Algorithm Selection in Short-Range Molecular Dynamics Simulations

TL;DR

This work tackles dynamic algorithm selection for short-range molecular dynamics simulations by extending AutoPas with three strategies—Predictive, Expert fuzzy-logic, and Random Forest—aimed at avoiding costly trial-and-error searches. It demonstrates that data-driven RF selection yields the strongest practical gains, achieving up to $4.05\times$ speedups over naive strategies and $1.25\times$ over an oracle-like best static configuration, across three varied MD scenarios. The study analyzes practicality, dataset design, and hardware-dependent behavior, concluding that RF offers the best balance of performance and ease of use, while suggesting online learning with confidence as a path toward further improvements. The results underscore the importance of adaptive configuration in diverse MD workloads and provide a framework for deploying dynamic algorithm selection in particle simulations.

Abstract

Numerous algorithms and parallelisations have been developed for short-range particle simulations; however, none are optimally performant for all scenarios. Such a concept led to the prior development of the particle simulation library AutoPas, which implemented many of these algorithms and parallelisations and could select and tune these over the course of the simulation as the scenario changed. Prior works have, however, used only naive approaches to the algorithm selection problem, which can lead to significant overhead from trialling poorly performing algorithmic configurations. In this work, we investigate this problem in the case of Molecular Dynamics simulations. We present three algorithm selection strategies: an approach which makes performance predictions from past data, an expert-knowledge fuzzy logic-based approach, and a data-driven random forest-based approach. We demonstrate that these approaches can achieve speedups of up to 4.05 compared to prior approaches and 1.25 compared to a perfect configuration selection without dynamic algorithm selection. In addition, we discuss the practicality of the strategies in comparison to their performance, to highlight the tractability of such solutions.

Figures (1)

• Figure 1: Comparison of the total time spent calculating forces for each thread across the entire simulation for each tuning strategy relative to the optimal single configuration for each experiment. These are (a) LC-C04-N3L-AoS-CSF1, (b) LC-C04_HCP-N3L-SoA-CSF1, (c) LC-C08-N3L-SoA-CSF0.5.