As discovered by @schwarzmeier. I don't know whether this is a PFFT bug or an OpenMPI bug. PFFT internally creates a Cartesian communicator when using MPI_COMM_WORLD
, so maybe it's just related to that.
As discovered by @schwarzmeier. I don't know whether this is a PFFT bug or an OpenMPI bug. PFFT internally creates a Cartesian communicator when using MPI_COMM_WORLD
, so maybe it's just related to that.