Computational Resource Assignment for Parallel Execution

2014 Fiscal Year Final Research Report

• Project/Area Number: 25730077
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: High performance computing
• Research Institution: Waseda University
• Principal Investigator: KANEMITSU Hidehiro 早稲田大学, グローバルエデュケーションセンター, 助教 (60434362)
• Project Period (FY): 2013-04-01 – 2015-03-31
• Keywords: タスクスケジューリング / タスククラスタリング / 不均一分散環境

Outline of Final Research Achievements

In this research, I developed a method for processor utilization in a heterogeneous distributed system, which consists of : (i) derivation of the lower bound of the work load for each processor to be assigned for parallel execution, (ii) task assignment, (iii) task scheduling. Conventional task scheduling algorithms for heterogeneous systems have one serious drawback, i.e., they can not achieve a high degree of processor utilization because they assign an averaged scheduling priority for each task and then a good response time can not be obtained in a system having high heterogeneity. From experimental results by the simulation, it can be concluded that the proposed algorithm in this research outperforms other conventional task scheduling algorithms in terms of both the response time and efficiency.

Free Research Field

並列分散処理