[gmx-users] testing ligand-protein system in Release Candidate 3 of version 3.3

[mollica.luca at hsr.it]

dear users,

we are trying to set up a novel installation of GROMACS on an IBM 
eServer BladeCenter (56 procs, 28 nodes, myrinet) using the GMX 3.3 - 
rc3.
The first tests seem to be really ok, but only in case of systems with 
protein, water and ions. I have tested the set up also with a protein-
ligand system, but I have encountered a strange problem that has never 
occourred in 3.2.1 stable version. 

Running grompp on EXACTLY THE SAME sytem and files theat used to give 
successful reults and work fine on 3.2.1 version on the same machine
(a version that has been left to be alive on the first half of nodes), 
I got this error that seems to be a typical topology error, happening 
in coincidence with the presence (in the ligand topology file) of 
dihedral angles (since line 300 until the end of the file):

Generated 1284 of the 1485 non-bonded parameter combinations
ERROR 2 [file "ligand.itp", line 300]:
   Too many parameters

etc etc ... until reaching the linmit of 10 errors and stopping the 
grompping process invoking an error in source code fatal.c in line 416.

What's going on with my ligand ? Everything works pretty well with 
3.2.1 installation, but it is not working with this version 3.3.rc3.
Is a release related problem ? Or does it depend on some changes that 
I have to make in the set up of th files for running MD simulation ?

thanx a lot

LM