Hi,
did you set in maui.cfg:
JOBNODEMATCHPOLICY EXACTNODE
and specified the right np=.. in server_priv/nodes?
If you still need a fix (instead of a wrapper), you may try setting
resources_default.nodes = 1:ppn=1
in the torque configuration (with "qmgr"). This works for me on a
per-queue basis, but could even be set as a server-wide default (not
tested).
Regards,
Burkhard Bunk.
----------------------------------------------------------------------
[email protected] Physics Institute, Humboldt University
fax: ++49-30 2093 7628 Newtonstr. 15
phone: ++49-30 2093 7980 12489 Berlin, Germany
----------------------------------------------------------------------
On Mon, 13 Dec 2010, Abhishek Gupta wrote:
> Hi,
> I just able to work on a way-around by making a wrapper for qsub in
> which I put a condition that if no one is asking for any node/ppn
> specifically then it adds an option of nodes=1:ppn=1 on the command
> line. This has resolved the problem of jobs exceeding the number of cores.
> I am still looking for a way to do it on the configuration file.
> Can anyone has any suggestions?
> Thanks,
> Abhi.
>
> Abhishek Gupta wrote:
>> Hi,
>> We are experiencing a problem that the number of jobs are exceeding
>> the number of cores on the node. This is happening on our nodes with 4
>> cores. But out nodes with 8 cores are getting the right number of
>> jobs. We are currently using maui-3.2.6p21 with torque-2.3.10
>> Does anyone has any ideas about it? Is there any diagnostic
>> tools/commands for maui or anyway I could see what configuration it is
>> following?
>> Thanks,
>> Abhi.
> _______________________________________________
> mauiusers mailing list
> [email protected]
> http://www.supercluster.org/mailman/listinfo/mauiusers
>
_______________________________________________
mauiusers mailing list
[email protected]
http://www.supercluster.org/mailman/listinfo/mauiusers
