Function _mm256_mask_compress_ps 

pub fn _mm256_mask_compress_ps(src: __m256, k: __mmask8, a: __m256) -> __m256

Available on x86-64 and target feature avx512f and target feature avx512vl only.

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.

Intel’s documentation