Re: [ESPResSo-users] Changes in the Langevin thermostat

[Joost de Graaf]

Dear Paul,

I understand it must be frustrating that the behavior has changed, but there were good reasons for this. I believe Georg is right, but please write down the equations you want to solve, then identify the friction term and use that. ESPResSo is rather merciless when it comes to unit checking (it simply doesn't) so it's up to you to make sure that this is taken care of properly.

Best Wishes,

Joost

On 18 November 2015 at 16:41, Georg Rempfer <address@hidden> wrote:

I rather think that

    gamma_new = gamma_old * m

That is because the input parameter gamma_old denoted the inverse relaxation time (1/T), where 1/T=gamma/m with gamma the actual friction coefficient as in the Langevin equation

    d^2 x / dt^2 = -gamma/m * dx/dt + R

Espresso now internally divides by m, so that gamma_new=gamma.

On Wed, Nov 18, 2015 at 4:15 PM, Paul Peterson <address@hidden> wrote:

Thanks Georg, guess that's the way. 

Then, the new gamma coefficient for a particle of mass m should be (m / gamma_OLD), right?

Paul

2015-11-18 16:17 GMT+01:00 Georg Rempfer <address@hidden>:

Hello Paul,

you might want to check out the LANGEVIN_PER_PARTICLE feature in the User's Guide.

Greetings,

Georg

On Wed, Nov 18, 2015 at 2:59 PM, Paul Peterson <address@hidden> wrote:

Hello everybody,

the 30 october version of the User Manual warns that the meaning of the gamma parameter for the Langevin thermostat has been changed from a relaxation time, as it used to be, to a friction coefficient. 

Can anybody tell me what is the conversion formula from old to new parameters?

Reading the paragraph, I guess the mass's particle is involved; but what if I have two types of particles with different masses?

Thanks, best

Paul