[gmx-users] Slow Runs
Denny Frost
dsfrost at cableone.net
Fri Jan 28 21:43:47 CET 2011
I'm leaning toward the possibility that it is actually only running 8 copies
of the same job on different processors. My question is how does gromacs4.5
know how many processors it has available to parallelize a job? Is it
specified in grompp or does it just detect it?
On Fri, Jan 28, 2011 at 1:32 PM, Justin A. Lemkul <jalemkul at vt.edu> wrote:
>
>
> Denny Frost wrote:
>
>> Here's my grompp command:
>>
>> grompp_d -nice 0 -v -f md.mdp -c ReadyForMD.gro -o md.tpr -p top.top
>>
>> and my mdrun command is this:
>> time mpiexec mdrun_mpi -np 8 -cpt 30000 -nice 0 -nt 1 -s
>> $PBS_O_WORKDIR/md.tpr -o $PBS_O_WORKDIR/mdDone.trr -x
>> $PBS_O_WORKDIR/mdDone.xtc -c $PBS_O_WORKDIR/mdDone.gro -e
>> $PBS_O_WORKDIR/md.edr -g $PBS_O_WORKDIR/md.log 1> $PBS_JOBID.pgm.out 4>
>> $PBS_JOBID.pgm.err
>>
>>
> The -np option of mdrun is nonexistent, but mdrun does not check for proper
> command line arguments, so you won't get an error. But then you've said
> that 8 processors are active, so I still suspect that mdrun was compiled
> incorrectly or in such a way that it's incompatible with your system. The
> output from the .log file indicates that only one processor was used. Maybe
> your admins can help you on this one, if the jobs spit out any useful
> diagnostic information.
>
> For our cluster, we use e.g.:
>
> mpirun -np 8 mdrun_mpi -deffnm md
>
> -Justin
>
> I know the -cpt option is 30000 because I don't want a checkpoint file
>> because every time it tries to make it, it fails due to quota issues and
>> kills the job. I'm not sure why this happens, but I think it's a separate
>> issue to take up with my supercomputing facility.
>>
>> On Fri, Jan 28, 2011 at 1:18 PM, Justin A. Lemkul <jalemkul at vt.edu<mailto:
>> jalemkul at vt.edu>> wrote:
>>
>>
>>
>> Denny Frost wrote:
>>
>> all 8 nodes are running at full capacity, though
>>
>>
>> What is your mdrun command line? How did you compile it? What can
>> happen is something went wrong during installation, so you think you
>> have an MPI-enabled binary, but it is simply executing 8 copies of
>> the same job.
>>
>> -Justin
>>
>> On Fri, Jan 28, 2011 at 1:13 PM, Justin A. Lemkul
>> <jalemkul at vt.edu <mailto:jalemkul at vt.edu>
>> <mailto:jalemkul at vt.edu <mailto:jalemkul at vt.edu>>> wrote:
>>
>>
>>
>> Denny Frost wrote:
>>
>> Here's what I've got:
>>
>> M E G A - F L O P S A C C O U N T I N G
>>
>> RF=Reaction-Field FE=Free Energy SCFE=Soft-Core/Free
>> Energy
>> T=Tabulated W3=SPC/TIP3p W4=TIP4p (single or
>> pairs)
>> NF=No Forces
>>
>> Computing: M-Number
>> M-Flops % Flops
>>
>> -----------------------------------------------------------------------------
>> Coul(T) + VdW(T) 1219164.751609
>> 82903203.109 80.6
>> Outer nonbonded loop 25980.879385
>> 259808.794 0.3
>> Calc Weights 37138.271040
>> 1336977.757 1.3
>> Spread Q Bspline 792283.115520
>> 1584566.231 1.5
>> Gather F Bspline 792283.115520
>> 4753698.693 4.6
>> 3D-FFT 119163.856212
>> 953310.850 0.9
>> Solve PME 2527.465668
>> 161757.803 0.2
>> NS-Pairs 47774.705001
>> 1003268.805 1.0
>> Reset In Box 371.386080
>> 1114.158 0.0
>> Shift-X 24758.847360
>> 148553.084 0.1
>> CG-CoM 1237.953600
>> 3713.861 0.0
>> Angles 18569.135520
>> 3119614.767 3.0
>> Propers 14855.308416
>> 3401865.627 3.3
>> Impropers 3094.855920
>> 643730.031 0.6
>> Virial 1242.417375
>> 22363.513 0.0
>> Stop-CM 1237.953600
>> 12379.536 0.0
>> P-Coupling 12379.423680
>> 74276.542 0.1
>> Calc-Ekin 12379.436160
>> 334244.776 0.3
>> Lincs 11760.476208
>> 705628.572 0.7
>> Lincs-Mat 245113.083072
>> 980452.332 1.0
>> Constraint-V 23520.928704
>> 188167.430 0.2
>> Constraint-Vir 11760.452496
>> 282250.860 0.3
>>
>> -----------------------------------------------------------------------------
