[gmx-users] GROMACS with MPICH on Sun Cluster

[David van der Spoel]

Alok wrote:
> Thanks a lot Dr. David Spoel :)
> 
> I compiled GROMACS again with CFLAGS=-g , after that i ran mdrun_mpi. It 
> is running fine though with some warning. This time I specified the full 
> path of mdrun_mpi  binary.
> But now I am having one strange problem. Sometimes it runs smoothly and 
> sometimes it does not. I am not able to pin point this problem.Actually 
> i am running mdrun_mpi by using one wrapper script as follows:
> 
> 
> ######################################################################################################## 
> 
> 
> #!/bin/csh -f
> #
> #
> # (c) 2004 Sun Microsystems, Inc. Use is subject to license terms.
> 
> # ---------------------------
> # our name
> #$ -N MPI_Job
> #
> # pe request
> #$ -pe mpich* 2-20
> #
> # MPIR_HOME from submitting environment
> #$ -v MPIR_HOME=/opt/mpichdefault-1.2.6
> # ---------------------------
> 
> #
> # needs in
> #   $NSLOTS
> #       the number of tasks to be used
> #   $TMPDIR/machines
> #       a valid machine file to be passed to mpirun
> 
> echo "Got $NSLOTS slots."
> 
> # enables $TMPDIR/rsh to catch rsh calls if available
> set path=($TMPDIR $path)
> 
> $MPIR_HOME/bin/mpirun -np $NSLOTS -machinefile $TMPDIR/machines
> -nolocal  /to/install/gromacs/x86_64-unknown-linux-gnu/bin/mdrun_mpi -v
> -s full -o full -x full -c after_full -e full -g full  >& full.job &
> 
> 
> ################################################################################################################### 
> 
> 
> Then I have to execute this script by following commands
> 
> qsub -pe mpichpar <no of nodes> -q  par.q  <file name of above script>
> 
> 
> But the problem is: sometimes it works well and sometimes it gives me 
> the following error message:
> 
> ***********************************************
> Cannot read /tmp/28015.1.par.q/machines.
> Looked for files with extension LINUX in
> directory /opt/mpichdefault-1.2.6//share .
> ***********************************************
> 
> 
> It seems that there is some problem in reading machines.LINUX file, but 
> if there is some problem then it should not work sometimes, as it does, 
> so I am very much confused now !!!
> 
> I have also installed some other software package but they are running 
> fine with same script.
> 
> I also talked with my system adminstrator he told me there is some 
> problem in installing the GROMACS that why it is behaving so strangely.
> 
> Could you suggest me something about what am I doing wrong or where am I 
> going wrong????
> 
> Thanks again for your kind help (Atleast now I am able to run mdrun_mpi 
> in the SUN CLUSTER)
> 
> Warm regards,
> Alok Jain
> 
These messages come either from the queuing system or from MPICH or fro 
both but not from GROMACS.
If it works sometimes but not otherwise, I suspect some of the nodes in 
your cluster are confused or broken.

-- 
David.
________________________________________________________________________
David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,
Dept. of Cell and Molecular Biology, Uppsala University.
Husargatan 3, Box 596,  	75124 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://folding.bmc.uu.se
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++