[gmx-developers] implicit solvent free energy

[Erik Lindahl]

Hi David,

Let me phrase it this way: We would certainly be happy to make the GB code very efficient, and make sure that it works well in domain decomposition too, if other people/developers are interested in doing the work to make sure it results in accurate free energies, test and tune how different GB implementations work with screening, and in general help provide a clear set of cases where it is obviously useful.

The biggest problem with GB is that it is difficult to get it efficient for standard simulations where you want long trajectories. Part of this comes from the current efficiency of explicit-solvent, but the main issue is that GB by definition will never be able to scale as efficiently as explicit solvent simulations (both due to spatial inhomogeneity, longer-range interactions, and several communication and separate kernel executions during force evaluation where things cannot be perfectly load-balanced). Unfortunately, this will only get worse the more you parallelize, and that seems to be the road ahead.

The one place I see where GB might be competitive would be rapid screening applications, in particular when you can stay on a single node for each compound.

So, to sum it up: If somebody would be willing to help discuss the equations we want to implement, and assist with all the testing, I'm sure we could find an interesting joint project where we get it implemented!

Cheers,

Erik





On Aug 30, 2013, at 8:52 AM, Gerrit Groenhof <ggroenh at gwdg.de> wrote:

> Hi David,
>> 
>> In any case:
>> - Are other people similarly interested in this?
> 
> 
> Not in implementing, but having the GB code working in combination with free energy could be of interest when one wants to address the effects of periodicity and background charge on the free energy. Just not sure if tis would be worth the effort of maintaining the GB code, but I might use it.
> 
> Best,
> Gerrit
> 
>> - Are there any efforts underway to get the GB code to work properly with free energy?
>> - What are the problems/obstacles? 
>> 
>> Thanks!
>> David
>> 
>> 
>> -- 
>> David Mobley
>> dmobley at gmail.com
>> 949-385-2436
>> -- 
>> gmx-developers mailing list
>> gmx-developers at gromacs.org
>> http://lists.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.
> 
> -- 
> gmx-developers mailing list
> gmx-developers at gromacs.org
> http://lists.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.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-developers/attachments/20130830/073e0c9d/attachment.html>