pub fn _kadd_mask8(a: __mmask8, b: __mmask8) -> __mmask8
avx512dq
Add 8-bit masks a and b, and store the result in dst.
Intel’s Documentation