[gmx-developers] box center with triclinic coords

[Berk Hess]

Vishal Vaidyanathan wrote:

>Hi,
>   Quick question - is the box center always given by (boxx + boxy +
>boxz)/2 where boxx, boxy and boxz are the rows of the box matrix? I just
>need to be able to compute the distance of a particle from the box center
>during the nonbonded force calculations and was wondering if I'm missing
>something about the way PBC (or location of origin) is handled in gromacs.
>  
>
The center of the box is irrelevant in mdrun.
For nonbonded force calculations the particles are placed on a grid and 
shifted
around take PBC into account (see the manual).
For reasons of efficiency, before the nonbonded calculations the charge 
groups
are placed with their centers of mass inside a brick with:
0<=x<box_a(x)
0<=y<box_b(y)
0<=z<box_c(z)

Berk.