Does anyone know of a clean way to get binding of each MPI task to a subset of cores (to match the number of OpenMP threads) to work with openmpi 4.1.x? I've gotten it to work by writing a hostfile manually, but that seems cumbersome. None of the things I've tried based on the docs work when I have more than 1 node - they all claim I'm trying to oversubscribe cores.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
