Running R under Sun Grid Engine with OpenMPI tight integration

On Tue, Jan 27, 2009 at 11:04 AM, Marce <marcemb at gmail.com> wrote:
  

2009/1/27 Rainer M Krug <r.m.krug at gmail.com>:
    

On Tue, Jan 27, 2009 at 2:30 AM, Peter Waltman <peter.waltman at gmail.com> wrote:
      

Hi -

I saw your posting on the R-help mailing list.  Were you ever able to get
this working?  did you end up switching to use the rsge library?
        

Yes - that is exactly what I did - I am using rsge or, which is in
most cases sufficient for me, starting several instances of R and run
the whole simulation (array processing).

But I would still like to know how I can use the Rmpi and snow on the
Sun Grid Engine.

Please keep me posted,

Rainer
      

I'm trying to do the same, and not having very much luck getting it going.

Thanks!

Peter Waltman

        

--
Rainer M. Krug, PhD (Conservation Ecology, SUN), MSc (Conservation
Biology, UCT), Dipl. Phys. (Germany)

Centre of Excellence for Invasion Biology
Faculty of Science
Natural Sciences Building
Private Bag X1
University of Stellenbosch
Matieland 7602
South Africa

Hi all,
I have been installed R in SGE but with LAM tight integration, I don't
k now if this help you :s . The problem I had with OpenMPI was that
the load in all nodes waiting for a job and with an instance of R.
Their load up to 100 %, and there isn't a job running on them.
Even so, I will try to configure it with OpenMPI..when I have results
I will post them ok?
    

Yes please - could you please provide some info (and a small example)
on how you managed this with LAM?

I think we should move this thread to r-sig-hpc, as it is more
appropriate there. I CC'd it to there - please respond there and not
on r-help.

Thanks

Rainer

  

Saludos

Hi,

in Munich we use SGE with Lam/MPI and Rmpi (latest version). It is working
very well now. There were only some small changes to the SGE configuration
required.

Attached a document form REvolution Computing for using SGE with OpenMPI. It
is a nice installation guide. Thanks to REvolution for providing this
document to the R community.

Best
Markus


Rainer M Krug schrieb:

On Tue, Jan 27, 2009 at 11:04 AM, Marce <marcemb at gmail.com> wrote:

2009/1/27 Rainer M Krug <r.m.krug at gmail.com>:

On Tue, Jan 27, 2009 at 2:30 AM, Peter Waltman <peter.waltman at gmail.com>
wrote:

Hi -

I saw your posting on the R-help mailing list.  Were you ever able to
get
this working?  did you end up switching to use the rsge library?

Yes - that is exactly what I did - I am using rsge or, which is in
most cases sufficient for me, starting several instances of R and run
the whole simulation (array processing).

But I would still like to know how I can use the Rmpi and snow on the
Sun Grid Engine.

Please keep me posted,

Rainer

I'm trying to do the same, and not having very much luck getting it
going.

Thanks!

Peter Waltman

--
Rainer M. Krug, PhD (Conservation Ecology, SUN), MSc (Conservation
Biology, UCT), Dipl. Phys. (Germany)

Centre of Excellence for Invasion Biology
Faculty of Science
Natural Sciences Building
Private Bag X1
University of Stellenbosch
Matieland 7602
South Africa

Hi all,
I have been installed R in SGE but with LAM tight integration, I don't
k now if this help you :s . The problem I had with OpenMPI was that
the load in all nodes waiting for a job and with an instance of R.
Their load up to 100 %, and there isn't a job running on them.
Even so, I will try to configure it with OpenMPI..when I have results
I will post them ok?

Yes please - could you please provide some info (and a small example)
on how you managed this with LAM?

I think we should move this thread to r-sig-hpc, as it is more
appropriate there. I CC'd it to there - please respond there and not
on r-help.

