On Wed, Apr 14, 2004, Lubos Vrbka wrote:
>
> what (and how often) should i carefully inspect?
Different MD implementations generally match one another very closely
for about 100 steps or so; after that, there are inevitable divergences.
I would set ntpr=1 for about 100 steps, and just diff the files -- that's
what the test cases do.
The "benchmarks" are set up to do performance testing, and the number of
steps was chosen to obtain a convenient timing. Of course, what is
"convenient" changes as CPU's get faster and faster. The benchmarks are not
designed to look for program correctness.
You could also try to run very long (converged) simulations and compare the
average behaviors. Since you are changing the pressure algorithm, looking
at volumes and densities is probably the most important thing to check.
Given the very small changes you are making to pmemd, I would think there is a
good chance you can convince yourself things are OK by some combination of
these sorts of tests.
....dac
--
==================================================================
David A. Case | e-mail: case.scripps.edu
Dept. of Molecular Biology, TPC15 | fax: +1-858-784-8896
The Scripps Research Institute | phone: +1-858-784-9768
10550 N. Torrey Pines Rd. | home page:
La Jolla CA 92037 USA | http://www.scripps.edu/case
==================================================================
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Wed Apr 14 2004 - 18:53:00 PDT
