The Vectorization of the Tersoff Multi-Body Potential: An Exercise in Performance Portability

TL;DR

This paper develops a vectorization scheme for multi-body potentials in molecular dynamics, enhancing performance portability across CPUs and accelerators, with significant efficiency gains demonstrated on various architectures.

Contribution

The paper introduces a novel vectorization approach for multi-body potentials that is architecture-agnostic, improving performance portability in molecular dynamics simulations.

Findings

3-5x speedup on Intel Xeon Phi clusters

Efficiency gains for both kernels and large-scale simulations

Effective vectorization applicable to CPUs and accelerators

Abstract

Molecular dynamics simulations, an indispensable research tool in computational chemistry and materials science, consume a significant portion of the supercomputing cycles around the world. We focus on multi-body potentials and aim at achieving performance portability. Compared with well-studied pair potentials, multibody potentials deliver increased simulation accuracy but are too complex for effective compiler optimization. Because of this, achieving cross-platform performance remains an open question. By abstracting from target architecture and computing precision, we develop a vectorization scheme applicable to both CPUs and accelerators. We present results for the Tersoff potential within the molecular dynamics code LAMMPS on several architectures, demonstrating efficiency gains not only for computational kernels, but also for large-scale simulations. On a cluster of Intel Xeon Phi's, our optimized solver is between 3 and 5 times faster than the pure MPI reference.