Variable results for the same simulation

From: Nd S (navdeep79_at_gmail.com)
Date: Tue Feb 03 2009 - 12:18:24 CST

Hi all

I am running a simulation on NAMD installed on SGI Altix 3700 and the OS is
IA-64 version of Linux. I ran the same simulation in different queues (on 4
processors, 8 processors, and 32 processors), with same input and
configuration files, and to my surprise the results vary by large amount.
The results on the P4 queue are totally unrealistic. Also my simualtion
failed to run on the 16 processors queue. I recieved some memory error.
Presently, I am running the simuation running on a single processor to
compare the results, but it will take a very long time.

Can anyone please explain why it is so. The simualtion system consists of
nanotube inside a water box.

Thanks
Navdeep

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:52:19 CST