[gmx-users] discrepancy trjconv and gmxdump

[Gavin Melaugh]

Hi all

I have generated a gro file from a traj.trr file using trjconv. When I
use gmxdump on the same traj.trr file to output a generic format history
file it seems that there is a discrepancy in the coordinates of some
atoms in a particular frame. Essentially I output frame 5 using trjconv,
then I take the corresponding frame from the history file and covert it
to gro format. I then run vimdiff on the two files. There are sign
differences with the coordiante components. The atoms in question are at
the edge of the periodic box. I was just wondering why there is a
discrepancy in the two utilities?

Many thanks

Gavin