Connected component identification and cluster update on GPU

TL;DR

This paper explores how to efficiently implement connected component identification and cluster updates on GPUs, comparing various parallelization strategies to serial methods for improved performance in physics and engineering simulations.

Contribution

It evaluates different GPU parallelization approaches for cluster labeling and updates, providing insights into their effectiveness versus serial implementations.

Findings

GPU approaches significantly speed up cluster identification

Parallel algorithms outperform serial methods in large-scale problems

Certain strategies are more suitable for non-local cluster updates

Abstract

Cluster identification tasks occur in a multitude of contexts in physics and engineering such as, for instance, cluster algorithms for simulating spin models, percolation simulations, segmentation problems in image processing, or network analysis. While it has been shown that graphics processing units (GPUs) can result in speedups of two to three orders of magnitude as compared to serial codes on CPUs for the case of local and thus naturally parallelized problems such as single-spin flip update simulations of spin models, the situation is considerably more complicated for the non-local problem of cluster or connected component identification. I discuss the suitability of different approaches of parallelization of cluster labeling and cluster update algorithms for calculations on GPU and compare to the performance of serial implementations.