Enabling GPU acceleration with messaging middleware

• Main Authors: DURAN, Randall E., ZHANG, Li, HAYHURST, Tom
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2011
• Subjects: GPU; GPGPU; middleware messaging; CUDA; k-means clustering; ZeroMQ; stream processing; Databases and Information Systems; Software Engineering
• Online Access: https://ink.library.smu.edu.sg/sis_research/6446; https://ink.library.smu.edu.sg/context/sis_research/article/7449/viewcontent/enabling_gpu.pdf
• Institution: Singapore Management University

Graphics processing units (GPUs) offer great potential for accelerating processing for a wide range of scientific and business applications. However, complexities associated with using GPU technology have limited its use in applications. This paper reviews earlier approaches improving GPU accessibility, and explores how integration with middleware messaging technologies can further improve the accessibility and usability of GPU-enabled platforms. The results of a proof-of-concept integration between an open-source messaging middleware platform and a general-purpose GPU platform using the CUDA framework are presented. Additional applications of this technique are identified and discussed as potential areas for further research.