[gmx-users] Help Energy Diferences

[Justin Lemkul]

On 7/21/14, 1:30 PM, ANDRES ADOLFO ORTEGA GUERRERO wrote:
> Hi Dear Gromacs Users,
>
> I have been dealing with a problem, i run an MD of my sistem , and i see
> problems with my energies due to one of my molecules is a CNT and has no
> charges and its Coul SR are huge, well i though this could be of an error
> in my NDX file ,
> the thing is that i run my MD in my laptop (Mac OS) for really short time
> and there were no mistake, and when i do the MD in my Work (2 Linux
> Computer with GPU Nvidia) with the same files, my energy are totally
> different,
> the version is Gromacs 5-0 rc1,
>
> Could you help me with this?
>

IIRC multiple energygrps are not supported on GPU, so whatever the .edr file 
thinks is CNT-CNT or whatever else is just the energy of the whole system.  This 
is a known issue.  If you do a run on GPU, don't use multiple energygrps, just 
post-process the trajectory with mdrun -rerun and a .tpr file that has the 
desired groups.

-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

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