On 27.02.2015 07:02, Axel Arnold wrote:
(...)
Note also that you need to be careful when adding particles so often.
The Langevin thermostat needs a while to establish the desired
temperature, namely roughly 1/gamma/dt time steps. So, for gamma=1 and
dt=0.01 you need about 100 steps to “heal” the overall temperature. You
can accelerate that by increasing gamma, however, also the product of
gamma and dt can’t be too large. 20 time steps is already critical, and
10 time steps is from my experience not enough to get the temperature
correct.
I'm sorry for off-topic but that sounds very interesting. So it is possible to calculate how fast Langevin thermostat will bring system to desired temperature? Do you know any reference about that?
Well, of course that depends on how far you are off equilibrium. But what gamma (or gamma/m, if you have masses compiled in) tells you is the relaxation time of the resulting dynamics. That is, you can determine gamma from the decay constant of the velocity autocorrelation function. I don’t know any reference on that (nor on the Langevin thermostat, actually), but it is immediately obvious from the Langevin equation, which the Langevin thermostat approximates.
As Peter already pointed out, that does not guarantee that the system has equilibrated yet, but before the velocity autocorrelation has decayed, there is no chance that the thermostat has equilibrated even a single degree of freedom.
Best,
Axel