comparison operand specified by imm8, and store the result in mask vector k.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compares the lower double-precision (64-bit) floating-point element in
a and b based on the comparison operand specified by IMM5,
store the result in the lower element of returned vector,
and copies the upper element from a to the upper element of returned
vector.
Compare the lower double-precision (64-bit) floating-point element in a and b based on the comparison operand specified by imm8, and store the result in mask vector k.
Compares the lower single-precision (32-bit) floating-point element in
a and b based on the comparison operand specified by IMM5,
store the result in the lower element of returned vector,
and copies the upper 3 packed elements from a to the upper elements of
returned vector.
Compare the lower single-precision (32-bit) floating-point element in a and b based on the comparison operand specified by imm8, and store the result in mask vector k.
Compares each of the four floats in a to the corresponding element in b.
The result in the output vector will be 0xffffffff if the input elements
were equal, or 0 otherwise.
Compares the lowest f32 of both inputs for equality. The lowest 32 bits of
the result will be 0xffffffff if the two inputs are equal, or 0
otherwise. The upper 96 bits of the result are the upper 96 bits of a.
Compares packed strings in a and b with lengths la and lb
using the control in IMM8, and return 1 if b did not
contain a null character and the resulting mask was zero, and 0
otherwise.
Compares packed strings a and b with lengths la and lb using the
control in IMM8 and return the generated index. Similar to
_mm_cmpistri with the exception that _mm_cmpistri implicitly
determines the length of a and b.
Compares each of the four floats in a to the corresponding element in b.
The result in the output vector will be 0xffffffff if the input element
in a is greater than or equal to the corresponding element in b, or 0
otherwise.
Compares the lowest f32 of both inputs for greater than or equal. The
lowest 32 bits of the result will be 0xffffffff if a.extract(0) is
greater than or equal b.extract(0), or 0 otherwise. The upper 96 bits
of the result are the upper 96 bits of a.
Compares each of the four floats in a to the corresponding element in b.
The result in the output vector will be 0xffffffff if the input element
in a is greater than the corresponding element in b, or 0 otherwise.
Compares the lowest f32 of both inputs for greater than. The lowest 32
bits of the result will be 0xffffffff if a.extract(0) is greater
than b.extract(0), or 0 otherwise. The upper 96 bits of the result
are the upper 96 bits of a.
Compares packed strings with implicit lengths in a and b using the
control in IMM8, and return 1 if b did not contain a null
character and the resulting mask was zero, and 0 otherwise.
Compares packed strings with implicit lengths in a and b using the
control in IMM8 and return the generated index. Similar to
_mm_cmpestri with the exception that _mm_cmpestri requires the
lengths of a and b to be explicitly specified.
Compares each of the four floats in a to the corresponding element in b.
The result in the output vector will be 0xffffffff if the input element
in a is less than or equal to the corresponding element in b, or 0
otherwise.
Compares the lowest f32 of both inputs for less than or equal. The lowest
32 bits of the result will be 0xffffffff if a.extract(0) is less than
or equal b.extract(0), or 0 otherwise. The upper 96 bits of the result
are the upper 96 bits of a.
Compares each of the four floats in a to the corresponding element in b.
The result in the output vector will be 0xffffffff if the input element
in a is less than the corresponding element in b, or 0 otherwise.
Compares the lowest f32 of both inputs for less than. The lowest 32 bits
of the result will be 0xffffffff if a.extract(0) is less than
b.extract(0), or 0 otherwise. The upper 96 bits of the result are the
upper 96 bits of a.
Compares each of the four floats in a to the corresponding element in b.
The result in the output vector will be 0xffffffff if the input elements
are not equal, or 0 otherwise.
Compares the lowest f32 of both inputs for inequality. The lowest 32 bits
of the result will be 0xffffffff if a.extract(0) is not equal to
b.extract(0), or 0 otherwise. The upper 96 bits of the result are the
upper 96 bits of a.
Compares each of the four floats in a to the corresponding element in b.
The result in the output vector will be 0xffffffff if the input element
in a is not greater than or equal to the corresponding element in b,
or 0 otherwise.
Compares the lowest f32 of both inputs for not-greater-than-or-equal. The
lowest 32 bits of the result will be 0xffffffff if a.extract(0) is not
greater than or equal to b.extract(0), or 0 otherwise. The upper 96
bits of the result are the upper 96 bits of a.
Compares each of the four floats in a to the corresponding element in b.
The result in the output vector will be 0xffffffff if the input element
in a is not greater than the corresponding element in b, or 0
otherwise.
Compares the lowest f32 of both inputs for not-greater-than. The lowest 32
bits of the result will be 0xffffffff if a.extract(0) is not greater
than b.extract(0), or 0 otherwise. The upper 96 bits of the result are
the upper 96 bits of a.
Compares each of the four floats in a to the corresponding element in b.
The result in the output vector will be 0xffffffff if the input element
in a is not less than or equal to the corresponding element in b, or
0 otherwise.
Compares the lowest f32 of both inputs for not-less-than-or-equal. The
lowest 32 bits of the result will be 0xffffffff if a.extract(0) is not
less than or equal to b.extract(0), or 0 otherwise. The upper 96 bits
of the result are the upper 96 bits of a.
Compares each of the four floats in a to the corresponding element in b.
The result in the output vector will be 0xffffffff if the input element
in a is not less than the corresponding element in b, or 0
otherwise.
Compares the lowest f32 of both inputs for not-less-than. The lowest 32
bits of the result will be 0xffffffff if a.extract(0) is not less than
b.extract(0), or 0 otherwise. The upper 96 bits of the result are the
upper 96 bits of a.
Compares each of the four floats in a to the corresponding element in b.
Returns four floats that have one of two possible bit patterns. The element
in the output vector will be 0xffffffff if the input elements in a and
b are ordered (i.e., neither of them is a NaN), or 0 otherwise.
Returns a new vector with the low element of a replaced by the result
of comparing both of the lower elements of a and b to NaN. If
neither are equal to NaN then 0xFFFFFFFFFFFFFFFF is used and 0
otherwise.
Checks if the lowest f32 of both inputs are ordered. The lowest 32 bits of
the result will be 0xffffffff if neither of a.extract(0) or
b.extract(0) is a NaN, or 0 otherwise. The upper 96 bits of the result
are the upper 96 bits of a.
Compares each of the four floats in a to the corresponding element in b.
Returns four floats that have one of two possible bit patterns. The element
in the output vector will be 0xffffffff if the input elements in a and
b are unordered (i.e., at least on of them is a NaN), or 0 otherwise.
Returns a new vector with the low element of a replaced by the result of
comparing both of the lower elements of a and b to NaN. If either is
equal to NaN then 0xFFFFFFFFFFFFFFFF is used and 0 otherwise.
Checks if the lowest f32 of both inputs are unordered. The lowest 32 bits
of the result will be 0xffffffff if any of a.extract(0) or
b.extract(0) is a NaN, or 0 otherwise. The upper 96 bits of the result
are the upper 96 bits of a.
Compare the lower double-precision (64-bit) floating-point element in a and b based on the comparison operand specified by imm8, and return the boolean result (0 or 1).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower single-precision (32-bit) floating-point element in a and b based on the comparison operand specified by imm8, and return the boolean result (0 or 1).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compares two 32-bit floats from the low-order bits of a and b. Returns
1 if the value from a is greater than or equal to the one from b, or
0 otherwise.
Test each 32-bit element of a for equality with all other elements in a closer to the least significant bit. Each element’s comparison forms a zero extended bit vector in dst.
Test each 64-bit element of a for equality with all other elements in a closer to the least significant bit. Each element’s comparison forms a zero extended bit vector in dst.
Convert the signed 32-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.\
Convert the signed 64-bit integer b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the signed 64-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower single-precision (32-bit) floating-point element in a to a 32-bit integer, and store the result in dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower double-precision (64-bit) floating-point element in a to a 64-bit integer, and store the result in dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower double-precision (64-bit) floating-point element in a to a 32-bit integer, and store the result in dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower double-precision (64-bit) floating-point element in a to a 64-bit integer, and store the result in dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower double-precision (64-bit) floating-point element in b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower double-precision (64-bit) floating-point element in a to an unsigned 32-bit integer, and store the result in dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower double-precision (64-bit) floating-point element in a to an unsigned 64-bit integer, and store the result in dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the signed 32-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.\
Convert the signed 64-bit integer b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the signed 64-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower single-precision (32-bit) floating-point element in a to a 32-bit integer, and store the result in dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower single-precision (32-bit) floating-point element in a to a 64-bit integer, and store the result in dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower single-precision (32-bit) floating-point element in b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the lower single-precision (32-bit) floating-point element in a to a 32-bit integer, and store the result in dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower single-precision (32-bit) floating-point element in a to a 64-bit integer, and store the result in dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower single-precision (32-bit) floating-point element in a to an unsigned 32-bit integer, and store the result in dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower single-precision (32-bit) floating-point element in a to an unsigned 64-bit integer, and store the result in dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the unsigned 32-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the unsigned 64-bit integer b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the unsigned 64-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the signed 32-bit integer b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.
Convert the signed 32-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Convert the signed 64-bit integer b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.
Convert the signed 64-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Convert packed single-precision (32-bit) floating-point elements in two 128-bit vectors
a and b to packed BF16 (16-bit) floating-point elements, and store the results in a
128-bit wide vector.
Intel’s documentation
Convert packed BF16 (16-bit) floating-point even-indexed elements stored at memory locations starting at
location a to single precision (32-bit) floating-point elements, and store the results in dst.
Convert packed BF16 (16-bit) floating-point odd-indexed elements stored at memory locations starting at
location a to single precision (32-bit) floating-point elements, and store the results in dst.
Converts the 4 x 16-bit half-precision float values in the lowest 64-bit of
the 128-bit vector a into 4 x 32-bit float values stored in a 128-bit wide
vector.
Converts the 4 x 32-bit float values in the 128-bit vector a into 4 x
16-bit half-precision float values stored in the lowest 64-bit of a 128-bit
vector.
Converts the lower double-precision (64-bit) floating-point element in b
to a single-precision (32-bit) floating-point element, store the result in
the lower element of the return value, and copies the upper element from a
to the upper element the return value.
Converts the lower single-precision (32-bit) floating-point element in b
to a double-precision (64-bit) floating-point element, store the result in
the lower element of the return value, and copies the upper element from a
to the upper element the return value.
Convert the lower double-precision (64-bit) floating-point element in a to a 32-bit integer with truncation, and store the result in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the lower double-precision (64-bit) floating-point element in a to a 64-bit integer with truncation, and store the result in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the lower double-precision (64-bit) floating-point element in a to a 32-bit integer with truncation, and store the result in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the lower double-precision (64-bit) floating-point element in a to a 64-bit integer with truncation, and store the result in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the lower double-precision (64-bit) floating-point element in a to an unsigned 32-bit integer with truncation, and store the result in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the lower double-precision (64-bit) floating-point element in a to an unsigned 64-bit integer with truncation, and store the result in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the lower single-precision (32-bit) floating-point element in a to a 32-bit integer with truncation, and store the result in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the lower single-precision (32-bit) floating-point element in a to a 64-bit integer with truncation, and store the result in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the lower single-precision (32-bit) floating-point element in a to a 32-bit integer with truncation, and store the result in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the lower single-precision (32-bit) floating-point element in a to a 64-bit integer with truncation, and store the result in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the lower single-precision (32-bit) floating-point element in a to an unsigned 32-bit integer with truncation, and store the result in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the lower single-precision (32-bit) floating-point element in a to an unsigned 64-bit integer with truncation, and store the result in dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst.
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst.
Convert the unsigned 32-bit integer b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.
Convert the unsigned 32-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Convert the unsigned 64-bit integer b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.
Convert the unsigned 64-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst. Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets.
Divide the lower double-precision (64-bit) floating-point element in a by the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.\
Divide the lower single-precision (32-bit) floating-point element in a by the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.\
Compute dot-product of BF16 (16-bit) floating-point pairs in a and b,
accumulating the intermediate single-precision (32-bit) floating-point elements
with elements in src, and store the results in dst.
Intel’s documentation
Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding signed 8-bit
integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding
32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding signed 8-bit
integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding
32-bit integer in src with signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding unsigned 8-bit
integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding
32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding unsigned 8-bit
integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding
32-bit integer in src with signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding unsigned 8-bit
integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding
32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding unsigned 8-bit
integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding
32-bit integer in src with signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding unsigned 16-bit
integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding
32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding unsigned 16-bit
integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding
32-bit integer in src with signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding signed 16-bit
integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding
32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding signed 16-bit
integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding
32-bit integer in src with signed saturation, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding unsigned 16-bit
integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding
32-bit integer in src, and store the packed 32-bit results in dst.
Multiply groups of 2 adjacent pairs of unsigned 16-bit integers in a with corresponding unsigned 16-bit
integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding
32-bit integer in src with signed saturation, and store the packed 32-bit results in dst.
Extracts a single-precision (32-bit) floating-point element from a,
selected with IMM8. The returned i32 stores the float’s bit-pattern,
and may be converted back to a floating point number via casting.
Fix up packed double-precision (64-bit) floating-point elements in a and b using packed 64-bit integers in c, and store the results in dst. imm8 is used to set the required flags reporting.
Fix up packed single-precision (32-bit) floating-point elements in a and b using packed 32-bit integers in c, and store the results in dst. imm8 is used to set the required flags reporting.
Fix up the lower double-precision (64-bit) floating-point elements in a and b using the lower 64-bit integer in c, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst. imm8 is used to set the required flags reporting.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Fix up the lower single-precision (32-bit) floating-point elements in a and b using the lower 32-bit integer in c, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst. imm8 is used to set the required flags reporting.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Fix up the lower double-precision (64-bit) floating-point elements in a and b using the lower 64-bit integer in c, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst. imm8 is used to set the required flags reporting.
Fix up the lower single-precision (32-bit) floating-point elements in a and b using the lower 32-bit integer in c, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst. imm8 is used to set the required flags reporting.
Round the packed double-precision (64-bit) floating-point elements in a
down to an integer value, and stores the results as packed double-precision
floating-point elements.
Round the packed single-precision (32-bit) floating-point elements in a
down to an integer value, and stores the results as packed single-precision
floating-point elements.
Round the lower double-precision (64-bit) floating-point element in b
down to an integer value, store the result as a double-precision
floating-point element in the lower element of the intrinsic result,
and copies the upper element from a to the upper element of the intrinsic
result.
Round the lower single-precision (32-bit) floating-point element in b
down to an integer value, store the result as a single-precision
floating-point element in the lower element of the intrinsic result,
and copies the upper 3 packed elements from a to the upper elements
of the intrinsic result.
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and add the intermediate result to the lower element in c. Store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and add the intermediate result to the lower element in c. Store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiplies the lower double-precision (64-bit) floating-point elements in
a and b, and add the intermediate result to the lower element in c.
Stores the result in the lower element of the returned value, and copy the
upper element from a to the upper elements of the result.
Multiplies the lower single-precision (32-bit) floating-point elements in
a and b, and add the intermediate result to the lower element in c.
Stores the result in the lower element of the returned value, and copy the
3 upper elements from a to the upper elements of the result.
Multiplies packed double-precision (64-bit) floating-point elements in a
and b, and alternatively add and subtract packed elements in c to/from
the intermediate result.
Multiplies packed single-precision (32-bit) floating-point elements in a
and b, and alternatively add and subtract packed elements in c to/from
the intermediate result.
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and subtract the lower element in c from the intermediate result. Store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and subtract the lower element in c from the intermediate result. Store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiplies the lower double-precision (64-bit) floating-point elements in
a and b, and subtract the lower element in c from the intermediate
result. Store the result in the lower element of the returned value, and
copy the upper element from a to the upper elements of the result.
Multiplies the lower single-precision (32-bit) floating-point elements in
a and b, and subtract the lower element in c from the intermediate
result. Store the result in the lower element of the returned value, and
copy the 3 upper elements from a to the upper elements of the result.
Multiplies packed double-precision (64-bit) floating-point elements in a
and b, and alternatively subtract and add packed elements in c from/to
the intermediate result.
Multiplies packed single-precision (32-bit) floating-point elements in a
and b, and alternatively subtract and add packed elements in c from/to
the intermediate result.
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and add the negated intermediate result to the lower element in c. Store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and add the negated intermediate result to the lower element in c. Store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiplies the lower double-precision (64-bit) floating-point elements in
a and b, and add the negated intermediate result to the lower element
in c. Store the result in the lower element of the returned value, and
copy the upper element from a to the upper elements of the result.
Multiplies the lower single-precision (32-bit) floating-point elements in
a and b, and add the negated intermediate result to the lower element
in c. Store the result in the lower element of the returned value, and
copy the 3 upper elements from a to the upper elements of the result.
Multiplies packed double-precision (64-bit) floating-point elements in a
and b, and subtract packed elements in c from the negated intermediate
result.
Multiplies packed single-precision (32-bit) floating-point elements in a
and b, and subtract packed elements in c from the negated intermediate
result.
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and subtract the lower element in c from the negated intermediate result. Store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, subtract the lower element in c from the negated intermediate result, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiplies the lower double-precision (64-bit) floating-point elements in
a and b, and subtract packed elements in c from the negated
intermediate result. Store the result in the lower element of the returned
value, and copy the upper element from a to the upper elements of the
result.
Multiplies the lower single-precision (32-bit) floating-point elements in
a and b, and subtract packed elements in c from the negated
intermediate result. Store the result in the lower element of the
returned value, and copy the 3 upper elements from a to the upper
elements of the result.
Test packed double-precision (64-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k.
imm can be a combination of:
Test packed single-precision (32-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k.
imm can be a combination of:
Test the lower double-precision (64-bit) floating-point element in a for special categories specified
by imm8, and store the results in mask vector k.
imm can be a combination of:
Test the lower single-precision (32-bit) floating-point element in a for special categories specified
by imm8, and store the results in mask vector k.
imm can be a combination of:
Convert the exponent of each packed double-precision (64-bit) floating-point element in a to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in dst. This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of each packed single-precision (32-bit) floating-point element in a to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in dst. This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of the lower double-precision (64-bit) floating-point element in b to a double-precision (64-bit) floating-point number representing the integer exponent, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst. This intrinsic essentially calculates floor(log2(x)) for the lower element.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the exponent of the lower single-precision (32-bit) floating-point element in b to a single-precision (32-bit) floating-point number representing the integer exponent, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst. This intrinsic essentially calculates floor(log2(x)) for the lower element.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the exponent of the lower double-precision (64-bit) floating-point element in b to a double-precision (64-bit) floating-point number representing the integer exponent, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst. This intrinsic essentially calculates floor(log2(x)) for the lower element.
Convert the exponent of the lower single-precision (32-bit) floating-point element in b to a single-precision (32-bit) floating-point number representing the integer exponent, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst. This intrinsic essentially calculates floor(log2(x)) for the lower element.
Normalize the mantissas of packed double-precision (64-bit) floating-point elements in a, and store the results in dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of packed single-precision (32-bit) floating-point elements in a, and store the results in dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Performs an affine transformation on the packed bytes in x.
That is computes a*x+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs an affine transformation on the inverted packed bytes in x.
That is computes a*inv(x)+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
The inverse of a byte is defined with respect to the reduction polynomial x^8+x^4+x^3+x+1.
The inverse of 0 is 0.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs a multiplication in GF(2^8) on the packed bytes.
The field is in polynomial representation with the reduction polynomial
x^8 + x^4 + x^3 + x + 1.
Horizontally adds the adjacent pairs of values contained in 2 packed
128-bit vectors of [8 x i16]. Positive sums greater than 7FFFh are
saturated to 7FFFh. Negative sums less than 8000h are saturated to 8000h.
Horizontally subtract the adjacent pairs of values contained in 2
packed 128-bit vectors of [8 x i16]. Positive differences greater than
7FFFh are saturated to 7FFFh. Negative differences less than 8000h are
saturated to 8000h.
Stores 2 double-precision (64-bit) floating-point elements from a to memory starting at location base_addr
at packed 32-bit integer indices stored in vindex scaled by scale
Stores 4 single-precision (32-bit) floating-point elements from a to memory starting at location base_addr
at packed 32-bit integer indices stored in vindex scaled by scale
Stores 2 double-precision (64-bit) floating-point elements from a to memory starting at location base_addr
at packed 64-bit integer indices stored in vindex scaled by scale
Stores 2 single-precision (32-bit) floating-point elements from a to memory starting at location base_addr
at packed 64-bit integer indices stored in vindex scaled by scale
Inserts the len least-significant bits from the lower 64 bits of the 128-bit integer vector operand y into
the lower 64 bits of the 128-bit integer vector operand x at the index idx and of the length len.
Loads 128-bits of integer data from unaligned memory.
This intrinsic may perform better than _mm_loadu_si128
when the data crosses a cache line boundary.
Load 128-bits (composed of 4 packed 32-bit integers) from memory into dst. mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Load 128-bits (composed of 2 packed 64-bit integers) from memory into dst. mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Loads 128-bits (composed of 2 packed double-precision (64-bit)
floating-point elements) from memory into the returned vector.
mem_addr must be aligned on a 16-byte boundary or a general-protection
exception may be generated.
