{"created":"2023-06-20T15:10:44.112588+00:00","id":40309,"links":{},"metadata":{"_buckets":{"deposit":"fe309547-e8ba-4404-adc9-c50b40c28525"},"_deposit":{"created_by":1,"id":"40309","owners":[1],"pid":{"revision_id":0,"type":"depid","value":"40309"},"status":"published"},"_oai":{"id":"oai:jaxa.repo.nii.ac.jp:00040309","sets":["1887:1893","1896:1898:1913:1917"]},"author_link":["447263","447262"],"item_3_alternative_title_1":{"attribute_name":"その他のタイトル","attribute_value_mlt":[{"subitem_alternative_title":"グローバルメモリアクセスにおけるNWT計算機システムの特性パラメータ"}]},"item_3_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicIssueDates":{"bibliographicIssueDate":"1998-10","bibliographicIssueDateType":"Issued"},"bibliographicPageEnd":"21","bibliographicPageStart":"1","bibliographicVolumeNumber":"1362T","bibliographic_titles":[{"bibliographic_title":"航空宇宙技術研究所報告"},{"bibliographic_title":"Technical Report of National Aerospace Laboratory","bibliographic_titleLang":"en"}]}]},"item_3_description_16":{"attribute_name":"抄録","attribute_value_mlt":[{"subitem_description":"1993年2月に運用を開始した航空宇宙技術研究所の数値風洞(NWT)計算機システムは2台の制御計算機、n台の要素計算機(nは運用開始時において140台、現在は166台)並びにクロスバーネットワークから構成されていて分散メモリ型メッセージパッシング方式のMIMD(複数命令列複数データ処理)計算機として動作する。各要素計算機自体はベクトル計算機である。ただし、利用者の並列プログラミングを容易にすることを目的として、プログラミングのための階層メモリ型並列計算機システムを仮定した論理モデルを備えているので、利用者は仮想的グローバルメモリを利用することができる。従って、「グローバル」と宣言された変数は、並列処理のために選択された要素計算機の間において、論理的共有メモリ上に存在するものとしてプログラミングすることができる。本稿ではこのグローバル・メモリ・アクセスの場合における、NWTの運用開始初期の1993年4月から6月の間に測定したNWTの基本的処理性能を表わす2つの特性パラメーターの測定結果について報告する。その1つはクロスバーネットワークを媒体として、グローバルとローカル・メモリ空間の間での情報伝達処理性能を表す特性パラメータ(r(sub ∞)、n(sub 1/2))であり、他はグローバル・メモリ・アクセスにおけるNWTの並列処理性能を表す特性パラメータ(circumflex-r(sub ∞)、circumflex-s(sub 1/2)、f(sub 1/2))である。ここで、r(sub ∞)はグローバルとローカル・メモリ空間の間での情報伝達のバンビ幅を、n(sub 1/2)はr(sub ∞/2)を達成するに必要なメッセージ長を、circumflex-r(sub ∞)はピーク最大処理性能を、circumflex-s(sub 1/2)はr(sub ∞/2)を達成するに必要なグレン・サイズを、f(sub 1/2)はcircumflex-r(sub ∞/2)を達成するに必要なインテンシティ(1度参照されたグローバル変数が何度計算に利用されるかを表わす度合)を表す。さらに、得られた結果の意味について考察するとともに、いくつかのハードウェア・パラメータを推定し、別稿で示したローカル・メモリ・アクセスにおけるNWTの並列処理性能の結果として比較する。なお、本稿で示した値はその後のNWTシステム・ソフトウェアの改善努力によって、大幅に改善されたが、その成果は稿を改めて報告する。","subitem_description_type":"Abstract"}]},"item_3_description_17":{"attribute_name":"抄録（英）","attribute_value_mlt":[{"subitem_description":"The NWT (Numerical Wind Tunnel) computer system available at the NAL since February 1993 comprises two system administrators, n processing elements (where n was 140 at the beginning, and is 166 at present) and a crossbar network, and operates as a distributed-memory message-passing MIMD (Multiple Instruction and Multiple Data stream) computer. Each processing element itself is a vector computer. This paper reports measurements of two pairs of the characteristic parameters, (r(sub infinity), n(sub 1/2)) and (circumflex-r(sub infinity), circumflex-s(sub 1/2), f(sub 1/2)), of the NWT with the communication performance between the global and local memory spaces through the medium of the crossbar network and with MIMD computing in the global memory access, respectively. The significance of the results is interpreted, and several hardware parameters are estimated. The results in this paper apply only to the NWT system software during the period April to June 1993.","subitem_description_type":"Other"}]},"item_3_description_32":{"attribute_name":"資料番号","attribute_value_mlt":[{"subitem_description":"資料番号: AA0001669000","subitem_description_type":"Other"}]},"item_3_description_33":{"attribute_name":"レポート番号","attribute_value_mlt":[{"subitem_description":"レポート番号: NAL TR-1362T","subitem_description_type":"Other"}]},"item_3_publisher_8":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"航空宇宙技術研究所"}]},"item_3_publisher_9":{"attribute_name":"出版者（英）","attribute_value_mlt":[{"subitem_publisher":"National