[gmx-users] a) rlist vs rvdw/rcoulomb size confusion, and b) reduced units

[Berk Hess]

>From: "Frankie Montenegro" <frankie.montenegro at gmail.com>
>Reply-To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>To: "Discussion list for GROMACS users" <gmx-users at gromacs.org>
>Subject: Re: [gmx-users] a) rlist vs rvdw/rcoulomb size confusion,and b) 
>reduced units
>Date: Tue, 17 Jul 2007 04:53:40 -0400
>
>>
>>I think that what you want is inconsistent.
>>If you want dispersion correction this means that in principle you do not
>>want
>>cut-off's.
>
>Berk, you are right: I don't want any cutoffs, I am trying to avoid
>them. But I thought
>the way Gromacs works is, if you don't choose one, you get assigned
>the default one,
>which is rvdw=1, rcoulomb=1 etc Correct?
>That's why I figured I'll pick a large enough number. Otherwise that
>part of manual I quoted is very confusing, bordering on incorrect: why
>would I want to specify cutoff if
>the point of the table is to avoid one? (Well, I suppose there could
>be other reasons to
>have the table for NB interactions, although I can't think of any.)
>
>Marin

The point of a table is usually not to avoid cut-off's.
The point of a table is to allow for any possible functional form
for the interaction function.

Furthermore a table even requires a cut-off, since it can not be
infinitely large, as opposed to for instance an analytical LJ potential.

If you only want to avoid cut-off's you should use the Gromacs shift
or switch function.

I might need to clarify the cut-off point again, as this is not stated
very clearly in the manual.
An interaction function is NOT cut off at rvdw of rcoulomb.
The neighborlist is made with rvdw and rcoulomb. But since on
can have charge-group and since for nstlist>1 particles move
between neighborlist updates, the actual cut-off is not exactly
rvdw or rcoulomb.

Berk.

_________________________________________________________________
Don't get caught with egg on your face. Play Chicktionary!  
http://club.live.com/chicktionary.aspx?icid=chick_hotmailtextlink2