[gmx-users] Limitations of simulations?

Justin Lemkul jalemkul at vt.edu
Thu Jul 25 03:17:40 CEST 2013

On 7/24/13 9:16 PM, Jonathan Saboury wrote:
>> For the rest of us mere mortals who don't have access to specialized hardware
>> that allows for 10- or 20-microsecond simulations, the brute force approach is
>> rather futile.  Techniques like steered MD and Hamiltonian replica exchange MD
>> are probably more feasible.  Unbiased simulations of sufficient length, using
>> "standard" supercomputing hardware, would probably take years.
>> -Justin
> Do you think implicit water with 10-20 ligands would be able to do it?
> I see that the gpu accelerated implicit gets about 100-150 ns/day so a
> 10-20 day simulation would yield that time frame, no?
> Thanks :)

In an implicit, non-periodic system, it is more likely that the ligand will 
float away from the protein.  I've tried it and that's all that ever happens. 
Moreover, the current Gromacs version does not support implicit solvent on GPU 
and the previous version that did had very limited functionality.



Justin A. Lemkul, Ph.D.
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


More information about the gromacs.org_gmx-users mailing list