[gmx-users] Answering my own question

[Marc Baaden]

Hi,

meanwhile I found out, why I had problems with the position restraints
(see previous eMail). It turned out to be a Gromacs bug.

Here is a short description:



2. POSRES and - shuffle bug (Fri Nov 30 09:04:02 GMT 2001)
==========================================================
 
Description
-----------
 
When performing MD runs using multiple processors and the -shuffle
option to grompp, the usage of position restraints is "unstable". E.g.
some atoms which were not supposed to be position restraints happen to
be and vice versa.
 
The presumed reason is that shuffling renumbers the simulation system
for improved performance, but does not renumber the position
restraints. So errors occur when the atoms you restrain will be
renumbered.
 
This has been tested and verified for Gromacs 2.0, but not yet for
3.0.5.
 
Workaround
----------
 
Use grompp without the - shuffle option.



Cheers,
Marc

-- 
 Dr. Marc Baaden - Laboratory of Molecular Biophysics, Oxford University
 mailto:baaden at smplinux.de  - ICQ# 11466242 -  http://www.marc-baaden.de
 FAX/Voice +49 40333 968508  -  Tel: +44 1865 275380  or  +33 609 843217





-- 
 Dr. Marc Baaden - Laboratory of Molecular Biophysics, Oxford University
 mailto:baaden at smplinux.de  - ICQ# 11466242 -  http://www.marc-baaden.de
 FAX/Voice +49 40333 968508  -  Tel: +44 1865 275380  or  +33 609 843217