jalemkul at vt.edu
Tue Dec 3 15:25:38 CET 2013
On 12/3/13 2:17 AM, kiana moghaddam wrote:
> Dear GMX Users
> I want to run NVT equilibration and NPT equilibration (after NVT) and want to increase temperature gradually i.e. from 0 to 300 K over 100 ps, but I have some questions about this process:
> when I use simulated annealing in NVT as following
> annealing_time = 0 15 30 45 60 80 100
> annealing_temp = 0 50 100 150 200 250 300
> I don't know how to change gen_temp. Should gen_temp be 0 or 300 K?
Generating velocities and then immediately freezing (in the strictest of sense)
the system is likely not going to be stable or effective because the thermostat
will go haywire trying to compensate for such a drastic change in temperature.
It is an interesting question, because how does one generate velocities at 0 K?
Theoretically, there are none. I would suggest setting gen_temp to 0; you can
check the .tpr (via gmxdump) to see if any velocities are present, but even if
they are they should be very small and largely irrelevant. The annealing
protocol will warm the system relatively quickly, so even after a few dynamics
steps, you will have small velocities within the system.
> At the first time, I did not use simulated annealing. I prepared 7 mdp files that in the first mdp file, I set nsteps=0, gen_temp=0, ref_t=0, then I use the output from first NVT equilibration for the second input. I repeat this to reach 300 K (in the second step nsteps=7500(*0.002=15), gen_temp=50, ref_t=50). I 'm not sure whether this process is correct?
Well, it works, but it's laborious and unnecessary because you're just manually
doing what simulated annealing is doing.
Justin A. Lemkul, Ph.D.
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