Performance Evaluation of Golub-Kahan-Lanczos Algorithm with Reorthogonalization by Classical Gram-Schmidt Algorithm and OpenMP

• Item type: Trans(1)
• 公開日: 2016-08-10
• タイトル: Performance Evaluation of Golub-Kahan-Lanczos Algorithm with Reorthogonalization by Classical Gram-Schmidt Algorithm and OpenMP
• タイトル / 言語: en
• タイトル: Performance Evaluation of Golub-Kahan-Lanczos Algorithm with Reorthogonalization by Classical Gram-Schmidt Algorithm and OpenMP
• 言語: eng
• キーワード / 主題Scheme: Other
• キーワード / 主題: [オリジナル論文] subset computation of singular pairs, Golub-Kahan-Lanczos algorithm with reorthogonalization, classical Gram-Schmidt algorithm with reorthogonalization, OpenMP, sharedmemory multi-core processing
• 資源タイプ / 資源タイプ識別子: http://purl.org/coar/resource_type/c_6501
• 資源タイプ: journal article
• 著者所属: Research Group of Information and Communication Technology for Life, Nara Women's University
• 著者所属: Graduate School of Informatics, Kyoto University
• 著者所属: Graduate School of Informatics, Kyoto University
• 著者所属: Graduate School of Informatics, Kyoto University
• 著者所属: Graduate School of Informatics, Kyoto University
• 著者所属: Graduate School of Informatics, Kyoto University
• 著者所属(英): en
• 著者所属(英): Research Group of Information and Communication Technology for Life, Nara Women's University
• 著者所属(英): en
• 著者所属(英): Graduate School of Informatics, Kyoto University
• 著者所属(英): en
• 著者所属(英): Graduate School of Informatics, Kyoto University
• 著者所属(英): en
• 著者所属(英): Graduate School of Informatics, Kyoto University
• 著者所属(英): en
• 著者所属(英): Graduate School of Informatics, Kyoto University
• 著者所属(英): en
• 著者所属(英): Graduate School of Informatics, Kyoto University
• 著者名: Masami, Takata; Hiroyuki, Ishigami; Kinji, Kimura; Yuki, Fujii; Hiroki, Tanaka; Yoshimasa, Nakamura
• 著者名(英): Masami, Takata; Hiroyuki, Ishigami; Kinji, Kimura; Yuki, Fujii; Hiroki, Tanaka; Yoshimasa, Nakamura
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: The Golub-Kahan-Lanczos algorithm with reorthogonalization (GKLR algorithm) is an algorithm for computing a subset of singular triplets of large-scale sparse matrices. The reorthogonalization tends to become a bottleneck of the execution time, as the iteration number of the GKLR algorithm increases. In this paper, OpenMP-based parallel implementation of the classical Gram-Schmidt algorithm with reorthogonalization (OMP-CGS2 algorithm) is introduced. The OMP-CGS2 algorithm has the advantage of data reusability and is expected to achieve higher performance of the reorthogonalization computations on shared-memory multi-core processors with large caches than the conventional reorthogonalization algorithms. Numerical experiments on shared-memory multi-core processors show that the OMP-CGS2 algorithm accelerates the GKLR algorithm more effectively for computing a subset of singular triplets of some matrices than the conventional reorthogonalization algorithms. In addition, we discuss the cache utilization in the OMP-CGS2 algorithm and a condition that the OMP-CGS2 algorithm achieves higher performance than the CGS2 algorithm.
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: The Golub-Kahan-Lanczos algorithm with reorthogonalization (GKLR algorithm) is an algorithm for computing a subset of singular triplets of large-scale sparse matrices. The reorthogonalization tends to become a bottleneck of the execution time, as the iteration number of the GKLR algorithm increases. In this paper, OpenMP-based parallel implementation of the classical Gram-Schmidt algorithm with reorthogonalization (OMP-CGS2 algorithm) is introduced. The OMP-CGS2 algorithm has the advantage of data reusability and is expected to achieve higher performance of the reorthogonalization computations on shared-memory multi-core processors with large caches than the conventional reorthogonalization algorithms. Numerical experiments on shared-memory multi-core processors show that the OMP-CGS2 algorithm accelerates the GKLR algorithm more effectively for computing a subset of singular triplets of some matrices than the conventional reorthogonalization algorithms. In addition, we discuss the cache utilization in the OMP-CGS2 algorithm and a condition that the OMP-CGS2 algorithm achieves higher performance than the CGS2 algorithm.
• 書誌レコードID / 収録物識別子タイプ: NCID
• 書誌レコードID / 収録物識別子: AA11464803
• 書誌情報: 情報処理学会論文誌数理モデル化と応用（TOM）; 巻 9, 号 2, p. 1-8, 発行日 2016-08-10
• ISSN / 収録物識別子タイプ: ISSN
• ISSN / 収録物識別子: 1882-7780
• 出版者 / 言語: ja
• 出版者: 情報処理学会