Re: PME GridSize

From: Neelanjana Sengupta (senguptan_at_gmail.com)
Date: Wed Nov 08 2006 - 10:45:40 CST

Hi,
Thanks for your prompt reply. No, I am not doing wrapAll. The reason being
that solvent diffusion is one of the things I would like to investigate in
my system. (I understand that ideally wrapAll should not affect it, but
since I use the minimum image convention during my analyses, I feel
comfortable about not having to use wrapAll. Please do point out if you
think I am not on the right track..)

Regards,
Neelanjana

On 11/8/06, JC Gumbart <gumbart_at_ks.uiuc.edu> wrote:
>
> Are you not wrapping your coordinates? Is wrapAll on or off in the config
> file? not present (it defaults to off)?
> Anyway, the cell basis is the reality of the periodic system, not the
> coordinates of the atoms (i.e. the box size). Sorry for the confusion
> about that point.
>
> On Nov 8, 2006, at 11:26 AM, Neelanjana Sengupta wrote:
>
> Just a re-confirmation: The PMEGridSize(X, Y, Z) should be greater than or
> equal to the Box Size, and NOT greater than or equal to the Cell Basis?
>
> Upon loading a previous trajectory, "molinfo top get a", etc, give me ~
> 55.0.
>
> However, when I do a selection "set all [atomselect top all]",
> "measure minmax $all" gives,
> {-127.4 -109.5 -142.2} {107.6 155.9 137.6}
> corresponding to lengths 235, 266, 280 (approx) along X, Y, Z.
>
> So should my PMEGridSizes be ~ 60, or ~ 280 ?
>
> Your remarks are highly appreciated.
> Thanks,
> Neelanjana
>
> On 11/7/06, JC Gumbart < gumbart_at_ks.uiuc.edu> wrote:
> >
> > You are right, they should be around the size of your box (or a little
> > bigger). They should also ideally be multiples of 2, 3, and 5.
> > If they are too low, you will have a grid of charges that very poorly
> > mimics the actual distribution. I'm not sure how it will affect the
> > simulation but results will be inaccurate.
> >
> >
> > On Nov 7, 2006, at 6:57 PM, Neelanjana Sengupta wrote:
> >
> > Hello,
> >
> > As per my thinking, the PMEGridSizeX, PMEGridSizeY and PMEGridSizeZ should
> > be greater than (or at least equal to), the lengths along the X, Y and Z
> > axes of the simulation box (assuming cubical box).
> > Could somebody point out if my thinking is not correct, and how the
> > PMEGridSizes affects the accuracy of the simulation results?
> >
> > Thanks and regards,
> > --
> > ****************************************
> > Neelanjana Sengupta, PhD student
> > Dept. of Chemistry
> > Univ. of California-Irvine
> > Irvine, CA 92697
> > USA
> > Phone: 1-949-824 9921
> > email: sengupta_at_uci.edu
> > ****************************************
> >
> >
> >
>

-- 
****************************************
Neelanjana Sengupta, PhD student
Dept. of Chemistry
Univ. of California-Irvine
Irvine, CA 92697
USA
Phone: 1-949-824 9921
email: sengupta_at_uci.edu
****************************************

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:42:48 CST