Loads four f32 values from aligned memory into a __m128. If the
pointer is not aligned to a 128-bit boundary (16 bytes) a general
protection fault will be triggered (fatal program crash).
Loads a double-precision value into the high-order bits of a 128-bit
vector of [2 x double]. The low-order bits are copied from the low-order
bits of the first operand.
Load internal wrapping key (IWKey). The 32-bit unsigned integer control specifies IWKey’s KeySource
and whether backing up the key is permitted. IWKey’s 256-bit encryption key is loaded from key_lo
and key_hi.
Loads a double-precision value into the low-order bits of a 128-bit
vector of [2 x double]. The high-order bits are copied from the
high-order bits of the first operand.
Loads 2 double-precision (64-bit) floating-point elements from memory into
the returned vector in reverse order. mem_addr must be aligned on a
16-byte boundary or a general-protection exception may be generated.
Loads 128-bits (composed of 2 packed double-precision (64-bit)
floating-point elements) from memory into the returned vector.
mem_addr does not need to be aligned on any particular boundary.
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the high 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst.
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the high 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst.
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the low 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst.
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the low 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst.
Multiplies corresponding pairs of packed 8-bit unsigned integer
values contained in the first source operand and packed 8-bit signed
integer values contained in the second source operand, add pairs of
contiguous products with signed saturation, and writes the 16-bit sums to
the corresponding bits in the destination.
Shuffle 8-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set).
Shuffle 32-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set).
Shuffle 64-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set)
Shuffle single-precision (32-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and add the intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper element from c to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and add the intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper 3 packed elements from c to the upper elements of dst.\
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and add the intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper element from c to the upper element of dst.
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and add the intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper 3 packed elements from c to the upper elements of dst.
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and subtract the lower element in c from the intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper element from c to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and subtract the lower element in c from the intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper 3 packed elements from c to the upper elements of dst.\
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and subtract the lower element in c from the intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper element from c to the upper element of dst.
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and subtract the lower element in c from the intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper 3 packed elements from c to the upper elements of dst.
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and add the negated intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper element from c to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and add the negated intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper 3 packed elements from c to the upper elements of dst.\
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and add the negated intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper element from c to the upper element of dst.
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and add the negated intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper 3 packed elements from c to the upper elements of dst.
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from c when the corresponding mask bit is not set).
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and subtract the lower element in c from the negated intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper element from c to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, subtract the lower element in c from the negated intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper 3 packed elements from c to the upper elements of dst.\
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and subtract the lower element in c from the negated intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper element from c to the upper element of dst.
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and subtract the lower element in c from the negated intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper 3 packed elements from c to the upper elements of dst.
Compute the absolute value of packed signed 8-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set)
Compute the absolute value of packed signed 16-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the absolute value of packed signed 32-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the absolute value of packed signed 64-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed 8-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed 16-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed 32-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed 64-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add the lower double-precision (64-bit) floating-point element in a and b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Add the lower single-precision (32-bit) floating-point element in a and b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Add the lower double-precision (64-bit) floating-point element in a and b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Add the lower single-precision (32-bit) floating-point element in a and b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Add packed signed 8-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed signed 16-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed unsigned 8-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed unsigned 16-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate pairs of 16-byte blocks in a and b into a 32-byte temporary result, shift the result right by imm8 bytes, and store the low 16 bytes in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate a and b into a 32-byte immediate result, shift the result right by imm8 32-bit elements, and store the low 16 bytes (4 elements) in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate a and b into a 32-byte immediate result, shift the result right by imm8 64-bit elements, and store the low 16 bytes (2 elements) in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Performs element-by-element bitwise AND between packed 32-bit integer elements of a and b, storing the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise AND of packed 64-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise AND of packed double-precision (64-bit) floating point numbers in a and b
and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Compute the bitwise AND of packed single-precision (32-bit) floating point numbers in a and b
and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Compute the bitwise NOT of packed 32-bit integers in a and then AND with b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise NOT of packed 64-bit integers in a and then AND with b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise NOT of packed double-precision (64-bit) floating point numbers in a and then
bitwise AND with b and store the results in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Compute the bitwise NOT of packed single-precision (32-bit) floating point numbers in a and then
bitwise AND with b and store the results in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Average packed unsigned 8-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Average packed unsigned 16-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Considers the input b as packed 64-bit integers and c as packed 8-bit integers.
Then groups 8 8-bit values from cas indices into the bits of the corresponding 64-bit integer.
It then selects these bits and packs them into the output.
Broadcasts the lower 2 packed 32-bit integers from a to all elements of dst using writemask k
(elements are copied from src if the corresponding bit is not set).
Broadcast the low packed 8-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low packed 32-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low packed 64-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low single-precision (32-bit) floating-point element from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Broadcast the low packed 16-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare the lower double-precision (64-bit) floating-point element in a and b based on the comparison operand specified by imm8, and store the result in mask vector k using zeromask k1 (the element is zeroed out when mask bit 0 is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower single-precision (32-bit) floating-point element in a and b based on the comparison operand specified by imm8, and store the result in mask vector k using zeromask k1 (the element is zeroed out when mask bit 0 is not seti).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower double-precision (64-bit) floating-point element in a and b based on the comparison operand specified by imm8, and store the result in mask vector k using zeromask k1 (the element is zeroed out when mask bit 0 is not set).
Compare the lower single-precision (32-bit) floating-point element in a and b based on the comparison operand specified by imm8, and store the result in mask vector k using zeromask k1 (the element is zeroed out when mask bit 0 is not set).
Compare packed signed 8-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed 32-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed 64-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed 32-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set).
Contiguously store the active 8-bit integers in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active 16-bit integers in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active 32-bit integers in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active 64-bit integers in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active double-precision (64-bit) floating-point elements in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active single-precision (32-bit) floating-point elements in a (those with their respective bit set in writemask k) to dst, and pass through the remaining elements from src.
Contiguously store the active double-precision (64-bit) floating-point elements in a (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Contiguously store the active single-precision (32-bit) floating-point elements in a (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Test each 32-bit element of a for equality with all other elements in a closer to the least significant bit using writemask k (elements are copied from src when the corresponding mask bit is not set). Each element’s comparison forms a zero extended bit vector in dst.
Test each 64-bit element of a for equality with all other elements in a closer to the least significant bit using writemask k (elements are copied from src when the corresponding mask bit is not set). Each element’s comparison forms a zero extended bit vector in dst.
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Convert the lower double-precision (64-bit) floating-point element in b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower single-precision (32-bit) floating-point element in b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Sign extend packed 8-bit integers in a to packed 16-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 8-bit integers in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 8-bit integers in the low 2 bytes of a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 16-bit integers in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 16-bit integers in a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed 32-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 32-bit integers in a to packed 16-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Sign extend packed 32-bit integers in a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 32-bit integers in a to packed 8-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed 32-bit integers in a to packed 16-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed 64-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 16-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed signed 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed 64-bit integers in a to packed 8-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed 64-bit integers in a to packed 16-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed 64-bit integers in a to packed 32-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Zero extend packed unsigned 8-bit integers in a to packed 16-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 8-bit integers in the low 4 bytes of a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 8-bit integers in the low 2 bytes of a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 16-bit integers in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 16-bit integers in the low 4 bytes of a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Zero extend packed unsigned 32-bit integers in a to packed 64-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed unsigned 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed single-precision (32-bit) floating-point elements in two vectors
a and b to packed BF16 (16-bit) floating-point elements, and store the results
in single vector dst using writemask k (elements are copied from src when the
corresponding mask bit is not set).
Intel’s documentation
Converts packed single-precision (32-bit) floating-point elements in a to packed BF16 (16-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Converts packed BF16 (16-bit) floating-point elements in a to single-precision (32-bit) floating-point
elements, and store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using writemask k (elements are copied from src if the corresponding bit is
not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit) floating-point elements, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Convert the lower double-precision (64-bit) floating-point element in b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 32-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed 16-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed 8-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 32-bit integers in a to packed 16-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 64-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 16-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 32-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 8-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 64-bit integers in a to packed 16-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed signed 64-bit integers in a to packed 32-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert the lower single-precision (32-bit) floating-point element in b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Convert packed double-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst using writemask k (elements are copied from src if the
corresponding bit is not set).
Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 32-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed unsigned 16-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed 8-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 32-bit integers in a to packed unsigned 16-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 64-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 16-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 32-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed 8-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 64-bit integers in a to packed 16-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Convert packed unsigned 64-bit integers in a to packed 32-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr.
Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets.
Divide packed double-precision (64-bit) floating-point elements in a by packed elements in b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Divide packed single-precision (32-bit) floating-point elements in a by packed elements in b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Divide the lower double-precision (64-bit) floating-point element in a by the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Divide the lower single-precision (32-bit) floating-point element in a by the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Divide the lower double-precision (64-bit) floating-point element in a by the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Divide the lower single-precision (32-bit) floating-point element in a by the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Compute dot-product of BF16 (16-bit) floating-point pairs in a and b,
accumulating the intermediate single-precision (32-bit) floating-point elements
with elements in src, and store the results in dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
Intel’s documentation
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 8-bit integers from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 16-bit integers from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 32-bit integers from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 64-bit integers from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active double-precision (64-bit) floating-point elements from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active single-precision (32-bit) floating-point elements from a (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 8-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 16-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 32-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active 64-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active double-precision (64-bit) floating-point elements from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Load contiguous active single-precision (32-bit) floating-point elements from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Fix up packed double-precision (64-bit) floating-point elements in a and b using packed 64-bit integers in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.
Fix up packed single-precision (32-bit) floating-point elements in a and b using packed 32-bit integers in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.
Fix up the lower double-precision (64-bit) floating-point elements in a and b using the lower 64-bit integer in c, store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. imm8 is used to set the required flags reporting.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Fix up the lower single-precision (32-bit) floating-point elements in a and b using the lower 32-bit integer in c, store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. imm8 is used to set the required flags reporting.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Fix up the lower double-precision (64-bit) floating-point elements in a and b using the lower 64-bit integer in c, store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. imm8 is used to set the required flags reporting.
Fix up the lower single-precision (32-bit) floating-point elements in a and b using the lower 32-bit integer in c, store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. imm8 is used to set the required flags reporting.
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and add the intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and add the intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and add the intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and add the intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and subtract the lower element in c from the intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and subtract the lower element in c from the intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and subtract the lower element in c from the intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and subtract the lower element in c from the intermediate result. Store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and add the negated intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and add the negated intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and add the negated intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and add the negated intermediate result to the lower element in c. Store the result in the lower element of dst using writemask k (the element is copied from a when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and subtract the lower element in c from the negated intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and subtract the lower element in c from the negated intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and subtract the lower element in c from the negated intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and subtract the lower element in c from the negated intermediate result. Store the result in the lower element of dst using writemask k (the element is copied from c when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Test packed double-precision (64-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the
corresponding mask bit is not set).
imm can be a combination of:
Test packed single-precision (32-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the
corresponding mask bit is not set).
imm can be a combination of:
Test the lower double-precision (64-bit) floating-point element in a for special categories specified
by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the
corresponding mask bit is not set).
imm can be a combination of:
Test the lower single-precision (32-bit) floating-point element in a for special categories specified
by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the
corresponding mask bit is not set).
imm can be a combination of:
Convert the exponent of each packed double-precision (64-bit) floating-point element in a to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of each packed single-precision (32-bit) floating-point element in a to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of the lower double-precision (64-bit) floating-point element in b to a double-precision (64-bit) floating-point number representing the integer exponent, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. This intrinsic essentially calculates floor(log2(x)) for the lower element.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the exponent of the lower single-precision (32-bit) floating-point element in b to a single-precision (32-bit) floating-point number representing the integer exponent, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. This intrinsic essentially calculates floor(log2(x)) for the lower element.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the exponent of the lower double-precision (64-bit) floating-point element in b to a double-precision (64-bit) floating-point number representing the integer exponent, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. This intrinsic essentially calculates floor(log2(x)) for the lower element.
Convert the exponent of the lower single-precision (32-bit) floating-point element in b to a single-precision (32-bit) floating-point number representing the integer exponent, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. This intrinsic essentially calculates floor(log2(x)) for the lower element.
Normalize the mantissas of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Performs an affine transformation on the packed bytes in x.
That is computes a*x+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs an affine transformation on the inverted packed bytes in x.
That is computes a*inv(x)+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
The inverse of a byte is defined with respect to the reduction polynomial x^8+x^4+x^3+x+1.
The inverse of 0 is 0.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs a multiplication in GF(2^8) on the packed bytes.
The field is in polynomial representation with the reduction polynomial
x^8 + x^4 + x^3 + x + 1.
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Stores 4 32-bit integer elements from a to memory starting at location base_addr at packed 32-bit integer
indices stored in vindex scaled by scale using writemask k (elements whose corresponding mask bit is not set
are not written to memory).
Stores 2 64-bit integer elements from a to memory starting at location base_addr at packed 32-bit integer
indices stored in vindex scaled by scale using writemask k (elements whose corresponding mask bit is not set
are not written to memory).
Stores 2 double-precision (64-bit) floating-point elements from a to memory starting at location base_addr
at packed 32-bit integer indices stored in vindex scaled by scale using writemask k (elements whose corresponding
mask bit is not set are not written to memory).
Stores 4 single-precision (32-bit) floating-point elements from a to memory starting at location base_addr
at packed 32-bit integer indices stored in vindex scaled by scale using writemask k (elements whose corresponding
mask bit is not set are not written to memory).
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Returns values from slice at offsets determined by offsets * scale,
where
scale should be 1, 2, 4 or 8. If mask is set, load the value from src in
that position instead.
Stores 2 32-bit integer elements from a to memory starting at location base_addr at packed 64-bit integer
indices stored in vindex scaled by scale using writemask k (elements whose corresponding mask bit is not set
are not written to memory).
Stores 2 64-bit integer elements from a to memory starting at location base_addr at packed 64-bit integer
indices stored in vindex scaled by scale using writemask k (elements whose corresponding mask bit is not set
are not written to memory).
Stores 2 double-precision (64-bit) floating-point elements from a to memory starting at location base_addr
at packed 64-bit integer indices stored in vindex scaled by scale using writemask k (elements whose corresponding
mask bit is not set are not written to memory).
Stores 2 single-precision (32-bit) floating-point elements from a to memory starting at location base_addr
at packed 64-bit integer indices stored in vindex scaled by scale using writemask k (elements whose corresponding
Load packed 32-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Load packed 64-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Load packed double-precision (64-bit) floating-point elements from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Load packed single-precision (32-bit) floating-point elements from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Load a double-precision (64-bit) floating-point element from memory into the lower element of dst
using writemask k (the element is copied from src when mask bit 0 is not set), and set the upper
element of dst to zero. mem_addr must be aligned on a 16-byte boundary or a general-protection
exception may be generated.
Load a single-precision (32-bit) floating-point element from memory into the lower element of dst
using writemask k (the element is copied from src when mask bit 0 is not set), and set the upper
3 packed elements of dst to zero. mem_addr must be aligned on a 16-byte boundary or a general-protection
exception may be generated.
Load packed 8-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 16-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 32-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 64-bit integers from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed double-precision (64-bit) floating-point elements from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed single-precision (32-bit) floating-point elements from memory into dst using writemask k
(elements are copied from src when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Counts the number of leading zero bits in each packed 32-bit integer in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Counts the number of leading zero bits in each packed 64-bit integer in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the high 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst using writemask k (elements are copied
from k when the corresponding mask bit is not set).
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the low 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst using writemask k (elements are copied
from k when the corresponding mask bit is not set).
Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed unsigned 8-bit integers in a by packed signed 8-bit integers in b, producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare the lower double-precision (64-bit) floating-point elements in a and b, store the maximum value in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower single-precision (32-bit) floating-point elements in a and b, store the maximum value in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower double-precision (64-bit) floating-point elements in a and b, store the maximum value in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Compare the lower single-precision (32-bit) floating-point elements in a and b, store the maximum value in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Compare packed signed 8-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare the lower double-precision (64-bit) floating-point elements in a and b, store the minimum value in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower single-precision (32-bit) floating-point elements in a and b, store the minimum value in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower double-precision (64-bit) floating-point elements in a and b, store the minimum value in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Compare the lower single-precision (32-bit) floating-point elements in a and b, store the minimum value in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Move packed double-precision (64-bit) floating-point elements from a to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Move packed single-precision (32-bit) floating-point elements from a to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Move the lower double-precision (64-bit) floating-point element from b to the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Move the lower single-precision (32-bit) floating-point element from b to the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Duplicate even-indexed double-precision (64-bit) floating-point elements from a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Duplicate odd-indexed single-precision (32-bit) floating-point elements from a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Duplicate even-indexed single-precision (32-bit) floating-point elements from a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the low signed 32-bit integers from each packed 64-bit element in a and b, and store the signed 64-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the low unsigned 32-bit integers from each packed 64-bit element in a and b, and store the unsigned 64-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the lower double-precision (64-bit) floating-point element in a and b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point element in a and b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiply the lower double-precision (64-bit) floating-point element in a and b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Multiply the lower single-precision (32-bit) floating-point element in a and b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Multiply the packed signed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the packed unsigned 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Truncate each intermediate integer to the 18 most significant bits, round by adding 1, and store bits [16:1] to dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the packed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the low 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the packed 32-bit integers in a and b, producing intermediate 64-bit integers, and store the low 32 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed 64-bit integers in a and b, producing intermediate 128-bit integers, and store
the low 64 bits of the intermediate integers in dst using writemask k (elements are copied from
src if the corresponding bit is not set).
For each 64-bit element in b, select 8 unaligned bytes using a byte-granular shift control within the corresponding 64-bit element of a, and store the 8 assembled bytes to the corresponding 64-bit element of dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise OR of packed 32-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise OR of packed 64-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise OR of packed double-precision (64-bit) floating point numbers in a and b
and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Compute the bitwise OR of packed single-precision (32-bit) floating point numbers in a and b
and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Convert packed signed 16-bit integers from a and b to packed 8-bit integers using signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers from a and b to packed 16-bit integers using signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 16-bit integers from a and b to packed 8-bit integers using unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed signed 32-bit integers from a and b to packed 16-bit integers using unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 128-bit lanes using the control in b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 8-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 32-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 64-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 8-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
double-precision (64-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src to dst if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
single-precision (32-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src to dst if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for the lower
double-precision (64-bit) floating-point element in a and b, store the result in the lower element
of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the
upper element from a to the upper element of dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for the lower
single-precision (32-bit) floating-point element in a and b, store the result in the lower element
of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the
upper 3 packed elements from a to the upper elements of dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for the lower
double-precision (64-bit) floating-point element in a and b, store the result in the lower element
of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the
upper element from a to the upper element of dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for the lower
single-precision (32-bit) floating-point element in a and b, store the result in the lower element
of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the
upper 3 packed elements from a to the upper elements of dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. The maximum relative error for this approximation is less than 2^-14.
Extract the reduced argument of packed double-precision (64-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using writemask k (elements are
copied from src to dst if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed single-precision (32-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using writemask k (elements are
copied from src to dst if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of the lower double-precision (64-bit) floating-point element in b
by the number of bits specified by imm8, store the result in the lower element of dst using writemask
k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a
to the upper element of dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of the lower single-precision (32-bit) floating-point element in b
by the number of bits specified by imm8, store the result in the lower element of dst using writemask
k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a.
to the upper element of dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of the lower double-precision (64-bit) floating-point element in b
by the number of bits specified by imm8, store the result in the lower element of dst using writemask
k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a
to the upper element of dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of the lower single-precision (32-bit) floating-point element in b
by the number of bits specified by imm8, store the result in the lower element of dst using writemask
k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a.
to the upper element of dst.
Rounding is done according to the imm8 parameter, which can be one of:
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Round packed double-precision (64-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed single-precision (32-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round the lower double-precision (64-bit) floating-point element in b to the number of fraction bits specified by imm8, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round the lower single-precision (32-bit) floating-point element in b to the number of fraction bits specified by imm8, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round the lower double-precision (64-bit) floating-point element in b to the number of fraction bits specified by imm8, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round the lower single-precision (32-bit) floating-point element in b to the number of fraction bits specified by imm8, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. The maximum relative error for this approximation is less than 2^-14.
Scale the packed double-precision (64-bit) floating-point elements in a using values from b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Scale the packed double-precision (64-bit) floating-point elements in a using values from b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Scale the packed double-precision (64-bit) floating-point elements in a using values from b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by imm8 bits, and store the upper 16-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by imm8 bits, and store the upper 32-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by imm8 bits, and store the upper 64-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 16-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 32-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 64-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by imm8 bits, and store the lower 16-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by imm8 bits, and store the lower 32-bits in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by imm8 bits, and store the lower 64-bits in dst using writemask k (elements are copied from src“ when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 16-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 32-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 64-bits in dst using writemask k (elements are copied from a when the corresponding mask bit is not set).
