[gmx-users] Haphazard Results with LAM-MPI

[Dallas Warren]

>We are a bunch of undergraduate students and have put up a cluster of 16 
>IBM's using LAM for the purpose of running GROMACS. During our simulation 
>test runs we noticed that the running time increases as the number of 
>nodes in the LAM topology increases. We couldn't diagonise the 
>problem...plz help.

How many atoms in this simulation?

If you aren't already doing so, would be best to use the benchmark 
simulation files on the web site for running your benchmarks, unless there 
is something you have of particular interest to you.

If there is a small number, then it isn't going to scale very well at all 
since the system will spend most of the time trying to communicate with the 
other nodes.  Even with a large system I have found that GROMACS scales 
poorly beyond a single box due to the slower speed of communication over 
the networks.  You can see this by looking at the benchmark page on the web 
site too.

Catch ya,

Dr. Dallas Warren
Research Fellow
Department of Pharmaceutical Biology and Pharmacology
Victorian College of Pharmacy
Monash University
381 Royal Parade
Parkville VIC 3010
dallas.warren at vcp.monash.edu.au
+61 3 9903 9083
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20030409/d7a3fab9/attachment.html>