I'm trying to use the CPSL to solve a (separable) ODE, which happens to have a sum of four Helmholtz equation solutions on the RHS. This means I have to call the library often, as I do the time stepping for the ODE. Everything works just fine, until after a good many calls to dmv0_helmholtz_2d I suddenly get:
Fatal error in MPI_Comm_dup:
Other MPI error, error stack:
MPI_Comm_dup(216)..: MPI_Comm_dup(comm=0x84000005, new_comm=0x7fffffffdfb0) failed
MPIR_Comm_copy(655): Too many communicators
which makes we wonder if the library is missing a MPI_Comm_free call somewhere? Help on this would be much appreciated, since I don't see how I can fix it and I've not found a workaround so far.
Intel compiler icc 11.0/083, MKL is 10.2.1.017, MPI is MVAPICH2 1.2.