[Paraview] Parallel anomoly
Berk Geveci
berk.geveci at kitware.com
Fri Feb 9 12:34:32 EST 2007
Weird. I used mpich 1.2 in the past and never saw this. I recommend trying
mpich2 configured with mpd. Since that doesn't use ssh for connection, you
won't see this problem. I believe mpich1 can also be configured to use mpd.
-berk
On 2/9/07, James Galbraith <james.galbraith at inl.gov> wrote:
>
> Hi Berk,
>
> Thanks for your quick response. I tried cpi. I'm seeing the same stuff:
>
> 7719 7593 38 09:53 pts/0 00:00:02 /home/galbja/mpich-1.2.7p1/examples/basic/cpi
> -p4pg /home/galbja/mpich-1.2.7p1/examples/basic/PI7593 -p4wd
> /home/galbja/mpich-1.2.7p1/examples/basic
> 7720 7719 0 09:53 pts/0 00:00:00 /home/galbja/mpich-1.2.7p1/examples/basic/cpi
> -p4pg /home/galbja/mpich-1.2.7p1/examples/basic/PI7593 -p4wd
> /home/galbja/mpich-1.2.7p1/examples/basic
> 7721 7719 0 09:53 pts/0 00:00:00 /usr/bin/ssh smclinux -l galbja -n
> /home/galbja/mpich-1.2.7p1/examples/basic/cpi smclinux 6271 \-p4amslave
> \-p4yourname smclinux \-p4rmrank 1
> 7722 4031 0 09:53 ? 00:00:00 sshd: galbja
> [priv]
> 7724 7722 0 09:53 ? 00:00:00 sshd: galbja at notty
>
> 7725 7724 39 09:53 ? 00:00:02 /home/galbja/mpich-1.2.7p1/examples/basic/cpi
> smclinux 6271 4amslave -p4yourname smclinux -p4rmrank 1
> 7747 7725 0 09:53 ? 00:00:00 /home/galbja/mpich-1.2.7p1/examples/basic/cpi
> smclinux 6271 4amslave -p4yourname smclinux -p4rmrank 1
> 7748 7719 0 09:53 pts/0 00:00:00 /usr/bin/ssh smclinux -l galbja -n
> /home/galbja/mpich-1.2.7p1/examples/basic/cpi smclinux 6271 \-p4amslave
> \-p4yourname smclinux \-p4rmrank 2
> 7749 4031 0 09:53 ? 00:00:00 sshd: galbja
> [priv]
> 7751 7749 0 09:53 ? 00:00:00 sshd: galbja at notty
>
> 7752 7751 39 09:53 ? 00:00:02 /home/galbja/mpich-1.2.7p1/examples/basic/cpi
> smclinux 6271 4amslave -p4yourname smclinux -p4rmrank 2
> 7775 7719 0 09:53 pts/0 00:00:00 /usr/bin/ssh smclinux -l galbja -n
> /home/galbja/mpich-1.2.7p1/examples/basic/cpi smclinux 6271 \-p4amslave
> \-p4yourname smclinux \-p4rmrank 3
> 7774 7752 0 09:53 ? 00:00:00 /home/galbja/mpich-1.2.7p1/examples/basic/cpi
> smclinux 6271 4amslave -p4yourname smclinux -p4rmrank 2
> root 7776 4031 0 09:53 ? 00:00:00 sshd: galbja
> [priv]
> 7778 7776 0 09:53 ? 00:00:00 sshd: galbja at notty
>
> 7779 7778 39 09:53 ? 00:00:02 /home/galbja/mpich-1.2.7p1/examples/basic/cpi
> smclinux 6271 4amslave -p4yourname smclinux -p4rmrank 3
> 7801 7779 0 09:53 ? 00:00:00 /home/galbja/mpich-1.2.7p1/examples/basic/cpi
> smclinux 6271 4amslave -p4yourname smclinux -p4rmrank 3
>
> Any more ideas?? I didn't specify anything special to configure when I
> built mpich. I'll also run this by the mpich folks.
>
> Jim
>
> Berk Geveci wrote:
>
> Can you try a sample mpi program? I think mpich comes with cpi. The fact
> that ssh is escaping the arguments looks strange.
