You have an invalid communicator. I suspect you are overwriting memory<div>somewhere, since it works in one place but has corruption in another. For</div><div>this kind of error, the best thing to do is use valgrind to find it.</div>
<div><br></div><div> Matt<br><br><div class="gmail_quote">On Fri, Oct 29, 2010 at 10:16 AM, TAY Wee Beng <span dir="ltr"><<a href="mailto:zonexo@gmail.com">zonexo@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
Hi,<br>
<br>
I have a mpi code which works fine in my previous clusters.<br>
<br>
However, there's mpi problem when I use it in the new clusters, which use openmpi. I wonder if there's a relation.<br>
<br>
The error is:<br>
<br>
[n12-70:14429] *** An error occurred in MPI_comm_size<br>
[n12-70:14429] *** on communicator MPI_COMM_WORLD<br>
[n12-70:14429] *** MPI_ERR_COMM: invalid communicator<br>
[n12-70:14429] *** MPI_ERRORS_ARE_FATAL (your MPI job will now abort)<br>
2.48user 0.45system 0:03.99elapsed 73%CPU (0avgtext+0avgdata 2871936maxresident)k<br>
0inputs+0outputs (0major+185587minor)pagefaults 0swaps<br>
--------------------------------------------------------------------------<br>
mpiexec has exited due to process rank 3 with PID 14425 on<br>
node n12-70 exiting without calling "finalize". This may<br>
have caused other processes in the application to be<br>
terminated by signals sent by mpiexec (as reported here).<br>
--------------------------------------------------------------------------<br>
[n12-70:14421] 3 more processes have sent help message help-mpi-errors.txt / mpi_errors_are_fatal<br>
[n12-70:14421] Set MCA parameter "orte_base_help_aggregate" to 0 to see all help / error messages<br>
<br>
I am not sure what other information you will need. I will provide more information if required. I tried running on 1, 4 and 8 processors but all can't work.<br>
<br>
-- <br>
Yours sincerely,<br><font color="#888888">
<br>
TAY Wee Beng<br>
<br>
</font></blockquote></div><br><br clear="all"><br>-- <br>What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>-- Norbert Wiener<br>
</div>