[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ESPResSo] setmd skin

From: Mikheil Azatov
Subject: Re: [ESPResSo] setmd skin
Date: Thu, 24 Jun 2010 14:38:56 -0400


Thanks for explaining Olaf,

the reason I asked is because in my simulation something weird is happening when I change the time_step. So I have a system of polymers in langevin thermostat. At some time_steps the sytem behaves normally and all the particles experience just thermal motion. When I increase the time_step after some point the system basically explodes. And this transition is very sharp. For example for time-step 0.012 particles move slowly due to thermal motion and for 0.013 the system basically explodes and espresso terminates itself because the particles have such big coordinates. I thought before that it might be because of the skin parameter but after your email and trying different parameters for skin it's obviously not because of it. The reason I change the time step is to make the code run faster. For small time-steps the particles move just very slow and it doesn't make any sense to keep time-step that small.
If anyone had similar problems or has a guess of why could I have them I would really appreciate your response.

On Thu, Jun 24, 2010 at 3:37 AM, Olaf Lenz <address@hidden> wrote:

On 06/23/2010 11:45 PM, Mikheil Azatov wrote:
> Does anyone know what exactly does setmd skin set ? I'm trying to think
> what should it be equal to for my simulation but I cannot figure it out
> what exactly is it. In the manual it says skin for the Verlet list but
> if anyone can say something more I would really appreciate it .

The most important information first: no matter what value for skin you
choose, the *results* of the simulation will always be the same.
However, the speed may change dramatically. Usually, there is an optimal
skin value that minimizes the computation time, which depends on your
Typically, the optimum should be somewhere between 0.1 and 5 times the
maximal interaction range for typical systems. To get the optimum, you
just have to try out.

As for what exactly the skin does, read a description of Verlet lists
for example in Frenkel, Smit "Understanding Molecular Simulations".

Best regards
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?

ESPResSo mailing list

reply via email to

[Prev in Thread] Current Thread [Next in Thread]