[Mitgcm-support] adcroft at mit.edu

mitgcm-support at dev.mitgcm.org mitgcm-support at dev.mitgcm.org
Wed Jul 9 15:45:43 EDT 2003


alistair-

all mail addressed to 

  adcroft at mit.edu

has bounced for the last week.
has your email address changed ? 

[greg]

> Date: Fri, 07 Mar 2003 12:55:20 -0500
> From: Alistair Adcroft <adcroft at mit.edu>
> CC: Helen Hill <helen at plume.mit.edu>
> 
> TS,
> 
> There seems to be a general problem with MPI jobs on the cluster;
> there are several users complaining about this. The absence of
> output is a common theme - NFS? Perhaps time for a system wide reboot?
> 
> A.
> cc HH.
> 
> -------- Original Message --------
> From: Helen Hill <helen at plume.mit.edu>
> Subject: Help using MITgcm on the MITgcm Cluster Facility
> To: support at mitgcm.org
> 
> Hi
> I'm trying to run the following job on the cluster:
> 
> #!/bin/csh
> #
> # PBS script to run a job on the myrinet-3 cluster.
> # The lines beginning #PBS set various queuing parameters.
> # o -N Job Name
> #PBS -N traceNATL
> #
> #
> # o -l resource lists that control where job goes
> #      here we ask for 3 nodes, each with the attribute "p4".
> #PBS -l nodes=16:p4
> #
> # o Where to write output
> #PBS -e stderr
> 
> #PBS -o stdout
> 
> #
> # o Export all my environment variables to the job
> #PBS -V
> #
> set verbose
> cd /s00/helen/offlineNATLfromNATL/offline_model/helen-relax2-np16/run
> cat $PBS_NODEFILE > mf
> set ncpus = ( `wc -l mf | awk '{print $1}'` )
> /usr/local/pkg/mpi/mpi-1.2.4..8a-gm-1.5/intel/bin/mpirun.ch_gm
> -machinefile
> mf -s -v --gm-kill 5 -np $ncpus ./mitgcmuv
> exit
> 
> The job fails after writing mf. There is no stderr or stdout files.
> Without error messages I don't now where to look to determine what is
> wrong.
> 
> Checking, I actually don't seem to be getting stderr and stdout files
> anymore even for working runs that do produce output ...
> 
> Any comments/ideas/suggestions?
> 
> Helen
> 
> Enter details of the problem below - Thanks, the MITgcm support team!
> 



More information about the MITgcm-support mailing list