AMBER Archive (2007)

Subject: Re: AMBER: Slow Processor Loads when Using PMEMD

From: David A. Case (case_at_scripps.edu)
Date: Mon Jul 02 2007 - 20:24:09 CDT


On Mon, Jul 02, 2007, Jonathan Suever wrote:
>
> When I submit this job to the cluster, it runs fine with no errors and shows
> that all 16 processors are currently being used. However when the detailed
> status is viewed, it can be seen that the highest load placed on any of the
> processors is around 0.20 resulting in very slow calculation times.

Are you *sure* the calculation is really slower, or are you just interpreting
the output of "top" or some similar command?

As you seem to suspect, there is no obvious reason why a restarted calculation
should be slower than a previous one, especially if the differences are
dramatic.

Try some short tests (a few thousand steps) with both sets of inputs, and see
if there really are reproducible differences in the actual elapsed times. You
might see if the density has changed a lot between the first run and the
second -- that might lead to the behavior you describe.

...dac

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu