[torqueusers] problem PBS torque when submit job with MPICH

Greenseid, Joseph M (IS) Joseph.Greenseid at ngc.com
Mon Nov 30 10:22:17 MST 2009


Does /opt/mpich-mx/bin/mpirun exist on the compute nodes?  It exists on your head node, so when you run it directly from the command line, it works fine.  However, Torque will run the contents of the job script from the first node in the node set it gives you.  If that isn't installed on the back end nodes, that'd be your problem.
 
--Joe

________________________________

From: torqueusers-bounces at supercluster.org on behalf of Luca Nannipieri
Sent: Mon 11/30/2009 6:20 AM
To: torqueusers
Subject: [torqueusers] problem PBS torque when submit job with MPICH



we have a Rocks 5.0 installation with PBS/torque roll, myrinet-mx roll
and intel roll.
If compile my program with /opt/mpich-mx/bin/mpif90  and run with
/opt/mpich-mx/bin/mpirun without PBS, the job works fine without errors.
If we submitted the same job with PBS torque, received the follow error:

/opt/torque/mom_priv/jobs/1094.hpc2.pi.ingv.it.SC: line 12:
/opt/mpich-mx/bin/mpirun: No such file or directory


you can see the job script below:

#---------------------------------------------
#!/bin/sh
#PBS -l nodes=5:ppn=4
#PBS -l walltime=240:00:00
#PBS -A ongaro
#PBS -N prova_intel-mx
#PBS -m abe
#PBS -M nannipieri at pi.ingv.it
export PBS_RSHCOMMAND=/usr/bin/ssh
export RSHCOMMAND=/usr/bin/ssh
cd $PBS_O_WORKDIR
/opt/mpich-mx/bin/mpirun -np 20 -machinefile $PBS_NODEFILE
/home/nannipieri/pdac_intel-mx.x
#---------------------------------------------

we do not understand where is the problem.

--
Ing. Luca Nannipieri
Istituto Nazionale di Geofisica e Vulcanologia
Sezione di Pisa
Via della Faggiola, 32 - 56126 Pisa - Italy
Tel. +39 050 8311926
fax: +39 050 8311942
http://www.pi.ingv.it/Pagine_personali/nannipieri.html

_______________________________________________
torqueusers mailing list
torqueusers at supercluster.org
http://www.supercluster.org/mailman/listinfo/torqueusers


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.supercluster.org/pipermail/torqueusers/attachments/20091130/0a3b12e5/attachment-0001.html 


More information about the torqueusers mailing list