Project/Area Number |
19K11887
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Multi-year Fund |
Section | 一般 |
Review Section |
Basic Section 60040:Computer system-related
|
Research Institution | Kanagawa University |
Principal Investigator |
|
Co-Investigator(Kenkyū-buntansha) |
金子 敬一 東京農工大学, 工学(系)研究科(研究院), 教授 (20194904)
|
Project Period (FY) |
2019-04-01 – 2023-03-31
|
Project Status |
Completed (Fiscal Year 2022)
|
Budget Amount *help |
¥4,420,000 (Direct Cost: ¥3,400,000、Indirect Cost: ¥1,020,000)
Fiscal Year 2021: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2020: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2019: ¥2,210,000 (Direct Cost: ¥1,700,000、Indirect Cost: ¥510,000)
|
Keywords | interconnect / supercomputer / dependability / routing / graph / fault-tolerance / interconnection network / graph theory / algorithm / torus / cycles / cluster / interconnection / fault tolerance / 計算機システム関連 / 相互結合網 / 経路選択問題 / システムディペンダビリティ / 信頼性 |
Outline of Research at the Start |
現代のスーパーコンピュータは数百万を超えるプロセッサを含み、複数の計算を同時に実施する、いわゆる並列処理を行うことで、最終結果を得るまでの計算時間を減らすことを目的とする。この考え方を実現するに当たり、できるだけ高速かつ確実にプロセッサ間で情報を通信する必要がある。プロセッサの数は膨大なため、故障プロセッサの存在を前提にし、本研究では、耐故障性を持つプロセッサ間の情報通信方法の提案を目的とする。
|
Outline of Final Research Achievements |
Throughout this research project, we have conducted investigations regarding the interconnection network of massively parallel systems, such as supercomputers. We have especially focused on torus-based interconnects, and considered related theoretical problems such as routing algorithms and related graph theory problems. In addition, we have considered related practical problems, for instance thinking about how to reduce power consumption of interconnection networks since power consumption of modern supercomputers is an on-going issue. Concretely, we have investigated low-power devices and the corresponding technologies, and IoT devices in general to this end.
|
Academic Significance and Societal Importance of the Research Achievements |
Modern supercomputers include numerous compute nodes, called processors. In order to maximise the system performance, it is critical to efficiently connect all these nodes as it conditions efficient data transfers and usage in general of the processors.
|