Re: [AMBER] mmpbsa restart

From: Jason Swails <jason.swails.gmail.com>
Date: Wed, 14 Nov 2012 11:33:09 -0500

On Wed, Nov 14, 2012 at 11:14 AM, Giulia <giulia.palermo.iit.it> wrote:

> Dear Jason,
>
> Thank you very much for your quick reply.
> I am trying to ask 12 processor for staring.
>
> However my calculations are crashing.
> I am asking for 12 procs in the submission file.
> Should I ask the same number of processors in the mmpbsa.in (input) file??
> And, if yes, in which format??
>

MMPBSA.py is run in parallel the same way that sander and pmemd are run in
parallel. That is, you need to make sure that MMPBSA.py.MPI is properly
installed, and you would run it with a command like:

mpiexec -hostfile $PBS_NODEFILE MMPBSA.py.MPI [normal options]

Of course, running MPI programs depends on the queuing system you use (this
one assumes Torque/PBS) and the MPI you use (this one assumes mpiexec takes
a "-hostfile" argument).

Your sysadmin should be able to help you run an MPI program on the cluster.

Good luck,
Jason

-- 
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Nov 14 2012 - 09:00:02 PST
Custom Search