[AMBER] Problem related to the parallel installation of Amber11

From: Sindrila Dutta banik <sindrila.duttabanik.yahoo.com>
Date: Tue, 2 Oct 2012 04:21:16 +0800 (SGT)

OK! I will check it out.


Can I ask another technical question?


Which one is more important for the performance of a system, number of core of a processor or the speed of the processor? Or it depends on the job which I want to perform. Sorry for asking some technical question.


Sindrila 



________________________________
 From: Jason Swails <jason.swails.gmail.com>
To: Sindrila Dutta banik <sindrila.duttabanik.yahoo.com>; AMBER Mailing List <amber.ambermd.org>
Sent: Tuesday, 2 October 2012 1:40 AM
Subject: Re: [AMBER] Problem related to the parallel installation of Amber11
 
On Mon, Oct 1, 2012 at 3:57 PM, Sindrila Dutta banik <
sindrila.duttabanik.yahoo.com> wrote:

> Thanks once again.
>
> It now working properly.
>
> Again I want to know a technical question. We have a HP DL980 server with
> 4 CPU and octacore processor. That means total number of processor 32
> (4x8=32). So when I run a job in this system what command should I use to
> get maximum efficiency? Is the following command is correct for such a
> system or not:
>
>
> mpirun -np 32 sander.MPI
>

That depends on your simulation.  The easiest way to check this out
yourself is to run short tests with different numbers of processors and see
which one finishes fastest.

HTH,
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
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Oct 01 2012 - 13:30:02 PDT
Custom Search