Shuffle 8-bit integers in a within 128-bit lanes using the control in the corresponding 8-bit element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 32-bit integers in a within 128-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements within 128-bit lanes using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a using the control in imm8, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 16-bit integers in the high 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the high 64 bits of 128-bit lanes of dst, with the low 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shuffle 16-bit integers in the low 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the low 64 bits of 128-bit lanes of dst, with the high 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the square root of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Compute the square root of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Compute the square root of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Compute the square root of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Shift packed 16-bit integers in a right by count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Store packed 32-bit integers from a into memory using writemask k.
mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Store packed 64-bit integers from a into memory using writemask k.
mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Store packed double-precision (64-bit) floating-point elements from a into memory using writemask k.
mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Store packed single-precision (32-bit) floating-point elements from a into memory using writemask k.
mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Store a double-precision (64-bit) floating-point element from a into memory using writemask k. mem_addr
must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Store a single-precision (32-bit) floating-point element from a into memory using writemask k. mem_addr
must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Store packed double-precision (64-bit) floating-point elements from a into memory using writemask k.
mem_addr does not need to be aligned on any particular boundary.
Store packed single-precision (32-bit) floating-point elements from a into memory using writemask k.
mem_addr does not need to be aligned on any particular boundary.
Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed 16-bit integers in b from packed 16-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed 32-bit integers in b from packed 32-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed 64-bit integers in b from packed 64-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed double-precision (64-bit) floating-point elements in b from packed double-precision (64-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed single-precision (32-bit) floating-point elements in b from packed single-precision (32-bit) floating-point elements in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract the lower double-precision (64-bit) floating-point element in b from the lower double-precision (64-bit) floating-point element in a, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Subtract the lower single-precision (32-bit) floating-point element in b from the lower single-precision (32-bit) floating-point element in a, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Subtract the lower double-precision (64-bit) floating-point element in b from the lower double-precision (64-bit) floating-point element in a, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Subtract the lower single-precision (32-bit) floating-point element in b from the lower single-precision (32-bit) floating-point element in a, store the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Subtract packed signed 8-bit integers in b from packed 8-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed signed 16-bit integers in b from packed 16-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 32-bit integer, the corresponding bit from src, a, and b are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst using writemask k at 32-bit granularity (32-bit elements are copied from src when the corresponding mask bit is not set).
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 64-bit integer, the corresponding bit from src, a, and b are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst using writemask k at 64-bit granularity (64-bit elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise AND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero.
Compute the bitwise AND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero.
Compute the bitwise AND of packed 32-bit integers in a and b, producing intermediate 32-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero.
Compute the bitwise AND of packed 64-bit integers in a and b, producing intermediate 64-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero.
Compute the bitwise NAND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero.
Compute the bitwise NAND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero.
Compute the bitwise NAND of packed 32-bit integers in a and b, producing intermediate 32-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero.
Compute the bitwise NAND of packed 64-bit integers in a and b, producing intermediate 64-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero.
Unpack and interleave 8-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 16-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 32-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 64-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave double-precision (64-bit) floating-point elements from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave single-precision (32-bit) floating-point elements from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 8-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 16-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 32-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave 64-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave double-precision (64-bit) floating-point elements from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Unpack and interleave single-precision (32-bit) floating-point elements from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise XOR of packed 32-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise XOR of packed 64-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the bitwise XOR of packed double-precision (64-bit) floating point numbers in a and b
and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Compute the bitwise XOR of packed single-precision (32-bit) floating point numbers in a and b
and store the results in dst using writemask k (elements are copied from src if the corresponding
bit is not set).
Loads packed 32-bit integers from memory pointed by mem_addr using mask
(elements are zeroed out when the highest bit is not set in the
corresponding element).
Loads packed 64-bit integers from memory pointed by mem_addr using mask
(elements are zeroed out when the highest bit is not set in the
corresponding element).
Loads packed double-precision (64-bit) floating-point elements from memory
into result using mask (elements are zeroed out when the high bit of the
corresponding element is not set).
Loads packed single-precision (32-bit) floating-point elements from memory
into result using mask (elements are zeroed out when the high bit of the
corresponding element is not set).
Stores packed 32-bit integers from a into memory pointed by mem_addr
using mask (elements are not stored when the highest bit is not set
in the corresponding element).
Stores packed 64-bit integers from a into memory pointed by mem_addr
using mask (elements are not stored when the highest bit is not set
in the corresponding element).
Compute the absolute value of packed signed 8-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the absolute value of packed signed 16-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the absolute value of packed signed 32-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the absolute value of packed signed 64-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed 8-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed 16-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed 32-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed 64-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add the lower double-precision (64-bit) floating-point element in a and b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Add the lower single-precision (32-bit) floating-point element in a and b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Add the lower double-precision (64-bit) floating-point element in a and b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Add the lower single-precision (32-bit) floating-point element in a and b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Add packed signed 8-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed signed 16-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed unsigned 8-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed unsigned 16-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate pairs of 16-byte blocks in a and b into a 32-byte temporary result, shift the result right by imm8 bytes, and store the low 16 bytes in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate a and b into a 32-byte immediate result, shift the result right by imm8 32-bit elements, and store the low 16 bytes (4 elements) in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate a and b into a 32-byte immediate result, shift the result right by imm8 64-bit elements, and store the low 16 bytes (2 elements) in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise AND of packed 32-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise AND of packed 64-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise AND of packed double-precision (64-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Compute the bitwise AND of packed single-precision (32-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Compute the bitwise NOT of packed 32-bit integers in a and then AND with b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise NOT of packed 64-bit integers in a and then AND with b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise NOT of packed double-precision (64-bit) floating point numbers in a and then
bitwise AND with b and store the results in dst using zeromask k (elements are zeroed out if the
corresponding bit is not set).
Compute the bitwise NOT of packed single-precision (32-bit) floating point numbers in a and then
bitwise AND with b and store the results in dst using zeromask k (elements are zeroed out if the
corresponding bit is not set).
Average packed unsigned 8-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Average packed unsigned 16-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcasts the lower 2 packed 32-bit integers from a to all elements of dst using zeromask k
(elements are zeroed out if the corresponding bit is not set).
Broadcast the low packed 8-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low packed 32-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low packed 64-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low single-precision (32-bit) floating-point element from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Broadcast the low packed 16-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Contiguously store the active double-precision (64-bit) floating-point elements in a (those with their respective bit set in zeromask k) to dst, and set the remaining elements to zero.
Contiguously store the active single-precision (32-bit) floating-point elements in a (those with their respective bit set in zeromask k) to dst, and set the remaining elements to zero.
Test each 32-bit element of a for equality with all other elements in a closer to the least significant bit using zeromask k (elements are zeroed out when the corresponding mask bit is not set). Each element’s comparison forms a zero extended bit vector in dst.
Test each 64-bit element of a for equality with all other elements in a closer to the least significant bit using zeromask k (elements are zeroed out when the corresponding mask bit is not set). Each element’s comparison forms a zero extended bit vector in dst.
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Convert the lower double-precision (64-bit) floating-point element in b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Rounding is done according to the rounding[3:0] parameter, which can be one of:\
Convert the lower single-precision (32-bit) floating-point element in b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Sign extend packed 8-bit integers in a to packed 16-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 8-bit integers in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 8-bit integers in the low 2 bytes of a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 16-bit integers in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 16-bit integers in a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 32-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 32-bit integers in a to packed 16-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Sign extend packed 32-bit integers in a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 16-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed 64-bit integers in a to packed 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed signed 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Zero extend packed unsigned 8-bit integers in a to packed 16-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 8-bit integers in th elow 4 bytes of a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 8-bit integers in the low 2 bytes of a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 16-bit integers in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 16-bit integers in the low 4 bytes of a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Zero extend packed unsigned 32-bit integers in a to packed 64-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed double-precision (64-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed double-precision (64-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed unsigned 64-bit integers in a to packed single-precision (32-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed single-precision (32-bit) floating-point elements in two vectors
a and b to packed BF16 (16-bit) floating-point elements, and store the results
in single vector dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set).
Intel’s documentation
Converts packed single-precision (32-bit) floating-point elements in a to packed BF16 (16-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out
when the corresponding mask bit is not set).
Converts packed BF16 (16-bit) floating-point elements in a to single-precision (32-bit) floating-point
elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers,
and store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit) floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Convert the lower double-precision (64-bit) floating-point element in b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 16-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed 32-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert the lower single-precision (32-bit) floating-point element in b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Convert packed double-precision (64-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set).
Convert packed double-precision (32-bit) floating-point elements in a to packed unsigned 32-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed unsigned 64-bit integers
with truncation, and store the result in dst using zeromask k (elements are zeroed out if the corresponding
bit is not set).
Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed unsigned 16-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 16-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed unsigned 32-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets.
Divide packed double-precision (64-bit) floating-point elements in a by packed elements in b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Divide packed single-precision (32-bit) floating-point elements in a by packed elements in b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Divide the lower double-precision (64-bit) floating-point element in a by the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Divide the lower single-precision (32-bit) floating-point element in a by the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Divide the lower double-precision (64-bit) floating-point element in a by the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Divide the lower single-precision (32-bit) floating-point element in a by the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Compute dot-product of BF16 (16-bit) floating-point pairs in a and b,
accumulating the intermediate single-precision (32-bit) floating-point elements
with elements in src, and store the results in dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
Intel’s documentation
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding signed 8-bit integers in b, producing 4 intermediate signed 16-bit results. Sum these 4 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src, and store the packed 32-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply groups of 2 adjacent pairs of signed 16-bit integers in a with corresponding 16-bit integers in b, producing 2 intermediate signed 32-bit results. Sum these 2 results with the corresponding 32-bit integer in src using signed saturation, and store the packed 32-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 8-bit integers from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 16-bit integers from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 32-bit integers from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 64-bit integers from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active double-precision (64-bit) floating-point elements from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active single-precision (32-bit) floating-point elements from a (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 8-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 16-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 32-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active 64-bit integers from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active double-precision (64-bit) floating-point elements from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Load contiguous active single-precision (32-bit) floating-point elements from unaligned memory at mem_addr (those with their respective bit set in mask k), and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Fix up packed double-precision (64-bit) floating-point elements in a and b using packed 64-bit integers in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.
Fix up packed single-precision (32-bit) floating-point elements in a and b using packed 32-bit integers in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). imm8 is used to set the required flags reporting.
Fix up the lower double-precision (64-bit) floating-point elements in a and b using the lower 64-bit integer in c, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. imm8 is used to set the required flags reporting.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Fix up the lower single-precision (32-bit) floating-point elements in a and b using the lower 32-bit integer in c, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. imm8 is used to set the required flags reporting.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Fix up the lower double-precision (64-bit) floating-point elements in a and b using the lower 64-bit integer in c, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. imm8 is used to set the required flags reporting.
Fix up the lower single-precision (32-bit) floating-point elements in a and b using the lower 32-bit integer in c, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. imm8 is used to set the required flags reporting.
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and add the intermediate result to the lower element in c. Store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and add the intermediate result to the lower element in c. Store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and add the intermediate result to the lower element in c. Store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and add the intermediate result to the lower element in c. Store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and subtract the lower element in c from the intermediate result. Store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and subtract the lower element in c from the intermediate result. Store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and subtract the lower element in c from the intermediate result. Store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and subtract the lower element in c from the intermediate result. Store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Multiply packed double-precision (64-bit) floating-point elements in a and b, alternatively add and subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, alternatively subtract and add packed elements in c from/to the intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, add the negated intermediate result to packed elements in c, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and add the negated intermediate result to the lower element in c. Store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and add the negated intermediate result to the lower element in c. Store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and add the negated intermediate result to the lower element in c. Store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and add the negated intermediate result to the lower element in c. Store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Multiply packed double-precision (64-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, subtract packed elements in c from the negated intermediate result, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and subtract the lower element in c from the negated intermediate result. Store the result in dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and subtract the lower element in c from the negated intermediate result. Store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiply the lower double-precision (64-bit) floating-point elements in a and b, and subtract the lower element in c from the negated intermediate result. Store the result in dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Multiply the lower single-precision (32-bit) floating-point elements in a and b, and subtract the lower element in c from the negated intermediate result. Store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Convert the exponent of each packed double-precision (64-bit) floating-point element in a to a double-precision (64-bit) floating-point number representing the integer exponent, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of each packed single-precision (32-bit) floating-point element in a to a single-precision (32-bit) floating-point number representing the integer exponent, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of the lower double-precision (64-bit) floating-point element in b to a double-precision (64-bit) floating-point number representing the integer exponent, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. This intrinsic essentially calculates floor(log2(x)) for the lower element.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the exponent of the lower single-precision (32-bit) floating-point element in b to a single-precision (32-bit) floating-point number representing the integer exponent, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. This intrinsic essentially calculates floor(log2(x)) for the lower element.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Convert the exponent of the lower double-precision (64-bit) floating-point element in b to a double-precision (64-bit) floating-point number representing the integer exponent, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. This intrinsic essentially calculates floor(log2(x)) for the lower element.
Convert the exponent of the lower single-precision (32-bit) floating-point element in b to a single-precision (32-bit) floating-point number representing the integer exponent, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. This intrinsic essentially calculates floor(log2(x)) for the lower element.
Normalize the mantissas of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Normalize the mantissas of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Normalize the mantissas of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined by interv and the sign depends on sc and the source sign.
The mantissa is normalized to the interval specified by interv, which can take the following values:
_MM_MANT_NORM_1_2 // interval [1, 2)
_MM_MANT_NORM_p5_2 // interval [0.5, 2)
_MM_MANT_NORM_p5_1 // interval [0.5, 1)
_MM_MANT_NORM_p75_1p5 // interval [0.75, 1.5)
The sign is determined by sc which can take the following values:
_MM_MANT_SIGN_src // sign = sign(src)
_MM_MANT_SIGN_zero // sign = 0
_MM_MANT_SIGN_nan // dst = NaN if sign(src) = 1
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Performs an affine transformation on the packed bytes in x.
That is computes a*x+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs an affine transformation on the inverted packed bytes in x.
That is computes a*inv(x)+b over the Galois Field 2^8 for each packed byte with a being a 8x8 bit matrix
and b being a constant 8-bit immediate value.
The inverse of a byte is defined with respect to the reduction polynomial x^8+x^4+x^3+x+1.
The inverse of 0 is 0.
Each pack of 8 bytes in x is paired with the 64-bit word at the same position in a.
Performs a multiplication in GF(2^8) on the packed bytes.
The field is in polynomial representation with the reduction polynomial
x^8 + x^4 + x^3 + x + 1.
Load packed 32-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Load packed 64-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Load packed double-precision (64-bit) floating-point elements from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Load packed single-precision (32-bit) floating-point elements from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Load a double-precision (64-bit) floating-point element from memory into the lower element of dst
using zeromask k (the element is zeroed out when mask bit 0 is not set), and set the upper element
of dst to zero. mem_addr must be aligned on a 16-byte boundary or a general-protection exception
may be generated.
Load a single-precision (32-bit) floating-point element from memory into the lower element of dst
using zeromask k (the element is zeroed out when mask bit 0 is not set), and set the upper 3 packed
elements of dst to zero. mem_addr must be aligned on a 16-byte boundary or a general-protection
exception may be generated.
Load packed 8-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 16-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 32-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed 64-bit integers from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed double-precision (64-bit) floating-point elements from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Load packed single-precision (32-bit) floating-point elements from memory into dst using zeromask k
(elements are zeroed out when the corresponding mask bit is not set).
mem_addr does not need to be aligned on any particular boundary.
Counts the number of leading zero bits in each packed 32-bit integer in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Counts the number of leading zero bits in each packed 64-bit integer in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the high 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst using writemask k (elements are zeroed
out when the corresponding mask bit is not set).
Multiply packed unsigned 52-bit integers in each 64-bit element of
b and c to form a 104-bit intermediate result. Add the low 52-bit
unsigned integer from the intermediate result with the
corresponding unsigned 64-bit integer in a, and store the
results in dst using writemask k (elements are zeroed
out when the corresponding mask bit is not set).
Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed unsigned 8-bit integers in a by packed signed 8-bit integers in b, producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 8-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare the lower double-precision (64-bit) floating-point elements in a and b, store the maximum value in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower single-precision (32-bit) floating-point elements in a and b, store the maximum value in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower double-precision (64-bit) floating-point elements in a and b, store the maximum value in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Compare the lower single-precision (32-bit) floating-point elements in a and b, store the maximum value in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Compare packed signed 8-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 16-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 32-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed signed 64-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 8-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 16-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 32-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed unsigned 64-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed double-precision (64-bit) floating-point elements in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare the lower double-precision (64-bit) floating-point elements in a and b, store the minimum value in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower single-precision (32-bit) floating-point elements in a and b, store the minimum value in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower double-precision (64-bit) floating-point elements in a and b, store the minimum value in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Compare the lower single-precision (32-bit) floating-point elements in a and b, store the minimum value in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Move packed double-precision (64-bit) floating-point elements from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Move packed single-precision (32-bit) floating-point elements from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Move the lower double-precision (64-bit) floating-point element from b to the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Move the lower single-precision (32-bit) floating-point element from b to the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Duplicate even-indexed double-precision (64-bit) floating-point elements from a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Duplicate odd-indexed single-precision (32-bit) floating-point elements from a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Duplicate even-indexed single-precision (32-bit) floating-point elements from a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the low signed 32-bit integers from each packed 64-bit element in a and b, and store the signed 64-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the low unsigned 32-bit integers from each packed 64-bit element in a and b, and store the unsigned 64-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed double-precision (64-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed single-precision (32-bit) floating-point elements in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the lower double-precision (64-bit) floating-point element in a and b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point element in a and b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiply the lower double-precision (64-bit) floating-point element in a and b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Multiply the lower single-precision (32-bit) floating-point element in a and b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Multiply the packed signed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the packed unsigned 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Truncate each intermediate integer to the 18 most significant bits, round by adding 1, and store bits [16:1] to dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the packed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the low 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the packed 32-bit integers in a and b, producing intermediate 64-bit integers, and store the low 32 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed 64-bit integers in a and b, producing intermediate 128-bit integers, and store
the low 64 bits of the intermediate integers in dst using zeromask k (elements are zeroed out if
the corresponding bit is not set).
For each 64-bit element in b, select 8 unaligned bytes using a byte-granular shift control within the corresponding 64-bit element of a, and store the 8 assembled bytes to the corresponding 64-bit element of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise OR of packed 32-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise OR of packed 64-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise OR of packed double-precision (64-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Compute the bitwise OR of packed single-precision (32-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Convert packed signed 16-bit integers from a and b to packed 8-bit integers using signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers from a and b to packed 16-bit integers using signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 16-bit integers from a and b to packed 8-bit integers using unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers from a and b to packed 16-bit integers using unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a within 128-bit lanes using the control in b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a within 128-bit lanes using the control in b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 8-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 32-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 64-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 8-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
double-precision (64-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
single-precision (32-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out if the corresponding mask bit is not set).
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for the lower
double-precision (64-bit) floating-point element in a and b, store the result in the lower element
of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper
element from a to the upper element of dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for the lower
single-precision (32-bit) floating-point element in a and b, store the result in the lower element
of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper
3 packed elements from a to the upper elements of dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for the lower
double-precision (64-bit) floating-point element in a and b, store the result in the lower element
of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper
element from a to the upper element of dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for the lower
single-precision (32-bit) floating-point element in a and b, store the result in the lower element
of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper
3 packed elements from a to the upper elements of dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. The maximum relative error for this approximation is less than 2^-14.
Extract the reduced argument of packed double-precision (64-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using zeromask k (elements are
zeroed out if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed single-precision (32-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst using zeromask k (elements are
zeroed out if the corresponding mask bit is not set).
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of the lower double-precision (64-bit) floating-point element in b
by the number of bits specified by imm8, store the result in the lower element of dst using zeromask
k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a
to the upper element of dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of the lower single-precision (32-bit) floating-point element in b
by the number of bits specified by imm8, store the result in the lower element of dst using zeromask
k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a.
to the upper element of dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of the lower double-precision (64-bit) floating-point element in b
by the number of bits specified by imm8, store the result in the lower element of dst using zeromask
k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a
to the upper element of dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of the lower single-precision (32-bit) floating-point element in b
by the number of bits specified by imm8, store the result in the lower element of dst using zeromask
k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a.
to the upper element of dst.
Rounding is done according to the imm8 parameter, which can be one of:
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Round packed double-precision (64-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed single-precision (32-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round the lower double-precision (64-bit) floating-point element in b to the number of fraction bits specified by imm8, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round the lower single-precision (32-bit) floating-point element in b to the number of fraction bits specified by imm8, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round the lower double-precision (64-bit) floating-point element in b to the number of fraction bits specified by imm8, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round the lower single-precision (32-bit) floating-point element in b to the number of fraction bits specified by imm8, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst. The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst. The maximum relative error for this approximation is less than 2^-14.
