AMBER: sander and mpich

From: Thomas Exner <thomas.exner.uni-konstanz.de>
Date: Thu, 29 Jan 2004 09:26:55 +0100

Dear amber users:

I bought a pc cluster to run parallel amber jobs. Before installing the
cluster, I tried different linux distributions on a dual-Xeon
workstation to find one which can be easily been installed on the
cluster. At the same time I tested the performance of amber 7 on these
distributions. But when I tried the test examples of sander (make
test.sander) I ran into problems on some distributions especially SUSE
8.2 and Mandrake 9.2. When I run the examples using more then one
processor sander gives totally different results to the results provided
in the amber tar-file. Especially the temperature and pressure values
for the tip4p and tip5p examples are totally different and meaningless
in my version. The serial and the mpich version using only one processor
work totally fin. I also tried redhat 7.3, 8.0 and 9 and an older
mandrake version and there where no problems at all. Even using the
amber version compiled on redhat 9 on the SUSE distribution worked but
not on mandrake 9.2.

Has anyone else experienced these problems and is there a way around
them expect using only the serial version? Thank you very much.

Best wishes.
Thomas


-- 
________________________________________________________________________________
Dr. Thomas E. Exner
Juniorprofessur (des.) "Theoretische Chemische Dynamik"
Fachbereich Chemie
Universität Konstanz
78457 Konstanz
Tel.: +49-(0)7531-882015
Fax:  +49-(0)7531-883587
Email: thomas.exner.uni-konstanz.de
________________________________________________________________________________
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Fri Jan 30 2004 - 18:56:03 PST
Custom Search