Research Services
information technology services
NAMD
General Information
Short Description: NAMD is a parallel molecular dynamics code designed for high-performance simulationsof large bimolecular systems
Category: Computational Chemistry
Installation Information
Version: 2.6 Version Date: August, 2006
Technical support: Please contact Research Services (researchservices@bc.edu)
Terms of Use:
License Information
Number of Licenses: Licensed for use on the Scorpio Linux Cluster.
Type of License: Network
Installed Sites: Scorpio Linux Cluster.
Installed Platforms: Linux
Vendor/Developer: Theoretical and Computational Biophysics Group at the University of Illinois at Urbana-Champaign
Vendor Website: www.ks.uiuc.edu/Research/namd
Usage Information
To use interactively, first load the namd module by typing:
module load namd
To use NAMD with PBS, create a file called namd.pbs that contains:
#!/bin/tcsh
#PBS -l mem= 2gb,nodes=1:ppn=4,walltime=75:00:00
cd directory-that-contains-your-namd-script
./namd_script input.inp
The above will request 4 processors and 2 GB of memory on one node, and 75 hours of time. Use your input file instead of "input.inp".
You also need to create a file called namd_script with the following:
#!/bin/sh
/usr/public/Modules/3.2.3/init/sh
module load namd
# create node-list file
rm -rf $HOME/.mpd.host.*
NODELIST="$HOME/.mpd.host.$LSB_JOBID"
# use ssh instead of rsh
export CONV_RSH=ssh
output=`echo $LSB_MCPU_HOSTS | /bin/awk '
BEGIN {counter1=0;counter2=0}
{
size = split($0, a, " ");
for (i = 1; i <= size; i += 2) {
counter1 = counter1 + 1;
counter2 = counter2 + a[i + 1];
for (j = 0; j < a[i + 1]; ++j) {
print "host", a[i] > hfile;
}
}
}
END {print counter2}' hfile=$NODELIST`
# find full path of namd2
command=`which namd2`
echo "charmrun ++p $TOTAL_NUM_PROC ++nodelist $NODELIST namd2 $@"
charmrun ++p $output ++nodelist $NODELIST $command $@
To submit a job to the queues, type:
qsub namd.pbs
Documentation
For information about namd see www.ks.uiuc.edu/Research/namd