Scale the packed double-precision (64-bit) floating-point elements in a using values from b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Scale the packed double-precision (64-bit) floating-point elements in a using values from b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Scale the packed double-precision (64-bit) floating-point elements in a using values from b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Broadcast the low packed 16-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by imm8 bits, and store the upper 16-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by imm8 bits, and store the upper 32-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by imm8 bits, and store the upper 64-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 16-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 32-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 64-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by imm8 bits, and store the lower 16-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by imm8 bits, and store the lower 32-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by imm8 bits, and store the lower 64-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 16-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 32-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 64-bits in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle packed 8-bit integers in a according to shuffle control mask in the corresponding 8-bit element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 32-bit integers in a within 128-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle double-precision (64-bit) floating-point elements within 128-bit lanes using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle single-precision (32-bit) floating-point elements in a using the control in imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 16-bit integers in the high 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the high 64 bits of 128-bit lanes of dst, with the low 64 bits of 128-bit lanes being copied from a to dst, using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shuffle 16-bit integers in the low 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the low 64 bits of 128-bit lanes of dst, with the high 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the square root of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Compute the square root of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Compute the square root of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Compute the square root of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Shift packed 16-bit integers in a right by count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 32-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed 16-bit integers in b from packed 16-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed 32-bit integers in b from packed 32-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed 64-bit integers in b from packed 64-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed double-precision (64-bit) floating-point elements in b from packed double-precision (64-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed single-precision (32-bit) floating-point elements in b from packed single-precision (32-bit) floating-point elements in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract the lower double-precision (64-bit) floating-point element in b from the lower double-precision (64-bit) floating-point element in a, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.\
Subtract the lower single-precision (32-bit) floating-point element in b from the lower single-precision (32-bit) floating-point element in a, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.\
Subtract the lower double-precision (64-bit) floating-point element in b from the lower double-precision (64-bit) floating-point element in a, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Subtract the lower single-precision (32-bit) floating-point element in b from the lower single-precision (32-bit) floating-point element in a, store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements of dst.
Subtract packed signed 8-bit integers in b from packed 8-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed signed 16-bit integers in b from packed 16-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 32-bit integer, the corresponding bit from a, b, and c are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst using zeromask k at 32-bit granularity (32-bit elements are zeroed out when the corresponding mask bit is not set).
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 64-bit integer, the corresponding bit from a, b, and c are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst using zeromask k at 64-bit granularity (64-bit elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 8-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 16-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 32-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 64-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave double-precision (64-bit) floating-point elements from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave single-precision (32-bit) floating-point elements from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 8-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 16-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 32-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave 64-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave double-precision (64-bit) floating-point elements from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Unpack and interleave single-precision (32-bit) floating-point elements from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise XOR of packed 32-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise XOR of packed 64-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the bitwise XOR of packed double-precision (64-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Compute the bitwise XOR of packed single-precision (32-bit) floating point numbers in a and b and
store the results in dst using zeromask k (elements are zeroed out if the corresponding bit is not set).
Compare the lower double-precision (64-bit) floating-point elements in a and b, store the maximum value in the lower element of dst, and copy the upper element from a to the upper element of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower single-precision (32-bit) floating-point elements in a and b, store the maximum value in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compares the first single-precision (32-bit) floating-point element of a
and b, and return the maximum value in the first element of the return
value, the other elements are copied from a.
Compare the lower double-precision (64-bit) floating-point elements in a and b, store the minimum value in the lower element of dst , and copy the upper element from a to the upper element of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower single-precision (32-bit) floating-point elements in a and b, store the minimum value in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compares the first single-precision (32-bit) floating-point element of a
and b, and return the minimum value in the first element of the return
value, the other elements are copied from a.
Finds the minimum unsigned 16-bit element in the 128-bit __m128i vector,
returning a vector containing its value in its first position, and its
index
in its second position; all other elements are set to zero.
Loads 4 32-bit integer elements from memory starting at location base_addr at packed 32-bit integer
indices stored in vindex scaled by scale using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Loads 2 64-bit integer elements from memory starting at location base_addr at packed 32-bit integer
indices stored in vindex scaled by scale using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Loads 2 double-precision (64-bit) floating-point elements from memory starting at location base_addr
at packed 32-bit integer indices stored in vindex scaled by scale using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Loads 4 single-precision (32-bit) floating-point elements from memory starting at location base_addr
at packed 32-bit integer indices stored in vindex scaled by scale using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Loads 2 32-bit integer elements from memory starting at location base_addr at packed 64-bit integer
indices stored in vindex scaled by scale using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Loads 2 64-bit integer elements from memory starting at location base_addr at packed 64-bit integer
indices stored in vindex scaled by scale using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Loads 2 double-precision (64-bit) floating-point elements from memory starting at location base_addr
at packed 64-bit integer indices stored in vindex scaled by scale using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Loads 2 single-precision (32-bit) floating-point elements from memory starting at location base_addr
at packed 64-bit integer indices stored in vindex scaled by scale using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Constructs a 128-bit floating-point vector of [2 x double]. The lower
64 bits are set to the lower 64 bits of the second parameter. The upper
64 bits are set to the upper 64 bits of the first parameter.
Subtracts 8-bit unsigned integer values and computes the absolute
values of the differences to the corresponding bits in the destination.
Then sums of the absolute differences are returned according to the bit
fields in the immediate operand.
Multiply the lower double-precision (64-bit) floating-point element in a and b, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.\
Multiply the lower single-precision (32-bit) floating-point element in a and b, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.\
Multiplies packed 16-bit signed integer values, truncate the 32-bit
product to the 18 most significant bits by right-shifting, round the
truncated value by adding 1, and write bits [16:1] to the destination.
Multiplies the packed 32-bit integers in a and b, producing intermediate
64-bit integers, and returns the lowest 32-bit, whatever they might be,
reinterpreted as a signed integer. While pmulld __m128i::splat(2), __m128i::splat(2) returns the obvious __m128i::splat(4), due to wrapping
arithmetic pmulld __m128i::splat(i32::MAX), __m128i::splat(2) would
return a negative number.
For each 64-bit element in b, select 8 unaligned bytes using a byte-granular shift control within the corresponding 64-bit element of a, and store the 8 assembled bytes to the corresponding 64-bit element of dst.
Shuffle double-precision (64-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst.
Shuffle single-precision (32-bit) floating-point elements in a and b across lanes using the corresponding selector and index in idx, and store the results in dst.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
double-precision (64-bit) floating-point elements in a and b, and store the results in dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for packed
single-precision (32-bit) floating-point elements in a and b, and store the results in dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for the lower
double-precision (64-bit) floating-point element in a and b, store the result in the lower element
of dst, and copy the upper element from a to the upper element of dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Calculate the max, min, absolute max, or absolute min (depending on control in imm8) for the lower
single-precision (32-bit) floating-point element in a and b, store the result in the lower element
of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Lower 2 bits of IMM8 specifies the operation control:
00 = min, 01 = max, 10 = absolute min, 11 = absolute max.
Upper 2 bits of IMM8 specifies the sign control:
00 = sign from a, 01 = sign from compare result, 10 = clear sign bit, 11 = set sign bit.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compute the approximate reciprocal of packed double-precision (64-bit) floating-point elements in a, and store the results in dst. The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of packed single-precision (32-bit) floating-point elements in a, and store the results in dst. The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst. The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst. The maximum relative error for this approximation is less than 2^-14.
Extract the reduced argument of packed double-precision (64-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of packed single-precision (32-bit) floating-point elements in a by
the number of bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of the lower double-precision (64-bit) floating-point element in b
by the number of bits specified by imm8, store the result in the lower element of dst, and copy
the upper element from a to the upper element of dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of the lower single-precision (32-bit) floating-point element in b
by the number of bits specified by imm8, store the result in the lower element of dst, and copy
the upper element from a.
to the upper element of dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of the lower double-precision (64-bit) floating-point element in b
by the number of bits specified by imm8, store the result in the lower element of dst using, and
copy the upper element from a.
to the upper element of dst.
Rounding is done according to the imm8 parameter, which can be one of:
Extract the reduced argument of the lower single-precision (32-bit) floating-point element in b
by the number of bits specified by imm8, store the result in the lower element of dst, and copy
the upper element from a.
to the upper element of dst.
Rounding is done according to the imm8 parameter, which can be one of:
Rotate the bits in each packed 32-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst.
Rotate the bits in each packed 64-bit integer in a to the left by the number of bits specified in the corresponding element of b, and store the results in dst.
Rotate the bits in each packed 32-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst.
Rotate the bits in each packed 64-bit integer in a to the right by the number of bits specified in the corresponding element of b, and store the results in dst.
Round the packed double-precision (64-bit) floating-point elements in a
using the ROUNDING parameter, and stores the results as packed
double-precision floating-point elements.
Rounding is done according to the rounding parameter, which can be one of:
Round the packed single-precision (32-bit) floating-point elements in a
using the ROUNDING parameter, and stores the results as packed
single-precision floating-point elements.
Rounding is done according to the rounding parameter, which can be one of:
Round the lower double-precision (64-bit) floating-point element in b
using the ROUNDING parameter, store the result as a double-precision
floating-point element in the lower element of the intrinsic result,
and copies the upper element from a to the upper element of the intrinsic
result.
Rounding is done according to the rounding parameter, which can be one of:
Round the lower single-precision (32-bit) floating-point element in b
using the ROUNDING parameter, store the result as a single-precision
floating-point element in the lower element of the intrinsic result,
and copies the upper 3 packed elements from a to the upper elements
of the intrinsic result.
Rounding is done according to the rounding parameter, which can be one of:
Round packed double-precision (64-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round packed single-precision (32-bit) floating-point elements in a to the number of fraction bits specified by imm8, and store the results in dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round the lower double-precision (64-bit) floating-point element in b to the number of fraction bits specified by imm8, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round the lower single-precision (32-bit) floating-point element in b to the number of fraction bits specified by imm8, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round the lower double-precision (64-bit) floating-point element in b to the number of fraction bits specified by imm8, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Round the lower single-precision (32-bit) floating-point element in b to the number of fraction bits specified by imm8, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Rounding is done according to the imm8[2:0] parameter, which can be one of:\
Compute the approximate reciprocal square root of packed double-precision (64-bit) floating-point elements in a, and store the results in dst. The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of packed single-precision (32-bit) floating-point elements in a, and store the results in dst. The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst. The maximum relative error for this approximation is less than 2^-14.
Compute the approximate reciprocal square root of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst. The maximum relative error for this approximation is less than 2^-14.
Scale the packed double-precision (64-bit) floating-point elements in a using values from b, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.\
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.\
Scale the packed double-precision (64-bit) floating-point elements in a using values from b, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.
Performs a serializing operation on all non-temporal (“streaming”) store instructions that
were issued by the current thread prior to this instruction.
Performs an intermediate calculation for the next four SHA1 message values
(unsigned 32-bit integers) using previous message values from a and b,
and returning the result.
Performs the final calculation for the next four SHA1 message values
(unsigned 32-bit integers) using the intermediate result in a and the
previous message values in b, and returns the result.
Calculate SHA1 state variable E after four rounds of operation from the
current SHA1 state variable a, add that value to the scheduled values
(unsigned 32-bit integers) in b, and returns the result.
Performs four rounds of SHA1 operation using an initial SHA1 state (A,B,C,D)
from a and some pre-computed sum of the next 4 round message values
(unsigned 32-bit integers), and state variable E from b, and return the
updated SHA1 state (A,B,C,D). FUNC contains the logic functions and round
constants.
Performs an intermediate calculation for the next four SHA256 message values
(unsigned 32-bit integers) using previous message values from a and b,
and return the result.
Performs the final calculation for the next four SHA256 message values
(unsigned 32-bit integers) using previous message values from a and b,
and return the result.
Performs 2 rounds of SHA256 operation using an initial SHA256 state
(C,D,G,H) from a, an initial SHA256 state (A,B,E,F) from b, and a
pre-computed sum of the next 2 round message values (unsigned 32-bit
integers) and the corresponding round constants from k, and store the
updated SHA256 state (A,B,E,F) in dst.
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by imm8 bits, and store the upper 16-bits in dst).
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by imm8 bits, and store the upper 32-bits in dst.
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by imm8 bits, and store the upper 64-bits in dst).
Concatenate packed 16-bit integers in a and b producing an intermediate 32-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 16-bits in dst.
Concatenate packed 32-bit integers in a and b producing an intermediate 64-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 32-bits in dst.
Concatenate packed 64-bit integers in a and b producing an intermediate 128-bit result. Shift the result left by the amount specified in the corresponding element of c, and store the upper 64-bits in dst.
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by imm8 bits, and store the lower 16-bits in dst.
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by imm8 bits, and store the lower 32-bits in dst.
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by imm8 bits, and store the lower 64-bits in dst.
Concatenate packed 16-bit integers in b and a producing an intermediate 32-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 16-bits in dst.
Concatenate packed 32-bit integers in b and a producing an intermediate 64-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 32-bits in dst.
Concatenate packed 64-bit integers in b and a producing an intermediate 128-bit result. Shift the result right by the amount specified in the corresponding element of c, and store the lower 64-bits in dst.
Constructs a 128-bit floating-point vector of [2 x double] from two
128-bit vector parameters of [2 x double], using the immediate-value
parameter as a specifier.
Negates packed 8-bit integers in a when the corresponding signed 8-bit
integer in b is negative, and returns the result.
Elements in result are zeroed out when the corresponding element in b
is zero.
Negates packed 16-bit integers in a when the corresponding signed 16-bit
integer in b is negative, and returns the results.
Elements in result are zeroed out when the corresponding element in b
is zero.
Negates packed 32-bit integers in a when the corresponding signed 32-bit
integer in b is negative, and returns the results.
Element in result are zeroed out when the corresponding element in b
is zero.
Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst.
This is one of the two SM3 message scheduling intrinsics. The intrinsic performs
an initial calculation for the next four SM3 message words. The calculated results
are stored in dst.
This is one of the two SM3 message scheduling intrinsics. The intrinsic performs
the final calculation for the next four SM3 message words. The calculated results
are stored in dst.
The intrinsic performs two rounds of SM3 operation using initial SM3 state (C, D, G, H)
from a, an initial SM3 states (A, B, E, F) from b and a pre-computed words from the
c. a with initial SM3 state of (C, D, G, H) assumes input of non-rotated left variables
from previous state. The updated SM3 state (A, B, E, F) is written to a. The imm8
should contain the even round number for the first of the two rounds computed by this instruction.
The computation masks the imm8 value by ANDing it with 0x3E so that only even round numbers
from 0 through 62 are used for this operation. The calculated results are stored in dst.
This intrinsic performs four rounds of SM4 key expansion. The intrinsic operates on independent
128-bit lanes. The calculated results are stored in dst.
Compute the square root of the lower double-precision (64-bit) floating-point element in b, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.\
Compute the square root of the lower single-precision (32-bit) floating-point element in b, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.\
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst.
Shift packed 64-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst.
Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst.
Stores the lower double-precision (64-bit) floating-point element from a
into 2 contiguous elements in memory. mem_addr must be aligned on a
16-byte boundary or a general-protection exception may be generated.
Store 128-bits (composed of 4 packed 32-bit integers) from a into memory. mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Store 128-bits (composed of 2 packed 64-bit integers) from a into memory. mem_addr must be aligned on a 16-byte boundary or a general-protection exception may be generated.
Stores 128-bits (composed of 2 packed double-precision (64-bit)
floating-point elements) from a into memory. mem_addr must be aligned
on a 16-byte boundary or a general-protection exception may be generated.
Stores the lower double-precision (64-bit) floating-point element from a
into 2 contiguous elements in memory. mem_addr must be aligned on a
16-byte boundary or a general-protection exception may be generated.
Stores 2 double-precision (64-bit) floating-point elements from a into
memory in reverse order.
mem_addr must be aligned on a 16-byte boundary or a general-protection
exception may be generated.
Stores 128-bits (composed of 2 packed double-precision (64-bit)
floating-point elements) from a into memory.
mem_addr does not need to be aligned on any particular boundary.
Load 128-bits of integer data from memory into dst. mem_addr must be aligned on a 16-byte
boundary or a general-protection exception may be generated. To minimize caching, the data
is flagged as non-temporal (unlikely to be used again soon)
Stores a 128-bit floating point vector of [2 x double] to a 128-bit
aligned memory location.
To minimize caching, the data is flagged as non-temporal (unlikely to be
used again soon).
Stores a 32-bit integer value in the specified memory location.
To minimize caching, the data is flagged as non-temporal (unlikely to be
used again soon).
Stores a 64-bit integer value in the specified memory location.
To minimize caching, the data is flagged as non-temporal (unlikely to be
used again soon).
Stores a 128-bit integer vector to a 128-bit aligned memory location.
To minimize caching, the data is flagged as non-temporal (unlikely to be
used again soon).
Subtract the lower double-precision (64-bit) floating-point element in b from the lower double-precision (64-bit) floating-point element in a, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.\
Subtract the lower single-precision (32-bit) floating-point element in b from the lower single-precision (32-bit) floating-point element in a, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.\
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 32-bit integer, the corresponding bit from a, b, and c are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst.
Bitwise ternary logic that provides the capability to implement any three-operand binary function; the specific binary function is specified by value in imm8. For each bit in each packed 64-bit integer, the corresponding bit from a, b, and c are used to form a 3 bit index into imm8, and the value at that bit in imm8 is written to the corresponding bit in dst.
Compute the bitwise AND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero.
Compute the bitwise AND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero.
Compute the bitwise AND of packed 32-bit integers in a and b, producing intermediate 32-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero.
Compute the bitwise AND of packed 64-bit integers in a and b, producing intermediate 64-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero.
Computes the bitwise AND of 128 bits (representing double-precision (64-bit)
floating-point elements) in a and b, producing an intermediate 128-bit
value, and set ZF to 1 if the sign bit of each 64-bit element in the
intermediate value is zero, otherwise set ZF to 0. Compute the bitwise
NOT of a and then AND with b, producing an intermediate value, and set
CF to 1 if the sign bit of each 64-bit element in the intermediate value
is zero, otherwise set CF to 0. Return the CF value.
Computes the bitwise AND of 128 bits (representing single-precision (32-bit)
floating-point elements) in a and b, producing an intermediate 128-bit
value, and set ZF to 1 if the sign bit of each 32-bit element in the
intermediate value is zero, otherwise set ZF to 0. Compute the bitwise
NOT of a and then AND with b, producing an intermediate value, and set
CF to 1 if the sign bit of each 32-bit element in the intermediate value
is zero, otherwise set CF to 0. Return the CF value.
Compute the bitwise NAND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k if the intermediate value is zero.
Compute the bitwise NAND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k if the intermediate value is zero.
Compute the bitwise NAND of packed 32-bit integers in a and b, producing intermediate 32-bit values, and set the corresponding bit in result mask k if the intermediate value is zero.
Compute the bitwise NAND of packed 64-bit integers in a and b, producing intermediate 64-bit values, and set the corresponding bit in result mask k if the intermediate value is zero.
Computes the bitwise AND of 128 bits (representing double-precision (64-bit)
floating-point elements) in a and b, producing an intermediate 128-bit
value, and set ZF to 1 if the sign bit of each 64-bit element in the
intermediate value is zero, otherwise set ZF to 0. Compute the bitwise
NOT of a and then AND with b, producing an intermediate value, and set
CF to 1 if the sign bit of each 64-bit element in the intermediate value
is zero, otherwise set CF to 0. Return 1 if both the ZF and CF values
are zero, otherwise return 0.
Computes the bitwise AND of 128 bits (representing single-precision (32-bit)
floating-point elements) in a and b, producing an intermediate 128-bit
value, and set ZF to 1 if the sign bit of each 32-bit element in the
intermediate value is zero, otherwise set ZF to 0. Compute the bitwise
NOT of a and then AND with b, producing an intermediate value, and set
CF to 1 if the sign bit of each 32-bit element in the intermediate value
is zero, otherwise set CF to 0. Return 1 if both the ZF and CF values
are zero, otherwise return 0.
Computes the bitwise AND of 128 bits (representing double-precision (64-bit)
floating-point elements) in a and b, producing an intermediate 128-bit
value, and set ZF to 1 if the sign bit of each 64-bit element in the
intermediate value is zero, otherwise set ZF to 0. Compute the bitwise
NOT of a and then AND with b, producing an intermediate value, and set
CF to 1 if the sign bit of each 64-bit element in the intermediate value
is zero, otherwise set CF to 0. Return the ZF value.
Computes the bitwise AND of 128 bits (representing single-precision (32-bit)
floating-point elements) in a and b, producing an intermediate 128-bit
value, and set ZF to 1 if the sign bit of each 32-bit element in the
intermediate value is zero, otherwise set ZF to 0. Compute the bitwise
NOT of a and then AND with b, producing an intermediate value, and set
CF to 1 if the sign bit of each 32-bit element in the intermediate value
is zero, otherwise set CF to 0. Return the ZF value.
Compares two 32-bit floats from the low-order bits of a and b. Returns
1 if they are equal, or 0 otherwise. This instruction will not signal
an exception if either argument is a quiet NaN.
Compares two 32-bit floats from the low-order bits of a and b. Returns
1 if the value from a is greater than or equal to the one from b, or
0 otherwise. This instruction will not signal an exception if either
argument is a quiet NaN.
Compares two 32-bit floats from the low-order bits of a and b. Returns
1 if the value from a is greater than the one from b, or 0
otherwise. This instruction will not signal an exception if either argument
is a quiet NaN.
Compares two 32-bit floats from the low-order bits of a and b. Returns
1 if the value from a is less than or equal to the one from b, or 0
otherwise. This instruction will not signal an exception if either argument
is a quiet NaN.
Compares two 32-bit floats from the low-order bits of a and b. Returns
1 if the value from a is less than the one from b, or 0 otherwise.
