[gmx-users] shuffle/sort problem

[Lynne E. Bilston]

Hi,

I'm simulating a largish system using the mpi version of Gromacs, on RH7.2 
linux dual athlon machines. My system runs just fine if I run it over 
multiple cpus, without the shuffle and sort options, with no error 
messages, albeit more slowly. The results look sensible. But when I use the 
shuffle/sort options, I get 1-4 interaction messages at the start of the 
simulation, and nan energies. This happens with the exact same input 
options and coordinate files.

I'm running a protein/lipid/water system, with position restraints on just 
the protein, no pressure coupling, and ordinary cutoffs (no PME yet). This 
simulation is for equilibration before the final production runs without 
pr. I've run energy minimisation on the system prior to the position 
restrained runs.

Any ideas?

-Lynne
________________________________________________________________________________
Lynne E. Bilston, PhD
Senior Scientist, Prince of Wales Medical Research Institute, and
Conjoint Associate Professor, Faculty of Medicine, University of New South 
Wales
Barker St, Randwick, NSW 2031 Australia
Tel: +61-2-9382-7924	Fax: +61-2-9382-2643
________________________________________________________________________________