[gmx-users] LJ simulation

[Berk Hess]

>From: Manoj <justkewlguy at yahoo.com>
>Reply-To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>To: gmx-users at gromacs.org
>Subject: [gmx-users] LJ simulation
>Date: Thu, 6 Jan 2005 13:25:41 -0800 (PST)
>
>Hi,
>
>I am simulating a simple LJ fluid at a particular
>state point (NVT run) and I am using the option
>'vdwtype = Shift'. Now, whether I specify
>'rvdw_switch' or not drastically changes the dynamics
>I get. The pressure obtained in both the cases is
>different but still not accurate and its closer to
>accurate if rvdw_switch is specified. On the other
>hand, when I use the option 'vdwtype=cutoff' along
>with 'DispCorr = EnerPres' (without specifying
>rvdw_switch), the pressure I get is pretty accurate
>and compares very well with the Monte Carlo result I
>already have.
>
>My questions are:
>
>1. What is 'rvdw_switch' fundamentally and what does
>it actually do?

This is r_1 on page 54 of the manual.
Below r_1 the potential is purely shifted and the forces are unmodified

>2. When Shift potentials are used, there is no long
>range correction, right? The log file shows Dispersion
>correction to be 0.0

Indeed.
The next release will support this, it is implemented in the CVS version.

>3. Does gromacs code actually correct the shifted
>potential back to true potentials and if it does then
>why are the results so different for cut-off and Shift
>options?

I don't really understand this question.
Maybe things will become clearer when you read the manual.

>4. If rvdw_switch is so important, what is the optimum
>value for it?

When rvdw is sufficiently large, rvdw_switch should have little
influence on the results. (I don't know what the sigma's of
your particles are.)

>5. Is there any standard ideal run input file which
>can be used for pure LJ simulation for all kinds;
>cut-off, cut-off and shifted ?

No.

Berk.

_________________________________________________________________
Talk with your online friends with MSN Messenger http://messenger.msn.nl/