[torqueusers] mpiexec not running on requested # of processors

Frank Mietke frank.mietke at informatik.tu-chemnitz.de
Fri Oct 10 05:45:44 MDT 2008


Carlos,

On Fri, Oct 10, 2008 at 01:04:51PM +0200, Frank Mietke wrote:
> Hi Carlos,
> 
> On Thu, Oct 09, 2008 at 05:26:32PM +0200, carlos vasco wrote:
> > We have been using OSC mpiexec with the --comm=pmi parameter for mpich2 codes:
> > mpiexec --comm=pmi  yourcode
> > 
> > But now I am trying mvapich2 1.2 with a new Infiniband cluster and I
> > have not get success with OSC mpiexec.
> 
> I think the problem is that they changed the startup mechanism in MVAPICH-1.2
> (see changelog). It is necessary to adapt OSC mpiexec to handle this new
> mechanism. I did this several times with MVAPICH1 and also MVAPICH2. I'll
> looking at it in the next days.

now I've tried MVAPICH-1.2RC2 with OSC mpiexec (fresh svn checkout) and
everything is working fine on our system. So, have you tried the svn checkout?

It seems that the new startup mechanism of MVAPICH2-1.2 is an alternative to the others known
from previous MVAPICH2 versions and exists beside them.

Best Regards,
Frank


> 
> 
> Regards,
> Frank
> 
> 
> > 
> > Regards,
> > Carlos
> > 
> > On Thu, Oct 9, 2008 at 5:20 PM, Bogdan Costescu
> > <Bogdan.Costescu at iwr.uni-heidelberg.de> wrote:
> > > On Thu, 9 Oct 2008, Justin Finnerty wrote:
> > >
> > >> I think the problem is you are not really running your job under PBS. As
> > >> far as I know the mpiexec program most people use with MPICH knows about
> > >> torque but the MPICH2 mpiexec does not.
> > >
> > > OSC's mpiexec knows how to start MPICH2 jobs as well. You need to specify a
> > > different mechanism through a command line parameter or you can have a
> > > different OSC mpiexec binary with a different default mechanism.
> > >
> > > --
> > > Bogdan Costescu
> > >
> > > IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
> > > Phone: +49 6221 54 8869/8240, Fax: +49 6221 54 8868/8850
> > > E-mail: bogdan.costescu at iwr.uni-heidelberg.de
> > > _______________________________________________
> > > torqueusers mailing list
> > > torqueusers at supercluster.org
> > > http://www.supercluster.org/mailman/listinfo/torqueusers
> > >
> > _______________________________________________
> > torqueusers mailing list
> > torqueusers at supercluster.org
> > http://www.supercluster.org/mailman/listinfo/torqueusers
> > 
> 
> -- 
> Dipl.-Inf. Frank Mietke     |     Fakultätsrechen- und Informationszentrum
> Tel.: 0371 - 531 - 35538    |     Fak. für Informatik
> Fax:  0371 - 531 8 35538    |     TU-Chemnitz
> Key-ID: 60F59599            |     frank.mietke at informatik.tu-chemnitz.de
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers
> 

-- 
Dipl.-Inf. Frank Mietke     |     Fakultätsrechen- und Informationszentrum
Tel.: 0371 - 531 - 35538    |     Fak. für Informatik
Fax:  0371 - 531 8 35538    |     TU-Chemnitz
Key-ID: 60F59599            |     frank.mietke at informatik.tu-chemnitz.de


More information about the torqueusers mailing list