[GE users] submitt Open MPI jobs thru SGE

Reuti reuti at staff.uni-marburg.de
Thu Mar 16 15:37:35 GMT 2006


    [ The following text is in the "ISO-8859-1" character set. ]
    [ Your display is set for the "ISO-8859-10" character set.  ]
    [ Some special characters may be displayed incorrectly. ]

Hi,

for now there is no Tight Integration possible with OpenMPI and SGE. 
But you can
get at least a Loose Integration. Just start with the typcial MPICH 
Integration,
then change inside the start-procedure:

PeHostfile2MachineFile()
{
   cat $1 | while read line; do
      # echo $line
      host=`echo $line|cut -f1 -d" "|cut -f1 -d"."`
      nslots=`echo $line|cut -f2 -d" "`
      echo $host slots=$nslots
   done
}


and adding -V to the rsh-wrapper rsh:

echo $SGE_ROOT/bin/$ARC/qrsh -V -inherit -nostdin $rhost $cmd
exec $SGE_ROOT/bin/$ARC/qrsh -V -inherit -nostdin $rhost $cmd
else
echo $SGE_ROOT/bin/$ARC/qrsh -V -inherit $rhost $cmd
exec $SGE_ROOT/bin/$ARC/qrsh -V -inherit $rhost $cmd


With the jobscript (adjust to your installation):

#!/bin/sh
export PATH=/home/reuti/local/openmpi-1.0/bin:$PATH
export LD_LIBRARY_PATH=/home/reuti/local/openmpi-1.0/lib:$LD_LIBRARY_PATH
mpiexec -n $NSLOTS -machinefile $TMPDIR/machines ./mpihello
exit 0

you should get it working then.

HTH - Reuti


Quoting Yusuf Sun <yusuf.sun at gmail.com>:

> Dear all,
>
> I just installed SGE on a small 5-node cluster.
> "qstat -f" looks like:
> queuename                      qtype used/tot. load_avg arch          states
> ----------------------------------------------------------------------------
> all.q at node-1                   BIP   0/1       0.00     lx26-x86
> ----------------------------------------------------------------------------
> all.q at node-2                   BIP   0/1       0.00     lx26-x86
> ----------------------------------------------------------------------------
> all.q at node-3                   BIP   0/1       0.00     lx26-x86
> ----------------------------------------------------------------------------
> all.q at node-4                   BIP   0/1       0.00     lx26-x86
>
> I also installed Open MPI 1.0.1. Now I can run MPI jobs directly by:
> mpirun -np 4 --hostfile myhosts ~/bin/myMPI.x
> with "myhosts" having contents like
> node-1
> node-2
> node-3
> node-4
>
> Now I hope to submit my MPI jobs thru SGE.
> Any reference that can help?
>
> Thanks and regards,
> YY
>



---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net




More information about the gridengine-users mailing list