Runtime Systems with Abstraction of High-Performance Computing Resources

• Project/Area Number: 25280043
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Partial Multi-year Fund
• Section: 一般
• Research Field: High performance computing
• Research Institution: Nagoya University
• Principal Investigator: Kato Shinpei 名古屋大学, 情報科学研究科, 准教授 (70631894)
• Co-Investigator(Kenkyū-buntansha): KOIBUCHI Michihiro 国立情報学研究所, アーキテクチャ科学研究系, 准教授 (40413926) ; KAWASHIMA Hideyuki 筑波大学, 大学院システム情報工学研究科, 准教授 (90407148)
• Co-Investigator(Renkei-kenkyūsha): FUJIWARA Ikki 国立情報学研究所, アーキテクチャ科学研究系, 特任研究員 (90648023) ; YUI Makoto 独立行政法人産業技術総合研究所, 情報技術部門, 研究員 (10586712) ; OHNO Kinji 名古屋大学, 大学院医学系研究科, 教授 (80397455) ; SUGIYAMA Yuki 名古屋大学, 大学院情報科学研究科, 教授 (20196778) ; ISHII Katsuya 名古屋大学, 情報連携基盤センター, 教授 (60134441)
• Project Period (FY): 2013-04-01 – 2016-03-31
• Project Status: Completed (Fiscal Year 2015)
• Budget Amount: ¥17,940,000 (Direct Cost: ¥13,800,000、Indirect Cost: ¥4,140,000); Fiscal Year 2015: ¥4,940,000 (Direct Cost: ¥3,800,000、Indirect Cost: ¥1,140,000); Fiscal Year 2014: ¥6,760,000 (Direct Cost: ¥5,200,000、Indirect Cost: ¥1,560,000); Fiscal Year 2013: ¥6,240,000 (Direct Cost: ¥4,800,000、Indirect Cost: ¥1,440,000)
• Keywords: 高性能計算 / データベース / 抽象化

Outline of Final Research Achievements

This project aimed at simplifying programming and data management of scientific applications. We developed runtime systems for high-performance computing clusters providing (i) virtualization of many-core accelerators (processor abstraction), (ii) virtualization of software-defined networks (network abstraction), and (iii) database encapsulation of scientific data (data abstraction).

