[gmx-users] distance restraints in Gromacs 4.0.4

[Ivan Gladich]

Dear users
I have found a problem for the  distance constraint in the version
4.0.4 of gromacs.
I am trying to simulate a water droplet around an ion. To do that I use
a distance restraint armonic potential (pag.66, manual version 4).
During my simulation the shape of my sphere fluctuating too much and
many molecules leave the sphere even if the restoring force is big (I
use for restraint armonic potential a constant k=1e4 KJ mol^-1 nm^-2 ).
The problem disappear if I run the exactly same simulation with gromacs
3.3.3.:the spherical shape is always conserve and no molecules leave the
sphere.

Furthermore, if I am trying to run droplet larger than 1.0 nm radius
(e.g. 1.6 nm). These  simulation at a certain point stop, giving me the
following error

** glibc detected *** Dmdrun: munmap_chunk(): invalid pointer:
0x0000000000a72bd0 ***

The same simulation in the version 3.3.3 runs without any problem.

Do you know what kind of problem could be?
Has the version 4.0.4  some problems with distance restrains?

Thank you in advance for any help or suggestion
     Ivan Gladich
-- 
-------
Ivan Gladich

Dreyfus Postdoctoral Fellow
Department of Chemistry
Purdue University
560 Oval Drive, box 140
West Lafayette
IN 47906 USA

Room: 265C
Tel:765-494-5225
e-mail: igladich at purdue.edu
web page: http://web.ics.purdue.edu/~igladich/

--------