Thanks

Rainer

Saludos

--
Dipl.-Tech. Math. Markus Schmidberger

Ludwig-Maximilians-Universit?t M?nchen
IBE - Institut f?r medizinische Informationsverarbeitung,
Biometrie und Epidemiologie
Marchioninistr. 15, D-81377 Muenchen
URL: http://www.ibe.med.uni-muenchen.de Mail: Markus.Schmidberger [at]
ibe.med.uni-muenchen.de
Tel: +49 (089) 7095 - 4497

On Wed, Jan 28, 2009 at 10:54 AM, Markus Schmidberger
<schmidb at ibe.med.uni-muenchen.de> wrote:
  

Hi,

in Munich we use SGE with Lam/MPI and Rmpi (latest version). It is working
very well now. There were only some small changes to the SGE configuration
required.

Attached a document form REvolution Computing for using SGE with OpenMPI. It
is a nice installation guide. Thanks to REvolution for providing this
document to the R community.
    

Thanks a lot for the document and the info - that sounds fantastic,
and I will try it out very soon.

Could you provide the link to the original document from Revolution,
so that I can check for updates on the document?

Cheers

Rainer

  

Best
Markus


Rainer M Krug schrieb:
    

On Tue, Jan 27, 2009 at 11:04 AM, Marce <marcemb at gmail.com> wrote:

      

2009/1/27 Rainer M Krug <r.m.krug at gmail.com>:

        

On Tue, Jan 27, 2009 at 2:30 AM, Peter Waltman <peter.waltman at gmail.com>
wrote:

          

Hi -

I saw your posting on the R-help mailing list.  Were you ever able to
get
this working?  did you end up switching to use the rsge library?

            

Yes - that is exactly what I did - I am using rsge or, which is in
most cases sufficient for me, starting several instances of R and run
the whole simulation (array processing).

But I would still like to know how I can use the Rmpi and snow on the
Sun Grid Engine.

Please keep me posted,

Rainer

          

I'm trying to do the same, and not having very much luck getting it
going.

Thanks!

Peter Waltman


            

--
Rainer M. Krug, PhD (Conservation Ecology, SUN), MSc (Conservation
Biology, UCT), Dipl. Phys. (Germany)

Centre of Excellence for Invasion Biology
Faculty of Science
Natural Sciences Building
Private Bag X1
University of Stellenbosch
Matieland 7602
South Africa

Hi all,
I have been installed R in SGE but with LAM tight integration, I don't
k now if this help you :s . The problem I had with OpenMPI was that
the load in all nodes waiting for a job and with an instance of R.
Their load up to 100 %, and there isn't a job running on them.
Even so, I will try to configure it with OpenMPI..when I have results
I will post them ok?

        

Yes please - could you please provide some info (and a small example)
on how you managed this with LAM?

I think we should move this thread to r-sig-hpc, as it is more
appropriate there. I CC'd it to there - please respond there and not
on r-help.

Thanks

Rainer


      

Saludos

--
Dipl.-Tech. Math. Markus Schmidberger

Ludwig-Maximilians-Universit?t M?nchen
IBE - Institut f?r medizinische Informationsverarbeitung,
Biometrie und Epidemiologie
Marchioninistr. 15, D-81377 Muenchen
URL: http://www.ibe.med.uni-muenchen.de Mail: Markus.Schmidberger [at]
ibe.med.uni-muenchen.de
Tel: +49 (089) 7095 - 4497


    

Thats a internal wiki from Revolution. I only got the pdf. But this should be the latest version.

Best
Markus

Yes, that is the latest version.  I just checked the document in our
internal Google Site.  There's no externally accessible URL, but feel
free to circulate it. It may become a white paper one day.

# David Smith

On Wed, Jan 28, 2009 at 1:43 AM, Markus Schmidberger
<schmidb at ibe.med.uni-muenchen.de> wrote:

Thats a internal wiki from Revolution. I only got the pdf. But this should be the latest version.

Best
Markus

--
David M Smith <david at revolution-computing.com>
Director of Community, REvolution Computing www.revolution-computing.com
Tel: +1 (206) 577-4778 x3203 (Seattle, USA)