[gmx-developers] Setting pull-group1-pbcatom

[Vedran Miletić]

Hi,

I am having an issue with pull-group1-pbcatom and custom slice pulling
code [1], intended for simulating pulling of a slice of a water slab.
For example, let's consider a 5 x 1 x 1 box with just water and a slice
from x=1.0 to x=3.0. Periodic boundary conditions are on.

With pbcatom set to the default value of 0, the center of mass will most
of the time be computed x ~= 2.0, but will occasionally start to move in
the direction of the pull force until it goes until the end of the box
through the periodic boundary and eventually reaches ~=2.0 again. This
repeats occasionally for as long as the simulation runs. Why does this
happen?

With pbcatom set to -1, the center of mass is always computed correctly.

The check which atoms from the group are in the slice is done in
update_sliced_pull_groups() (line 1703 in pull.cpp [2]) as x[ii][0] >=
x_slice_min && x[ii][0] <= x_slice_max. Is there anything I'm missing here?

Thanks in advance.

Vedran

[1] https://gerrit.gromacs.org/#/c/6283/
[2] https://gerrit.gromacs.org/#/c/6283/9/src/gromacs/pulling/pull.cpp

-- 
Vedran Miletić
vedran.miletic.net