Optimizing Event-Driven Simulations

TL;DR

This paper presents new optimization techniques for event-driven molecular dynamics simulations, significantly improving computational speed for modeling hard particles of various shapes.

Contribution

It introduces a generalized linked list method and an improved nearest neighbor list approach for faster event-driven simulations.

Findings

Significant speedup over previous methods

Effective generalization for various particle shapes

Enhanced efficiency in event-driven simulation algorithms

Abstract

Event-driven molecular dynamics is a valuable tool in condensed and soft matter physics when particles can be modeled as hard objects or more generally if their interaction potential can be modeled in a stepwise fashion. Hard spheres model has been indeed widely used both for computational and theoretical description of physical systems. Recently further developments of computational techniques allow simulations of hard rigid objects of generic shape. In present paper we will present some optimizations for event-driven simulations that offered significant speedup over previous methods. In particular we will describe a generalization of well known linked list method and an improvement on nearest neighbor lists method recently proposed by us.