CFLAG

The current version of the PR is looking ok from my perspective. I think we could avoid duplicate and have some force calculation shared between the 2D and 3D parts....

Yes, the underlying idea is the removal of the older fftw3 backend later to keep only the fftw3_f03 one May 26, 2024 09:29:12 Paul Bartholomew ***@***.***>: > > Is this...

The issue should be clarified. Asynchronous MPI IO is possible using [mpi_file_iread_all](https://docs.open-mpi.org/en/v5.0.x/man-openmpi/man3/MPI_File_iread_all.3.html) and [mpi_file_iwrite_all](https://docs.open-mpi.org/en/v5.0.x/man-openmpi/man3/MPI_File_iwrite_all.3.html)

I think this capability is available using interpolation : https://github.com/2decomp-fft/2decomp-fft/blob/dev/src/interp.f90 Maybe one example is needed to show how to use interpolation.

I think there is no need for an enum. The enum was removed in the dev branch not long ago. Dec 21, 2023 16:34:59 Paul Bartholomew ***@***.***>: > > I'm...

I think the main conclusion is that our profiler/caliper submodule is not compatible with aocc. If the library is working without caliper, this is not a major issue. In the...

This is the PR when we moved from a submodule to a module : https://github.com/2decomp-fft/2decomp-fft/pull/304/files

This is related to https://github.com/xcompact3d/2decomp-fft/issues/6. But not a duplicate because the other one is more generic.

Are you planning to split 2decomp into multiple smaller modules ? Or have many submodules and a 2decomp module mostly limited to the declaration of interfaces ?

I think this is a good idea. But this move should occur after we release a clean version of the library with backward-compatibility.