GROMACS on AMD GPU-Based HPC Platforms: Using SYCL for Performance and Portability

TL;DR

This paper demonstrates how GROMACS, a molecular dynamics software, can effectively use SYCL to achieve performance and portability on AMD GPUs, facilitating its deployment on diverse HPC platforms.

Contribution

The paper presents the adaptation of GROMACS to SYCL for AMD GPUs, showing that performance can be maintained while enhancing portability across multiple hardware vendors.

Findings

SYCL enables GROMACS to run efficiently on AMD GPUs.

Portability is achieved without significant performance loss.

Performance analysis provides best practices for SYCL GPU programming.

Abstract

GROMACS is a widely-used molecular dynamics software package with a focus on performance, portability, and maintainability across a broad range of platforms. Thanks to its early algorithmic redesign and flexible heterogeneous parallelization, GROMACS has successfully harnessed GPU accelerators for more than a decade. With the diversification of accelerator platforms in HPC and no obvious choice for a multi-vendor programming model, the GROMACS project found itself at a crossroads. The performance and portability requirements, and a strong preference for a standards-based solution, motivated our choice to use SYCL on both new HPC GPU platforms: AMD and Intel. Since the GROMACS 2022 release, the SYCL backend has been the primary means to target AMD GPUs in preparation for exascale HPC architectures like LUMI and Frontier. SYCL is a cross-platform, royalty-free, C++17-based standard for programming hardware accelerators. It allows using the same code to target GPUs from all three major vendors with minimal specialization. While SYCL implementations build on native toolchains, performance of such an approach is not immediately evident. Biomolecular simulations have challenging performance characteristics: latency sensitivity, the need for strong scaling, and typical iteration times as short as hundreds of microseconds. Hence, obtaining good performance across the range of problem sizes and scaling regimes is particularly challenging. Here, we share the results of our work on readying GROMACS for AMD GPU platforms using SYCL, and demonstrate performance on Cray EX235a machines with MI250X accelerators. Our findings illustrate that portability is possible without major performance compromises. We provide a detailed analysis of node-level kernel and runtime performance with the aim of sharing best practices with the HPC community on using SYCL as a performance-portable GPU framework.