Effects of Process/Thread Allocation for Optimization of Communication-Computation Overlapping in Parallel Multigrid Methods

• Item type: SIG Technical Reports(1)
• 公開日: 2024-12-09
• タイトル: Effects of Process/Thread Allocation for Optimization of Communication-Computation Overlapping in Parallel Multigrid Methods
• タイトル / 言語: en
• タイトル: Effects of Process/Thread Allocation for Optimization of Communication-Computation Overlapping in Parallel Multigrid Methods
• 言語: eng
• キーワード / 主題Scheme: Other
• キーワード / 主題: アーキテクチャ
• 資源タイプ / 資源タイプ識別子: http://purl.org/coar/resource_type/c_18gh
• 資源タイプ: technical report
• 著者所属: Information Technology Center, The University of Tokyo／RIKEN, Center for Computational Science (R-CCS)
• 著者所属(英): en
• 著者所属(英): Information Technology Center, The University of Tokyo / RIKEN, Center for Computational Science (R-CCS)
• 著者名: Kengo, Nakajima
• 著者名(英): Kengo, Nakajima
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: Preconditioned iterative methods based on the Krylov subspace technique are widely employed in various scientific and technical computing. When utilizing large-scale parallel computing systems, the communication overhead tends to increase with the growth in the number of nodes, making its reduction a crucial challenge. In parallel FEM/FVM, halo communication and computation overlapping (CC-Overlapping) are commonly employed, often in conjunction with the dynamic loop scheduling feature of OpenMP. In the previous work, the author proposes a method to apply CC-Overlapping to the forward and backward substitutions of the IC(0) smoother of the parallel Conjugate Gradient method preconditioned by Multigrid (MGCG). Using up to 4,096 nodes on Wisteria/BDEC-01 (Odyssey) with A64FX, performance improvement of approximately 40+% was achieved compared to the original implementation. In the present work, effects of process/thread allocation within a compute node in OpenMP/MPI Hybrid parallel programming model has been conducted for optimization of CC-Overlapping.
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: Preconditioned iterative methods based on the Krylov subspace technique are widely employed in various scientific and technical computing. When utilizing large-scale parallel computing systems, the communication overhead tends to increase with the growth in the number of nodes, making its reduction a crucial challenge. In parallel FEM/FVM, halo communication and computation overlapping (CC-Overlapping) are commonly employed, often in conjunction with the dynamic loop scheduling feature of OpenMP. In the previous work, the author proposes a method to apply CC-Overlapping to the forward and backward substitutions of the IC(0) smoother of the parallel Conjugate Gradient method preconditioned by Multigrid (MGCG). Using up to 4,096 nodes on Wisteria/BDEC-01 (Odyssey) with A64FX, performance improvement of approximately 40+% was achieved compared to the original implementation. In the present work, effects of process/thread allocation within a compute node in OpenMP/MPI Hybrid parallel programming model has been conducted for optimization of CC-Overlapping.
• 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: AN10463942
• 書誌情報: 研究報告ハイパフォーマンスコンピューティング（HPC）; 巻 2024-HPC-197, 号 23, p. 1-10, 発行日 2024-12-09
• ISSN / 収録物識別子タイプ: ISSN
• ISSN / 収録物識別子: 2188-8841
• Notice: SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc.
• 出版者 / 言語: ja
• 出版者: 情報処理学会