Job Parallelism using Graphical Processing Unit Individual Multi-Processors and Localised Memory

摘要

Graphical Processing Units (GPUs) are usually programmed to provide data-parallel acceleration to a host processor. Modern GPUs typically have an internal multi-processor (MP) structure that can be exploited in an unusual way to offer semi-independent task parallelism providing the MPs can operate within their own localised memory and apply data-parallelism to their own problem subset. We describe a combined simulation and statistical analysis application using component labelling and benchmark it on a range of modern GPU and CPU devices with various numbers of cores. As well as demonstrating a high degree of job parallelism and throughput we find a typical GPU MP outperforms a conventional CPU core.