High Speed Computation for Eigenproblems and Singular Value Decomposition Using Multiple Division Divide-and-Conquer Method

Research Project

Project/Area Number	23760072
Research Category	Grant-in-Aid for Young Scientists (B)
Allocation Type	Multi-year Fund
Research Field	Engineering fundamentals
Research Institution	Saitama University
Principal Investigator	KUWAJIMA Yutaka 埼玉大学, 理工学研究科, 助教 (40451736)
Project Period (FY)	2011-04-28 – 2015-03-31
Project Status	Completed (Fiscal Year 2014)
Budget Amount *help	¥1,690,000 (Direct Cost: ¥1,300,000、Indirect Cost: ¥390,000) Fiscal Year 2014: ¥260,000 (Direct Cost: ¥200,000、Indirect Cost: ¥60,000) Fiscal Year 2013: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000) Fiscal Year 2012: ¥390,000 (Direct Cost: ¥300,000、Indirect Cost: ¥90,000) Fiscal Year 2011: ¥390,000 (Direct Cost: ¥300,000、Indirect Cost: ¥90,000)
Keywords	固有値問題 / 特異値分解 / 分割統治法 / 多分割の分割統治法 / 並列計算
Outline of Final Research Achievements	We implemented divide-and-conquer method with multiple division for singular value decomposition on parallel computers. It is achieved by applying the divide-and-conquer method with multiple division for symmetric eigenvalue problems. We improved algorithm for eigenproblem of real diagonal with low-rank perturbation which is main part in terms of accuracy. As a result, it can be expected to reduce the calculation time in the worst case.

Report

(5 results)