Function core::arch::x86_64::_mm256_blendv_epi8

pub unsafe fn _mm256_blendv_epi8(
    a: __m256i, 
    b: __m256i, 
    mask: __m256i
) -> __m256i

This is supported on x86-64 and target feature avx2 only.

Blends packed 8-bit integers from a and b using mask.

Intel’s documentation