Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Add AVX512 accelerated 1D/3D LUTS (#1932)
* Initial AVX512 support Signed-off-by: Mark Reid <[email protected]> * Lut1DOp add AVX512 implementation Signed-off-by: Mark Reid <[email protected]> * Lut3DOp add AVX512 implementation Signed-off-by: Mark Reid <[email protected]> * Don't use SIMD if only 1 pixel is requested Signed-off-by: Mark Reid <[email protected]> * Remove #if, f16c is always available with AVX512 Signed-off-by: Mark Reid <[email protected]> * Cast pointers to __m512 instead of __m256 Signed-off-by: Mark Reid <[email protected]> * Use size method from vector being tested Signed-off-by: Mark Reid <[email protected]> * Add to help message that f16c is only used with AVX/AVX2 Signed-off-by: Mark Reid <[email protected]> * Clarify test case by using uint8 maxValue Signed-off-by: Mark Reid <[email protected]> --------- Signed-off-by: Mark Reid <[email protected]>
- Loading branch information