This instruction will not signal an exception if either argument is a quiet
NaN.
Compares two 32-bit floats from the low-order bits of a and b. Returns
1 if they are not equal, or 0 otherwise. This instruction will not
signal an exception if either argument is a quiet NaN.
Returns vector of type __m128d with indeterminate elements.with indetermination elements.
Despite using the word “undefined” (following Intel’s naming scheme), this non-deterministically
picks some valid value and is not equivalent to mem::MaybeUninit.
In practice, this is typically equivalent to mem::zeroed.
Returns vector of type __m128 with indeterminate elements.with indetermination elements.
Despite using the word “undefined” (following Intel’s naming scheme), this non-deterministically
picks some valid value and is not equivalent to mem::MaybeUninit.
In practice, this is typically equivalent to mem::zeroed.
Returns vector of type __m128i with indeterminate elements.with indetermination elements.
Despite using the word “undefined” (following Intel’s naming scheme), this non-deterministically
picks some valid value and is not equivalent to mem::MaybeUninit.
In practice, this is typically equivalent to mem::zeroed.
Adds unsigned 32-bit integers a and b with unsigned 8-bit carry-in c_in
(carry or overflow flag), and store the unsigned 32-bit result in out, and
the carry-out is returned (carry or overflow flag).
Adds unsigned 64-bit integers a and b with unsigned 8-bit carry-in c_in.
(carry or overflow flag), and store the unsigned 64-bit result in out, and
the carry-out is returned (carry or overflow flag).
Convert scalar BF16 (16-bit) floating point element stored at memory locations starting at location
a to single precision (32-bit) floating-point, broadcast it to packed single precision (32-bit) floating-point
elements, and store the results in dst.
Convert scalar half-precision (16-bit) floating-point element stored at memory locations starting
at location a to a single-precision (32-bit) floating-point, broadcast it to packed single-precision
(32-bit) floating-point elements, and store the results in dst.
Cast vector of type __m256d to type __m256h. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m128h to type __m256h. The upper 8 elements of the result are undefined.
In practice, the upper elements are zeroed. This intrinsic can generate the vzeroupper instruction,
but most of the time it does not generate any instructions.
Cast vector of type __m256h to type __m128h. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m256h to type __m256d. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m256h to type __m256. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m256h to type __m256i. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m256 to type __m256h. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m256i to type __m256h. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Compare packed half-precision (16-bit) floating-point elements in a and b based on the comparison
operand specified by imm8, and store the results in mask vector k.
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Compute the complex conjugates of complex numbers in a, and store the results in dst. Each complex number
is composed of two adjacent half-precision (16-bit) floating-point elements, which defines the complex
number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Convert packed signed 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst. The upper 64 bits of dst are zeroed out.
Convert packed unsigned 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst. The upper 64 bits of dst are zeroed out.
Convert packed half-precision (16-bit) floating-point even-indexed elements stored at memory locations starting at
location a to single precision (32-bit) floating-point elements, and store the results in dst.
Convert packed half-precision (16-bit) floating-point odd-indexed elements stored at memory locations starting at
location a to single precision (32-bit) floating-point elements, and store the results in dst.
Convert packed double-precision (64-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst. The upper 64 bits of dst are zeroed out.
Convert packed half-precision (16-bit) floating-point elements in a to packed double-precision (64-bit)
floating-point elements, and store the results in dst.
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst.
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst.
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c, and store the results in dst. Each complex number is composed
of two adjacent half-precision (16-bit) floating-point elements, which defines the complex number
complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst.
Multiply packed complex numbers in a and b, and store the results in dst. Each complex number is
composed of two adjacent half-precision (16-bit) floating-point elements, which defines the complex
number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst.
Test packed half-precision (16-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k.
imm can be a combination of:
Convert the exponent of each packed half-precision (16-bit) floating-point element in a to a half-precision
(16-bit) floating-point number representing the integer exponent, and store the results in dst.
This intrinsic essentially calculates floor(log2(x)) for each element.
Normalize the mantissas of packed half-precision (16-bit) floating-point elements in a, and store
the results in dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends
on the interval range defined by norm and the sign depends on sign and the source sign.
Load 256-bits (composed of 16 packed half-precision (16-bit) floating-point elements) from memory into
a new vector. The address must be aligned to 32 bytes or a general-protection exception may be generated.
Load 256-bits (composed of 16 packed half-precision (16-bit) floating-point elements) from memory into
a new vector. The address does not need to be aligned to any particular boundary.
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c, and store the results in dst using writemask k (the element is
copied from c when the corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst using writemask k (the element is copied from c when the corresponding
mask bit is not set). Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst using writemask k (the element is copied
from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst using writemask k (the element is copied
from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst using writemask k (the element is copied
from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst using writemask k (the element is
copied from c when the corresponding mask bit is not set).
Add packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Compare packed half-precision (16-bit) floating-point elements in a and b based on the comparison
operand specified by imm8, and store the results in mask vector k using zeromask k (elements are
zeroed out when the corresponding mask bit is not set).
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using writemask k (the element is copied from src when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Compute the complex conjugates of complex numbers in a, and store the results in dst using writemask k
(the element is copied from src when corresponding mask bit is not set). Each complex number is composed of two
adjacent half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Convert packed signed 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed signed 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed signed 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set). The upper 64 bits of dst are zeroed out.
Convert packed unsigned 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set). The upper 64 bits of dst are zeroed out.
Convert packed double-precision (64-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to dst
when the corresponding mask bit is not set). The upper 64 bits of dst are zeroed out.
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers, and
store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers, and store the
results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers,
and store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers, and store
the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed double-precision (64-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to
dst when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers with
truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers with
truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to
dst when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to dst
when the corresponding mask bit is not set).
Divide packed half-precision (16-bit) floating-point elements in a by b, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c, and store the results in dst using writemask k (the element is
copied from a when the corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using writemask k (the element is copied from src when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst using writemask k (the element is copied from a when the corresponding mask
bit is not set). Each complex number is composed of two adjacent half-precision (16-bit) floating-point
elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst using writemask k (the element is copied
from a when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from a when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst using writemask k (the element is copied
from a when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from a when the corresponding mask bit is not set).
Multiply packed complex numbers in a and b, and store the results in dst using writemask k (the element
is copied from src when corresponding mask bit is not set). Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst using writemask k (the element is copied
from a when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst using writemask k (the element is
copied from a when the corresponding mask bit is not set).
Test packed half-precision (16-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
imm can be a combination of:
Convert the exponent of each packed half-precision (16-bit) floating-point element in a to a half-precision
(16-bit) floating-point number representing the integer exponent, and store the results in dst using writemask k
(elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates
floor(log2(x)) for each element.
Normalize the mantissas of packed half-precision (16-bit) floating-point elements in a, and store
the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined
by norm and the sign depends on sign and the source sign.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed maximum
values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum value when inputs are
NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed minimum
values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value when inputs are
NaN or signed-zero values.
Multiply packed complex numbers in a and b, and store the results in dst using writemask k (the element
is copied from src when corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the approximate reciprocal of packed 16-bit floating-point elements in a and stores the results in dst
using writemask k (elements are copied from src when the corresponding mask bit is not set).
The maximum relative error for this approximation is less than 1.5*2^-12.
Extract the reduced argument of packed half-precision (16-bit) floating-point elements in a by the
number of bits specified by imm8, and store the results in dst using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Round packed half-precision (16-bit) floating-point elements in a to the number of fraction bits
specified by imm8, and store the results in dst using writemask k (elements are copied from src when
the corresponding mask bit is not set).
Compute the approximate reciprocal square root of packed half-precision (16-bit) floating-point
elements in a, and store the results in dst using writemask k (elements are copied from src when
the corresponding mask bit is not set).
The maximum relative error for this approximation is less than 1.5*2^-12.
Scale the packed half-precision (16-bit) floating-point elements in a using values from b, and store
the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the square root of packed half-precision (16-bit) floating-point elements in a, and store the
results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed half-precision (16-bit) floating-point elements in b from a, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using zeromask k (the element is zeroed out when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Compute the complex conjugates of complex numbers in a, and store the results in dst using zeromask k
(the element is zeroed out when corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Convert packed signed 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
The upper 64 bits of dst are zeroed out.
Convert packed unsigned 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
The upper 64 bits of dst are zeroed out.
Convert packed double-precision (64-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set). The upper 64 bits of dst are zeroed out.
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers, and store the
results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers, and store
the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed double-precision (64-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers with
truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers with
truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
Divide packed half-precision (16-bit) floating-point elements in a by b, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c, and store the results in dst using zeromask k (the element is
zeroed out when the corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using zeromask k (the element is zeroed out when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst using zeromask k (the element is zeroed out when the corresponding mask
bit is not set). Each complex number is composed of two adjacent half-precision (16-bit) floating-point
elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst using zeromask k (the element is zeroed
out when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k
(the element is zeroed out when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst using zeromask k (the element is zeroed
out when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst using zeromask k
(the element is zeroed out when the corresponding mask bit is not set).
Multiply packed complex numbers in a and b, and store the results in dst using zeromask k (the element
is zeroed out when corresponding mask bit is not set). Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst using zeromask k (the element is zeroed
out when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst using zeromask k (the element is
zeroed out when the corresponding mask bit is not set).
Convert the exponent of each packed half-precision (16-bit) floating-point element in a to a half-precision
(16-bit) floating-point number representing the integer exponent, and store the results in dst using zeromask
k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates
floor(log2(x)) for each element.
Normalize the mantissas of packed half-precision (16-bit) floating-point elements in a, and store
the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined
by norm and the sign depends on sign and the source sign.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed maximum
values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum value when inputs are
NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed minimum
values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value when inputs are
NaN or signed-zero values.
Multiply packed complex numbers in a and b, and store the results in dst using zeromask k (the element
is zeroed out when corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the approximate reciprocal of packed 16-bit floating-point elements in a and stores the results in dst
using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
The maximum relative error for this approximation is less than 1.5*2^-12.
Extract the reduced argument of packed half-precision (16-bit) floating-point elements in a by the
number of bits specified by imm8, and store the results in dst using zeromask k (elements are zeroed
out when the corresponding mask bit is not set).
Round packed half-precision (16-bit) floating-point elements in a to the number of fraction bits
specified by imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set).
Compute the approximate reciprocal square root of packed half-precision (16-bit) floating-point
elements in a, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
The maximum relative error for this approximation is less than 1.5*2^-12.
Scale the packed half-precision (16-bit) floating-point elements in a using values from b, and store
the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the square root of packed half-precision (16-bit) floating-point elements in a, and store the
results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed half-precision (16-bit) floating-point elements in b from a, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed maximum
values in dst. Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum
value when inputs are NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed minimum
values in dst. Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value
when inputs are NaN or signed-zero values.
Multiply packed complex numbers in a and b, and store the results in dst. Each complex number is
composed of two adjacent half-precision (16-bit) floating-point elements, which defines the complex
number complex = vec.fp16[0] + i * vec.fp16[1].
Compute the approximate reciprocal of packed 16-bit floating-point elements in a and stores the results in dst.
The maximum relative error for this approximation is less than 1.5*2^-12.
Extract the reduced argument of packed half-precision (16-bit) floating-point elements in a by the
number of bits specified by imm8, and store the results in dst.
Compute the approximate reciprocal square root of packed half-precision (16-bit) floating-point
elements in a, and store the results in dst.
The maximum relative error for this approximation is less than 1.5*2^-12.
Store 256-bits (composed of 16 packed half-precision (16-bit) floating-point elements) from a into memory.
The address must be aligned to 32 bytes or a general-protection exception may be generated.
Store 256-bits (composed of 16 packed half-precision (16-bit) floating-point elements) from a into memory.
The address does not need to be aligned to any particular boundary.
Return vector of type __m256h with indetermination elements.
Despite using the word “undefined” (following Intel’s naming scheme), this non-deterministically
picks some valid value and is not equivalent to mem::MaybeUninit.
In practice, this is typically equivalent to mem::zeroed.
Cast vector of type __m256h to type __m128h. The upper 8 elements of the result are zeroed.
This intrinsic can generate the vzeroupper instruction, but most of the time it does not generate
any instructions.
Add packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst.
Rounding is done according to the rounding parameter, which can be one of:
Cast vector of type __m512d to type __m512h. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m128h to type __m512h. The upper 24 elements of the result are undefined.
In practice, the upper elements are zeroed. This intrinsic can generate the vzeroupper instruction,
but most of the time it does not generate any instructions.
Cast vector of type __m256h to type __m512h. The upper 16 elements of the result are undefined.
In practice, the upper elements are zeroed. This intrinsic can generate the vzeroupper instruction,
but most of the time it does not generate any instructions.
Cast vector of type __m512h to type __m128h. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m512h to type __m256h. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m512h to type __m512d. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m512h to type __m512. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m512h to type __m512i. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m512 to type __m512h. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m512i to type __m512h. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Compare packed half-precision (16-bit) floating-point elements in a and b based on the comparison
operand specified by imm8, and store the results in mask vector k.
Compare packed half-precision (16-bit) floating-point elements in a and b based on the comparison
operand specified by imm8, and store the results in mask vector k.
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Compute the complex conjugates of complex numbers in a, and store the results in dst. Each complex number
is composed of two adjacent half-precision (16-bit) floating-point elements, which defines the complex
number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Convert packed double-precision (64-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst.
Convert packed half-precision (16-bit) floating-point elements in a to packed double-precision (64-bit)
floating-point elements, and store the results in dst.
Convert packed double-precision (64-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst.
Convert packed half-precision (16-bit) floating-point elements in a to packed double-precision (64-bit)
floating-point elements, and store the results in dst.
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst.
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst.
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst.
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst.
Divide packed half-precision (16-bit) floating-point elements in a by b, and store the results in dst.
Rounding is done according to the rounding parameter, which can be one of:
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c, and store the results in dst. Each complex number is composed
of two adjacent half-precision (16-bit) floating-point elements, which defines the complex number
complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c, and store the results in dst. Each complex number is composed
of two adjacent half-precision (16-bit) floating-point elements, which defines the complex number
complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst.
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst.
Multiply packed complex numbers in a and b, and store the results in dst. Each complex number is composed
of two adjacent half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed complex numbers in a and b, and store the results in dst. Each complex number is composed
of two adjacent half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Rounding is done according to the rounding parameter, which can be one of:
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst.
Test packed half-precision (16-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k.
imm can be a combination of:
Convert the exponent of each packed half-precision (16-bit) floating-point element in a to a half-precision
(16-bit) floating-point number representing the integer exponent, and store the results in dst.
This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of each packed half-precision (16-bit) floating-point element in a to a half-precision
(16-bit) floating-point number representing the integer exponent, and store the results in dst.
This intrinsic essentially calculates floor(log2(x)) for each element. Exceptions can be suppressed
by passing _MM_FROUND_NO_EXC in the sae parameter
Normalize the mantissas of packed half-precision (16-bit) floating-point elements in a, and store
the results in dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends
on the interval range defined by norm and the sign depends on sign and the source sign.
Normalize the mantissas of packed half-precision (16-bit) floating-point elements in a, and store
the results in dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends
on the interval range defined by norm and the sign depends on sign and the source sign. Exceptions can
be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter
Load 512-bits (composed of 32 packed half-precision (16-bit) floating-point elements) from memory into
a new vector. The address must be aligned to 64 bytes or a general-protection exception may be generated.
Load 512-bits (composed of 32 packed half-precision (16-bit) floating-point elements) from memory into
a new vector. The address does not need to be aligned to any particular boundary.
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c, and store the results in dst using writemask k (the element is
copied from c when the corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c using writemask k (the element is copied from c when the corresponding
mask bit is not set), and store the results in dst. Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1, or the complex
conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst using writemask k (the element is copied from c when the corresponding
mask bit is not set). Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst using writemask k (the element is copied
from c when the corresponding mask bit is not set).
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst using writemask k (the element is copied from c when the corresponding
mask bit is not set). Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst using writemask k (the element is copied
from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst using writemask k (the element is copied
from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst using writemask k (the element is copied
from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst using writemask k (the element is copied
from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst using writemask k (the element is copied
from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst using writemask k (the element is
copied from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst using writemask k (the element is
copied from c when the corresponding mask bit is not set).
Add packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Add packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the rounding parameter, which can be one of:
Compare packed half-precision (16-bit) floating-point elements in a and b based on the comparison
operand specified by imm8, and store the results in mask vector k using zeromask k (elements are
zeroed out when the corresponding mask bit is not set).
Compare packed half-precision (16-bit) floating-point elements in a and b based on the comparison
operand specified by imm8, and store the results in mask vector k using zeromask k (elements are
zeroed out when the corresponding mask bit is not set).
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using writemask k (the element is copied from src when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using writemask k (the element is copied from src when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Compute the complex conjugates of complex numbers in a, and store the results in dst using writemask k
(the element is copied from src when corresponding mask bit is not set). Each complex number is composed of two
adjacent half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Convert packed signed 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed signed 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed signed 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed unsigned 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to dst
when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers, and
store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers, and store the
results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers,
and store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers, and store
the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed double-precision (64-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to
dst when the corresponding mask bit is not set).
Convert packed signed 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed signed 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed signed 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed unsigned 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to dst
when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers, and
store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers, and store the
results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers,
and store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers, and store
the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed double-precision (64-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to
dst when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers with
truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers with
truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers with
truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers with
truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to
dst when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to dst
when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to
dst when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to dst
when the corresponding mask bit is not set).
Divide packed half-precision (16-bit) floating-point elements in a by b, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Divide packed half-precision (16-bit) floating-point elements in a by b, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the rounding parameter, which can be one of:
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c, and store the results in dst using writemask k (the element is
copied from a when the corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c, and store the results in dst using writemask k (the element is
copied from a when the corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using writemask k (the element is copied from src when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using writemask k (the element is copied from src when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst using writemask k (the element is copied from a when the corresponding mask
bit is not set). Each complex number is composed of two adjacent half-precision (16-bit) floating-point
elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst using writemask k (the element is copied
from a when the corresponding mask bit is not set).
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst using writemask k (the element is copied from a when the corresponding mask
bit is not set). Each complex number is composed of two adjacent half-precision (16-bit) floating-point
elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst using writemask k (the element is copied
from a when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from a when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from a when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst using writemask k (the element is copied
from a when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst using writemask k (the element is copied
from a when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from a when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from a when the corresponding mask bit is not set).
Multiply packed complex numbers in a and b, and store the results in dst using writemask k (the element
is copied from src when corresponding mask bit is not set). Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed complex numbers in a and b, and store the results in dst using writemask k (the element
is copied from src when corresponding mask bit is not set). Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Rounding is done according to the rounding parameter, which can be one of:
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst using writemask k (the element is copied
from a when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst using writemask k (the element is copied
from a when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst using writemask k (the element is
copied from a when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst using writemask k (the element is
copied from a when the corresponding mask bit is not set).
Test packed half-precision (16-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
imm can be a combination of:
Convert the exponent of each packed half-precision (16-bit) floating-point element in a to a half-precision
(16-bit) floating-point number representing the integer exponent, and store the results in dst using writemask k
(elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates
floor(log2(x)) for each element.
Convert the exponent of each packed half-precision (16-bit) floating-point element in a to a half-precision
(16-bit) floating-point number representing the integer exponent, and store the results in dst using writemask k
(elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates
floor(log2(x)) for each element. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter
Normalize the mantissas of packed half-precision (16-bit) floating-point elements in a, and store
the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined
by norm and the sign depends on sign and the source sign.
Normalize the mantissas of packed half-precision (16-bit) floating-point elements in a, and store
the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined
by norm and the sign depends on sign and the source sign. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC
in the sae parameter
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed maximum
values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum value when inputs are
NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed maximum
values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter. Does not follow the
IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum value when inputs are NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed minimum
values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value when inputs are
NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed minimum
values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter. Does not follow the
IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value when inputs are NaN or signed-zero values.
Multiply packed complex numbers in a and b, and store the results in dst using writemask k (the element
is copied from src when corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the packed complex numbers in a and b, and store the results in dst using writemask k (the element
is copied from src when corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the rounding parameter, which can be one of:
Compute the approximate reciprocal of packed 16-bit floating-point elements in a and stores the results in dst
using writemask k (elements are copied from src when the corresponding mask bit is not set).
The maximum relative error for this approximation is less than 1.5*2^-12.
Extract the reduced argument of packed half-precision (16-bit) floating-point elements in a by the
number of bits specified by imm8, and store the results in dst using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Extract the reduced argument of packed half-precision (16-bit) floating-point elements in a by the
number of bits specified by imm8, and store the results in dst using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Round packed half-precision (16-bit) floating-point elements in a to the number of fraction bits
specified by imm8, and store the results in dst using writemask k (elements are copied from src when
the corresponding mask bit is not set).
Round packed half-precision (16-bit) floating-point elements in a to the number of fraction bits
specified by imm8, and store the results in dst using writemask k (elements are copied from src when
the corresponding mask bit is not set). Exceptions can be suppressed by passing _MM_FROUND_NO_EXC
in the sae parameter
Compute the approximate reciprocal square root of packed half-precision (16-bit) floating-point
elements in a, and store the results in dst using writemask k (elements are copied from src when
the corresponding mask bit is not set).
