[gmx-users] problem with g_dist (probably solved)

[Thomas Schlesier]

Hi all,
i had problems with g_dist. My system consists of 15 molecules urea with
a short alkane linker. With constraints and restraintss (distance,
angles and dihedrals) made it that the molecules are in one line (the
last methyl-group of the alkan linker).
Here a little picture:
________  ...
|      |       |    ...
X    X     X   ...

| is the alkan linker; X is the urea-group and _____  is the line which
conacts the methyl-groups.
The distance between each molecule is 2nm, so (with 15 molecules) the
distance between the first and the last molecule should be 28nm. The
size of the simulation box was 40 10 10 (all nm), with no pbc (periodic
boundary conditions) (the 'line' was along the x-axis).
When i determined the distance between the first and the last molecule
with g_dist, the program gave some errors:  "There were 1 inconsistent
shifts. Check your topology" , but the program ended normally. The
calculeted distance was around 12nm (versus the 28nm which is the
correct value; i calculated it from the .gro file).

So my idea was the g_dist uses pbc to calculate the distance (even if
the simulation had no pbc). The distance between the first and the last
methyl-group (in the two nearest boxes) should be around 40nm (box size)
-28nm (distance in one box) = 12 nm (distaance over two boxes).
Is this idea right?!?

In another simulation i increased the box size to 60 60 60 (nm). Here
g_dist had no troubles. I get no error messages ("There were 1
inconsistent shifts. Check your topology") and the distance is right.

The other question is: What does the error message mean?

Thanks for your help.
Thomas