AMBER Archive (2005)

Subject: Re: AMBER: meaningful tax for sander question

From: David A. Case (case_at_scripps.edu)
Date: Tue May 31 2005 - 14:59:20 CDT


On Tue, May 31, 2005, Cordova, Luis E. wrote:
>
> I am trying to benchmark/tax an experimental processor running Amber8,
> in particular the sander link.
>
> Does anyone know/could point me to other meaningful examples or
> benchmarks that tax the sander link significantly?
>
> One example coming with the
> source code of Amber is /benchmarks/dhfr but it takes only seconds to
> run on a single processor and there is no documentation for how to make
> it bigger or how to extend it or if it is meaningful to the
> biochemistry/&c. community.

There are a number of benchmarks that come with Amber: jac (23,000 atoms),
hb (about 40,000 atoms), factor_ix (about 91,000 atoms), rt (about 140,000
atoms.) Any of these can be made to run longer by simply increasing the
number of md steps.

The "dhfr" benchmark only differs in minor ways from jac. Take a look at the
benchmarks link at the Amber web site for more information. Also, note that
for solvated periodic simulations, pmemd will (always or almost always)
outperform sander, and is the recommended program for problems that it can
handle.

...dac

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