<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 6.5.7655.7">
<TITLE>RE: [gmx-users] mpi run in Gromacs 4.5.3</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/plain format -->
<P><FONT SIZE=2>Thank you, Justin and Mark. I'll recompile gromacs. In my current version -nt is not available (using mdrun -h).<BR>
Thanks again,<BR>
JT<BR>
<BR>
<BR>
-----Original Message-----<BR>
From: gmx-users-bounces@gromacs.org on behalf of Justin A. Lemkul<BR>
Sent: Mon 12/13/2010 5:49 PM<BR>
To: Discussion list for GROMACS users<BR>
Subject: Re: [gmx-users] mpi run in Gromacs 4.5.3<BR>
<BR>
<BR>
<BR>
Te, Jerez A., Ph.D. wrote:<BR>
> Hi Mark,<BR>
><BR>
> Thank you for your reply. Just to confirm, mdrun_mpi is still being used in<BR>
> Gromacs 4.5.3? The gromacs manual (Appendix A.5- Running Gromacs in parallel)<BR>
> suggested using "mdrun -np 8 -s topol -v -N 8" in running a single machine<BR>
> with multiple (8) processors and "mpirun -p goofus,doofus,fred 10 mdrun -s<BR>
> topol -v -N 30" to run three machines with ten processors each. You probably<BR>
> already know this but I am confused as to whether these commands are correct<BR>
<BR>
Those commands are outdated. I am updating the manual to fix this.<BR>
<BR>
> in running parallel simulations (thus the assumption that mdrun_mpi is no<BR>
> longer applicable in gromacs 4.5.3). So far I have not been successful in<BR>
> running parallel simulations (as I mentioned before the two options above<BR>
> would only start X identical serial processes). My bottom line is I want to<BR>
> run Gromacs simulation in parallel (regardless of whether it is running on<BR>
> one silicon with a number of processors or on different machines or nodes<BR>
> with a specified number of processors).<BR>
><BR>
<BR>
The exact implementation depends on the nature of your cluster setup. For<BR>
instance, our aging supercomputer does not support threading, so we have to<BR>
compile with --enable-mpi to produce an mdrun_mpi binary that is called via<BR>
mpirun. My laptop, however, supports threading, so I can initiate a process<BR>
over both cores with mdrun -nt, no external MPI support necessary.<BR>
<BR>
> I assume that in order to get mdrun_mpi, Gromacs has to be recompiled using<BR>
> the --enable-mpi option because currently our gromacs executable bin does not<BR>
> have mdrun_mpi. Our system administrator said that all mpi-related options<BR>
> have been enabled in compiling gromacs and still the mdrun_mpi is not found<BR>
> as one of the exe files. Please help.<BR>
><BR>
<BR>
Just because your /bin subdirectory does not have mdrun_mpi does not necessarily<BR>
mean the binary was not compiled with MPI support. Based on what you have<BR>
reported, it sounds like you indeed only have a serial mdrun, but it is possible<BR>
to suppress the default suffix. If you have threading support, the -nt option<BR>
will be printed if you issue mdrun -h.<BR>
<BR>
-Justin<BR>
<BR>
> Thanks, JT<BR>
><BR>
><BR>
> -----Original Message----- From: gmx-users-bounces@gromacs.org on behalf of<BR>
> Mark Abraham Sent: Mon 12/13/2010 4:38 PM To: Discussion list for GROMACS<BR>
> users Subject: Re: [gmx-users] mpi run in Gromacs 4.5.3<BR>
><BR>
> On 14/12/2010 7:48 AM, Te, Jerez A., Ph.D. wrote:<BR>
>> Hi, I have been trying to run Gromacs 4.5.3 parallel simulations using<BR>
>> openmpi 1.4.2. From my understanding, mdrun_mpi is not used in this version<BR>
>> of Gromacs.<BR>
>><BR>
><BR>
> I don't understand what (you think) you mean. You can use thread-based<BR>
> parallelism for processors that share common silicon, or MPI-based<BR>
> parallelism if a network connection is involved, but not both. The latter is<BR>
> named mdrun_mpi by default.<BR>
><BR>
>> Our system administrator told me that all mpi related options have been<BR>
>> turned on while installing Gromacs. With either commands: mdrun -np X<BR>
>> -deffnm topol -N X (run in an 8-cpu node)<BR>
>><BR>
><BR>
> This won't run in parallel at all. mdrun ignores -np and -N<BR>
><BR>
>> or mpirun -np X mdrun -deffnm topol -N X (submitted in a number of nodes<BR>
>> depending on availability)<BR>
>><BR>
><BR>
> This will get you the symptoms below, but -N is still ignored.<BR>
><BR>
>> I get X identical simulations instead of a parallel run. If X=4, I get 4<BR>
>> identical simulations (the same simulation ran 4 times) instead of 1<BR>
>> parallel simulation in 4 processors. The performance difference between a<BR>
>> single-processor run and the X=4 run are also similar (no marked difference<BR>
>> in the time it takes to finish the simulation). Has anyone encountered this<BR>
>> problem?<BR>
>><BR>
><BR>
> You're using a serial mdrun. Use a parallel mdrun.<BR>
><BR>
> Mark<BR>
><BR>
><BR>
<BR>
--<BR>
========================================<BR>
<BR>
Justin A. Lemkul<BR>
Ph.D. Candidate<BR>
ICTAS Doctoral Scholar<BR>
MILES-IGERT Trainee<BR>
Department of Biochemistry<BR>
Virginia Tech<BR>
Blacksburg, VA<BR>
jalemkul[at]vt.edu | (540) 231-9080<BR>
<A HREF="http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin">http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin</A><BR>
<BR>
========================================<BR>
--<BR>
gmx-users mailing list gmx-users@gromacs.org<BR>
<A HREF="http://lists.gromacs.org/mailman/listinfo/gmx-users">http://lists.gromacs.org/mailman/listinfo/gmx-users</A><BR>
Please search the archive at <A HREF="http://www.gromacs.org/Support/Mailing_Lists/Search">http://www.gromacs.org/Support/Mailing_Lists/Search</A> before posting!<BR>
Please don't post (un)subscribe requests to the list. Use the<BR>
www interface or send it to gmx-users-request@gromacs.org.<BR>
Can't post? Read <A HREF="http://www.gromacs.org/Support/Mailing_Lists">http://www.gromacs.org/Support/Mailing_Lists</A><BR>
<BR>
</FONT>
</P>
</BODY>
</HTML>