Preprint
Computer Vision
Featured

LAMMPS - a flexible simulation tool for particle-based materials modeling at the atomic, meso, and continuum scales

Aidan P. Thompson(Sandia National Laboratories), Hasan Metin Aktulga(Michigan State University), Richard A. Berger(Temple University), Dan Bolintineanu(Sandia National Laboratories), W. Michael Brown(Intel (United States)), Paul Crozier(Sandia National Laboratories), Pieter J. in ’t Veld(BASF (Germany)), Axel Kohlmeyer(Temple University), Stan Moore(Sandia National Laboratories), Trung Dac Nguyen(Northwestern University), Ray Shan, Mark J. Stevens(Sandia National Laboratories), Julien Tranchida(Sandia National Laboratories), Christian Robert Trott(Sandia National Laboratories), Steven J. Plimpton(Sandia National Laboratories)
September 22, 2021Computer Physics Communications12,521 citations
PDFDOI

13k

Citations

0

Influential Citations

Computer Physics Communications

Venue

2021

Year

Abstract

Since the classical molecular dynamics simulator LAMMPS was released as an open source code in 2004, it has become a widely-used tool for particle-based modeling of materials at length scales ranging from atomic to mesoscale to continuum. Reasons for its popularity are that it provides a wide variety of particle interaction models for different materials, that it runs on any platform from a single CPU core to the largest supercomputers with accelerators, and that it gives users control over simulation details, either via the input script or by adding code for new interatomic potentials, constraints, diagnostics, or other features needed for their models. As a result, hundreds of people have contributed new capabilities to LAMMPS and it has grown from fifty thousand lines of code in 2004 to a million lines today. In this paper several of the fundamental algorithms used in LAMMPS are described along with the design strategies which have made it flexible for both users and developers. We also highlight some capabilities recently added to the code which were enabled by this flexibility, including dynamic load balancing, on-the-fly visualization, magnetic spin dynamics models, and quantum-accuracy machine learning interatomic potentials. Program Title: Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) CPC Library link to program files: https://doi.org/10.17632/cxbxs9btsv.1 Developer's repository link: https://github.com/lammps/lammps Licensing provisions: GPLv2 Programming language: C++, Python, C, Fortran Supplementary material: https://www.lammps.org Nature of problem: Many science applications in physics, chemistry, materials science, and related fields require parallel, scalable, and efficient generation of long, stable classical particle dynamics trajectories. Within this common problem definition, there lies a great diversity of use cases, distinguished by different particle interaction models, external constraints, as well as timescales and lengthscales ranging from atomic to mesoscale to macroscopic. Solution method: The LAMMPS code uses parallel spatial decomposition, distributed neighbor lists, and parallel FFTs for long-range Coulombic interactions [1]. The time integration algorithm is based on the Størmer-Verlet symplectic integrator [2], which provides better stability than higher-order non-symplectic methods. In addition, LAMMPS supports a wide range of interatomic potentials, constraints, diagnostics, software interfaces, and pre- and post-processing features. Additional comments including restrictions and unusual features: This paper serves as the definitive reference for the LAMMPS code. S. Plimpton, Fast parallel algorithms for short-range molecular dynamics. J. Comp. Phys. 117 (1995) 1–19. L. Verlet, Computer experiments on classical fluids: I. Thermodynamical properties of Lennard–Jones molecules, Phys. Rev. 159 (1967) 98–103.

Analysis

Why This Paper Matters

LAMMPS (Large-scale Atomic/Molecular Massively Parallel Simulator) is a cornerstone open-source tool for particle-based materials modeling, widely adopted across physics, chemistry, and materials science. Since its release in 2004, it has become the de facto standard for classical molecular dynamics simulations, evidenced by over 12,500 citations. This paper serves as the definitive reference, documenting the algorithms, design philosophy, and recent advancements that have cemented LAMMPS's role in computational science.

The importance of LAMMPS lies in its unparalleled flexibility and scalability. It supports a vast array of interatomic potentials, runs efficiently on platforms from laptops to the world's largest supercomputers, and allows users to easily extend its capabilities. This adaptability has fostered a large contributor community, growing the codebase from 50,000 to one million lines. As materials modeling increasingly incorporates machine learning and multiscale approaches, LAMMPS provides a robust, extensible platform for integrating cutting-edge methods.

Moreover, the paper highlights how LAMMPS has evolved to meet modern simulation demands, including dynamic load balancing, on-the-fly visualization, magnetic spin dynamics, and quantum-accuracy machine learning potentials. These features demonstrate the code's ongoing relevance and its role in enabling AI-driven materials discovery.

Technical Contributions

  • Parallel Algorithms: LAMMPS employs parallel spatial decomposition, distributed neighbor lists, and parallel FFTs for long-range Coulombic interactions, ensuring scalability to extreme scales.
  • Symplectic Integrator: The time integration uses the Størmer-Verlet symplectic integrator, providing superior stability for long trajectories compared to higher-order non-symplectic methods.
  • Extensible Architecture: Users can control simulations via input scripts or add custom code for new interatomic potentials, constraints, diagnostics, and other features, facilitating community contributions.
  • Recent Capabilities: The paper details newly added features enabled by the flexible design, including dynamic load balancing, on-the-fly visualization, magnetic spin dynamics models, and machine learning interatomic potentials with quantum accuracy.
  • Community Growth: The code has grown from 50,000 to one million lines, with hundreds of contributors, reflecting its open-source success and adaptability.

Results

The paper does not present new experimental results but rather documents LAMMPS's capabilities and growth. Key metrics include its 12,521 citations, one million lines of code, and support for a wide range of platforms from single CPU cores to supercomputers with accelerators. The scalability and performance are implied through its widespread adoption in large-scale simulations. The integration of machine learning potentials is a notable result, showcasing the code's ability to incorporate AI methods for quantum-accuracy simulations.

Significance

LAMMPS has had a profound impact on computational materials science, enabling simulations that were previously infeasible. Its open-source nature and extensibility have fostered a vibrant ecosystem, accelerating research in areas such as nanotechnology, drug discovery, and materials design. The incorporation of machine learning potentials bridges classical and quantum simulations, opening avenues for AI-enhanced materials modeling. As the field moves toward more complex, multiscale simulations, LAMMPS remains a critical infrastructure, empowering researchers to push the boundaries of scientific discovery.

molecular-dynamicssimulationparallel-computingmaterials-sciencemachine-learning-potentialsopen-sourcescientific-computing
Computer Vision

Related Papers