Research Products

• [Journal Article] GPUvm: GPU Virtualization at the Hypervisor (2016)
  • Author(s): Yusuke Suzuki, Shinpei Kato, Hiroshi Yamada and Kenji Kono
  • Journal Title: IEEE Transactions on Computers Volume: 印刷中 Issue: 9 Pages: 1-1
  • DOI: 10.1109/tc.2015.2506582
  • Peer Reviewed / Acknowledgement Compliant
• [Journal Article] Fast, multi-channel real-time processing of signals with microsecond latency using graphics processing units (2014)
  • Author(s): N. Rath, S. Kato, J.P. Levesque, M.E. Mauel, G.A. Navratil, and Q. Peng
  • Journal Title: Review of Scientific Instruments Volume: 85 Issue: 4
  • DOI: 10.1063/1.4870901
  • Peer Reviewed / Open Access
• [Presentation] An Empirical Study on the NoC Architecture Based on Bidirectional Ring and Mesh Topologies (2016)
  • Author(s): Jie Yin, Ye Liu, Shinpei Kato, Hiroshi Sasaki, Hiroaki Takada
  • Organizer: The 2016 Workshop on Multicore and Rack-scale Systems
  • Place of Presentation: London, UK
  • Year and Date: 2016-04-18
  • Int'l Joint Research
• [Presentation] Towards Multi-tenant GPGPU: Event-driven Programming Model for System-wide Scheduling on Shared GPUs (2016)
  • Author(s): Yusuke Suzuki, Hiroshi Yamada, Shinpei Kato, Kenji Kono
  • Organizer: The 2016 Workshop on Multicore and Rack-scale Systems
  • Place of Presentation: London, UK
  • Year and Date: 2016-04-18
  • Int'l Joint Research
• [Presentation] Recursive Incremental Computation for Efficient Window Aggregate over Array Database (2016)
  • Author(s): Jiang Li, Hideyuki Kawashima, Osamu Tatebe
  • Organizer: 情報処理学会研究報告システ ムソフトウェアとオペレーティング・システム
  • Place of Presentation: 理化学研究所計算科学研究機構
  • Year and Date: 2016-02-29
• [Presentation] Singularity of Future Computer-System Networks (2015)
  • Author(s): Michihiro Koibuchi
  • Organizer: ACM International Symposium on Information and Communication Technology
  • Place of Presentation: Hue, Vietnam
  • Year and Date: 2015-12-03
• [Presentation] Connected Smartphones and High-Performance Servers for Remote Object Detection (2014)
  • Author(s): Y. Iida, M. Hirabayashi, T. Azumi, N. Nishio, and S. Kato
  • Organizer: IEEE International Conference on Cyber-Physical Systems, Networks, and Applications
  • Place of Presentation: 香港
  • Year and Date: 2014-08-25 – 2014-08-26
• [Presentation] Reducing Data Copies between GPUs and NICs (2014)
  • Author(s): A. Nguyen, Y. Fujii, Y. Iida, T. Azumi, N. Nishio, and S. Kato
  • Organizer: IEEE International Conference on Cyber-Physical Systems, Networks, and Applications
  • Place of Presentation: 香港
  • Year and Date: 2014-08-25 – 2014-08-26
• [Presentation] Implementing Incremental Aggregate Computation on SciDB (2014)
  • Author(s): Li Jiang, Hideyuki Kawashima, Osamu Tatebe
  • Organizer: 情報処理学会研究報告システムソフトウェアとオペレーティング・システム（OS）
  • Place of Presentation: 新潟
  • Year and Date: 2014-07-21
• [Presentation] GPUvm: Why Not Virtualizing GPUs at the Hypervisor? (2014)
  • Author(s): Y. Suzuki, S. Kato, H. Yamada, and K. Kono
  • Organizer: USENIX Annual Technical Conference
  • Place of Presentation: 米国フィラデルフィア
  • Year and Date: 2014-06-17 – 2014-06-20
• [Presentation] GPUvm: Why Not Virtualizing GPUs at the Hypervisor? (2014)
  • Author(s): Yusuke Suzuki, Shinpei Kato, Hiroshi Yamada, Kenji Kono
  • Organizer: USENIX Annual Technical Conference
  • Place of Presentation: Philadelphia, USA
• [Presentation] GDM: Device Memory Management for GPGPU Computing (2014)
  • Author(s): Kaibo Wang, Xiaoning Ding, Rubao Lee, Shinpei Kato, Xiaodong Zhang
  • Organizer: ACM SIGMETRICS
  • Place of Presentation: Texas, USA
• [Presentation] Power and Performance Characterization and Modeling of GPU-accelerated Systems (2014)
  • Author(s): Yuki Abe, Hiroshi Sasaki, Shinpei Kato, Koji Inoue, Masato Edahiro, Martin Peres
  • Organizer: IEEE International Parallel & Distributed Processing Symposium
  • Place of Presentation: Phoenix, USA
• [Presentation] リアルタイム系ビッグデータ利活用のための アクセラレータ型オンボードデー タベース (2014)
  • Author(s): 加藤真平
  • Organizer: 情報処理学会 第76回全国大会 企画セッション「先進的ハードウェアによる 超 高性能データ処理へ向けて」
  • Place of Presentation: 東京
  • Invited
• [Presentation] Data Transfer Matters for GPU Computing (2013)
  • Author(s): Yusuke Fujii, Takuya Azumi, Shinpei Kato, Nobuhiko Nishio
  • Organizer: IEEE International Conference on Parallel and Distributed Systems
  • Place of Presentation: Seoul, Korea
• [Presentation] GPU Implementations of Object Detection using HOG Features and Deformable Models (2013)
  • Author(s): Manato Hirabayashi, Shinpei Kato, Masato Edahiro, Kazuya Takeda, Taiki Kawano, Seiichi Mita
  • Organizer: IEEE International Conference on Cyber-Physical Systems, Networks, and Applications
  • Place of Presentation: Taipei, Taiwan
• [Presentation] Zero-Copy I/O Processing for Low-Latency GPU Computing (2013)
  • Author(s): Shinpei Kato, Jason Aumiller, Scott Brandt
  • Organizer: ACM/IEEE International Conference on Cyber-Physical Systems
  • Place of Presentation: Philadelphia, USA
• [Presentation] Performance Degradation by Deactivated Cores in 2-D Mesh NoCs (2013)
  • Author(s): Ikki Fujiwara, Michihiro Koibuchi, Hiroki Matsutani
  • Organizer: IEEE International Symposium on Embedded Multicore Systems-on-chip
  • Place of Presentation: Tokyo, Japan
• [Presentation] Investigating Performance Advantages of Random Topologies on Network-on-Chip (2013)
  • Author(s): Sarat Yoowattana, Ikki Fujiwara, Michihiro Koibuchi
  • Organizer: Workshop on Synthesis And System Integration of Mixed Information technologies
  • Place of Presentation: Sapporo, Japan
• [Presentation] An Incremental Computation Scheme over Array Database (2013)
  • Author(s): Jiang Li , Hideyuki Kawashima
  • Organizer: 情報処理学会研究報告データベースシステム
  • Place of Presentation: 京都
• [Presentation] サイバーフィジカルとメニーコアが創出するイノベーション (2013)
  • Author(s): 加藤真平
  • Organizer: CEATEC JAPAN 2013
  • Place of Presentation: 東京
  • Invited
• [Remarks] Gdev: First-Class GPU Resource Management
  • URL: https://github.com/cpfl/gdev
• [Remarks] Gdev
  • URL: https://github.com/CPFL/gdev
• [Remarks] GXen
  • URL: https://github.com/CPFL/gxen
• [Remarks] Gdevオープンソースソフトウェア
  • URL: https://github.com/CS005/gdev
• [Remarks] GPUvmオープンソースソフトウェア
  • URL: https://github.com/CS005/gxen