|
Highly Efficient FFT for Exascale: HeFFTe v2.4
|
In CUDA mode, the default GPU backend is cufft. More...
#include <heffte_backend_cuda.h>
Public Types | |
| using | type = cufft |
| Set the cufft tag. | |
| using | type = onemkl |
| Set the onemkl tag. | |
| using | type = rocfft |
| Set the rocfft tag. | |
In CUDA mode, the default GPU backend is cufft.
In ROCm mode, the default GPU backend is rocfft.
In oneAPI mode, the default GPU backend is onemkl.