[MMTK] Bug in MMTK.Universe.largestDistance()?

[Konrad Hinsen]

On 04.01.2008, at 13:11, Andreas Kring wrote:

> I think there is a bug in MMTK.Universe.largestDistance(). An  
> example of
> the error can be seen in the script below.

There is indeed a bug, but it is not where you think it is. The  
calculation for orthogonal universes (including cubic ones) is  
correct, but the one for non-orthogonal boxes isn't.

What largestDistance() should return is the largest interparticle  
distance that can be represented in the universe independent of  
orientation. In other words, the largest distance L is defined as the  
largest L such that an atom A2 placed at a distance L from a given  
atom A1 in an arbitrarily chosen direction will
make the minimum-image distance between A1 and A2 equal to L. For a  
rectangular box, this makes L equal to half the shortest edge length.

This method is used in MMTK for checking cutoff radii for particle  
interactions. For consistency, all real-space cutoffs should be less  
than the number returned by universe.largestDistance(). For any  
larger cutoff, the effective interaction volume of each atom is no  
longer spherical.

Konrad.
--
---------------------------------------------------------------------
Konrad Hinsen
Centre de Biophysique Moléculaire, CNRS Orléans
Synchrotron Soleil - Division Expériences
Saint Aubin - BP 48
91192 Gif sur Yvette Cedex, France
Tel. +33-1 69 35 97 15
E-Mail: hinsen at cnrs-orleans.fr
---------------------------------------------------------------------