>> Total
>> 102874947.133 100.0
>>
>> -----------------------------------------------------------------------------
>>
>>
>> R E A L C Y C L E A N D T I M E A C C O U N T
>> I N G
>>
>> Computing: Nodes Number G-Cycles
>> Seconds %
>>
>> -----------------------------------------------------------------------
>> Neighbor search 1 99195 8779.027 3300.3
>> 3.8
>> Force 1 991941 188562.885
>> 70886.8 81.7
>> PME mesh 1 991941 18012.830 6771.6
>> 7.8
>> Write traj. 1 41 16.835
>> 6.3
>> 0.0
>> Update 1 991941 2272.379
>> 854.3 1.0
>> Constraints 1 991941 11121.146
>> 4180.8 4.8
>> Rest 1 2162.628
>> 813.0 0.9
>>
>> -----------------------------------------------------------------------
>> Total 1 230927.730
>> 86813.1 100.0
>>
>> -----------------------------------------------------------------------
>>
>> -----------------------------------------------------------------------
>> PME spread/gather 1 1983882 17065.384
>> 6415.4 7.4
>> PME 3D-FFT 1 1983882 503.340
>> 189.2
>> 0.2
>> PME solve 1 991941 427.136
>> 160.6 0.2
>>
>> -----------------------------------------------------------------------
>>
>> Does that mean it's only using 1 node? That would
>> explain the
>> speed issues.
>>
>>
>> That's what it looks like to me.
>>
>>
>> -Justin
>>
>> -- ========================================
>>
>> Justin A. Lemkul
>> Ph.D. Candidate
>> ICTAS Doctoral Scholar
>> MILES-IGERT Trainee
>> Department of Biochemistry
>> Virginia Tech
>> Blacksburg, VA
>> jalemkul[at]vt.edu <http://vt.edu> <http://vt.edu> | (540)
>>
>> 231-9080
>>
>> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>>
>> ========================================
>> -- gmx-users mailing list gmx-users at gromacs.org
>> <mailto:gmx-users at gromacs.org>
>> <mailto:gmx-users at gromacs.org <mailto:gmx-users at gromacs.org>>
>>
>>
>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>> Please search the archive at
>> http://www.gromacs.org/Support/Mailing_Lists/Search before
>> posting!
>> Please don't post (un)subscribe requests to the list. Use the
>> www
>> interface or send it to gmx-users-request at gromacs.org
>> <mailto:gmx-users-request at gromacs.org>
>> <mailto:gmx-users-request at gromacs.org
>> <mailto:gmx-users-request at gromacs.org>>.
>>
>> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>
>>
>>
>> -- ========================================
>>
>> Justin A. Lemkul
>> Ph.D. Candidate
>> ICTAS Doctoral Scholar
>> MILES-IGERT Trainee
>> Department of Biochemistry
>> Virginia Tech
>> Blacksburg, VA
>> jalemkul[at]vt.edu <http://vt.edu> | (540) 231-9080
>> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>>
>> ========================================
>> -- gmx-users mailing list gmx-users at gromacs.org
>> <mailto:gmx-users at gromacs.org>
>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>> Please search the archive at
>> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>> Please don't post (un)subscribe requests to the list. Use the www
>> interface or send it to gmx-users-request at gromacs.org
>> <mailto:gmx-users-request at gromacs.org>.
>> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>
>>
>>
> --
> ========================================
>
> Justin A. Lemkul
> Ph.D. Candidate
> ICTAS Doctoral Scholar
> MILES-IGERT Trainee
> Department of Biochemistry
> Virginia Tech
> Blacksburg, VA
> jalemkul[at]vt.edu | (540) 231-9080
> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>
> ========================================
> --
> gmx-users mailing list gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> Please don't post (un)subscribe requests to the list. Use the www interface
> or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20110128/cf8c0d70/attachment.html>
More information about the gromacs.org_gmx-users
mailing list