Grant-in-Aid for International Scientific Research.
|Section||University-to-University Cooperative Research|
|Research Institution||Gunma University, Faculty of Engineering|
IGARASHI Yoshihide Department of Computer Science, Gunma University, Professor, 工学部, 教授 (60006260)
WILKINSON Ti シティ大学, 研究員
MCCANN Julie City University, 研究員
MURRAY Kevin City University, 講師
DOWLING Geof シティ大学, 講師
OSMON Peter City University, 教授
茂木 和弘 群馬大学, 工学部, 助手 (00251124)
阿久津 達也 群馬大学, 工学部, 助教授 (90261859)
白石 洋一 群馬大学, 工学部, 助教授 (80261858)
西谷 泰昭 群馬大学, 工学部, 助教授 (60198463)
WILKINSON Tim City University
DOWLING Geoff City University
MOTEGI Kazuhiro Department of Computer Science, Gunma University, Research Assistant
NISHITANI Yasuyuki Department of Computer Science, Gunma University, Associate Professor
SHIRAISHI Yoichi Department of Computer Science, Gunma University, Associate Professor
|Project Fiscal Year
1995 – 1996
Completed(Fiscal Year 1996)
|Budget Amount *help
¥3,700,000 (Direct Cost : ¥3,700,000)
Fiscal Year 1996 : ¥1,700,000 (Direct Cost : ¥1,700,000)
Fiscal Year 1995 : ¥2,000,000 (Direct Cost : ¥2,000,000)
|Keywords||parallel / distribution / algorithm / computer / semi-conductor / device / communication / synchronization / fault-tolerant / 並列 / 分散 / アルゴリズム / 計算機 / 半導体 / デバイス / 通信 / 同期 / 並列・分散処理 / 並列オペレーティングシステム / 並列・分散アルゴリズム / 耐故障性 / VLSIレイアウト / AND-EXOR論理式 / 遺伝情報解析 / デバイスシミュレーション|
The design and the analysis of the parallel and distributed algorithms and the parallelising method of the very time-consuming computations are researched and developed in the group of the computer science department of Gunma University. In the systems architecture research center of the City University, the scaleable parallel computer and the parallel operating system are researched and developed. Then, the joint research project between these universities is established in order to research and develop the practical parallel and distributed computation method by exchanging the information between their main research fields.
In summary, the results (5) - (8) are attained corresponding to the objectives (1) - (4), respectively.
(1) The efficiency verification of the speed-up and the performance of the parallel algorithm solving the layout problem how to place the components and to route the interconnections between them in the VLSI chip.
(2) The efficiency verification of the ac
celeration of the parallel computation method in the device simulator based on the data distribution and the algorithm solving simultaneous linear equations.
(3) The theoretical analyzes and the experimental verifications of the communication, synchronization, broadcasting, load-balancing in the parallel and distributed operating system. Moreover, the fault-tolerance in the data transmission in the fundamental parallel and distributed algorithms should be theoretically analyzed and be experimentally verified.
(4) The development of the platform on a parallel computer which is used for the application program development and the utilization of the experience of the parallel application program development.
(5) The efficiency of the parallel placement algorithm which hierarchically places a set of components on the placement area of the VLSI chip. Moreover, various kinds of problems in the speed-up and in the performance of the placement results are found.
(6) The prototype of the parallel device simulator is developed and the experimentation environment is partly established.
(7) The communication, synchronization, broadcasting and load-balancing, etc.of the parallel operating system are theoretically analyzed and experimentally verified. This operating system is categorized into the same type as that developed in the group of the City University.
(8) A part of the platform has been developed on the operating system, Angel, which is being developed in the group of the City University. Actually, the most important function, mutual exclusion control, is implemented and is experimented. Less