[mpich-discuss] Fatal error in PMPI_Reduce: Other MPI error, error stac

Pavan Balaji balaji at mcs.anl.gov
Mon Feb 21 18:18:18 CST 2011


Here's an entry on the FAQ that describes this:

http://wiki.mcs.anl.gov/mpich2/index.php/Frequently_Asked_Questions#Q:_My_MPI_program_aborts_with_an_error_saying_it_cannot_communicate_with_other_processes

  -- Pavan

On 02/21/2011 06:03 PM, T.R. Sanderson wrote:
> Hello, I've been testing my MPICH2 installation with cpi and am having some
> issues. Cpi runs fine on either computer. If I add two nodes to my hosts
> file I receive the error below, if I only have one node in the host file it
> runs happily even if executed via MPI.
>
> I would be very grateful for any advice, if you would like the verbose
> output just let me know.
>
> Many thanks,
> Theo
>
> trs38 at node0:~$ mpiexec.hydra -l -n 2 /root/mpich2-1.3.2/examples/cpi [1]
> Process 1 of 2 is on node2 [0] Process 0 of 2 is on node1 [0] Fatal error
> in PMPI_Reduce: Other MPI error, error stack: [0]
> PMPI_Reduce(1322)...............: MPI_Reduce(sbuf=0x7fffbfe9d028,
> rbuf=0x7fffbfe9d020, count=1, MPI_DOUBLE, MPI_SUM, root=0, MPI_COMM_WORLD)
> failed [0] MPIR_Reduce_impl(1139)..........: [0]
> MPIR_Reduce_intra(947)..........: [0] MPIR_Reduce_binomial(176).......: [0]
> MPIDI_CH3U_Recvq_FDU_or_AEP(380): Communication error with rank 1
> [mpiexec at node0] ONE OF THE PROCESSES TERMINATED BADLY: CLEANING UP
> [proxy:0:1 at node2] HYD_pmcd_pmip_control_cmd_cb
> (./pm/pmiserv/pmip_cb.c:868): assert (!closed) failed [proxy:0:1 at node2]
> HYDT_dmxu_poll_wait_for_event (./tools/demux/demux_poll.c:77): callback
> returned error status [proxy:0:1 at node2] main (./pm/pmiserv/pmip.c:208):
> demux engine error waiting for event APPLICATION TERMINATED WITH THE EXIT
> STRING: Hangup (signal 1)
>
> _______________________________________________
> mpich-discuss mailing list
> mpich-discuss at mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss

-- 
Pavan Balaji
http://www.mcs.anl.gov/~balaji


More information about the mpich-discuss mailing list