pub unsafe fn _mm256_ceil_ps(a: __m256) -> __m256
Rounds packed single-precision (32-bit) floating point elements in a toward positive infinity.
a
Intel’s documentation