Leveraging GPUDirect Storage for Efficient Image Reconstruction

• Item type: SIG Technical Reports(1)
• 公開日: 2024-08-01
• タイトル: Leveraging GPUDirect Storage for Efficient Image Reconstruction
• タイトル / 言語: en
• タイトル: Leveraging GPUDirect Storage for Efficient Image Reconstruction
• 言語: eng
• キーワード / 主題Scheme: Other
• キーワード / 主題: 分散システム
• 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: Tokyo Institute of Technology／RIKEN Center for Computational Science
• 著者所属: National Institute of Advanced Industrial Science and Technology (AIST)／RIKEN Center for Computational Science
• 著者所属: Tokyo Institute of Technology
• 著者所属: RIKEN Center for Computational Science
• 著者所属: RIKEN Center for Computational Science
• 著者所属(英): en
• 著者所属(英): Tokyo Institute of Technology / RIKEN Center for Computational Science
• 著者所属(英): en
• 著者所属(英): National Institute of Advanced Industrial Science and Technology (AIST) / RIKEN Center for Computational Science
• 著者所属(英): en
• 著者所属(英): Tokyo Institute of Technology
• 著者所属(英): en
• 著者所属(英): RIKEN Center for Computational Science
• 著者所属(英): en
• 著者所属(英): RIKEN Center for Computational Science
• 著者名: Du, Wu; Peng, Chen; Toshio, Endo; Satoshi, Matsuoka; Mohamed, Wahib
• 著者名(英): Du, Wu; Peng, Chen; Toshio, Endo; Satoshi, Matsuoka; Mohamed, Wahib
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: GPUDirect Storage, a novel tool provided by Nvidia, facilitates better utilization of GPUs by avoiding extra copies through a bounce buffer in the CPU host memory and enabling direct memory access. This technology offers significant advantages, particularly its high throughput capabilities. However, it also presents challenges due to high instruction latency. In image reconstruction, input data transfer often involves multiple small data blocks, which prevents optimal utilization of GPUDirect Storage, while output data transfer is large data blocks, which is well suited for GPUDirect Storage. This paper experimentally confirms the high throughput and high instruction latency of GPUDirect Storage. We also propose a novel solution to improve the efficiency of small data block transfers in image reconstruction, leveraging GPUDirect Storage's strengths while mitigating its latency issues.
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: GPUDirect Storage, a novel tool provided by Nvidia, facilitates better utilization of GPUs by avoiding extra copies through a bounce buffer in the CPU host memory and enabling direct memory access. This technology offers significant advantages, particularly its high throughput capabilities. However, it also presents challenges due to high instruction latency. In image reconstruction, input data transfer often involves multiple small data blocks, which prevents optimal utilization of GPUDirect Storage, while output data transfer is large data blocks, which is well suited for GPUDirect Storage. This paper experimentally confirms the high throughput and high instruction latency of GPUDirect Storage. We also propose a novel solution to improve the efficiency of small data block transfers in image reconstruction, leveraging GPUDirect Storage's strengths while mitigating its latency issues.
• 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-195, 号 5, p. 1-6, 発行日 2024-08-01
• 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