AMBER Archive (2009)

Subject: Re: Re: [AMBER] sander.MPI parallel run problem

From: Carlos Simmerling (carlos.simmerling_at_gmail.com)
Date: Mon Apr 06 2009 - 06:02:13 CDT

why kill them? did you check to see how many are actually accumulating time?
use something like the unix "top" command. it's not clear that there is a
problem.

2009/4/6 Dechang Li <li.dc06_at_gmail.com>

> Dear Dr. Simmerling,
>
> Thank you for your reply!
> I check the .out file, it displayed "Running AMBER/MPI version on 4
> nodes".
> If there some things not actually doing work, however, when I killed either
> one
> of them, the simulation down immediately.
>
>
>
>
>
> >how many nodes does sander say it is using (it really means mpi threads,
> not
> >nodes)? also, are all 7 collecting cpu time? some may be things not
> actually
> >doing work, such as one that starts up the other mpi jobs.
> >
> >2009/4/6 Dechang Li <li.dc06_at_gmail.com>
> >
> >> Dear all,
> >>
> >> I run the parallel simulation using sander.MPI in a cluster.
> >> The command I used was:
> >>
> >> mpirun -np 4 -machinefile myhosts ...
> >>
> >> I required 4 cpus to do the parallel simulation, but finally
> >> there were 7 programes run at the node:
> >>
> >> 9279 user1 25 0 37928 6088 1592 R 101 0.1 11:03.15
> >> /hpexport2/home/user1/amber9/exe/sander.MPI cn27 56237 -p4amslave
> >> -p4yourname cn27
> >> 9281 user1 25 0 38864 6128 1592 R 101 0.1 11:03.07
> >> /hpexport2/home/user1/amber9/exe/sander.MPI cn27 56237 -p4amslave
> >> -p4yourname cn27
> >> 9253 user1 16 0 37928 6088 1592 R 99 0.1 10:45.95
> >> /hpexport2/home/user1/amber9/exe/sander.MPI cn27 56237 -p4amslave
> >> -p4yourname cn27
> >> 9280 user1 25 0 37928 6068 1592 R 99 0.1 11:03.23
> >> /hpexport2/home/user1/amber9/exe/sander.MPI cn27 56237 -p4amslave
> >> -p4yourname cn27
> >> 9207 user1 16 0 38864 6128 1592 R 97 0.1 10:44.25
> >> /hpexport2/home/user1/amber9/exe/sander.MPI cn27 56237 -p4amslave
> >> -p4yourname cn27
> >> 9230 user1 16 0 37928 6068 1592 R 97 0.1 10:41.93
> >> /hpexport2/home/user1/amber9/exe/sander.MPI cn27 56237 -p4amslave
> >> -p4yourname cn27
> >> 9202 user1 16 0 35872 4848 2184 R 91 0.1 10:09.37
> >> /hpexport2/home/user1/amber9/exe/sander.MPI -O -i
> >> /hpexport2/home/user1/abc/water/
> >>
> >>
> >> What is the problem?
> >>
> >> Best regards,
> >> 2009-4-6
> >>
> >>
> >>
> >> =========================================
> >> Dechang Li, Ph.D Candidate
> >> Department of Engineering Mechanics
> >> Tsinghua University
> >> Beijing 100084
> >> P.R. China
> >>
> >> Tel: +86-10-62773574(O)
> >> Email: lidc02 at mails.tsinghua.edu.cn
> >> =========================================
> >>
> >>
> >>
> >>
> >> _______________________________________________
> >> AMBER mailing list
> >> AMBER_at_ambermd.org
> >> http://lists.ambermd.org/mailman/listinfo/amber
> >>
> >_______________________________________________
> >AMBER mailing list
> >AMBER_at_ambermd.org
> >http://lists.ambermd.org/mailman/listinfo/amber
>
>
>
>
> _______________________________________________
> AMBER mailing list
> AMBER_at_ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber