Function _mm512_mask_rcp14_ps
#[target_feature(enable = "avx512f")]
pub fn _mm512_mask_rcp14_ps(src: __m512, k: __mmask16, a: __m512) -> __m512
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.