jeffhammond / mukautuva

An MPI ABI compatibility layer
MIT License
29 stars 0 forks source link

comm spawn issues #24

Open jeffhammond opened 1 year ago

jeffhammond commented 1 year ago
  1. It only works with Open-MPI as the default SO. MPI_LIB is ignored by spawned processes and thus things don't work.

  2. This:

    mpirun.openmpi --oversubscribe -n 1    ./spawninfo1
    .--------------------------------------------------------------------------
    Unexpected output in spawninfo1: --------------------------------------------------------------------------
    mpirun.openmpi was unable to find the specified executable file, and therefore
    Unexpected output in spawninfo1: mpirun.openmpi was unable to find the specified executable file, and therefore
    did not launch the job.  This error was first reported for process
    Unexpected output in spawninfo1: did not launch the job.  This error was first reported for process
    rank 0; it may have occurred for other processes as well.
    Unexpected output in spawninfo1: rank 0; it may have occurred for other processes as well.
  3. This:

    mpirun.openmpi --oversubscribe -n 2    ./spaiccreate
    .[nuclear:4044987] *** An error occurred in MPI_Intercomm_create
    Unexpected output in spaiccreate: [nuclear:4044987] *** An error occurred in MPI_Intercomm_create
    [nuclear:4044987] *** reported by process [1230700545,0]
    Unexpected output in spaiccreate: [nuclear:4044987] *** reported by process [1230700545,0]
    [nuclear:4044987] *** on communicator MPI_COMM_WORLD
    Unexpected output in spaiccreate: [nuclear:4044987] *** on communicator MPI_COMM_WORLD
    [nuclear:4044987] *** MPI_ERR_COMM: invalid communicator
    Unexpected output in spaiccreate: [nuclear:4044987] *** MPI_ERR_COMM: invalid communicator
    [nuclear:4044987] *** MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort,
    Unexpected output in spaiccreate: [nuclear:4044987] *** MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort,
    [nuclear:4044987] ***    and potentially your MPI job)
    Unexpected output in spaiccreate: [nuclear:4044987] ***    and potentially your MPI job)
    [nuclear:4044982] 1 more process has sent help message help-mpi-errors.txt / mpi_errors_are_fatal
    Unexpected output in spaiccreate: [nuclear:4044982] 1 more process has sent help message help-mpi-errors.txt / mpi_errors_are_fatal
    [nuclear:4044982] Set MCA parameter "orte_base_help_aggregate" to 0 to see all help / error messages
    Unexpected output in spaiccreate: [nuclear:4044982] Set MCA parameter "orte_base_help_aggregate" to 0 to see all help / error messages
    Program spaiccreate exited without No Errors
jeffhammond commented 1 year ago

spawninfo1 is related to the working directory in info.