More bang for your buck: Improved use of GPU nodes for GROMACS 2018

TLDR

The study aims to identify the hardware that best produces molecular dynamics trajectories on Linux compute clusters using GROMACS 2018. The authors benchmark GROMACS 2018 performance across diverse compute nodes and compare results to node costs, including lifetime energy and cooling expenses. Consumer GPU nodes deliver a markedly higher performance‑to‑price ratio than CPU nodes, and the optimal CPU‑to‑GPU balance has shifted further toward GPUs, so that nodes optimized for GROMACS 2018 (or upgraded with recent GPUs) achieve substantially better performance‑to‑price than older‑version‑optimized nodes, enabling cost‑effective upgrades of legacy hardware. © 2019 Wiley Periodicals, Inc.

Abstract

We identify hardware that is optimal to produce molecular dynamics (MD) trajectories on Linux compute clusters with the GROMACS 2018 simulation package. Therefore, we benchmark the GROMACS performance on a diverse set of compute nodes and relate it to the costs of the nodes, which may include their lifetime costs for energy and cooling. In agreement with our earlier investigation using GROMACS 4.6 on hardware of 2014, the performance to price ratio of consumer GPU nodes is considerably higher than that of CPU nodes. However, with GROMACS 2018, the optimal CPU to GPU processing power balance has shifted even more toward the GPU. Hence, nodes optimized for GROMACS 2018 and later versions enable a significantly higher performance to price ratio than nodes optimized for older GROMACS versions. Moreover, the shift toward GPU processing allows to cheaply upgrade old nodes with recent GPUs, yielding essentially the same performance as comparable brand‐new hardware. © 2019 Wiley Periodicals, Inc.