pub unsafe fn _mm256_sub_pd(a: __m256d, b: __m256d) -> __m256d
Subtracts packed double-precision (64-bit) floating-point elements in b from packed elements in a.
b
a
Intel’s documentation