Hi Rachit,
To understand better the case, can you elaborate how you are launching the
1000's simulations? Is this done under the FePEST framework? Or are you using
your own script to do the job? The issue with the MPI can be related to several
things, for example, you already have an MPI environment in place, which could
conflict with FEFLOW's one. Or the script is not properly terminating the MPI
instance.
Best regards
Carlos Rivera