The LAMMPS particle simulation package: Bringing together innovative physics models, machine-learning interatomic potentials, and extreme-scale computing resources

The molecular dynamics method (MD), as implemented in the LAMMPS[1] particle simulation code, is a powerful tool for explicitly sampling the phase space distribution of hundreds or billions of interacting particles. It provides a wealth of information on how particular microscopic interactions lead to a vast range of emergent behaviors on much larger length and time scales, complementing theory and experiment. The overall usefulness of the method is contingent upon choosing interaction potentials, initial conditions, and external constraints that adequately approximate the true physical situation. Over many decades, this has led to the emergence of many prominent interaction potentials that lie on the accuracy-cost Pareto frontier. At the one extreme of minimal complexity are well-established models for simple fluids (Lennard Jones particles), polymer melts (FENE chains), and metals (EAM). More recently, the Pareto frontier has extended rapidly to higher accuracy models, with the emergence of complex machine-learning (ML) surrogates for expensive quantum methods. These are trained to reproduce the energy and forces of many small configurations of atoms obtained from quantum electronic structure calculations e.g. Density Functional Theory. Many of these ML potentials have been implemented in LAMMPS, including Behler-Parinello potentials, GAP, SNAP[2], and the Atomic Cluster Expansion (ACE)[3], which has recently emerged as a powerful generalization of earlier approaches.

In this talk, I will review some recent capabilities added to the LAMMPS code, such as the ability to estimate local continuum thermomechanical properties of condensed phases from time- and volume-averaged microscopic observables for arbitrary interatomic potentials. I will then describe several recent scientific applications of LAMMPS that combine innovative physics models[4], machine-learning interatomic potentials, and extreme scale computing resources.

[1] Thompson et al., Comp. Phys. Comm., 271:108171, 2022. DOI 10.1016/j.cpc.2021.108171 [2] Thompson et al., J. Comp. Phys., 285:316, 2015. DOI 10.1016/j.jcp.2014.12.018 [3] Lysogorskiy, npj Comp. Mat. 7:1, 2021. DOI 10.1038/s41524-021-00559-9 [4] Tranchida, Plimpton, Thibaudeau, and Thompson, J. Comp. Phys., 372:406, 2018. DOI 10.1016/j.jcp.2018.06.042