pub fn _mm_cvtsi16_si128(a: i16) -> __m128i
stdarch_x86_avx512_f16
Copy 16-bit integer a to the lower elements of dst, and zero the upper elements of dst.
Intel’s documentation