General Information
| Home page | Version | Module dependencies | |
| Compiler | MPT | ||
| lammps.sandia.gov | 2014.09.05 | intelcomp/15.0.1 | mpt/2.10 |
| 2014.12.09 | intelcomp/15.0.1 | mpt/2.11 | |
LAMMPS is a classical molecular dynamics code. It includes potentials for solid-state materials (metals, semiconductors) and soft matter (biomolecules, polymers) and coarse-grained or mesoscopic systems. It can be used to model atoms or, more generically, as a parallel particle simulator at the atomic, meso, or continuum scale.
Usage on Vilje
Load the intelcomp, mpt, and lammps modules to use LAMMPS on Vilje, e.g.:
$ module load intelcomp/15.0.1 mpt/2.11 lammps/141209
Sample LAMMPS Job Script
#PBS -N LAMMPS_test
#PBS -l walltime=00:10:00
#PBS -l select=2:ncpus=32:mpiprocs=1:ompthreads=16
#PBS -j oe
#PBS -A nn1234k
# Load program modules
module load intelcomp/15.0.1 mpt/2.11 lammps/141209
# Create and use a temporary work directory
export WORKDIR=/work/${USER}/${PBS_JOBNAME}.${PBS_JOBID}
mkdir -p ${WORKDIR}
cd ${WORKDIR}
# Copy sample problem
cp /sw/sdev/Modules/lammps/lammps-141209/examples/kim/* .
# Run simulation
mpiexec_mpt lmp_intel < in.kim.lj.lmp
# Move temporary work directory to qsub work directory
mv ${WORKDIR} ${PBS_O_WORKDIR}/
Replace nn1234k with your project account number in the lammps.pbs job script, and submit the job to the queuing system
$ qsub lammps.pbs
Note that LAMMPS runs in hybrid MPI / OpenMP mode; 'mpiprocs' and 'ompthreads' in the sample script control the balance between processes and threads. Their values should multiply to 16 to match the number of physical cores on a Vilje node, the optimal balance depends on details of the input case.