Performance optimizations for porting the openQ$^\star$D package to GPUs

TL;DR

This paper discusses optimizing the openQ$^ op$D package for GPU platforms, focusing on solver improvements and a new adaptive CPU/GPU hybrid method to enhance performance in lattice QCD simulations.

Contribution

It introduces a novel adaptive CPU/GPU hybrid implementation and analyzes solver optimizations for GPU acceleration in openQ$^ op$D.

Findings

Improved solver performance on GPU platforms.

Effective adaptive CPU/GPU hybrid approach.

Enhanced support for multiple GPU vendors.

Abstract

OpenQ$^\star$D code has been used by the RC$^\star$ collaboration for the generation of fully dynamical QCD+QED gauge configurations with C$^\star$ boundary conditions. In this talk, optimization of solvers provided with the openQ$^\star$D package relevant for porting the code on GPU-accelerated supercomputing platforms is discussed. We present the analysis of the current implementations of the GCR solver preconditioned with Schwarz alternating procedure for ill-conditioned Dirac-operators. With the goal of enabling support for GPUs from various vendors, a novel method of adaptive CPU/GPU-hybrid implementation is proposed.