Function _mm512_cmp_ph_mask 

pub fn _mm512_cmp_ph_mask(a: __m512h, b: __m512h, const IMM5: i32) -> __mmask32

Available on x86 and target feature avx512fp16 only.

Compare packed half-precision (16-bit) floating-point elements in a and b based on the comparison operand specified by imm8, and store the results in mask vector k.

Intel’s documentation