Portable Lattice QCD implementation based on OpenCL

TL;DR

This paper presents a portable OpenCL implementation of Lattice QCD, benchmarking its performance across different GPU architectures and comparing it with CUDA-based implementations to evaluate portability and efficiency.

Contribution

The authors developed an OpenCL backend for Lattice QCD simulations, enabling cross-architecture compatibility and providing performance benchmarks against CUDA implementations.

Findings

OpenCL implementation performs comparably to CUDA on Nvidia GPUs

Significant portability achieved across AMD and Nvidia hardware

Benchmark results demonstrate the viability of OpenCL for Lattice QCD

Abstract

The presence of GPU from different vendors demands the Lattice QCD codes to support multiple architectures. To this end, Open Computing Language (OpenCL) is one of the viable frameworks for writing a portable code. It is of interest to find out how the OpenCL implementation performs as compared to the code based on a dedicated programming interface such as CUDA for Nvidia GPUs. We have developed an OpenCL backend for our already existing code of the Wuppertal-Budapest collaboration. In this contribution, we show benchmarks of the most time consuming part of the numerical simulation, namely, the inversion of the Dirac operator. We present the code performance on the JUWELS and LUMI Supercomputers based on Nvidia and AMD graphics cards, respectively, and compare with the CUDA backend implementation.