Re: namd-openmosix

From: Nicholas M Glykos (glykos_at_mbg.duth.gr)
Date: Tue May 19 2009 - 11:19:51 CDT

Hi Jose,

> We are trying to compile namd 2.6 on openmosix (
> http://openmosix.sourceforge.net/) using a cluster of 20 nodes, any
> suggestion or information related to it.

We have got openmosix+SGE working with NAMD (and even added a small
paragraph at http://www.ks.uiuc.edu/Research/namd/wiki/index.cgi?NamdOnGridEngine )

The crux of the matter is that you do not want allow openmosix getting
anywhere near NAMD: all NAMD jobs should be locked on their respective
nodes (and behave as if openmosix didn't exist). If you don't do this you
will be spending most of your time migrating jobs around (instead of
calculating). A primitive job submitted via 'at' would look like this:

/bin/mosrun -L /usr/local/NAMD/charmrun /bin/mosrun -L /usr/local/NAMD/namd2 +p9 heat.namd > LOG

Having said these, please do note that (i) openmosix is dead (see the
message at the link you provided), (ii) although I had fun setting-up a
NOW cluster with openmosix in the past, I wouldn't even contemplate using
openmosix in a production environment these days. Get a queue manager and
forget all about openmosix.

My twopence,
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/

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:50:59 CST