[gmx-users] Dppc Benchmark
Andrea Carotti
and.carotti at farmchim.uniba.it
Mon Jun 21 13:15:42 CEST 2004
Dear David,
only today I've finished my benchmarks. A little clarification:
i've 2 dual xeon (SMP) 2,66 GHz whith 1GB ram running linux FC2 on a gigabit ethernet.
I've used mpich 1.2.5.2 and gromacs 3.2.1.
these are the commands i've used:
grompp -f -c -p -po -shuffle -sort -np # -o
mpirun -np # /home/Application/gromacs3.2.1/i686-pc-linux-gnu/bin/mdrun -s -x -c -g -e -np # >& log.log
(#=2,4)
On the dppc system i've obtained this results:
(Mnbf/s) (MFlops) (ps/NODE hour) (NODE hour/ns)
1 CPU 20.015 721.825 6.184 161.710
2 CPUs 20.747 747.955 6.410 156.000
4 CPUs 34.735 1.252 10.730 93.194
For the compilation of mpich i've not used any flag. For fftw and gromacs
I've used the howto indications (http://www.gromacs.org/documentation/howtos/mpich_howto.html).
So my question was...do I need to recompile mpich with some flags?
Thanks to everyone will help
Andrea
----- Original Message -----
From: "David" <spoel at xray.bmc.uu.se>
To: "Discussion list for GROMACS users" <gmx-users at gromacs.org>
Sent: Friday, June 18, 2004 5:08 PM
Subject: Re: [gmx-users] Dppc Benchmark
> On Fri, 2004-06-18 at 16:43, Andrea Carotti wrote:
> > I've run:
> > grompp -f -c -p -po -shuffle -sort -np 4 -o
> > and then
> > mpirun -np 4
> > /home/Application/gromacs3.2.1/i686-pc-linux-gnu/bin/mdrun -s -x -c
> > conf_wat.gro -g -e -np 4 >& log.log &
>
> and what does the performance look like for 1,2,4 processors?
> Are they SMP or single CPU nodes?
> >
> > ----- Original Message -----
> > From: "David" <spoel at xray.bmc.uu.se>
> > To: "Discussion list for GROMACS users" <gmx-users at gromacs.org>
> > Sent: Friday, June 18, 2004 4:35 PM
> > Subject: Re: [gmx-users] Dppc Benchmark
> >
> >
> > > On Fri, 2004-06-18 at 13:22, Andrea Carotti wrote:
> > > > Hi all.
> > > > I' ve the same problem: same hardware and bad performance..but I've
> > > > MPICH 1.2.5.2, Fedora Core2, gromacs 3.2.1.
> > > > (I've used the howto:
> > > > http://www.gromacs.org/documentation/howtos/mpich_howto.html)
> > > > There is a solution also for me? ...like a flag in the compilation of
> > mpich
> > > > that i've missed....
> > > same question then, have you used grompp -shuffle ?
> > > > TNX in advance
> > > > Andrea
> > > >
> > > >
> > > > ----- Original Message -----
> > > > From: "David" <spoel at xray.bmc.uu.se>
> > > > To: "Seonah Kim" <kim at qtp.ufl.edu>; "Discussion list for GROMACS users"
> > > > <gmx-users at gromacs.org>
> > > > Sent: Friday, June 18, 2004 10:43 AM
> > > > Subject: Re: [gmx-users] Dppc Benchmark
> > > >
> > > >
> > > > > On Thu, 2004-06-17 at 23:23, Seonah Kim wrote:
> > > > > > Dear all,
> > > > > >
> > > > > > We are testing our linux cluster now, using gromacs 3.2.1, with the
> > dppc
> > > > > > benchmark.
> > > > > >
> > > > > > We have dual Xeon 2.88 Ghz, 1 GB memory per box. We run RedHat 9.0,
> > and
> > > > > > use LAM. 512 Mb L2 cache as far as I can tell.
> > > > > >
> > > > > > We see 135 ps/day for a single processor, but only 250 for two.
> > > > > > This is not the superscaling reported in the gromacs web site.
> > > > > >
> > > > > > The two processor run really works in the same box, using the 2
> > procs,
> > > > > > we checked, so it is not the network. We read/write only at the
> > > > > > beggining and end, so it is not the disk.
> > > > > >
> > > > > > Can anyone help us ? Is this a LAM issue ?
> > > > > You may need to recompile using some flags for LAM (short message size
> > > > > or something like that and set it to 512 kb).
> > > > >
> > > > > Did you run grompp -shuffle?
> > > > > >
> > > > > > Thanks !!!
> > > > > > a.
> > > > > >
> > > > > > _______________________________________________
> > > > > > gmx-users mailing list
> > > > > > gmx-users at gromacs.org
> > > > > > http://www.gromacs.org/mailman/listinfo/gmx-users
> > > > > > Please don't post (un)subscribe requests to the list. Use the
> > > > > > www interface or send it to gmx-users-request at gromacs.org.
> > > > > --
> > > > > David.
> > > > >
> > ________________________________________________________________________
> > > > > David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,
> > > > > Dept. of Cell and Molecular Biology, Uppsala University.
> > > > > Husargatan 3, Box 596, 75124 Uppsala, Sweden
> > > > > phone: 46 18 471 4205 fax: 46 18 511 755
> > > > > spoel at xray.bmc.uu.se spoel at gromacs.org http://xray.bmc.uu.se/~spoel
> > > > >
> > ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> > > > >
> > > > > _______________________________________________
> > > > > gmx-users mailing list
> > > > > gmx-users at gromacs.org
> > > > > http://www.gromacs.org/mailman/listinfo/gmx-users
> > > > > Please don't post (un)subscribe requests to the list. Use the
> > > > > www interface or send it to gmx-users-request at gromacs.org.
> > > > >
> > > >
> > > >
> > > > _______________________________________________
> > > > gmx-users mailing list
> > > > gmx-users at gromacs.org
> > > > http://www.gromacs.org/mailman/listinfo/gmx-users
> > > > Please don't post (un)subscribe requests to the list. Use the
> > > > www interface or send it to gmx-users-request at gromacs.org.
> > > --
> > > David.
> > > ________________________________________________________________________
> > > David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,
> > > Dept. of Cell and Molecular Biology, Uppsala University.
> > > Husargatan 3, Box 596, 75124 Uppsala, Sweden
> > > phone: 46 18 471 4205 fax: 46 18 511 755
> > > spoel at xray.bmc.uu.se spoel at gromacs.org http://xray.bmc.uu.se/~spoel
> > > ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> > >
> > > _______________________________________________
> > > gmx-users mailing list
> > > gmx-users at gromacs.org
> > > http://www.gromacs.org/mailman/listinfo/gmx-users
> > > Please don't post (un)subscribe requests to the list. Use the
> > > www interface or send it to gmx-users-request at gromacs.org.
> >
> >
> > _______________________________________________
> > gmx-users mailing list
> > gmx-users at gromacs.org
> > http://www.gromacs.org/mailman/listinfo/gmx-users
> > Please don't post (un)subscribe requests to the list. Use the
> > www interface or send it to gmx-users-request at gromacs.org.
> --
> David.
> ________________________________________________________________________
> David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,
> Dept. of Cell and Molecular Biology, Uppsala University.
> Husargatan 3, Box 596, 75124 Uppsala, Sweden
> phone: 46 18 471 4205 fax: 46 18 511 755
> spoel at xray.bmc.uu.se spoel at gromacs.org http://xray.bmc.uu.se/~spoel
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20040621/a54bec9e/attachment.html>
More information about the gromacs.org_gmx-users
mailing list