A Discrete Event Simulation Framework for Distributed GPGPU Environments

2015 Fiscal Year Final Research Report

• Project/Area Number: 26730044
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Software
• Research Institution: Tokyo Metropolitan Industrial Technology Research Institute
• Principal Investigator: Ohara Mamoru 地方独立行政法人東京都立産業技術研究センター, 開発本部開発第一部, 主任研究員 (80463008)
• Co-Investigator(Renkei-kenkyūsha): FUKUMOTO Satoshi 首都大学東京, システムデザイン研究科, 教授 (50247590)
• Project Period (FY): 2014-04-01 – 2016-03-31
• Keywords: GPGPU / 離散イベントシミュレーション / タイムワープ法 / 空間的冗長化 / ソフトウェア若化 / 信頼性解析モデル

Outline of Final Research Achievements

GPGPU techniques, which utilize graphic processors to accelerate general purpose computing, have become popular in the past several years. In this study, we developed a technique for efficiently executing discrete event simulations (DES) using GPGPU. Processing branches lead performance degradation in GPGPU. Thus we reduced the branches by introducing temporal redundancy. Moreover, we developed a number of software rejuvenation schemes for parallel DES. We also constructed an analytical model for evaluating reliability of a simple software rejuvenation technique.

Free Research Field

ディペンダブルコンピューティング