{"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00217481","sets":["1164:2240:10902:10903"]},"path":["10903"],"owner":"44499","recid":"217481","title":["タンパク質立体構造予測システムAlphaFoldのTSUBAME3.0上での高速化"],"pubdate":{"attribute_name":"公開日","attribute_value":"2022-03-10"},"_buckets":{"deposit":"9da0ec9c-c987-4d54-8806-e4eaa63b8839"},"_deposit":{"id":"217481","pid":{"type":"depid","value":"217481","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"タンパク質立体構造予測システムAlphaFoldのTSUBAME3.0上での高速化","author_link":["563594","563595","563592","563593"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"タンパク質立体構造予測システムAlphaFoldのTSUBAME3.0上での高速化"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"高性能計算","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":"東京工業大学情報理工学院"},{"subitem_text_value":"東京工業大学学術国際情報センター"},{"subitem_text_value":"東京工業大学学術国際情報センター／東京工業大学情報理工学院"},{"subitem_text_value":"東京工業大学情報理工学院"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"School of Computing, Tokyo Institute of Technology","subitem_text_language":"en"},{"subitem_text_value":"Global Scientific Information and Computing Center, Tokyo Institute of Technology","subitem_text_language":"en"},{"subitem_text_value":"Global Scientific Information and Computing Center, Tokyo Institute of Technology / School of Computing, Tokyo Institute of Technology","subitem_text_language":"en"},{"subitem_text_value":"School of Computing, Tokyo Institute of Technology","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/217481/files/IPSJ-HPC22183003.pdf","label":"IPSJ-HPC22183003.pdf"},"date":[{"dateType":"Available","dateValue":"2024-03-10"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-HPC22183003.pdf","filesize":[{"value":"821.0 kB"}],"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":"d8e9bfd2-ad03-4cd9-bd1e-9b681f081c64","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":"藤田, 隼斗"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"野村, 哲弘"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"遠藤, 敏夫"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"関嶋, 政和"}],"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":"タンパク質の立体構造の理解はタンパク質の機能の理解に役立ち，人体の働きを理解する基礎となる．ただし既知のタンパク質のアミノ酸配列に対して，実験的に立体構造が決定されているタンパク質は多いとは言えない．これらのギャップを埋めるために計算機によるアプローチが必要であり，そのアプローチの 1 つとして開発されたのが AlphaFold である．AlphaFold は hhblits 等の既存のツールを使い，BFD などの巨大な遺伝子データベースから MSA を取得する．しかし，そのうちの 1 つの hhblits の実行に多大な時間がかかるという問題がある．この実行時間の一番の要因は I/O にあり，遺伝子データベースをどういったストレージに保管するかによって実行時間が大きく異なってくる．実際に TSUBAME3.0 の高速ストレージ領域にデフォルトの設定で保管した場合と Stripe 設定をして保管した場合とで実行時間が大きく異なっている．そこで，本研究では TSUBAME3.0 上で選択できる遺伝子データベースの保管方法を変更し，実行時間を計測した．さらに，ボトルネックとなっている hhblits に対してプロファイリングを行い，その結果から並列化数のチューニングやデータベースの保存方法の変更，ソートの最適化を行った．また，AlphaFold の MSA 取得ツール間のデータ依存関係からツール実行を非同期に行うようにした．その結果，TSUBAME3.0 上で単一のアミノ酸配列から立体構造を予測する際の実行時間を平均して 2 分の 1 に短縮することに成功した．","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"7","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":"3","bibliographicVolumeNumber":"2022-HPC-183"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":217481,"updated":"2025-01-19T15:28:57.838902+00:00","links":{},"created":"2025-01-19T01:17:58.333186+00:00"}