Open MPI logo

Open MPI User's Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Open MPI User's mailing list

Subject: Re: [OMPI users] mpirun is using one PBS node only
From: Gus Correa (gus_at_[hidden])
Date: 2009-12-01 22:40:04


Hi Belaid

PBS loves to read the nodes' list backwards.
If you want to start with WN1,
put it last on the Torque/PBS "nodes" file.

Gus Correa
---------------------------------------------------------------------
Gustavo Correa
Lamont-Doherty Earth Observatory - Columbia University
Palisades, NY, 10964-8000 - USA
---------------------------------------------------------------------

Belaid MOA wrote:
> I actually tried both:
> -- in the interactive mode, as soon as I hit enter, the PBS sends me to
> a worker node (WN2) that does not have tm support.
> I guess if I added the head node to the list of PBS nodes, I would
> not run into the problem. However, I am glad I did run into
> the problem. You guys helped me understand how mpirun interacts with
> PBS.
> -- in the batch mode, the PBS chooses WN2 as the root node (rank 0) to
> run mpirun.
> So, Both behaved in the same in my case.
>
> With many thanks Ralph for the clarification. I am now getting into the
> openMPI spirit. By the way, I love it :)
>
> ~Belaid.
>
>
>
>
>
> ------------------------------------------------------------------------
> From: rhc_at_[hidden]
> Date: Tue, 1 Dec 2009 19:44:38 -0700
> To: users_at_[hidden]
> Subject: Re: [OMPI users] mpirun is using one PBS node only
>
> Just to further show my confusion (since I wrote much of the TM support):
>
> If you get an interactive allocation and then type "mpirun ....", mpirun
> will execute on the node upon which you are sitting. Jeff's statement is
> -only- true if you "qsub" the job - i.e., you run it in batch mode.
>
> From your prior notes, I was under the impression you were running
> interactively. If this is true, then you only need OMPI to be built with
> TM support on the node which you are interactively working on.
>
> If you are running it in batch mode, then Jeff is correct - you will
> need OMPI built with TM support on all nodes since you (in general)
> cannot know which node PBS will launch you on.
>
> Hope that is clear
> Ralph
>
> On Dec 1, 2009, at 7:35 PM, Belaid MOA wrote:
>
> > Yes, this page is definitely incorrect if you want to run with PBS/TM
> > support -- you definitely need to install with TM support on all
> nodes.
> >
> > The reason is that PBS will launch your script (and therefore
> > "mpirun") on the first node of the job. This node must have an Open
> > MPI mpirun with TM support so that it can query PBS to get the nodes
> > to launch on, use the TM API to launch processes on the nodes
> > allocated to the job, etc.
> That's exactly the sound logic I understood from OpenMPI
> documentation but
> I thought the website is reliable in the sense it configured the
> server node with
> tm support.
>
> I am very grateful to subscribe to this list. It is one of the most
> active list I ever saw.
>
> Thanks a lot Jeff.
>
> With Best Regards.
> ~Belaid.
>
> ------------------------------------------------------------------------
> Get a great deal on Windows 7 and see how it works the way you
> want. See the Windows 7 offers now.
> <http://go.microsoft.com/?linkid=9691813>_______________________________________________
> users mailing list
> users_at_[hidden] <mailto:users_at_[hidden]>
> http://www.open-mpi.org/mailman/listinfo.cgi/users
>
>
>
> ------------------------------------------------------------------------
> Get a great deal on Windows 7 and see how it works the way you want. See
> the Windows 7 offers now. <http://go.microsoft.com/?linkid=9691813>
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> users mailing list
> users_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/users