3 ms·Is CUDA not more analogous to using MKL, rather than AVX?by ip26 2y agoIs CUDA not more analogous to using MKL, rather than AVX?