[gmx-users] microcanonical

[Mark Abraham]

karamyog singh wrote:
> I almost removed everything. I am generating initial velocity at 0.01 K 
> so the system should freeze. Its a dilute gas. But somehow I am getting 
> very large velocites i.e. very large KE of the whole system.

That .mdp looks fine for NVE. Use g_energy -w to see the evolution of KE 
over time. It is well known that the use of cut-offs induces heating in 
the system - i.e. energy is not being conserved. Yours are much longer 
than you would use in protein-in-water MD, which is a step in the right 
direction, but I don't know how long they are for the system you're 
looking at - literature may help here. You would want cut-offs that get 
more than a few atoms into the interaction sphere, of course, and that 
depends how dilute the gas is.

Another possibility is that the system is not equilibrated straight 
after grompp does the gen_vel - it's just sampling velocities from a 
distribution, so the result is stochastic. I'd suggest doing a period of 
NVT and/or NPT to get the temperature and density correct for the system 
you want, and only then switch off the regulation algorithms to get NVE.

Mark