The maximum relative error for this approximation is less than 1.5*2^-12.
Scale the packed half-precision (16-bit) floating-point elements in a using values from b, and store
the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Scale the packed half-precision (16-bit) floating-point elements in a using values from b, and store
the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the square root of packed half-precision (16-bit) floating-point elements in a, and store the
results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the square root of packed half-precision (16-bit) floating-point elements in a, and store the
results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the rounding parameter, which can be one of:
Subtract packed half-precision (16-bit) floating-point elements in b from a, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract packed half-precision (16-bit) floating-point elements in b from a, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Rounding is done according to the rounding parameter, which can be one of:
Add packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the rounding parameter, which can be one of:
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using zeromask k (the element is zeroed out when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using zeromask k (the element is zeroed out when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Compute the complex conjugates of complex numbers in a, and store the results in dst using zeromask k
(the element is zeroed out when corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Convert packed signed 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers, and store the
results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers, and store
the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed double-precision (64-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
Convert packed signed 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed double-precision (64-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers, and store the
results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers, and store
the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed double-precision (64-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers with
truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers with
truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers with
truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers with
truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
Divide packed half-precision (16-bit) floating-point elements in a by b, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Divide packed half-precision (16-bit) floating-point elements in a by b, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the rounding parameter, which can be one of:
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c, and store the results in dst using zeromask k (the element is
zeroed out when the corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c using zeromask k (the element is zeroed out when the corresponding
mask bit is not set), and store the results in dst. Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1, or the complex
conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using zeromask k (the element is zeroed out when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using zeromask k (the element is zeroed out when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst using zeromask k (the element is zeroed out when the corresponding mask
bit is not set). Each complex number is composed of two adjacent half-precision (16-bit) floating-point
elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst using zeromask k (the element is zeroed
out when the corresponding mask bit is not set).
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst using zeromask k (the element is zeroed out when the corresponding mask
bit is not set). Each complex number is composed of two adjacent half-precision (16-bit) floating-point
elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst using zeromask k (the element is zeroed
out when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k
(the element is zeroed out when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k
(the element is zeroed out when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst using zeromask k (the element is zeroed
out when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst using zeromask k (the element is zeroed
out when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst using zeromask k
(the element is zeroed out when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst using zeromask k
(the element is zeroed out when the corresponding mask bit is not set).
Multiply packed complex numbers in a and b, and store the results in dst using zeromask k (the element
is zeroed out when corresponding mask bit is not set). Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed complex numbers in a and b, and store the results in dst using zeromask k (the element
is zeroed out when corresponding mask bit is not set). Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Rounding is done according to the rounding parameter, which can be one of:
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst using zeromask k (the element is zeroed
out when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst using zeromask k (the element is zeroed
out when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst using zeromask k (the element is
zeroed out when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst using zeromask k (the element is
zeroed out when the corresponding mask bit is not set).
Convert the exponent of each packed half-precision (16-bit) floating-point element in a to a half-precision
(16-bit) floating-point number representing the integer exponent, and store the results in dst using zeromask
k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates
floor(log2(x)) for each element.
Convert the exponent of each packed half-precision (16-bit) floating-point element in a to a half-precision
(16-bit) floating-point number representing the integer exponent, and store the results in dst using zeromask
k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates
floor(log2(x)) for each element. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter
Normalize the mantissas of packed half-precision (16-bit) floating-point elements in a, and store
the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined
by norm and the sign depends on sign and the source sign.
Normalize the mantissas of packed half-precision (16-bit) floating-point elements in a, and store
the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined
by norm and the sign depends on sign and the source sign. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC
in the sae parameter
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed maximum
values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum value when inputs are
NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed maximum
values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter. Does not follow the
IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum value when inputs are NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed minimum
values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value when inputs are
NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed minimum
values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter. Does not follow the
IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value when inputs are NaN or signed-zero values.
Multiply packed complex numbers in a and b, and store the results in dst using zeromask k (the element
is zeroed out when corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the packed complex numbers in a and b, and store the results in dst using zeromask k (the element
is zeroed out when corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the rounding parameter, which can be one of:
Compute the approximate reciprocal of packed 16-bit floating-point elements in a and stores the results in dst
using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
The maximum relative error for this approximation is less than 1.5*2^-12.
Extract the reduced argument of packed half-precision (16-bit) floating-point elements in a by the
number of bits specified by imm8, and store the results in dst using zeromask k (elements are zeroed
out when the corresponding mask bit is not set).
Extract the reduced argument of packed half-precision (16-bit) floating-point elements in a by the
number of bits specified by imm8, and store the results in dst using zeromask k (elements are zeroed
out when the corresponding mask bit is not set).
Round packed half-precision (16-bit) floating-point elements in a to the number of fraction bits
specified by imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set).
Round packed half-precision (16-bit) floating-point elements in a to the number of fraction bits
specified by imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set). Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter
Compute the approximate reciprocal square root of packed half-precision (16-bit) floating-point
elements in a, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
The maximum relative error for this approximation is less than 1.5*2^-12.
Scale the packed half-precision (16-bit) floating-point elements in a using values from b, and store
the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Scale the packed half-precision (16-bit) floating-point elements in a using values from b, and store
the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the square root of packed half-precision (16-bit) floating-point elements in a, and store the
results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the square root of packed half-precision (16-bit) floating-point elements in a, and store the
results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the rounding parameter, which can be one of:
Subtract packed half-precision (16-bit) floating-point elements in b from a, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract packed half-precision (16-bit) floating-point elements in b from a, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Rounding is done according to the rounding parameter, which can be one of:
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed maximum
values in dst. Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum
value when inputs are NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed maximum
values in dst. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum value when inputs are
NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed minimum
values in dst. Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value
when inputs are NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed minimum
values in dst. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter. Does not
follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value when inputs are NaN or signed-zero values.
Multiply packed complex numbers in a and b, and store the results in dst. Each complex number is
composed of two adjacent half-precision (16-bit) floating-point elements, which defines the complex
number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the packed complex numbers in a and b, and store the results in dst. Each complex number is
composed of two adjacent half-precision (16-bit) floating-point elements, which defines the complex
number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst.
Rounding is done according to the rounding parameter, which can be one of:
Compute the approximate reciprocal of packed 16-bit floating-point elements in a and stores the results in dst.
The maximum relative error for this approximation is less than 1.5*2^-12.
Extract the reduced argument of packed half-precision (16-bit) floating-point elements in a by the
number of bits specified by imm8, and store the results in dst.
Extract the reduced argument of packed half-precision (16-bit) floating-point elements in a by the
number of bits specified by imm8, and store the results in dst.
Round packed half-precision (16-bit) floating-point elements in a to the number of fraction bits
specified by imm8, and store the results in dst. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC
in the sae parameter
Compute the approximate reciprocal square root of packed half-precision (16-bit) floating-point
elements in a, and store the results in dst.
The maximum relative error for this approximation is less than 1.5*2^-12.
Compute the square root of packed half-precision (16-bit) floating-point elements in a, and store the
results in dst.
Rounding is done according to the rounding parameter, which can be one of:
Store 512-bits (composed of 32 packed half-precision (16-bit) floating-point elements) from a into memory.
The address must be aligned to 64 bytes or a general-protection exception may be generated.
Store 512-bits (composed of 32 packed half-precision (16-bit) floating-point elements) from a into memory.
The address does not need to be aligned to any particular boundary.
Subtract packed half-precision (16-bit) floating-point elements in b from a, and store the results in dst.
Rounding is done according to the rounding parameter, which can be one of:
Return vector of type __m512h with indetermination elements.
Despite using the word “undefined” (following Intel’s naming scheme), this non-deterministically
picks some valid value and is not equivalent to mem::MaybeUninit.
In practice, this is typically equivalent to mem::zeroed.
Cast vector of type __m128h to type __m512h. The upper 24 elements of the result are zeroed.
This intrinsic can generate the vzeroupper instruction, but most of the time it does not generate
any instructions.
Cast vector of type __m256h to type __m512h. The upper 16 elements of the result are zeroed.
This intrinsic can generate the vzeroupper instruction, but most of the time it does not generate
any instructions.
Add the lower half-precision (16-bit) floating-point elements in a and b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst.
Rounding is done according to the rounding parameter, which can be one of:
Add the lower half-precision (16-bit) floating-point elements in a and b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst.
Convert scalar BF16 (16-bit) floating point element stored at memory locations starting at location
a to single precision (32-bit) floating-point, broadcast it to packed single precision (32-bit)
floating-point elements, and store the results in dst.
Convert scalar half-precision (16-bit) floating-point element stored at memory locations starting
at location a to a single-precision (32-bit) floating-point, broadcast it to packed single-precision
(32-bit) floating-point elements, and store the results in dst.
Cast vector of type __m128d to type __m128h. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m128h to type __m128d. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m128h to type __m128. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m128h to type __m128i. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m128 to type __m128h. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Cast vector of type __m128i to type __m128h. This intrinsic is only used for compilation and
does not generate any instructions, thus it has zero latency.
Compare packed half-precision (16-bit) floating-point elements in a and b based on the comparison
operand specified by imm8, and store the results in mask vector k.
Compare the lower half-precision (16-bit) floating-point elements in a and b based on the comparison
operand specified by imm8, and store the result in mask vector k. Exceptions can be suppressed by
passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower half-precision (16-bit) floating-point elements in a and b based on the comparison
operand specified by imm8, and store the result in mask vector k.
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex numbers in a by the complex conjugates of the lower complex numbers in b,
and store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
Multiply the lower complex numbers in a by the complex conjugates of the lower complex numbers in b,
and store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
Compare the lower half-precision (16-bit) floating-point elements in a and b based on the comparison
operand specified by imm8, and return the boolean result (0 or 1).
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower half-precision (16-bit) floating-point elements in a and b based on the comparison
operand specified by imm8, and return the boolean result (0 or 1).
Compute the complex conjugates of complex numbers in a, and store the results in dst. Each complex
number is composed of two adjacent half-precision (16-bit) floating-point elements, which defines
the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate
conjugate = vec.fp16[0] - i * vec.fp16[1].
Convert the signed 32-bit integer b to a half-precision (16-bit) floating-point element, store the
result in the lower element of dst, and copy the upper 7 packed elements from a to the upper elements
of dst.
Convert the signed 64-bit integer b to a half-precision (16-bit) floating-point element, store the
result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements
of dst.
Convert the lower double-precision (64-bit) floating-point element in b to a half-precision (16-bit)
floating-point elements, store the result in the lower element of dst, and copy the upper 7 packed
elements from a to the upper elements of dst.
Convert the lower half-precision (16-bit) floating-point element in b to a double-precision (64-bit)
floating-point element, store the result in the lower element of dst, and copy the upper element from a
to the upper element of dst.
Convert the lower half-precision (16-bit) floating-point element in b to a single-precision (32-bit)
floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements
from a to the upper elements of dst.
Convert the lower single-precision (32-bit) floating-point element in b to a half-precision (16-bit)
floating-point elements, store the result in the lower element of dst, and copy the upper 7 packed
elements from a to the upper elements of dst.
Convert the unsigned 32-bit integer b to a half-precision (16-bit) floating-point element, store the
result in the lower element of dst, and copy the upper 7 packed elements from a to the upper elements
of dst.
Convert the unsigned 64-bit integer b to a half-precision (16-bit) floating-point element, store the
result in the lower element of dst, and copy the upper 1 packed elements from a to the upper elements
of dst.
Convert packed signed 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst. The upper 64 bits of dst are zeroed out.
Convert packed signed 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst. The upper 96 bits of dst are zeroed out.
Convert packed unsigned 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst. The upper 64 bits of dst are zeroed out.
Convert packed unsigned 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst. The upper 96 bits of dst are zeroed out.
Convert the signed 32-bit integer b to a half-precision (16-bit) floating-point element, store the
result in the lower element of dst, and copy the upper 7 packed elements from a to the upper elements
of dst.
Convert the signed 64-bit integer b to a half-precision (16-bit) floating-point element, store the
result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements
of dst.
Convert packed half-precision (16-bit) floating-point even-indexed elements stored at memory locations starting at
location a to single precision (32-bit) floating-point elements, and store the results in dst.
Convert packed half-precision (16-bit) floating-point odd-indexed elements stored at memory locations starting at
location a to single precision (32-bit) floating-point elements, and store the results in dst.
Convert packed double-precision (64-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst. The upper 96 bits of dst are zeroed out.
Convert packed half-precision (16-bit) floating-point elements in a to packed double-precision (64-bit)
floating-point elements, and store the results in dst.
Convert the lower double-precision (64-bit) floating-point element in b to a half-precision (16-bit)
floating-point elements, store the result in the lower element of dst, and copy the upper 7 packed
elements from a to the upper elements of dst.
Convert the lower half-precision (16-bit) floating-point element in b to a double-precision (64-bit)
floating-point element, store the result in the lower element of dst, and copy the upper element
from a to the upper element of dst.
Convert the lower half-precision (16-bit) floating-point element in b to a single-precision (32-bit)
floating-point element, store the result in the lower element of dst, and copy the upper 3 packed
elements from a to the upper elements of dst.
Convert the lower single-precision (32-bit) floating-point element in b to a half-precision (16-bit)
floating-point elements, store the result in the lower element of dst, and copy the upper 7 packed
elements from a to the upper elements of dst.
Convert the unsigned 32-bit integer b to a half-precision (16-bit) floating-point element, store the
result in the lower element of dst, and copy the upper 7 packed elements from a to the upper elements
of dst.
Convert the unsigned 64-bit integer b to a half-precision (16-bit) floating-point element, store the
result in the lower element of dst, and copy the upper 1 packed elements from a to the upper elements
of dst.
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst.
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst.
Divide the lower half-precision (16-bit) floating-point elements in a by b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst.
Rounding is done according to the rounding parameter, which can be one of:
Divide the lower half-precision (16-bit) floating-point elements in a by b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst.
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c, and store the results in dst. Each complex number is composed
of two adjacent half-precision (16-bit) floating-point elements, which defines the complex number
complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex number in a by the complex conjugate of the lower complex number in b,
accumulate to the lower complex number in c, and store the result in the lower elements of dst,
and copy the upper 6 packed elements from a to the upper elements of dst. Each complex number is
composed of two adjacent half-precision (16-bit) floating-point elements, which defines the complex
number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex number in a by the complex conjugate of the lower complex number in b,
accumulate to the lower complex number in c, and store the result in the lower elements of dst,
and copy the upper 6 packed elements from a to the upper elements of dst. Each complex number is
composed of two adjacent half-precision (16-bit) floating-point elements, which defines the complex
number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex numbers in a by the complex conjugates of the lower complex numbers in b,
and store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
Multiply the lower complex numbers in a by the complex conjugates of the lower complex numbers in b,
and store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst.
Multiply the lower complex numbers in a and b, accumulate to the lower complex number in c, and
store the result in the lower elements of dst. Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and add the intermediate
result to the lower element in c. Store the result in the lower element of dst, and copy the upper
7 packed elements from a to the upper elements of dst.
Multiply the lower complex numbers in a and b, accumulate to the lower complex number in c, and
store the result in the lower elements of dst, and copy the upper 6 packed elements from a to the
upper elements of dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and add the intermediate
result to the lower element in c. Store the result in the lower element of dst, and copy the upper
7 packed elements from a to the upper elements of dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract packed elements
in c from the intermediate result. Store the result in the lower element of dst, and copy the upper
7 packed elements from a to the upper elements of dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract packed elements
in c from the intermediate result. Store the result in the lower element of dst, and copy the upper
7 packed elements from a to the upper elements of dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst.
Multiply packed complex numbers in a and b, and store the results in dst. Each complex number is
composed of two adjacent half-precision (16-bit) floating-point elements, which defines the complex
number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower complex numbers in a and b, and store the results in dst. Each complex number is composed
of two adjacent half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower complex numbers in a and b, and store the results in dst. Each complex number is
composed of two adjacent half-precision (16-bit) floating-point elements, which defines the complex
number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst, and copy the upper 7 packed
elements from a to the upper elements of dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst, and copy the upper 7 packed
elements from a to the upper elements of dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst, and copy the upper 7 packed
elements from a to the upper elements of dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst, and copy the upper 7 packed
elements from a to the upper elements of dst.
Test packed half-precision (16-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k.
imm can be a combination of:
Test the lower half-precision (16-bit) floating-point element in a for special categories specified
by imm8, and store the result in mask vector k.
imm can be a combination of:
Convert the exponent of each packed half-precision (16-bit) floating-point element in a to a half-precision
(16-bit) floating-point number representing the integer exponent, and store the results in dst.
This intrinsic essentially calculates floor(log2(x)) for each element.
Convert the exponent of the lower half-precision (16-bit) floating-point element in b to a half-precision
(16-bit) floating-point number representing the integer exponent, store the result in the lower element
of dst, and copy the upper 7 packed elements from a to the upper elements of dst. This intrinsic essentially
calculates floor(log2(x)) for the lower element. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC
in the sae parameter
Convert the exponent of the lower half-precision (16-bit) floating-point element in b to a half-precision
(16-bit) floating-point number representing the integer exponent, store the result in the lower element
of dst, and copy the upper 7 packed elements from a to the upper elements of dst. This intrinsic essentially
calculates floor(log2(x)) for the lower element.
Normalize the mantissas of packed half-precision (16-bit) floating-point elements in a, and store
the results in dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends
on the interval range defined by norm and the sign depends on sign and the source sign.
Normalize the mantissas of the lower half-precision (16-bit) floating-point element in b, store
the result in the lower element of dst, and copy the upper 7 packed elements from a to the upper
elements of dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends
on the interval range defined by norm and the sign depends on sign and the source sign. Exceptions can
be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter
Normalize the mantissas of the lower half-precision (16-bit) floating-point element in b, store
the result in the lower element of dst, and copy the upper 7 packed elements from a to the upper
elements of dst. This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends
on the interval range defined by norm and the sign depends on sign and the source sign.
Load 128-bits (composed of 8 packed half-precision (16-bit) floating-point elements) from memory into
a new vector. The address must be aligned to 16 bytes or a general-protection exception may be generated.
Load 128-bits (composed of 8 packed half-precision (16-bit) floating-point elements) from memory into
a new vector. The address does not need to be aligned to any particular boundary.
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c, and store the results in dst using writemask k (the element is
copied from c when the corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex number in a by the complex conjugate of the lower complex number in b,
accumulate to the lower complex number in c, and store the result in the lower elements of dst using
writemask k (the element is copied from c when the corresponding mask bit is not set), and copy the upper
6 packed elements from a to the upper elements of dst. Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex number in a by the complex conjugate of the lower complex number in b,
accumulate to the lower complex number in c, and store the result in the lower elements of dst using
writemask k (the element is copied from c when the corresponding mask bit is not set), and copy the upper
6 packed elements from a to the upper elements of dst. Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst using writemask k (the element is copied from c when the corresponding
mask bit is not set). Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst using writemask k (the element is copied
from c when the corresponding mask bit is not set).
Multiply the lower complex numbers in a and b, accumulate to the lower complex number in c, and
store the result in the lower elements of dst using writemask k (elements are copied from c when
mask bit 0 is not set), and copy the upper 6 packed elements from a to the upper elements of dst.
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements,
which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and add the intermediate
result to the lower element in c. Store the result in the lower element of dst using writemask k (the element
is copied from c when the mask bit 0 is not set), and copy the upper 7 packed elements from c to the
upper elements of dst.
Multiply the lower complex numbers in a and b, accumulate to the lower complex number in c, and
store the result in the lower elements of dst using writemask k (elements are copied from c when
mask bit 0 is not set), and copy the upper 6 packed elements from a to the upper elements of dst.
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements,
which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and add the intermediate
result to the lower element in c. Store the result in the lower element of dst using writemask k (the element
is copied from c when the mask bit 0 is not set), and copy the upper 7 packed elements from c to the
upper elements of dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst using writemask k (the element is copied
from c when the corresponding mask bit is not set).
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract packed elements
in c from the intermediate result. Store the result in the lower element of dst using writemask k (the element
is copied from c when the mask bit 0 is not set), and copy the upper 7 packed elements from c to the
upper elements of dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract packed elements
in c from the intermediate result. Store the result in the lower element of dst using writemask k (the element
is copied from c when the mask bit 0 is not set), and copy the upper 7 packed elements from c to the
upper elements of dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from c when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst using writemask k (the element is copied
from c when the corresponding mask bit is not set).
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst using writemask k (the element
is copied from c when the mask bit 0 is not set), and copy the upper 7 packed elements from c to the upper
elements of dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst using writemask k (the element
is copied from c when the mask bit 0 is not set), and copy the upper 7 packed elements from c to the upper
elements of dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst using writemask k (the element is
copied from c when the corresponding mask bit is not set).
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst using writemask k (the element
is copied from c when the mask bit 0 is not set), and copy the upper 7 packed elements from c to the upper
elements of dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst using writemask k (the element
is copied from c when the mask bit 0 is not set), and copy the upper 7 packed elements from c to the upper
elements of dst.
