[gmx-users] parallel run breaks !!

[delara aghaie]

Dear Gromacs users
I run DPPC monolayer on Tip4p2005 water layer system with gromacs 3.3
I had no problem before but when I run the system with command
grompp_mpi -c .gro -f .mdp -n .ndx -p .top -o .tpr -np 8
and then qsub .ll
 
the run starts but it has break down several times, each ime after for ex 987654 steps or more steps or less.
 
I get in .o.ii file this message:
 
----
PBS has allocated the following nodes:
cx1-25-8-4
cx1-25-8-4
cx1-25-8-4
cx1-25-8-4
cx1-25-8-4
cx1-25-8-4
cx1-25-8-4
cx1-25-8-4
---
mpiexec.intel -genv I_MPI_DEVICE ssm -genv I_MPI_DEBUG 0 -genv I_MPI_PIN yes -genv I_MPI_PIN_MODE lib -genv I_MPI_FALLBACK_DEVICE disable -genv DAPL_MAX_CM_RETRIES 500 -genv DAPL_MAX_CM_RESPONSE_TIME 300 -genv I_MPI_DAPL_CONNECTION_TIMEOUT 300 -machinefile /tmp/pbs.5455562.cx1/tmp.VRhkM11825 -n 8 -wdir /tmp/pbs.5455562.cx1 mdrun_mpi -v -s topol.tpr -np 8
Job output begins below
-----------------------
mpdallexit: cannot connect to local mpd (/tmp/pbs.5455562.cx1/mpd2.console_dmohamma_110411.074928); possible causes:
  1. no mpd is running on this host
  2. an mpd is running but was started without a "console" (-n option)
---
what can be the solvation of the possible problem?
 
Thanks
D. Aghaie
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20110417/b63c34c3/attachment.html>