Re: NAMD2.7b1 performance

From: Axel Kohlmeyer (akohlmey_at_gmail.com)
Date: Fri Sep 25 2009 - 18:07:14 CDT

On Fri, Sep 25, 2009 at 5:53 PM, Nicholas M Glykos <glykos_at_mbg.duth.gr> wrote:
>
> Ooops, I erroneously ignored the CPU count from the first part of the log
> you've showed ...
>
>> Info: Benchmark time: 86 CPUs 0.0318438 s/step 0.184281 days/ns 70.0844 MB memory
>
> .. in favor of the second part:
>
>> Charm++> cpu topology info is being gathered!
>> Charm++> 17 unique compute nodes detected!
>
> Apologies. 86 is indeed 86 ;-)

yes, but that raises a different issue. to get to 86 there have to
be nodes with different counts of cpu cores within the machine
being used. and while NAMD does have the ability to do load
balancing, its efficiency should decrease with increasing number
of cpu cores being used (since there are less tasks to distribute
across the cores).

in any case, there is nothing but running systematic benchmarks
that can tell how well this all works out. using a homo-/heterogeneous
set of nodes is just adding to the complexity of the tests needed.

cheers,
    axel.
>
> Nicholas
>
>
>
> --
>
>
> Dr Nicholas M. Glykos, Department of Molecular Biology
> and Genetics, Democritus University of Thrace, University Campus,
> Dragana, 68100 Alexandroupolis, Greece, Tel/Fax (office) +302551030620,
> Ext.77620, Tel (lab) +302551030615, http://utopia.duth.gr/~glykos/
>
>

-- 
Dr. Axel Kohlmeyer    akohlmey_at_gmail.com
Institute for Computational Molecular Science
College of Science and Technology
Temple University, Philadelphia PA, USA.

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:53:18 CST