{"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00023509","sets":["1164:1579:1616:1620"]},"path":["1620"],"owner":"1","recid":"23509","title":["WindowsMPIによる非均－PCクラスタ並列分子動力学法"],"pubdate":{"attribute_name":"公開日","attribute_value":"2002-03-07"},"_buckets":{"deposit":"18b704be-f08f-4aab-b307-ad3a83cd23f5"},"_deposit":{"id":"23509","pid":{"type":"depid","value":"23509","revision_id":0},"owners":[1],"status":"published","created_by":1},"item_title":"WindowsMPIによる非均－PCクラスタ並列分子動力学法","author_link":["0","0"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"WindowsMPIによる非均－PCクラスタ並列分子動力学法"},{"subitem_title":"Parallel Molecular Dynamics Method on Heterogeneous PC Cluster with WindowsMPI","subitem_title_language":"en"}]},"item_type_id":"4","publish_date":"2002-03-07","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"茨城大学大学院理工学研究科"},{"subitem_text_value":"茨城大学工学部"},{"subitem_text_value":"茨城大学大学院理工学研究科"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Graduate School of Science and Engineering, Ibaraki University","subitem_text_language":"en"},{"subitem_text_value":"Faculty of Engineering, Ibaraki University","subitem_text_language":"en"},{"subitem_text_value":"Graduate School of Science and Engineering, Ibaraki University","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/23509/files/IPSJ-ARC01147003.pdf"},"date":[{"dateType":"Available","dateValue":"2004-03-07"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-ARC01147003.pdf","filesize":[{"value":"956.8 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":"16"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"0041788d-91e7-4937-bda4-6580026e346f","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2002 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"橋本, 孝一"},{"creatorName":"石黒, 美佐子"},{"creatorName":"絵幡, 大輔"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Koichi, Hashimoto","creatorNameLang":"en"},{"creatorName":"Misako, Ishiguro","creatorNameLang":"en"},{"creatorName":"Daisuke, Ebata","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10096105","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_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"分子動力学法は、本来、多量の粒子のタイムステップからなる膨大な計算で構成される。本研究では、MD法プログラムを非均一な速度のPCから成るクラスタを用いて並列化する。さらに、2プロセッサ搭載のPCではマルチスレッドを用いて並列化する。OSは，PC上の並列処理でよく使われているLinuxではなく，通常のWindows NT/2000とし，通信は WMPI通信ライブラリ（Windows Message Passing Interface)を使用する。並列の方法には，空間分割法を用い，自動不可分散を考慮する。空間分割法では相互作用の非常に小さい遠距離粒子をカットすれば，隣接PCとの通信のみで処理できる。その上、PC台数が多いほどカットオフの判断に要する計算が減少し、並列化効果が台数効果を上回るという隠れた利点があることがわかった。","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"Molecular dynamics (MD) method consists of the calculation of enormous particles and enormous time steps. In this study, we try to parallelize the MD program using a PC cluster composed of heterogeneous PCs. The program is also executed in parallel on dual processor PC by multi-threads of Visual C++. Here, usual Windows NT/2000 operating system is used rather than Linux, and accordingly, Windows MPI(WMPI) communication library is applied as the message passing interface. Cut-off method is introduced to calculate particle-particle force. Domain decomposition method is applied on the basis of an automatic load balancing policy. As the result, it is found that the cut-off brings into an additional parallelization effect under the domain decomposition, where the speed-up ratio exceeds the CPUs number effect because of saving of computing time required for the cut-off judgment.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"18","bibliographic_titles":[{"bibliographic_title":"情報処理学会研究報告計算機アーキテクチャ（ARC）"}],"bibliographicPageStart":"13","bibliographicIssueDates":{"bibliographicIssueDate":"2002-03-07","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"22(2001-ARC-147)","bibliographicVolumeNumber":"2002"}]},"relation_version_is_last":true,"weko_creator_id":"1"},"id":23509,"updated":"2025-01-22T20:20:21.603933+00:00","links":{},"created":"2025-01-18T22:54:53.855076+00:00"}