Re: Problems with the nodelist file !

From: Edward Patrick Obrien (edobrien_at_Glue.umd.edu)
Date: Wed Mar 02 2005 - 17:22:02 CST

Hi Parag,
   This has always worked for me:

1. In file called "nodelist", located in working directory:

group main
   host n1
   host n1
   host n2
   host n2

2. to run

charmrun namd2 ++nodelist nodelist +p4 file.namd >& file.log

3. If running with PBS queing system, I also specify the nodes:

> qsub -l nodes=n1+n2 file.pbs

Note on step 1:
   This example runs on 2 nodes using 4 processors (2 cpu's per node).
It is necessary to duplicate the computer name to specify that 2 cpu's
will be used on one node.
   If you have 1 cpu per node, but still running parallel, the nodelist
file should instead contain:

group main
   host n1
   host n2

and step 2 would be:

charmrun namd2 ++nodelist nodelist +p2 file.namd >& file.log

Hope that helps,
Ed

On Wed, 2 Mar 2005, Parag Adhangale wrote:

>
> This mailing list is of great help. Thanks for all the
> responses.
>
> I tried running NAMD using charmrun by setting up a
> node list file and using the ++nodelist option.
> Charmrun however, seems to ignore it and goes to the
> nodes
> that are up. Is there a way to force charmrun to only
> use nodes specified in the nodelist file ?
>
> Thanks
>
> Parag
>
>
>
>
> __________________________________
> Celebrate Yahoo!'s 10th Birthday!
> Yahoo! Netrospective: 100 Moments of the Web
> http://birthday.yahoo.com/netrospective/
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 05:18:36 CST