MPI run crashes on more than one node

MPI run crashes on more than one node

Sravan K.'s picture


KInd Attn: James Tullos (Intel)

I am facing a similar problem like and

My mpirun crashes giving the below error message.

[proxy:0:4@cn004] HYDT_dmxu_poll_wait_for_event (./tools/demux/demux_poll.c:70): assert (!(pollfds[i].revents & ~POLLIN & ~POLLOUT & ~POLLHUP)) failed
[proxy:0:4@cn004] main (./pm/pmiserv/pmip.c:387): demux engine error waiting for event
[mpiexec@hn1] HYDT_bscu_wait_for_completion (./tools/bootstrap/utils/bscu_wait.c:101): one of the processes terminated badly; aborting
[mpiexec@hn1] HYDT_bsci_wait_for_completion (./tools/bootstrap/src/bsci_wait.c:18): bootstrap device returned error waiting for completion
[mpiexec@hn1] HYD_pmci_wait_for_completion (./pm/pmiserv/pmiserv_pmci.c:521): bootstrap server returned error waiting for completion
[mpiexec@hn1] main (./ui/mpich/mpiexec.c:548): process manager error waiting for completion

I have also tried the command [bash]mpirun -np N -check_mpi ./wrf.exe which is not giving any output.

Thanks for your help and support.




2 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.
James Tullos (Intel)'s picture

Hi Sravan,

Are you able to run the provided test program (in the $I_MPI_ROOT/test/ folder, compile any one of the files there)?

Using -check_mpi links to the MPI correctness checking library, which is for checking the MPI calls within your code.  It won't help diagnosing a problem with getting your code started, which appears to be the case here.

James Tullos
Technical Consulting Engineer
Intel® Cluster Tools

Login to leave a comment.