LAMMPS

Overview

LAMMPS (Large-scale Atomic/Molecular Massively Parallel Simulator) is a molecular-dynamics code The code is designed to be easy to modify or extend with new functionality.  LAMMPS can run on a single core but is designed to be highly efficient running on a large number of cores in parallel using message-passing techniques and a spatial decomposition of the simulation domain.  It solves systems ranging from single atoms through polymers and proteins to rigid collections of particles.  A variety of force fields is supported.

Local support is not available.  LAMMPS has documentation and tutorials at its Website.   It has a large and active community of users; to search or join the mailing list see the instructions here.

LAMMPS on Rivanna

ARCS builds a version of LAMMPS incorporating the most popular optional packages.  To find the available versions and how to load them, run

module spider lammps

This will return a result such as

  For detailed information about a specific "lammps" module (including how to load the modules) use the module's full name.
  For example:

     $ module spider lammps/2017-08-11

Running this command provides information about the software and shows

    This module can be loaded directly: module load lammps/2017-08-11

Users may build their own versions of LAMMPS if they wish to use a different compiler+MPI combination, or to choose individual optional packages.  Instructions are available at the LAMMPS "Getting Started" page.

Sample SLURM script

To run the system version of LAMMPS, a script similar to the following can be used.  LAMMPS has many options so only the most basic is shown.

#!/bin/bash
#SBATCH --account my_acct
#SBATCH --nodes=2
#SBATCH --ntasks-per-node=16
#SBATCH --time=3-00:00:00
#SBATCH --output=thermo.out
#SBATCH --partition=parallel

module load lammps/2017-08-11

srun lmp_intel_openmpi -in run.in.npt