pub fn _mm512_cvttps_epi64(a: __m256) -> __m512i
Convert packed single-precision (32-bit) floating-point elements in a to packed signed 64-bit integers with truncation, and store the result in dst.
Intel’s Documentation