[gmx-users] hints for core/shell optimization?

Justin Lemkul jalemkul at vt.edu
Fri Jul 25 20:01:27 CEST 2014



On 7/25/14, 12:40 PM, Tamas Karpati wrote:
> Dear Justin,
>
> Thanks for your educational answers.
>
>> Coulombic interactions fail at short distances; you probably need to apply
>
> I was afraid of that... somehow removing shells from the cores in the
> initial structure have let it "functionally" work (with not yet
> reasonable results).
>
>> Thole screening to avoid polarization catastrophe.  Ions are particularly
>> problematic in this regard.
>
> I've seen this mentioned in the Manual, but hadn't ever hit GROMACS
> specific details on how to apply polarization in the input files.
> The only source I could locate is within the GROMACS package,
> under the name "sw.itp". It does exclusively implement the
> so called "water polarization" model -at least I think so.
>

The "water polarization" function is a water-specific anisotropy function. 
Don't try to use it for anything else; the interpretation of the atom numbers 
for local axis construction are very specific.

> Can you please direct me to a source from which I could learn
> how to polarize GROMACS? I was'not lucky on the Internet and,
> indeed, even at the GROMACS site or Manual (neighter application
> examples nor file format descriptions).
>

The Thole screening function is (in the released version) not used by anything, 
so it's not documented.  In its present incarnation, you need a 
[thole_polarization] directive that lists atom-shell/Drude pairs as follows:

atom_i shell_i atom_j shell_j 2 a alpha_i alpha_j

The "2" is a required function type.  My implementation of the CHARMM Drude FF 
is nearly done, and there are changes to the way the Thole directive is laid out 
in the future, but at the moment (up through version 5.0), this is the way it 
works.  The code is in src/gromacs/gmxlib/bondfree.c.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================


More information about the gromacs.org_gmx-users mailing list