>
> -berk
>
> On 2/9/07, James Galbraith <james.galbraith at inl.gov> wrote:
> >
> > Hi,
> >
> > I have built ParaView 2.4.4 enabling mpi and offscreen rendering to be
> > used on some of our servers. I am seeing some an anomoly in the command
> > line of the resulting processes when I do a ps.
> >
> > I initially start pvclient -rc. I then runn pvserver with the following
> >
> > command line:
> > mpirun -np 2 /home/galbja/paraview-2.4.4-bin/bin/pvserver -rc
> > --use-offscreen-rendering
> >
> > Everything connects and the GUI comes up. But when I do a ps on the
> > server, here is what I see (cleaned out a bit so it is readable):
> > 10535 19740 1 09:20 pts/2 00:00:00 /bin/sh mpirun -np 2
> > paraview-2.4.4-bin/bin/pvserver -rc --use-offscreen-rendering
> > 10619 10535 14 09:20 pts/2 00:00:00
> > paraview-2.4.4-bin/bin/pvserver-real -rc --use-offscreen-rendering -p4pg
> >
> > paraview-2.4.4-bin/PI10535 -p4wd paraview-2.4.4-bin
> > 10620 10619 0 09:20 pts/2 00:00:00
> > paraview-2.4.4-bin/bin/pvserver-real -rc --use-offscreen-rendering -p4pg
> > paraview-2.4.4-bin/PI10535 -p4wd paraview-2.4.4-bin
> > 10621 10619 0 09:20 pts/2 00:00:00 ssh smclinux -l galbja -n
> > paraview-2.4.4-bin/bin/pvserver smclinux 14663 \-p4amslave \-p4yourname
> > smclinux \-p4rmrank 1
> > 10622 19492 0 09:20 ? 00:00:00 sshd: galbja
> > [priv]
> > 10624 10622 0 09:20 ? 00:00:00 sshd:
> > galbja at notty
> > 10625 10624 16 09:20 ? 00:00:00
> > paraview-2.4.4-bin/bin/pvserver-real smclinux 14663 4amslave
> > -p4yourname smclinux -p4rmrank 1
> > 10647 10625 0 09:20 ? 00:00:00
> > paraview-2.4.4-bin/bin/pvserver-real smclinux 14663 4amslave
> > -p4yourname smclinux -p4rmrank 1
> >
> > My question is this. On the last two lines, the "-p4amslave" argments
> > have been changed to "4amslave" somewhere along the line.
> >
> > When I open a file and accept it, the head node appears to be doing most
> > if not all the work. The "slave" process isn't doing squat. It renders
> >
> > but it appears to be that the parallel capabilities aren't being used.
> >
> > Where have I gone awry? This is happening on our cluster, and on a
> > dual-quadcore machine. When I run on our smp machine, everything is
> > fine - but it is using a different (SGI) mpi implementation. Everything
> > else is using mpich.
> >
> > Thanks in advance for your help.
> >
> > Jim
> >
> > --
> > "To Do Is To Be" - Plato
> > "To Be is To Do" - Descartes
> > "Do Be Do Be Do" - Sinatra
> >
> > James A. Galbraith
> > Idaho National Laboratory (INL)
> > Battelle Energy Alliance (BEA)
> > P.O. box 1625
> > Idaho Falls, ID 83415-3779
> > James.Galbraith at inl.gov
> > (208)526-1864
> >
> > _______________________________________________
> > ParaView mailing list
> > ParaView at paraview.org
> > http://www.paraview.org/mailman/listinfo/paraview
> >
>
>
>
> --
> Berk Geveci
> Kitware Inc.
> 28 Corporate Drive
> Clifton Park, NY, 12065
>
>
>
> --
> "To Do Is To Be" - Plato
> "To Be is To Do" - Descartes
> "Do Be Do Be Do" - Sinatra
>
> James A. Galbraith
> Idaho National Laboratory (INL)
> Battelle Energy Alliance (BEA)
> P.O. box 1625
> Idaho Falls, ID 83415-3779
> James.Galbraith at inl.gov
> (208)526-1864
>
>
--
Berk Geveci
Kitware Inc.
28 Corporate Drive
Clifton Park, NY, 12065
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://public.kitware.com/pipermail/paraview/attachments/20070209/3609e584/attachment.htm
More information about the ParaView
mailing list