Open tkoskela opened 1 month ago
In https://github.com/OrderN/CONQUEST-release/blob/f424a27d322011e6c131293ba204a87bc50e0472/src/fft_fftw3.f90#L56-L58
in fftx_exec_wrapper there is a loop over columns. It should be possible to thread it trivially. If @davidbowler can provide a test case where this becomes a bottleneck, we can try threading it and measure the speedup.
In https://github.com/OrderN/CONQUEST-release/blob/f424a27d322011e6c131293ba204a87bc50e0472/src/fft_fftw3.f90#L56-L58
in fftx_exec_wrapper there is a loop over columns. It should be possible to thread it trivially. If @davidbowler can provide a test case where this becomes a bottleneck, we can try threading it and measure the speedup.