pub unsafe fn _mm_cvtepi32_ps(a: __m128i) -> __m128
Converts packed 32-bit integers in a to packed single-precision (32-bit) floating-point elements.
a
Intel’s documentation