[gmx-users] low CPU load percentages on 16 core box with CentOS4.4

[David van der Spoel]

Makoto Yoneya wrote:
> Dear Mark:
> 
> Thank you for your quick comment.
> 
> 
>>Well this looks like a classic problem of incompatibility of your
>>hardware and these lam versions that is not arising for MPICH. You
>>should look for help there, since the problem is not arising because of
>>gromacs - unless you can demonstrate lam works correctly on your
>>hardware for other parallel software.
> 
> 
> OK, I should check our installation of lam itself as suggested.
> 
> 
>>>However, the test simulation results looked different (e.g. densities of
> 
> the
> 
>>>system) with lam and mpich2 (I should check more).
>>>How can I solve this situation?
> 
> 
>>It sounds like gromacs+MPICH is working fine...
> 
> 
> Then, MPICH (same as MPICH2?) is safer to use (lam is dangerous) for a new
> hardware?
MPICH version 1.x has always given trouble with gromac, whereas LAM 
usually works fine. MPICH v 2.x should be OK though. The difference 
between LAM and MPICH in load could also be problem related. MPICH2 is 
somewhat better optimized.

Nevertheless, things like average density should be independent of the 
MPI library and number of processors. So please, if you find 
inconsistencies there report them back to the list, or,preferably, to a 
bugzilla.
-- 
David.
________________________________________________________________________
David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,
Dept. of Cell and Molecular Biology, Uppsala University.
Husargatan 3, Box 596,  	75124 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://folding.bmc.uu.se
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++