Re: Does NAMD have some way to test the result?

From: Jérôme Hénin (jerome.henin_at_ibpc.fr)
Date: Sat Oct 04 2014 - 04:20:57 CDT

It's a good question, in a way... it's the main issue in the whole business.

That was a little like emailing a violin-makers' list and saying: "Hi, I've
been making violins, but they sound horrible. Is there any way to make
violins that sound good?"

And Axel's answer was basically, "Yes: study violin-making".

Cheers,
Jerome

On 4 October 2014 05:17, Axel Kohlmeyer <akohlmey_at_gmail.com> wrote:

> On Fri, Oct 3, 2014 at 10:50 PM, ukulililixl <ukulililixl_at_gmail.com>
> wrote:
> > hello!
> > each time we run the same sample we will have different results,is there
> any
> > way to test whether the result is acceptable?
>
> please update your understanding of MD and particularly statistical
> mechanics.
> MD is chaotic (like any system of coupled partial differential
> equations) and thus trajectories will diverge exponentially, yet they
> will sample the same phase space and converged ensemble averages will
> be the same.
>
> axel.
>
> >
> >
>
>
>
> --
> Dr. Axel Kohlmeyer akohlmey_at_gmail.com http://goo.gl/1wk0
> College of Science & Technology, Temple University, Philadelphia PA, USA
> International Centre for Theoretical Physics, Trieste. Italy.
>
>

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:22:54 CST