[gmx-users] rhombic dodecahedron and very high pressures

[Brian Curtis Stephenson]

Dear Gromacs Users,

I recently created a micelle and placed it in a rhombic dodecahedron box. 
Visualization with VMD showed clearly that the micelle was solvated and was
placed within a dodecahedron volume of water as expected.  The initial density
is very close to what would be expected at equilibration, as has always been
the case in the past when I use genbox -cs.  However, as soon as I begin
simulation the dodecahedron increases in size very rapidly with both berendsen
and Parrinello-Rahman pressure coupling (with both isotropic and anisotropic
options, and with pbc=xyz or pbc=full).  g_energy on the output reveals a very
high internal pressure in the cell.  What could be going on?  I've read a paper
by a group in China stating that they used pressure coupling in GROMACS with a
rhombic dodecahedron, so this should be supported.  There doesn't appear to be
anything on the forums related to this, so I thought I would throw it out to
all of you users.  Any help would be much appreciated!

Btw, I tested this with Gromacs-3.2.1 and a CVS version of GROMACS I use
regularly and both versions give the same problem.

My command sequence was:

genbox -cp start_file.gro -cs -o solv.gro -d 1.0 -p test.top
editconf -f solv.gro -bt dodecahedron -o dodec.gro

Visualization was done with:
trjconv -f dodec.gro -s topol.tpr -o vis.gro -ur compact
and everything showed up as expected.

Here's what g_energy gave for the pressures:

Pressure XX - bar
Pressure YY - bar
Pressure ZZ - bar
0.0000   35691   40510   42191
0.3000   9310    9656    9636
0.6000   and it keeps decreasing with time as the box expands....