LAMMPS is a Molecular Dynamics simulator built and maintained at Sandia National Labs. It has many versions and recently underwent a rewrite in C++. We only support the later versions of this code.

LAMMPS (as it is installed at CSE) requires the following libraries and software:

• Pathscale compiler
• Open MPI
• FFTW library

We try to support a recent version. Here are the versions that we currently have installed and tested:

• LAMMPS C++ with all patches through September 9, 2008.

To load LAMMPS you must first load all the prerequisite modules, then load the LAMMPS module. To do this in one command type:

module load compilers/pathscale-3.2 mpi/openmpi-1.2.6-pathscale-3.2 math/fftw-2.1.5 md/lammps-current

You should put this command in your submit script. Here is an example:

#!/bin/bash
#
#$ -cwd
#$ -j y
#$ -S /bin/bash
 
module load compilers/pathscale-3.2 mpi/openmpi-1.2.6-pathscale-3.2 math/fftw-2.1.5 md/lammps-current
 
INPUT=~/src/lammps/input
OUTPUT=~/src/lammps/output
 
# need other data files
cd $INPUT
 
mpirun lmp_cse -log $OUTPUT/lj-2.log -in in.lj

Then change to your data directory and run the following command where N is the number of CPUs you want to use.

$ qsub -pe mpi N submit.sh

LAMMPS is designed to work with FFTW 2.1.5 (not version 3). It has been built against the double-precision fftw 2.1.5 library (with pthreads and mpi enabled - but that doesn't seem to make a difference). The following packages have been enabled:

$ make package-status
Installed YES: package ASPHERE
Installed YES: package CLASS2
Installed YES: package COLLOID
Installed YES: package DIPOLE
Installed YES: package DPD
Installed YES: package GRANULAR
Installed YES: package KSPACE
Installed YES: package MANYBODY
Installed  NO: package MEAM
Installed YES: package MOLECULE
Installed YES: package OPT
Installed YES: package PERI
Installed  NO: package POEMS
Installed YES: package XTC

Installed  NO: package USER-ACKLAND
Installed  NO: package USER-CG-CMM
Installed  NO: package USER-EWALDN
Installed  NO: package USER-SMD

There is a makefile with the following configuration called src/MAKE/Makefile.cse:

$ export F77=pathf90
$ export CC=pathcc
$ head MAKE/Makefile.cse
CC =            mpic++
CCFLAGS =       -O3 -OPT:Ofast -DFFT_FFTW -DLAMMPS_GZIP \
                -fstrict-aliasing -Wall -W -Wno-uninitialized \
                -I/share/apps/fftw-2.1.5/include
NOALIAS =       -fno-strict-aliasing
DEPFLAGS =      -M
LINK =          mpic++
LINKFLAGS =     -L/share/apps/fftw-2.1.5/lib
USRLIB =        -ldfftw -ldfftw_threads -ldfftw_mpi
SYSLIB =
ARCHIVE =       ar
ARFLAGS =       -rc
SIZE =          size

We ran a few benchmarks that are included with LAMMPS. You can compare these numbers to other hardware benchmarks for LAMMPS.

We used this input file.

Atomic fluid:

• 32,000 atoms for 100 timesteps
• reduced density = 0.8442 (liquid)
• force cutoff = 2.5 sigma
• neighbor skin = 0.3 sigma
• neighbors/atom = 55 (within force cutoff)
• NVE time integration

<format gnuplot> set title “LAMMPS LJ Fixed Benchmark” set style line set log x set xtic (1,2,4,8,16) set yrange [0:110] set ytic (0,10,20,30,40,50,60,70,80,90,100,110) set xlabel “CPUs” set ylabel “Parallel Efficiency (%)” set grid plot '-' using 1:2 lw 2 with lp title “tribe (% of 5.895s)”, \

   '-' using 1:2 lw 2 with lp title "urdarbrunnr (% of 5.964s)"

1 100 2 91.13 4 90.80 8 67.11 16 47.83 e 1 100 2 97.16 4 92.94 8 91.97 16 86.73 32 83.11 e </format>

We used this input file.

Bead-spring polymer melt with 100-mer chains and FENE bonds:

• 32,000 atoms for 100 timesteps
• reduced density 0.8442 (liquid)
• force cutoff of 2^(1/6) sigma
• neighbor skin = 0.4 sigma
• neighbors/atom = 5 (within force cutoff)
• NVE time integration

<format gnuplot> set title “LAMMPS Chain Fixed Benchmark” set style line set log x set yrange [0:120] set xtic (1,2,4,8,16) set ytic (0,10,20,30,40,50,60,70,80,90,100,110,120) set xlabel “CPUs” set ylabel “Parallel Efficiency (%)” set grid plot '-' using 1:2 lw 2 with lp title “tribe (% of 3.367s)”, \

   '-' using 1:2 lw 2 with lp title "urdarbrunnr (% of 2.758s)"

1 100 2 89.30 4 112.30 8 71.89 16 45.56 e 1 100 2 99.65 4 98.23 8 104.37 16 92.92 32 90.91 e </format>

The LAMMPS user manual is available on Sandia's website.