|
* wrap fft libraries
* implementation in cpp to avoid duplicate symbole
* delete some fftw code
* further cleaning
* typo
* pass fft plans by reference
* fix bug due to typo. Dammit, macros sometimes make it hard
* FFT wrapper also support cuFFT (not tested yet)
* eol
* further cleaning
* fix cuFFT, tested on Summit
* clean WarpX Complex checks
* pass directly the IntVect instead of the components (what were you thinking?)
* add some amrex prefix
* add a few const
* Should not need an FFT library to compile FDTD
* gather all FFT header files into one.
* eol
* minor changes in code comments.
* let dimension be chosen at runtime instead of at compile-time
* fix compilation on GPU
* add description to the wrapper namespace
|
MaxThevenet
