[Pw_forum] mpi-pwcond error

[ambavale sagar]

Dear PwScf users,
I am getting MPI error during PWCOND run at particular point. The standard output gives:

MPI_Allreduce: invalid communicator: Unknown error 2064 (rank 0, comm 16)
Rank (0, MPI_COMM_WORLD) : Call stack within LAM:
Rank (0, MPI_COMM_WORLD) : -MPI_Allreduce()
Rank (0, MPI_COMM_WORLD) : -main()
------------------------------------------------------
One of the processes started by mpirun has exited with a nonzero exit........
...
..
..
PID 16077 failed on node n1 (192.168.0.12) with exit status 1.
-----------------------------


I am running espresso-4.0 installed with lam-mpi 7.1.4 and ifort 10.1 on dual cpu quad core xeon processors. There are 2 machines running in parallel. Thus 4 cpus or 16 cores are available. However, this calculation was ran on 4 cores only. The scf and relax calculations do not crash. One more thing to note is: recently I have enhanced RAM on both the machines. And after enhnacement did not do any reinstallation or recompilation of any software including lam-mpi. 

Note : Through search on forum archives I found a post by Derek :
http://www.democritos.it/pipermail/pw_forum/2005-November/003219.html
which indicates this might be a problem related to mpi or def.h......


Thank you.

Regards
Sagar Ambavale
PhD Student
The M.S. Uni. of Baroda
India



      Yahoo! India has a new look. Take a sneak peek http://in.yahoo.com/trynew
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.democritos.it/pipermail/pw_forum/attachments/20091029/ec9ec657/attachment.htm