[gmx-developers] Testing an integrator in MD

[David van der Spoel]

Nick Schafer wrote:
> Dear Gromacs Developers -
> 
> Thanks for your responses regarding my question about implementing a new 
> integrator in GROMACS and other MD codes.  My follow up question is 
> this: If you were reading a paper that was introducing a new integrator, 
> what kind of tests/data/analysis on what kind of systems would you need 
> to see in order to decide about its validity and usefulness as a general 
> purpose MD integrator?  As you can probably tell, I'm looking forward to 
> designing and running these experiments, but I thought I had better ask 
> the people who would potentially be using the results, which results in 
> particular they would like to see, first, before I went off and did 
> something that was irrelevant/unhelpful on accident.  I'd be open to 
> biological, as well as materials, examples.  I've heard many people say 
> energy conservation is important.  Is that always true?  What other 
> parameters/phenomena should I be monitoring? Please try to be as 
> specific as possible.

if you are going to implement this in a large package like gromacs, you 
first want to test energy conservation. Therefore you need to construct 
a test system that does not have any other approximations. The easiest 
(and maybe only) system that does that is an cluster of molecules in 
vacuum without cut-offs. For such a system you should test energy 
conservation as a function of integration time step in both single and 
double precision.

We find that already with the current integrator we get very good energy 
conservation in DP, but not good enough in single. Note that we have 
performed simulations of > 1 billion 2 fs  time steps with very good 
energy conservation (in DP).

Hence, if you can develop an integrator that does better in SP we could 
in principle save half the CPU time by switching from DP to SP.
> 
> Thanks in advance for any advice.
> 
> Nick Schafer
> http://sbel.wisc.edu/People/schafer/mainframeset.html
> 
> _______________________________________________
> gmx-developers mailing list
> gmx-developers at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-developers
> Please don't post (un)subscribe requests to the list. Use the www 
> interface or send it to gmx-developers-request at gromacs.org.


-- 
David.
________________________________________________________________________
David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,
Dept. of Cell and Molecular Biology, Uppsala University.
Husargatan 3, Box 596,  	75124 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://folding.bmc.uu.se
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++