Data transfer matters for GPU computing

Yusuke Fujii, Takuya Azumi, Nobuhiko Nishio, Shinpei Kato, Masato Edahiro

Research output: Chapter in Book/Report/Conference proceedingConference contribution

40 Citations (Scopus)

Abstract

Graphics processing units (GPUs) embrace many-core compute devices where massively parallel compute threads are offloaded from CPUs. This heterogeneous nature of GPU computing raises non-trivial data transfer problems especially against latency-critical real-time systems. However even the basic characteristics of data transfers associated with GPU computing are not well studied in the literature. In this paper, we investigate and characterize currently-achievable data transfer methods of cutting-edge GPU technology. We implement these methods using open-source software to compare their performance and latency for real-world systems. Our experimental results show that the hardware-assisted direct memory access (DMA) and the I/O read-and-write access methods are usually the most effective, while on-chip micro controllers inside the GPU are useful in terms of reducing the data transfer latency for concurrent multiple data streams. We also disclose that CPU priorities can protect the performance of GPU data transfers.

Original languageEnglish
Title of host publicationProceedings - 2013 19th IEEE International Conference on Parallel and Distributed Systems, ICPADS 2013
PublisherIEEE Computer Society
Pages275-282
Number of pages8
ISBN (Print)9781479920815
DOIs
Publication statusPublished - 2013
Externally publishedYes
Event2013 19th IEEE International Conference on Parallel and Distributed Systems, ICPADS 2013 - Seoul, Korea, Republic of
Duration: 2013 Dec 152013 Dec 18

Publication series

NameProceedings of the International Conference on Parallel and Distributed Systems - ICPADS
ISSN (Print)1521-9097

Conference

Conference2013 19th IEEE International Conference on Parallel and Distributed Systems, ICPADS 2013
CountryKorea, Republic of
CitySeoul
Period13/12/1513/12/18

Keywords

  • Data Transfer
  • GPGPU
  • Latency
  • OS
  • Performance

ASJC Scopus subject areas

  • Hardware and Architecture

Fingerprint Dive into the research topics of 'Data transfer matters for GPU computing'. Together they form a unique fingerprint.

  • Cite this

    Fujii, Y., Azumi, T., Nishio, N., Kato, S., & Edahiro, M. (2013). Data transfer matters for GPU computing. In Proceedings - 2013 19th IEEE International Conference on Parallel and Distributed Systems, ICPADS 2013 (pp. 275-282). [6808184] (Proceedings of the International Conference on Parallel and Distributed Systems - ICPADS). IEEE Computer Society. https://doi.org/10.1109/ICPADS.2013.47