Add packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Add the lower half-precision (16-bit) floating-point elements in a and b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
writemask k (the element is copied from src when mask bit 0 is not set).
Rounding is done according to the rounding parameter, which can be one of:
Add the lower half-precision (16-bit) floating-point elements in a and b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
writemask k (the element is copied from src when mask bit 0 is not set).
Compare packed half-precision (16-bit) floating-point elements in a and b based on the comparison
operand specified by imm8, and store the results in mask vector k using zeromask k (elements are
zeroed out when the corresponding mask bit is not set).
Compare the lower half-precision (16-bit) floating-point elements in a and b based on the comparison
operand specified by imm8, and store the result in mask vector k using zeromask k1. Exceptions can be
suppressed by passing _MM_FROUND_NO_EXC in the sae parameter.
Compare the lower half-precision (16-bit) floating-point elements in a and b based on the comparison
operand specified by imm8, and store the result in mask vector k using zeromask k1.
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using writemask k (the element is copied from src when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex numbers in a by the complex conjugates of the lower complex numbers in b,
and store the results in dst using writemask k (the element is copied from src when mask bit 0 is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex numbers in a by the complex conjugates of the lower complex numbers in b,
and store the results in dst using writemask k (the element is copied from src when mask bit 0 is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
Compute the complex conjugates of complex numbers in a, and store the results in dst using writemask k
(the element is copied from src when corresponding mask bit is not set). Each complex number is composed of two
adjacent half-precision (16-bit) floating-point elements, which defines the complex number
complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Convert the lower double-precision (64-bit) floating-point element in b to a half-precision (16-bit)
floating-point elements, store the result in the lower element of dst using writemask k (the element
if copied from src when mask bit 0 is not set), and copy the upper 7 packed elements from a to the
upper elements of dst.
Convert the lower half-precision (16-bit) floating-point element in b to a double-precision (64-bit)
floating-point element, store the result in the lower element of dst using writemask k (the element is
copied from src to dst when mask bit 0 is not set), and copy the upper element from a to the upper element
of dst.
Convert the lower half-precision (16-bit) floating-point element in b to a single-precision (32-bit)
floating-point element, store the result in the lower element of dst using writemask k (the element is
copied from src to dst when mask bit 0 is not set), and copy the upper 3 packed elements from a to the
upper elements of dst.
Convert the lower single-precision (32-bit) floating-point element in b to a half-precision (16-bit)
floating-point elements, store the result in the lower element of dst using writemask k (the element
if copied from src when mask bit 0 is not set), and copy the upper 7 packed elements from a to the
upper elements of dst.
Convert packed signed 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed signed 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set). The upper 64 bits of dst are zeroed out.
Convert packed signed 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set). The upper 96 bits of dst are zeroed out.
Convert packed unsigned 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set). The upper 64 bits of dst are zeroed out.
Convert packed unsigned 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using writemask k (elements are copied from src to dst when the corresponding
mask bit is not set). The upper 96 bits of dst are zeroed out.
Convert packed double-precision (64-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to dst
when the corresponding mask bit is not set). The upper 96 bits of dst are zeroed out.
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers, and
store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers, and store the
results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers,
and store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers, and store
the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed double-precision (64-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to
dst when the corresponding mask bit is not set).
Convert the lower double-precision (64-bit) floating-point element in b to a half-precision (16-bit)
floating-point elements, store the result in the lower element of dst using writemask k (the element
if copied from src when mask bit 0 is not set), and copy the upper 7 packed elements from a to the
upper elements of dst.
Convert the lower half-precision (16-bit) floating-point element in b to a double-precision (64-bit)
floating-point element, store the result in the lower element of dst using writemask k (the element is
copied from src to dst when mask bit 0 is not set), and copy the upper element from a to the upper element
of dst.
Convert the lower half-precision (16-bit) floating-point element in b to a single-precision (32-bit)
floating-point element, store the result in the lower element of dst using writemask k (the element is
copied from src to dst when mask bit 0 is not set), and copy the upper 3 packed elements from a to the
upper elements of dst.
Convert the lower single-precision (32-bit) floating-point element in b to a half-precision (16-bit)
floating-point elements, store the result in the lower element of dst using writemask k (the element
if copied from src when mask bit 0 is not set), and copy the upper 7 packed elements from a to the
upper elements of dst.
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers with
truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers with
truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers with truncation, and
store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to
dst when the corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using writemask k (elements are copied from src to dst
when the corresponding mask bit is not set). The upper 64 bits of dst are zeroed out.
Divide packed half-precision (16-bit) floating-point elements in a by b, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Divide the lower half-precision (16-bit) floating-point elements in a by b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
writemask k (the element is copied from src when mask bit 0 is not set).
Rounding is done according to the rounding parameter, which can be one of:
Divide the lower half-precision (16-bit) floating-point elements in a by b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
writemask k (the element is copied from src when mask bit 0 is not set).
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c, and store the results in dst using writemask k (the element is
copied from a when the corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex number in a by the complex conjugate of the lower complex number in b,
accumulate to the lower complex number in c, and store the result in the lower elements of dst using
writemask k (the element is copied from a when the corresponding mask bit is not set), and copy the upper
6 packed elements from a to the upper elements of dst. Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex number in a by the complex conjugate of the lower complex number in b,
accumulate to the lower complex number in c, and store the result in the lower elements of dst using
writemask k (the element is copied from a when the corresponding mask bit is not set), and copy the upper
6 packed elements from a to the upper elements of dst. Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using writemask k (the element is copied from src when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex numbers in a by the complex conjugates of the lower complex numbers in b,
and store the results in dst using writemask k (the element is copied from src when mask bit 0 is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex numbers in a by the complex conjugates of the lower complex numbers in b,
and store the results in dst using writemask k (the element is copied from src when mask bit 0 is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst using writemask k (the element is copied from a when the corresponding
mask bit is not set). Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst using writemask k (the element is copied
from a when the corresponding mask bit is not set).
Multiply the lower complex numbers in a and b, accumulate to the lower complex number in c, and
store the result in the lower elements of dst using writemask k (elements are copied from a when
mask bit 0 is not set), and copy the upper 6 packed elements from a to the upper elements of dst.
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements,
which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and add the intermediate
result to the lower element in c. Store the result in the lower element of dst using writemask k (the element
is copied from a when the mask bit 0 is not set), and copy the upper 7 packed elements from a to the
upper elements of dst.
Multiply the lower complex numbers in a and b, accumulate to the lower complex number in c, and
store the result in the lower elements of dst using writemask k (elements are copied from a when
mask bit 0 is not set), and copy the upper 6 packed elements from a to the upper elements of dst.
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements,
which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and add the intermediate
result to the lower element in c. Store the result in the lower element of dst using writemask k (the element
is copied from a when the mask bit 0 is not set), and copy the upper 7 packed elements from a to the
upper elements of dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from a when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst using writemask k (the element is copied
from a when the corresponding mask bit is not set).
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract packed elements
in c from the intermediate result. Store the result in the lower element of dst using writemask k (the element
is copied from a when the mask bit 0 is not set), and copy the upper 7 packed elements from a to the
upper elements of dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract packed elements
in c from the intermediate result. Store the result in the lower element of dst using writemask k (the element
is copied from a when the mask bit 0 is not set), and copy the upper 7 packed elements from a to the
upper elements of dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst using writemask k
(the element is copied from a when the corresponding mask bit is not set).
Multiply packed complex numbers in a and b, and store the results in dst using writemask k (the element
is copied from src when corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower complex numbers in a and b, and store the results in dst using writemask k (the element
is copied from src when mask bit 0 is not set). Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower complex numbers in a and b, and store the results in dst using writemask k (the element
is copied from src when mask bit 0 is not set). Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst using writemask k (the element is copied
from a when the corresponding mask bit is not set).
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst using writemask k (the element
is copied from a when the mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper
elements of dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst using writemask k (the element
is copied from a when the mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper
elements of dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst using writemask k (the element is
copied from a when the corresponding mask bit is not set).
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst using writemask k (the element
is copied from a when the mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper
elements of dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst using writemask k (the element
is copied from a when the mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper
elements of dst.
Test packed half-precision (16-bit) floating-point elements in a for special categories specified
by imm8, and store the results in mask vector k using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
imm can be a combination of:
Test the lower half-precision (16-bit) floating-point element in a for special categories specified
by imm8, and store the result in mask vector k using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
imm can be a combination of:
Convert the exponent of each packed half-precision (16-bit) floating-point element in a to a half-precision
(16-bit) floating-point number representing the integer exponent, and store the results in dst using writemask k
(elements are copied from src when the corresponding mask bit is not set). This intrinsic essentially calculates
floor(log2(x)) for each element.
Convert the exponent of the lower half-precision (16-bit) floating-point element in b to a half-precision
(16-bit) floating-point number representing the integer exponent, store the result in the lower element
of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 7
packed elements from a to the upper elements of dst. This intrinsic essentially calculates floor(log2(x))
for the lower element. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter
Convert the exponent of the lower half-precision (16-bit) floating-point element in b to a half-precision
(16-bit) floating-point number representing the integer exponent, store the result in the lower element
of dst using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 7
packed elements from a to the upper elements of dst. This intrinsic essentially calculates floor(log2(x))
for the lower element.
Normalize the mantissas of packed half-precision (16-bit) floating-point elements in a, and store
the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined
by norm and the sign depends on sign and the source sign.
Normalize the mantissas of the lower half-precision (16-bit) floating-point element in b, store
the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set),
and copy the upper 7 packed elements from a to the upper elements of dst. This intrinsic essentially calculates
±(2^k)*|x.significand|, where k depends on the interval range defined by norm and the sign depends on sign and
the source sign. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter
Normalize the mantissas of the lower half-precision (16-bit) floating-point element in b, store
the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set),
and copy the upper 7 packed elements from a to the upper elements of dst. This intrinsic essentially calculates
±(2^k)*|x.significand|, where k depends on the interval range defined by norm and the sign depends on sign and
the source sign.
Load a half-precision (16-bit) floating-point element from memory into the lower element of a new vector
using writemask k (the element is copied from src when mask bit 0 is not set), and zero the upper elements.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed maximum
values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum value when inputs are
NaN or signed-zero values.
Compare the lower half-precision (16-bit) floating-point elements in a and b, store the maximum value
in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set),
and copy the upper 7 packed elements from a to the upper elements of dst. Exceptions can be suppressed by
passing _MM_FROUND_NO_EXC in the sae parameter. Does not follow the IEEE Standard for Floating-Point Arithmetic
(IEEE 754) maximum value when inputs are NaN or signed-zero values.
Compare the lower half-precision (16-bit) floating-point elements in a and b, store the maximum
value in the lower element of dst using writemask k (the element is copied from src when mask bit 0
is not set), and copy the upper 7 packed elements from a to the upper elements of dst. Does not follow
the IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum value when inputs are NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed minimum
values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value when inputs are
NaN or signed-zero values.
Compare the lower half-precision (16-bit) floating-point elements in a and b, store the minimum value
in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set),
and copy the upper 7 packed elements from a to the upper elements of dst. Exceptions can be suppressed by
passing _MM_FROUND_NO_EXC in the sae parameter. Does not follow the IEEE Standard for Floating-Point Arithmetic
(IEEE 754) minimum value when inputs are NaN or signed-zero values.
Compare the lower half-precision (16-bit) floating-point elements in a and b, store the minimum
value in the lower element of dst using writemask k (the element is copied from src when mask bit 0
is not set), and copy the upper 7 packed elements from a to the upper elements of dst. Does not follow
the IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value when inputs are NaN or signed-zero values.
Move the lower half-precision (16-bit) floating-point element from b to the lower element of dst
using writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper
7 packed elements from a to the upper elements of dst.
Multiply packed complex numbers in a and b, and store the results in dst using writemask k (the element
is copied from src when corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Multiply the lower complex numbers in a and b, and store the result in the lower elements of dst using
writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 6 packed
elements from a to the upper elements of dst. Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower half-precision (16-bit) floating-point elements in a and b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
writemask k (the element is copied from src when mask bit 0 is not set).
Rounding is done according to the rounding parameter, which can be one of:
Multiply the lower complex numbers in a and b, and store the result in the lower elements of dst using
writemask k (the element is copied from src when mask bit 0 is not set), and copy the upper 6 packed
elements from a to the upper elements of dst. Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower half-precision (16-bit) floating-point elements in a and b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
writemask k (the element is copied from src when mask bit 0 is not set).
Compute the approximate reciprocal of packed 16-bit floating-point elements in a and stores the results in dst
using writemask k (elements are copied from src when the corresponding mask bit is not set).
The maximum relative error for this approximation is less than 1.5*2^-12.
Compute the approximate reciprocal of the lower half-precision (16-bit) floating-point element in b,
store the result in the lower element of dst using writemask k (the element is copied from src when
mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper elements of dst.
The maximum relative error for this approximation is less than 1.5*2^-12.
Extract the reduced argument of packed half-precision (16-bit) floating-point elements in a by the
number of bits specified by imm8, and store the results in dst using writemask k (elements are copied
from src when the corresponding mask bit is not set).
Extract the reduced argument of the lower half-precision (16-bit) floating-point element in b by
the number of bits specified by imm8, store the result in the lower element of dst using writemask k
(the element is copied from src when mask bit 0 is not set), and copy the upper 7 packed elements from a
to the upper elements of dst.
Extract the reduced argument of the lower half-precision (16-bit) floating-point element in b by
the number of bits specified by imm8, store the result in the lower element of dst using writemask k
(the element is copied from src when mask bit 0 is not set), and copy the upper 7 packed elements from
a to the upper elements of dst.
Round packed half-precision (16-bit) floating-point elements in a to the number of fraction bits
specified by imm8, and store the results in dst using writemask k (elements are copied from src when
the corresponding mask bit is not set).
Round the lower half-precision (16-bit) floating-point element in b to the number of fraction bits
specified by imm8, store the result in the lower element of dst using writemask k (the element is copied
from src when mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper elements of dst.
Round the lower half-precision (16-bit) floating-point element in b to the number of fraction bits
specified by imm8, store the result in the lower element of dst using writemask k (the element is copied
from src when mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper elements of dst.
Compute the approximate reciprocal square root of packed half-precision (16-bit) floating-point
elements in a, and store the results in dst using writemask k (elements are copied from src when
the corresponding mask bit is not set).
The maximum relative error for this approximation is less than 1.5*2^-12.
Compute the approximate reciprocal square root of the lower half-precision (16-bit) floating-point
element in b, store the result in the lower element of dst using writemask k (the element is copied from src
when mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper elements of dst.
The maximum relative error for this approximation is less than 1.5*2^-12.
Scale the packed half-precision (16-bit) floating-point elements in a using values from b, and store
the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, store
the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set),
and copy the upper 7 packed elements from a to the upper elements of dst.
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, store
the result in the lower element of dst using writemask k (the element is copied from src when mask bit 0 is not set),
and copy the upper 7 packed elements from a to the upper elements of dst.
Compute the square root of packed half-precision (16-bit) floating-point elements in a, and store the
results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set).
Compute the square root of the lower half-precision (16-bit) floating-point element in b, store
the result in the lower element of dst using writemask k (the element is copied from src when mask
bit 0 is not set), and copy the upper 7 packed elements from a to the upper elements of dst.
Rounding is done according to the rounding parameter, which can be one of:
Compute the square root of the lower half-precision (16-bit) floating-point element in b, store
the result in the lower element of dst using writemask k (the element is copied from src when mask
bit 0 is not set), and copy the upper 7 packed elements from a to the upper elements of dst.
Subtract packed half-precision (16-bit) floating-point elements in b from a, and store the results in dst using
writemask k (elements are copied from src when the corresponding mask bit is not set).
Subtract the lower half-precision (16-bit) floating-point elements in b from a, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
writemask k (the element is copied from src when mask bit 0 is not set).
Rounding is done according to the rounding parameter, which can be one of:
Subtract the lower half-precision (16-bit) floating-point elements in b from a, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
writemask k (the element is copied from src when mask bit 0 is not set).
Add packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Add the lower half-precision (16-bit) floating-point elements in a and b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
zeromask k (the element is zeroed out when mask bit 0 is not set).
Rounding is done according to the rounding parameter, which can be one of:
Add the lower half-precision (16-bit) floating-point elements in a and b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
zeromask k (the element is zeroed out when mask bit 0 is not set).
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using zeromask k (the element is zeroed out when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex numbers in a by the complex conjugates of the lower complex numbers in b,
and store the results in dst using zeromask k (the element is zeroed out when mask bit 0 is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex numbers in a by the complex conjugates of the lower complex numbers in b,
and store the results in dst using zeromask k (the element is zeroed out when mask bit 0 is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
Compute the complex conjugates of complex numbers in a, and store the results in dst using zeromask k
(the element is zeroed out when corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Convert the lower double-precision (64-bit) floating-point element in b to a half-precision (16-bit)
floating-point elements, store the result in the lower element of dst using zeromask k (the element
is zeroed out when mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper
elements of dst.
Convert the lower half-precision (16-bit) floating-point element in b to a double-precision (64-bit)
floating-point element, store the result in the lower element of dst using zeromask k (the element is
zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Convert the lower half-precision (16-bit) floating-point element in b to a single-precision (32-bit)
floating-point element, store the result in the lower element of dst using zeromask k (the element is
zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements
of dst.
Convert the lower single-precision (32-bit) floating-point element in b to a half-precision (16-bit)
floating-point elements, store the result in the lower element of dst using zeromask k (the element
is zeroed out when mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper
elements of dst.
Convert packed signed 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed signed 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
The upper 64 bits of dst are zeroed out.
Convert packed signed 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
The upper 96 bits of dst are zeroed out.
Convert packed unsigned 16-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed unsigned 32-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
The upper 64 bits of dst are zeroed out.
Convert packed unsigned 64-bit integers in a to packed half-precision (16-bit) floating-point elements,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
The upper 96 bits of dst are zeroed out.
Convert packed double-precision (64-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set). The upper 96 bits of dst are zeroed out.
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers, and store the
results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers,
and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers, and store
the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed double-precision (64-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
Convert the lower double-precision (64-bit) floating-point element in b to a half-precision (16-bit)
floating-point elements, store the result in the lower element of dst using zeromask k (the element
is zeroed out when mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper
elements of dst.
Convert the lower half-precision (16-bit) floating-point element in b to a double-precision (64-bit)
floating-point element, store the result in the lower element of dst using zeromask k (the element is
zeroed out when mask bit 0 is not set), and copy the upper element from a to the upper element of dst.
Convert the lower half-precision (16-bit) floating-point element in b to a single-precision (32-bit)
floating-point element, store the result in the lower element of dst using zeromask k (the element is
zeroed out when mask bit 0 is not set), and copy the upper 3 packed elements from a to the upper elements
of dst.
Convert the lower single-precision (32-bit) floating-point element in b to a half-precision (16-bit)
floating-point elements, store the result in the lower element of dst using zeromask k (the element
is zeroed out when mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper
elements of dst.
Convert packed half-precision (16-bit) floating-point elements in a to packed 16-bit integers with
truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed unsigned 16-bit integers with
truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 32-bit unsigned integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed 64-bit unsigned integers with truncation, and
store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Convert packed half-precision (16-bit) floating-point elements in a to packed single-precision (32-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
Convert packed single-precision (32-bit) floating-point elements in a to packed half-precision (16-bit)
floating-point elements, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set). The upper 64 bits of dst are zeroed out.
Divide packed half-precision (16-bit) floating-point elements in a by b, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Divide the lower half-precision (16-bit) floating-point elements in a by b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
zeromask k (the element is zeroed out when mask bit 0 is not set).
Rounding is done according to the rounding parameter, which can be one of:
Divide the lower half-precision (16-bit) floating-point elements in a by b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
zeromask k (the element is zeroed out when mask bit 0 is not set).
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, accumulate
to the corresponding complex numbers in c, and store the results in dst using zeromask k (the element is
zeroed out when the corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex number in a by the complex conjugate of the lower complex number in b,
accumulate to the lower complex number in c using zeromask k (the element is zeroed out when the corresponding
mask bit is not set), and store the result in the lower elements of dst, and copy the upper 6 packed elements
from a to the upper elements of dst. Each complex number is composed of two adjacent half-precision (16-bit)
floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1, or the complex
conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex number in a by the complex conjugate of the lower complex number in b,
accumulate to the lower complex number in c, and store the result in the lower elements of dst using
zeromask k (the element is zeroed out when the corresponding mask bit is not set), and copy the upper
6 packed elements from a to the upper elements of dst. Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1],
or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a by the complex conjugates of packed complex numbers in b, and
store the results in dst using zeromask k (the element is zeroed out when corresponding mask bit is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex numbers in a by the complex conjugates of the lower complex numbers in b,
and store the results in dst using zeromask k (the element is zeroed out when mask bit 0 is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply the lower complex numbers in a by the complex conjugates of the lower complex numbers in b,
and store the results in dst using zeromask k (the element is zeroed out when mask bit 0 is not set).
Each complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1], or the complex conjugate conjugate = vec.fp16[0] - i * vec.fp16[1].
