Probably the best option to parallelize and distribute is mpi4py.
This issue is opened to leave comments and lessons to be learned while this is implemented.
Other options as multiprocessing, phatos, multithreading... do not really work well for this. It could be useful to run in a single node with threads but this can also be covered with mpi4py.
Probably the best option to parallelize and distribute is mpi4py. This issue is opened to leave comments and lessons to be learned while this is implemented.