[MITgcm-support] Mpirun Problem

Andrew Keats wakeats at gmail.com
Fri Jul 9 05:43:01 EDT 2010


Hi Ming-jie,

Those errors are not coming from mitgcm, they are coming from mpirun.

> Program --app either does not exist, is not
> executable, or is an erroneous argument to mpirun.

You should check with your cluster administrator about how to run a
multi-processor jobs...

Regards

Andrew




2010/7/9 limingjie07 <limingjie07 at qdio.ac.cn>:
> Hi everyone,
>
>    I have run an experiment successfully with 2 cores on my desktop
> computer. But when I am trying to run it in cluster, I meet a problem. After
> I type “bsub –q normal –n 2 –a openmpi mpirun.lsf ./mitgcmuv”, the error as
> follows appears. I do not know what I have missed. Any advice will be
> appreciated. Thanks in advance.
>
> PS: I choose “linux_amd64_gfortran+mpi_generic” as the opt file, some
> warning (not error) come out when I “make”. Does this cause the problem?
>
>
>
> Li Ming-jie
>
> << output from stdout >>
>
> Warning: Command line arguments for program should be given
>
> after the program name.  Assuming that
> /public/home/limingjie/.openmpi_appfile_290860 is a
>
> command line argument for the program.
>
> Missing: program name
>
> Program --app either does not exist, is not
>
> executable, or is an erroneous argument to mpirun.
>
> Job  /public/lsf/7.0/linux2.6-glibc2.3-x86_64/bin/openmpi_wrapper ./mitgcmuv
>
>
>
> TID   HOST_NAME   COMMAND_LINE            STATUS            TERMINATION_TIME
>
> ===== ========== ================  =======================
> ===================
>
> 00001 node79                       Undefined
>
> 00002 node79                       Undefined
>
> 2010-07-09
> ________________________________
> Li Mingjie                                                            李明杰 敬上
> Institute of Oceanology,Chinese Academy of Sciences
>  (http://www.qdio.ac.cn)
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support
>
>



More information about the MITgcm-support mailing list