pub unsafe fn _mm256_blend_ps<const IMM8: i32>(a: __m256, b: __m256) -> __m256
Blends packed single-precision (32-bit) floating-point elements from a and b using control mask imm8.
a
b
imm8
Intel’s documentation