[gmx-users] mpich parallel system errors

[Carsten Kutzner]

On Sun, 18 Dec 2005, David van der Spoel wrote:

> Rongliang Wu wrote:
> > Hello yangye,
> >            i have 5 md.log files but just two of them gave data calculated. the other three stop just at the two "thanks"
> > the job ended and don't know whether the results from just 2 nodes are trustable,for the atoms have been divided to seperate nodes, will the two normal nodes calculate the atoms divides to the other three atoms.
> >
> It is the combination MPICH/gromacs that causes problems. Of course we'd
> like to blame MPICH since GROMACS works with LAM and other libraries,
> but on the other hand MPICH works with most software as well.

MPICH-2 works with GROMACS, though LAM still performs slightly
better!

   Carsten


---------------------------------------------------
Dr. Carsten Kutzner
Max Planck Institute for Biophysical Chemistry
Theoretical and Computational Biophysics Department
Am Fassberg 11
37077 Goettingen, Germany
Tel. +49-551-2012313, Fax: +49-551-2012302
eMail ckutzne at gwdg.de
http://www.gwdg.de/~ckutzne