Speeding up Large-Scale Combinatorial Search by Parallel Accelerator with Fast Mutual Communication Function

2016 Fiscal Year Final Research Report

• Project/Area Number: 23300014
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Single-year Grants
• Section: 一般
• Research Field: Computer system/Network
• Research Institution: The University of Tokyo
• Principal Investigator: Mary Inaba 東京大学, 大学院情報理工学系研究科, 准教授 (60282711)
• Co-Investigator(Renkei-kenkyūsha): Hiroshi Imai 東京大学, 大学院情報理工学系研究科, 教授 (80183010) ; Ayumu Nagai 群馬大学, 工学部情報工学科, 助教 (70375567)
• Project Period (FY): 2011-04-01 – 2016-03-31
• Keywords: リコンフィギャラブルシステム / 離散最適化 / ハートウェアアルゴリズム

Outline of Final Research Achievements

This research aims at enlarging the application fields of parallel accelerators to large-scale combinatorial optimization by efficiently searching optimal solutions in a distributed manner. For this purpose, we propose and implement a unified system which equips the mutual communication function, combining programmable hardware accelerators and host servers, and software layer bridging application layer and accelerator with hiding complicataed communication control. Using this system, we expand applications from the constraint satisfaction problem, CSP, to planned serendipity.

Free Research Field

計算機システム・ネットワーク