Multiply packed complex numbers in a and b, accumulate to the corresponding complex numbers in c,
and store the results in dst using zeromask k (the element is zeroed out when the corresponding mask
bit is not set). Each complex number is composed of two adjacent half-precision (16-bit) floating-point
elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, add the intermediate
result to packed elements in c, and store the results in dst using zeromask k (the element is zeroed
out when the corresponding mask bit is not set).
Multiply the lower complex numbers in a and b, accumulate to the lower complex number in c, and
store the result in the lower elements of dst using zeromask k (elements are zeroed out when mask
bit 0 is not set), and copy the upper 6 packed elements from a to the upper elements of dst. Each
complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and add the intermediate
result to the lower element in c. Store the result in the lower element of dst using zeromask k (the element
is zeroed out when mask bit 0 is not set), and copy the upper 7 packed elements from a to the
upper elements of dst.
Multiply the lower complex numbers in a and b, accumulate to the lower complex number in c, and
store the result in the lower elements of dst using zeromask k (elements are zeroed out when mask
bit 0 is not set), and copy the upper 6 packed elements from a to the upper elements of dst. Each
complex number is composed of two adjacent half-precision (16-bit) floating-point elements, which
defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and add the intermediate
result to the lower element in c. Store the result in the lower element of dst using zeromask k (the element
is zeroed out when mask bit 0 is not set), and copy the upper 7 packed elements from a to the
upper elements of dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
subtract packed elements in c to/from the intermediate result, and store the results in dst using zeromask k
(the element is zeroed out when the corresponding mask bit is not set).
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the intermediate result, and store the results in dst using zeromask k (the element is zeroed
out when the corresponding mask bit is not set).
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract packed elements
in c from the intermediate result. Store the result in the lower element of dst using zeromask k (the element
is zeroed out when mask bit 0 is not set), and copy the upper 7 packed elements from a to the
upper elements of dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract packed elements
in c from the intermediate result. Store the result in the lower element of dst using zeromask k (the element
is zeroed out when mask bit 0 is not set), and copy the upper 7 packed elements from a to the
upper elements of dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
and add packed elements in c to/from the intermediate result, and store the results in dst using zeromask k
(the element is zeroed out when the corresponding mask bit is not set).
Multiply packed complex numbers in a and b, and store the results in dst using zeromask k (the element
is zeroed out when corresponding mask bit is not set). Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower complex numbers in a and b, and store the results in dst using zeromask k (the element
is zeroed out when mask bit 0 is not set). Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower complex numbers in a and b, and store the results in dst using zeromask k (the element
is zeroed out when mask bit 0 is not set). Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract the intermediate
result from packed elements in c, and store the results in dst using zeromask k (the element is zeroed
out when the corresponding mask bit is not set).
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst using zeromask k (the element
is zeroed out when the mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper
elements of dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst using zeromask k (the element
is zeroed out when the mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper
elements of dst.
Multiply packed half-precision (16-bit) floating-point elements in a and b, subtract packed elements
in c from the negated intermediate result, and store the results in dst using zeromask k (the element is
zeroed out when the corresponding mask bit is not set).
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst using zeromask k (the element
is zeroed out when the mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper
elements of dst.
Multiply the lower half-precision (16-bit) floating-point elements in a and b, and subtract the intermediate
result from the lower element in c. Store the result in the lower element of dst using zeromask k (the element
is zeroed out when the mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper
elements of dst.
Convert the exponent of each packed half-precision (16-bit) floating-point element in a to a half-precision
(16-bit) floating-point number representing the integer exponent, and store the results in dst using zeromask
k (elements are zeroed out when the corresponding mask bit is not set). This intrinsic essentially calculates
floor(log2(x)) for each element.
Convert the exponent of the lower half-precision (16-bit) floating-point element in b to a half-precision
(16-bit) floating-point number representing the integer exponent, store the result in the lower element
of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 7 packed
elements from a to the upper elements of dst. This intrinsic essentially calculates floor(log2(x)) for the
lower element. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter
Convert the exponent of the lower half-precision (16-bit) floating-point element in b to a half-precision
(16-bit) floating-point number representing the integer exponent, store the result in the lower element
of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 7 packed
elements from a to the upper elements of dst. This intrinsic essentially calculates floor(log2(x)) for the
lower element.
Normalize the mantissas of packed half-precision (16-bit) floating-point elements in a, and store
the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
This intrinsic essentially calculates ±(2^k)*|x.significand|, where k depends on the interval range defined
by norm and the sign depends on sign and the source sign.
Normalize the mantissas of the lower half-precision (16-bit) floating-point element in b, store
the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set),
and copy the upper 7 packed elements from a to the upper elements of dst. This intrinsic essentially calculates
±(2^k)*|x.significand|, where k depends on the interval range defined by norm and the sign depends on sign and
the source sign. Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter
Normalize the mantissas of the lower half-precision (16-bit) floating-point element in b, store
the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set),
and copy the upper 7 packed elements from a to the upper elements of dst. This intrinsic essentially calculates
±(2^k)*|x.significand|, where k depends on the interval range defined by norm and the sign depends on sign and
the source sign.
Load a half-precision (16-bit) floating-point element from memory into the lower element of a new vector
using zeromask k (the element is zeroed out when mask bit 0 is not set), and zero the upper elements.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed maximum
values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum value when inputs are
NaN or signed-zero values.
Compare the lower half-precision (16-bit) floating-point elements in a and b, store the maximum value
in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and
copy the upper 7 packed elements from a to the upper elements of dst. Exceptions can be suppressed by
passing _MM_FROUND_NO_EXC in the sae parameter. Does not follow the IEEE Standard for Floating-Point Arithmetic
(IEEE 754) maximum value when inputs are NaN or signed-zero values.
Compare the lower half-precision (16-bit) floating-point elements in a and b, store the maximum value
in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and
copy the upper 7 packed elements from a to the upper elements of dst. Does not follow the IEEE Standard
for Floating-Point Arithmetic (IEEE 754) maximum value when inputs are NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed minimum
values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value when inputs are
NaN or signed-zero values.
Compare the lower half-precision (16-bit) floating-point elements in a and b, store the minimum value
in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and
copy the upper 7 packed elements from a to the upper elements of dst. Exceptions can be suppressed by
passing _MM_FROUND_NO_EXC in the sae parameter. Does not follow the IEEE Standard for Floating-Point Arithmetic
(IEEE 754) minimum value when inputs are NaN or signed-zero values.
Compare the lower half-precision (16-bit) floating-point elements in a and b, store the minimum value
in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set), and
copy the upper 7 packed elements from a to the upper elements of dst. Does not follow the IEEE Standard
for Floating-Point Arithmetic (IEEE 754) minimum value when inputs are NaN or signed-zero values.
Move the lower half-precision (16-bit) floating-point element from b to the lower element of dst
using zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 7 packed
elements from a to the upper elements of dst.
Multiply packed complex numbers in a and b, and store the results in dst using zeromask k (the element
is zeroed out when corresponding mask bit is not set). Each complex number is composed of two adjacent
half-precision (16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply packed half-precision (16-bit) floating-point elements in a and b, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Multiply the lower complex numbers in a and b, and store the result in the lower elements of dst using
zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 6 packed elements
from a to the upper elements of dst. Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower half-precision (16-bit) floating-point elements in a and b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
zeromask k (the element is zeroed out when mask bit 0 is not set).
Rounding is done according to the rounding parameter, which can be one of:
Multiply the lower complex numbers in a and b, and store the result in the lower elements of dst using
zeromask k (the element is zeroed out when mask bit 0 is not set), and copy the upper 6 packed elements
from a to the upper elements of dst. Each complex number is composed of two adjacent half-precision
(16-bit) floating-point elements, which defines the complex number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower half-precision (16-bit) floating-point elements in a and b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
zeromask k (the element is zeroed out when mask bit 0 is not set).
Compute the approximate reciprocal of packed 16-bit floating-point elements in a and stores the results in dst
using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
The maximum relative error for this approximation is less than 1.5*2^-12.
Compute the approximate reciprocal of the lower half-precision (16-bit) floating-point element in b,
store the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0
is not set), and copy the upper 7 packed elements from a to the upper elements of dst.
The maximum relative error for this approximation is less than 1.5*2^-12.
Extract the reduced argument of packed half-precision (16-bit) floating-point elements in a by the
number of bits specified by imm8, and store the results in dst using zeromask k (elements are zeroed
out when the corresponding mask bit is not set).
Extract the reduced argument of the lower half-precision (16-bit) floating-point element in b by
the number of bits specified by imm8, store the result in the lower element of dst using zeromask k
(the element is zeroed out when mask bit 0 is not set), and copy the upper 7 packed elements from a
to the upper elements of dst.
Extract the reduced argument of the lower half-precision (16-bit) floating-point element in b by
the number of bits specified by imm8, store the result in the lower element of dst using zeromask k
(the element is zeroed out when mask bit 0 is not set), and copy the upper 7 packed elements from a
to the upper elements of dst.
Round packed half-precision (16-bit) floating-point elements in a to the number of fraction bits
specified by imm8, and store the results in dst using zeromask k (elements are zeroed out when the corresponding
mask bit is not set).
Round the lower half-precision (16-bit) floating-point element in b to the number of fraction bits
specified by imm8, store the result in the lower element of dst using zeromask k (the element is zeroed
out when mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper elements of dst.
Round the lower half-precision (16-bit) floating-point element in b to the number of fraction bits
specified by imm8, store the result in the lower element of dst using zeromask k (the element is zeroed
out when mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper elements of dst.
Compute the approximate reciprocal square root of packed half-precision (16-bit) floating-point
elements in a, and store the results in dst using zeromask k (elements are zeroed out when the
corresponding mask bit is not set).
The maximum relative error for this approximation is less than 1.5*2^-12.
Compute the approximate reciprocal square root of the lower half-precision (16-bit) floating-point
element in b, store the result in the lower element of dst using zeromask k (the element is zeroed out when
mask bit 0 is not set), and copy the upper 7 packed elements from a to the upper elements of dst.
The maximum relative error for this approximation is less than 1.5*2^-12.
Scale the packed half-precision (16-bit) floating-point elements in a using values from b, and store
the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, store
the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set),
and copy the upper 7 packed elements from a to the upper elements of dst.
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, store
the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0 is not set),
and copy the upper 7 packed elements from a to the upper elements of dst.
Compute the square root of packed half-precision (16-bit) floating-point elements in a, and store the
results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Compute the square root of the lower half-precision (16-bit) floating-point element in b, store
the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0
is not set), and copy the upper 7 packed elements from a to the upper elements of dst.
Rounding is done according to the rounding parameter, which can be one of:
Compute the square root of the lower half-precision (16-bit) floating-point element in b, store
the result in the lower element of dst using zeromask k (the element is zeroed out when mask bit 0
is not set), and copy the upper 7 packed elements from a to the upper elements of dst.
Subtract packed half-precision (16-bit) floating-point elements in b from a, and store the results in dst using
zeromask k (elements are zeroed out when the corresponding mask bit is not set).
Subtract the lower half-precision (16-bit) floating-point elements in b from a, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
zeromask k (the element is zeroed out when mask bit 0 is not set).
Rounding is done according to the rounding parameter, which can be one of:
Subtract the lower half-precision (16-bit) floating-point elements in b from a, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst using
zeromask k (the element is zeroed out when mask bit 0 is not set).
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed maximum
values in dst. Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum
value when inputs are NaN or signed-zero values.
Compare the lower half-precision (16-bit) floating-point elements in a and b, store the maximum value
in the lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter. Does not follow the
IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum value when inputs are NaN or signed-zero values.
Compare the lower half-precision (16-bit) floating-point elements in a and b, store the maximum
value in the lower element of dst, and copy the upper 7 packed elements from a to the upper elements
of dst. Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) maximum value
when inputs are NaN or signed-zero values.
Compare packed half-precision (16-bit) floating-point elements in a and b, and store packed minimum
values in dst. Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value
when inputs are NaN or signed-zero values.
Compare the lower half-precision (16-bit) floating-point elements in a and b, store the minimum value
in the lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst.
Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter. Does not follow the
IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value when inputs are NaN or signed-zero values.
Compare the lower half-precision (16-bit) floating-point elements in a and b, store the minimum
value in the lower element of dst, and copy the upper 7 packed elements from a to the upper elements
of dst. Does not follow the IEEE Standard for Floating-Point Arithmetic (IEEE 754) minimum value when
inputs are NaN or signed-zero values.
Move the lower half-precision (16-bit) floating-point element from b to the lower element of dst,
and copy the upper 7 packed elements from a to the upper elements of dst.
Multiply packed complex numbers in a and b, and store the results in dst. Each complex number is
composed of two adjacent half-precision (16-bit) floating-point elements, which defines the complex
number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower complex numbers in a and b, and store the result in the lower elements of dst,
and copy the upper 6 packed elements from a to the upper elements of dst. Each complex number is
composed of two adjacent half-precision (16-bit) floating-point elements, which defines the complex
number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower half-precision (16-bit) floating-point elements in a and b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst.
Rounding is done according to the rounding parameter, which can be one of:
Multiply the lower complex numbers in a and b, and store the result in the lower elements of dst,
and copy the upper 6 packed elements from a to the upper elements of dst. Each complex number is
composed of two adjacent half-precision (16-bit) floating-point elements, which defines the complex
number complex = vec.fp16[0] + i * vec.fp16[1].
Multiply the lower half-precision (16-bit) floating-point elements in a and b, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst.
Compute the approximate reciprocal of packed 16-bit floating-point elements in a and stores the results in dst.
The maximum relative error for this approximation is less than 1.5*2^-12.
Compute the approximate reciprocal of the lower half-precision (16-bit) floating-point element in b,
store the result in the lower element of dst, and copy the upper 7 packed elements from a to the
upper elements of dst.
The maximum relative error for this approximation is less than 1.5*2^-12.
Extract the reduced argument of packed half-precision (16-bit) floating-point elements in a by the
number of bits specified by imm8, and store the results in dst.
Extract the reduced argument of the lower half-precision (16-bit) floating-point element in b by
the number of bits specified by imm8, store the result in the lower element of dst, and copy the upper
7 packed elements from a to the upper elements of dst.
Extract the reduced argument of the lower half-precision (16-bit) floating-point element in b by
the number of bits specified by imm8, store the result in the lower element of dst, and copy the
upper 7 packed elements from a to the upper elements of dst.
Round the lower half-precision (16-bit) floating-point element in b to the number of fraction bits
specified by imm8, store the result in the lower element of dst, and copy the upper 7 packed elements
from a to the upper elements of dst.
Round the lower half-precision (16-bit) floating-point element in b to the number of fraction bits
specified by imm8, store the result in the lower element of dst, and copy the upper 7 packed elements
from a to the upper elements of dst.
Compute the approximate reciprocal square root of packed half-precision (16-bit) floating-point
elements in a, and store the results in dst.
The maximum relative error for this approximation is less than 1.5*2^-12.
Compute the approximate reciprocal square root of the lower half-precision (16-bit) floating-point
element in b, store the result in the lower element of dst, and copy the upper 7 packed elements from a
to the upper elements of dst.
The maximum relative error for this approximation is less than 1.5*2^-12.
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, store
the result in the lower element of dst, and copy the upper 7 packed elements from a to the upper
elements of dst.
Scale the packed single-precision (32-bit) floating-point elements in a using values from b, store
the result in the lower element of dst, and copy the upper 7 packed elements from a to the upper
elements of dst.
Compute the square root of the lower half-precision (16-bit) floating-point element in b, store
the result in the lower element of dst, and copy the upper 7 packed elements from a to the upper
elements of dst.
Rounding is done according to the rounding parameter, which can be one of:
Compute the square root of the lower half-precision (16-bit) floating-point element in b, store
the result in the lower element of dst, and copy the upper 7 packed elements from a to the upper
elements of dst.
Store 128-bits (composed of 8 packed half-precision (16-bit) floating-point elements) from a into memory.
The address must be aligned to 16 bytes or a general-protection exception may be generated.
Store 128-bits (composed of 8 packed half-precision (16-bit) floating-point elements) from a into memory.
The address does not need to be aligned to any particular boundary.
Subtract the lower half-precision (16-bit) floating-point elements in b from a, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst.
Rounding is done according to the rounding parameter, which can be one of:
Subtract the lower half-precision (16-bit) floating-point elements in b from a, store the result in the
lower element of dst, and copy the upper 7 packed elements from a to the upper elements of dst.
Compare the lower half-precision (16-bit) floating-point elements in a and b for equality, and
return the boolean result (0 or 1). This instruction will not signal an exception for QNaNs.
Compare the lower half-precision (16-bit) floating-point elements in a and b for greater-than-or-equal,
and return the boolean result (0 or 1). This instruction will not signal an exception for QNaNs.
Compare the lower half-precision (16-bit) floating-point elements in a and b for greater-than, and return
the boolean result (0 or 1). This instruction will not signal an exception for QNaNs.
Compare the lower half-precision (16-bit) floating-point elements in a and b for less-than-or-equal, and
return the boolean result (0 or 1). This instruction will not signal an exception for QNaNs.
Compare the lower half-precision (16-bit) floating-point elements in a and b for less-than, and return
the boolean result (0 or 1). This instruction will not signal an exception for QNaNs.
Compare the lower half-precision (16-bit) floating-point elements in a and b for not-equal, and return
the boolean result (0 or 1). This instruction will not signal an exception for QNaNs.
Return vector of type __m128h with indetermination elements.
Despite using the word “undefined” (following Intel’s naming scheme), this non-deterministically
picks some valid value and is not equivalent to mem::MaybeUninit.
In practice, this is typically equivalent to mem::zeroed.
Perform matrix multiplication of two tiles containing complex elements and accumulate the results into a packed single precision tile.
Each dword element in input tiles a and b is interpreted as a complex number with FP16 real part and FP16 imaginary part.
Calculates the imaginary part of the result. For each possible combination of (row of a, column of b),
it performs a set of multiplication and accumulations on all corresponding complex numbers (one from a and one from b).
The imaginary part of the a element is multiplied with the real part of the corresponding b element, and the real part of
the a element is multiplied with the imaginary part of the corresponding b elements. The two accumulated results are added,
and then accumulated into the corresponding row and column of dst.
Perform matrix multiplication of two tiles containing complex elements and accumulate the results into a packed single precision tile.
Each dword element in input tiles a and b is interpreted as a complex number with FP16 real part and FP16 imaginary part.
Calculates the real part of the result. For each possible combination of (row of a, column of b),
it performs a set of multiplication and accumulations on all corresponding complex numbers (one from a and one from b).
The real part of the a element is multiplied with the real part of the corresponding b element, and the negated imaginary part of
the a element is multiplied with the imaginary part of the corresponding b elements.
The two accumulated results are added, and then accumulated into the corresponding row and column of dst.
Compute dot-product of BF16 (16-bit) floating-point pairs in tiles a and b,
accumulating the intermediate single-precision (32-bit) floating-point elements
with elements in dst, and store the 32-bit result back to tile dst.
Compute dot-product of bytes in tiles with a source/destination accumulator.
Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding
signed 8-bit integers in b, producing 4 intermediate 32-bit results.
Sum these 4 results with the corresponding 32-bit integer in dst, and store the 32-bit result back to tile dst.
Compute dot-product of bytes in tiles with a source/destination accumulator.
Multiply groups of 4 adjacent pairs of signed 8-bit integers in a with corresponding
unsigned 8-bit integers in b, producing 4 intermediate 32-bit results.
Sum these 4 results with the corresponding 32-bit integer in dst, and store the 32-bit result back to tile dst.
Compute dot-product of bytes in tiles with a source/destination accumulator.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding
signed 8-bit integers in b, producing 4 intermediate 32-bit results.
Sum these 4 results with the corresponding 32-bit integer in dst, and store the 32-bit result back to tile dst.
Compute dot-product of bytes in tiles with a source/destination accumulator.
Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in a with corresponding
unsigned 8-bit integers in b, producing 4 intermediate 32-bit results.
Sum these 4 results with the corresponding 32-bit integer in dst, and store the 32-bit result back to tile dst.
Compute dot-product of FP16 (16-bit) floating-point pairs in tiles a and b,
accumulating the intermediate single-precision (32-bit) floating-point elements
with elements in dst, and store the 32-bit result back to tile dst.
Load tile configuration from a 64-byte memory location specified by mem_addr.
The tile configuration format is specified below, and includes the tile type pallette,
the number of bytes per row, and the number of rows. If the specified pallette_id is zero,
that signifies the init state for both the tile config and the tile data, and the tiles are zeroed.
Any invalid configurations will result in #GP fault.
Load tile rows from memory specifieid by base address and stride into destination tile dst using the tile configuration previously configured via _tile_loadconfig.
Stores the current tile configuration to a 64-byte memory location specified by mem_addr.
The tile configuration format is specified below, and includes the tile type pallette,
the number of bytes per row, and the number of rows. If tiles are not configured, all zeroes will be stored to memory.
Store the tile specified by src to memory specifieid by base address and stride using the tile configuration previously configured via _tile_loadconfig.
Load tile rows from memory specifieid by base address and stride into destination tile dst using the tile configuration
previously configured via _tile_loadconfig. This intrinsic provides a hint to the implementation that the data will
likely not be reused in the near future and the data caching can be optimized accordingly.
Queries whether the processor is executing in a transactional region identified by restricted
transactional memory (RTM) or hardware lock elision (HLE).