Open mikemhenry opened 11 months ago
Just to note here that part of the difficulty is that the perses CLI withthe whole pipeline is not really meant to be run using MPI (it doesn't handle an MPI context/environment). So what we should do is having an example for the small molecule pipeline that generates all the objects independently before the computation, such that only the computation part can be run with MPI.
We need to get mpi4py on the container @ijpulidos will work on the example