GLoop: An event-driven runtime for consolidating GPGPU applications

Yusuke Suzuki, Hiroshi Yamada, Shinpei Kato, Kenji Kono

研究成果: Conference contribution

5 被引用数 (Scopus)

抄録

Graphics processing units (GPUs) have become an attractive platform for general-purpose computing (GPGPU) in various domains. Making GPUs a time-multiplexing resource is a key to consolidating GPGPU applications (apps) in multi-tenant cloud platforms. However, advanced GPGPU apps pose a new challenge for consolidation. Such highly functional GPGPU apps, referred to as GPU eaters, can easily monopolize a shared GPU and starve collocated GPGPU apps. This paper presents GLoop, which is a software run-time that enables us to consolidate GPGPU apps including GPU eaters. GLoop offers an event-driven programming model, which allows GLoop-based apps to inherit the GPU eaters' high functionality while proportionally scheduling them on a shared GPU in an isolated manner. We implemented a prototype of GLoop and ported eight GPU eaters on it. The experimental results demonstrate that our prototype successfully schedules the consolidated GPGPU apps on the basis of its scheduling policy and isolates resources among them.

本文言語English
ホスト出版物のタイトルSoCC 2017 - Proceedings of the 2017 Symposium on Cloud Computing
出版社Association for Computing Machinery, Inc
ページ80-93
ページ数14
ISBN(電子版)9781450350280
DOI
出版ステータスPublished - 2017 9 24
イベント2017 Symposium on Cloud Computing, SoCC 2017 - Santa Clara, United States
継続期間: 2017 9 242017 9 27

出版物シリーズ

名前SoCC 2017 - Proceedings of the 2017 Symposium on Cloud Computing

Other

Other2017 Symposium on Cloud Computing, SoCC 2017
国/地域United States
CitySanta Clara
Period17/9/2417/9/27

ASJC Scopus subject areas

  • 計算理論と計算数学
  • 理論的コンピュータサイエンス

フィンガープリント

「GLoop: An event-driven runtime for consolidating GPGPU applications」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

引用スタイル