Aerospace Laboratory (NAL)"}]},"item_3_source_id_21":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"0389-4010","subitem_source_identifier_type":"ISSN"}]},"item_3_text_6":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"航空宇宙技術研究所 数理解析部"}]},"item_3_text_7":{"attribute_name":"著者所属（英）","attribute_value_mlt":[{"subitem_text_language":"en","subitem_text_value":"National Aerospace Laboratory Computational Sciences Division"}]},"item_creator":{"attribute_name":"著者","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"畑山, 茂樹"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Hatayama, Shigeki","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_files":{"attribute_name":"ファイル情報","attribute_type":"file","attribute_value_mlt":[{"accessrole":"open_date","date":[{"dateType":"Available","dateValue":"2020-01-27"}],"displaytype":"detail","filename":"naltr0001362t.pdf","filesize":[{"value":"453.8 kB"}],"format":"application/pdf","licensetype":"license_note","mimetype":"application/pdf","url":{"label":"naltr0001362t.pdf","url":"https://jaxa.repo.nii.ac.jp/record/40309/files/naltr0001362t.pdf"},"version_id":"7b4ae043-cfe5-43f6-a763-64bb40e850f4"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"ローカルメモリアクセス","subitem_subject_scheme":"Other"},{"subitem_subject":"数値風洞シミュレーション","subitem_subject_scheme":"Other"},{"subitem_subject":"マルチコンピュータシステム","subitem_subject_scheme":"Other"},{"subitem_subject":"並列計算機","subitem_subject_scheme":"Other"},{"subitem_subject":"計算機処理性能","subitem_subject_scheme":"Other"},{"subitem_subject":"ベンチマーク試験","subitem_subject_scheme":"Other"},{"subitem_subject":"並列処理","subitem_subject_scheme":"Other"},{"subitem_subject":"データ転送","subitem_subject_scheme":"Other"},{"subitem_subject":"通信オーバヘッド","subitem_subject_scheme":"Other"},{"subitem_subject":"MIMD計算","subitem_subject_scheme":"Other"},{"subitem_subject":"複数命令列複数データ処理計算","subitem_subject_scheme":"Other"},{"subitem_subject":"local memory access","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"numerical wind tunnel simulation","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"multicomputer system","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"parallel computer","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"computer processing performance","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"benchmark test","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"parallel processing","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"data transfer","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"communication overhead","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"MIMD computation","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"multiple instruction and multiple data stream computation","subitem_subject_language":"en","subitem_subject_scheme":"Other"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"eng"}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourcetype":"technical report","resourceuri":"http://purl.org/coar/resource_type/c_18gh"}]},"item_title":"Characteristic parameters of the NWT computer system in global memory access","item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"Characteristic parameters of the NWT computer system in global memory access","subitem_title_language":"en"}]},"item_type_id":"3","owner":"1","path":["1893","1917"],"pubdate":{"attribute_name":"公開日","attribute_value":"2015-03-26"},"publish_date":"2015-03-26","publish_status":"0","recid":"40309","relation_version_is_last":true,"title":["Characteristic parameters of the NWT computer system in global memory access"],"weko_creator_id":"1","weko_shared_id":1},"updated":"2023-06-20T19:15:46.357192+00:00"}