[gmx-users] energy debugging

[David]

On Mon, 2003-01-27 at 10:58, Anton Feenstra wrote:
> Hi,
> 
> 
> I'm trying to debug a 'nan' total energy for a large system (protein,
> bilayer, ligand and water, >120k atoms). I'd like to know which of
> the energy components are responsible. It used to be possible to get
> detailed output by setting
> nstenergy = 1
> nstlog    = 1
> energy_grps = Protein DMP HIM Sol
> and 'setenv LOG_BUFS 0' to turn off ouput file buffering. However,
> mdrun now refuses to print any energy group information in the log
> file (even with -nocompact), and the enrgy file refuses to be filled
> (it remains empty). Since the system is pretty large, we don't feel
> like waiting for the buffers to flush (if they ever will).
I don't think energy-group information was ever written to the log file,
except for averages at the very end.

LOG_BUFS should work though. Do also run mdrun -v, that gives you
explicit flushing of (some) files.

-- 
Groeten, David.
________________________________________________________________________
Dr. David van der Spoel, 	Dept. of Cell and Molecular Biology
Husargatan 3, Box 596,  	75124 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://xray.bmc.uu.se/~spoel
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++