{"links":{},"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00217483","sets":["1164:2240:10902:10903"]},"path":["10903"],"owner":"44499","recid":"217483","title":["Fortran標準規格do concurrentを用いたGPUオフローディング手法の評価"],"pubdate":{"attribute_name":"公開日","attribute_value":"2022-03-10"},"_buckets":{"deposit":"6adc8fb6-90eb-4e5d-a588-1f1a58b34df1"},"_deposit":{"id":"217483","pid":{"type":"depid","value":"217483","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"Fortran標準規格do concurrentを用いたGPUオフローディング手法の評価","author_link":["563601","563604","563603","563600","563602"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"Fortran標準規格do concurrentを用いたGPUオフローディング手法の評価"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"アクセラレータ","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2022-03-10","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"東京大学情報基盤センター"},{"subitem_text_value":"東京大学情報基盤センター"},{"subitem_text_value":"東京大学情報基盤センター"},{"subitem_text_value":"東京大学情報基盤センター"},{"subitem_text_value":"東京大学情報基盤センター"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Information Technology Center","subitem_text_language":"en"},{"subitem_text_value":"Information Technology Center","subitem_text_language":"en"},{"subitem_text_value":"Information Technology Center","subitem_text_language":"en"},{"subitem_text_value":"Information Technology Center","subitem_text_language":"en"},{"subitem_text_value":"Information Technology Center","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"item_publisher":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"情報処理学会","subitem_publisher_language":"ja"}]},"publish_status":"0","weko_shared_id":-1,"item_file_price":{"attribute_name":"Billing file","attribute_type":"file","attribute_value_mlt":[{"url":{"url":"https://ipsj.ixsq.nii.ac.jp/record/217483/files/IPSJ-HPC22183005.pdf","label":"IPSJ-HPC22183005.pdf"},"date":[{"dateType":"Available","dateValue":"2024-03-10"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-HPC22183005.pdf","filesize":[{"value":"1.1 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"660","billingrole":"5"},{"tax":["include_tax"],"price":"330","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"14"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"714a2a79-bdb6-48df-8ffa-b402cf3cb959","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2022 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"星野, 哲也"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"河合, 直聡"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"三木, 洋平"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"塙, 敏博"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"中島, 研吾"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10463942","subitem_source_identifier_type":"NCID"}]},"item_4_textarea_12":{"attribute_name":"Notice","attribute_value_mlt":[{"subitem_textarea_value":"SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc."}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_18gh","resourcetype":"technical report"}]},"item_4_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"2188-8841","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"Fortran 2008 の標準規格として導入された do concurrent 構文は，適用ループが並列実行可能であることを意味する．NVIDIA 社の nvfortran コンパイラは，do concurrent で記述されたループを NVIDIA GPU でオフロード実行する機能を提供しており，従って専用言語や指示文を用いない標準規格の Fortran プログラムの GPU 実行が可能となっている．本稿では，do concurrentによる GPU オフローディングの評価を目的とし，ICCG ソルバ，H-行列・ベクトル積，3 次元拡散方程式の do concurrent 実装を行った．OpenACC や OpenMP 5.x と比較を行い，NVIDIA Tesla A100 上で評価を行った．その結果，do concurrent の制約を受けないケースでは OpenACC や OpenMP 5.x と遜色ない性能が得られたものの，do concurrent の制約である縮約演算が必要なケースや明示的に CPU-GPU 間のデータ移動を行いたいケースにおいて，大きな性能ギャップが生じることや，制約を避けるための実装コストが大きいことがわかった．","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"8","bibliographic_titles":[{"bibliographic_title":"研究報告ハイパフォーマンスコンピューティング（HPC）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2022-03-10","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"5","bibliographicVolumeNumber":"2022-HPC-183"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"created":"2025-01-19T01:17:58.451389+00:00","updated":"2025-01-19T15:28:54.890738+00:00","id":217483}