[gmx-users] simulation acceleration of SWM4-NDP water and ions

[Justin Lemkul]

On 8/15/15 9:13 AM, 折晓会 wrote:
> Dear all,
> I am using polarizable water model (SWM4-NDP) and ions in my simulation. However, the simulation speed is really slow, and it will take several days to get the simulation results.
> I know the SWM4-NDP only supports 1 MPI thread at the moment. I am wondering if SWM4-NDP supports GPU acceleration, or is there other method to improve the simulation speed? I plan to upgrade my current workstation to speed up the simulation.
> Any help will be appreciated!

Please see http://dx.doi.org/10.1002/jcc.23937

As a result of that implementation, I also made SCF compatible with domain 
decomposition, so it is much faster than it was before.  The extended Lagrangian 
is 3x faster than the SCF, but only supports NVT at this point.

You can get the force field files and instructions for how to get the code from 
http://mackerell.umaryland.edu/charmm_drude_ff.shtml

There are no tutorials available (lack of time), but for a simple system like 
water and ions, the example inputs in the SI of the paper are all you should need.

-Justin

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

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

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

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

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