Project

General

Profile

Bug #207

input file for 2ns of simulation

Added by andrea spitaleri about 11 years ago. Updated about 11 years ago.

Status:
Closed
Priority:
Normal
Assignee:
Erik Lindahl
Category:
mdrun
Target version:
Affected version - extra info:
Affected version:
Difficulty:
uncategorized
Close

Description

Created an attachment (id=283)
input file for 2ns of simulation

Hi there,
I have made this input file using this command line:
grompp -f md.mdp -c file.gro -o md2ns.tpr
then i have run in on our cluster HP with this pbs script:
#PBS -S /bin/bash
#PBS -V
#PBS -q batch
#PBS -j oe
#PBS -l nodes=24
#PBS -o out
PBS_O_INITDIR=$PBS_O_WORKDIR
cd $PBS_O_WORKDIR

/usr/local/cluster/bin/mpirun --prefix /usr/local/cluster/ --mca btl tcp,self --mca pls_rsh_agent rsh /usr/local/cluster/gromacs-3.3.csv/bin/mdrun -s md2ns.tpr -v -deffnm md2ns >& md2nslog

Using 24 cpus the run goes fine. Using a different number of cpu (27 with -dd 3 3 3, 36 with no further option) the job crashes with this error:

"1 particles communicated to PME node 9 are more than a cell length out of the domain decomposition cell of their charge group"

Moreover, I am trying to extend my simulation using tpbconv with this comand:
1. tpbconv -s md2ns.tpr -f file.trr -extend 2000 -o new
2. mdrun -s new.tpr -v

but the simulation crashes. If I use this sequence:
1. tpbconv -s md2ns.tpr -f file.cpt -extend 2000 -o new
2. mdrun -s new.tpr -v

the simulation goes fine however it starts from 0.

thanks in advance

andrea

md2ns.tpr (8.86 MB) md2ns.tpr input file for 2ns of simulation andrea spitaleri, 09/11/2008 10:35 AM

History

#1 Updated by andrea spitaleri about 11 years ago

I have used this option:
1. tpbconv -s md2ns.tpr -f file.trr -extend 2000 -o new
2. mdrun -s new.tpr -v -cpi file.cpt

and the run goes with 27 cpus rather than 24.
It seems so much correlated to the cpu number.

#2 Updated by Berk Hess about 11 years ago

I fixed this bug several days ago.

Berk

Also available in: Atom PDF