[gmx-users] initial velocities Langevin dynamics

[Mark Abraham]

Loris Moretti wrote:
> Hello everybody,
> 
> Sorry to post again the same question but I've got no answer so far.
> 
> I have a technical question, it might be silly but I couldn't find an answer so far. How are assigned the initial velocities for the particles when Langevin dynamics is performed?
> 
> In normal MD the gen_vel/gen_seed are used for random generation of the initial velocities. I am used to vary this number with the same system in order to explore the effect of different initial conditions. I was wondering if there is such a possibility with Langevin dynamics, like changing the ld_seed number?

As you'll see in the manual (section 7.3.4) ld_seed is not used for 
generating velocities. I expect gen_vel and gen_seed do this, since the 
purpose of generating initial velocities doesn't change with the 
integrator. You're still wanting to sample a Maxwell distribution, no?

You could have experimented with gen_seed and used gmxcheck to compare 
pairs of .tpr files to answer your own question.

Mark