Re: CCL:Re: parallel md+mc
On Sat, 12 Sep 1998, Guido Germano wrote:
> Sorry, but you don't need to start your MD run from an *equilibrated*
> configuration. You can use MD itself to reach equilibrium. So the
> question should be: from which configuration start? There are two main
> possibilities: construct a regular crystal-like arrangement and melt it
> rising the temperature, or construct a random arrangement at lower
> density and compress it gradually during the first steps of the run.
> The first approach is easier and good enough for your case.
It really depends on the system; for uncoupled LJ particles with no
connectivity, I'd agree that MC equilibration is probably unnecessary.
For many (bio)polymer systems, especially for more exploratory work
where crystal and/or NMR structures aren't available, some form of
evaluating model built configurations is probably essential, and MC is a
reasonable tool to use.
> On Wed, 9 Sep 1998, Matthias Hloucha wrote:
> > One important task is to construct an equilibrated start configuration
> > before starting the md runs. Do you think it is resonable to use Monte
> > Carlo techniques to equilibrate very many simple particles? Are you
aware
> > of any existing code in this area?
The book "Computer Simulations of Liquids" by MP Allen and DJ
Tildesley
should have what you need for this simple case of a LJ fluid; the
program listings are online in a couple places (but, alas, not at the
ftp site listed in the book). They are on MP Allen's web pages at
http://toucan.phy.bris.ac.uk/AllenTildesley/home.html
--
Rick Venable =====\ |=| "Eschew Obfuscation"
FDA/CBER Biophysics Lab |____/ |=|
Bethesda, MD U.S.A. | \ / |=| ( Not an official statement or
rvenable -AatT- deimos.cber.nih.gov | \ / |=| position of the FDA; for
that,
http://www.erols.com/rvenable \/ |=| see http://www.fda.gov )