[torqueusers] MPI job not able to run after upgraded to 4.1.5.1

Steven Lo slo at cacr.caltech.edu
Wed Oct 16 14:55:34 MDT 2013


On 10/16/2013 01:53 AM, Matt Ismail wrote:
> On Tue, Oct 15, 2013 at 03:04:34PM -0700, Steven Lo wrote:
>> We just upgraded our Torque server to 4.1.5.1 and we are having
>> trouble of running a simple MPI program
> Hi Steven,
>
> The TM interface in Torque 4 is not entirely backward compatible
> with previous versions (I think that's a bug). If your Open MPI was
> configured --with-tm in a pre-Torque 4 environment, then this could
> be the problem you're running into. Does the issue go away if you
> rebuild Open MPI against your new Torque 4 installation?
>
> Best regards,
> Matt
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers
>

Hi Matt,

This is great information.  I will double check with the person who 
build the
Open MPI.

I was suspecting the MPI code since the "momctl" works both ways which means
that they are communicating fine.  The "momctl" on the maui/torque 
server returns
correct information as well.

Will let you know how it goes.

Thanks.

Steven.



More information about the torqueusers mailing list