[gmx-users] Brownian dynamics
michel.fodje at mbfys.lu.se
Fri Feb 22 14:16:43 CET 2002
I wish all of you for all your help with sorting out my problems with
brownian dynamics. The suggestion that worked was the following
On Wed, 2002-02-20 at 12:33, Berk Hess wrote:
> The timestep by itself says nothing in brownian dynamics.
> What matters is dt/bd_fric.
> In your case bd_fric = 0, which means that mass/tau_t is
> used as bd_fric. your tau_t is 0.1 which is very small,
> so you'll need a very small timestep.
> So there are three options:
> 1. make the timestep much smaller
> 2. increase tau_t (mass dependent friction coefficient)
> 3. set bd_fric to some reasonable value (3000 or such)
More information about the gromacs.org_gmx-users