{"links":{},"id":23481,"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00023481","sets":["1164:1579:1616:1618"]},"path":["1618"],"owner":"1","recid":"23481","title":["値予測における値履歴テーブルエントリ数の削減"],"pubdate":{"attribute_name":"公開日","attribute_value":"2002-08-22"},"_buckets":{"deposit":"a64f9bcf-5742-4c65-a2e0-388c842d2ebf"},"_deposit":{"id":"23481","pid":{"type":"depid","value":"23481","revision_id":0},"owners":[1],"status":"published","created_by":1},"item_title":"値予測における値履歴テーブルエントリ数の削減","author_link":["0","0"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"値予測における値履歴テーブルエントリ数の削減"},{"subitem_title":"Decreasing Value History Table Entry on Value Prediction","subitem_title_language":"en"}]},"item_type_id":"4","publish_date":"2002-08-22","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"東京大学大学院工学系研究科"},{"subitem_text_value":"東京大学大学院情報理工学系研究科"},{"subitem_text_value":"東京大学大学院情報理工学系研究科"},{"subitem_text_value":"東京大学大学院情報理工学系研究科"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Graduate School of Engineering, The University of Tokyo","subitem_text_language":"en"},{"subitem_text_value":"Graduate School of Information Science and Technology, The University of Tokyo","subitem_text_language":"en"},{"subitem_text_value":"Graduate School of Information Science and Technology, The University of Tokyo","subitem_text_language":"en"},{"subitem_text_value":"Graduate School of Information Science and Technology, The University of Tokyo","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/23481/files/IPSJ-ARC02149011.pdf"},"date":[{"dateType":"Available","dateValue":"2004-08-22"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-ARC02149011.pdf","filesize":[{"value":"223.5 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":"3903464a-6a3a-4e99-bcf6-057ed9edaa3e","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":"飯塚, 大介"},{"creatorName":"角田, 忠信"},{"creatorName":"坂井, 修一"},{"creatorName":"田中, 英彦"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Daisuke, Iizuka","creatorNameLang":"en"},{"creatorName":"Tadanobu, Tsunoda","creatorNameLang":"en"},{"creatorName":"Shuichi, Sakai","creatorNameLang":"en"},{"creatorName":"Hidehiko, Tanaka","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":"値予測は真の依存関係を解消することで並列度を抽出する手法である。性能向上率を向上させるために予測精度を高めた種々の予測方式が提案されているが、予測に必要なハードウェア量が膨大なため実装するのが困難となっている。特に予測履歴テーブルは大きな資源を必要とするため、タグや予測値のビット数、エントリ数を制限してハードウェア量を削減する手法が近年いくつか提案されている。本稿では静的に命令を分類し、予測候補から除外する命令を選択することですることで総予測命令数を削減し、値予測履歴テーブルのエントリ数を削減した場合の性能低下を抑制する方式を提案する。その結果、同じエントリ数の場合に性能向上率が若干向上することを確認した。","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"Value prediction is a technique to resolve true dependency and exploit more parallelism. Various value prediction mechanism have been proposed to achieve high prediction accuracy. However, those sophisticated mechanisms often require a large amount of hardware, especially for the Value History Table(VHT). Recently, several  researches have been done on the techniques to reduce the VHT size by cutting the length of tag or prediction values, or by limiting the number of table entries. This paper proposes a technique to reduce the VHT entries without loosing performance based on static  analysis. We propose to use VHT only for instructions  that have been selected by the compiler.  Evaluation results show that, using  our technique, a higher performance can be achieved with the VHT of  the same number of entries.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"66","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":"61","bibliographicIssueDates":{"bibliographicIssueDate":"2002-08-22","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"81(2002-ARC-149)","bibliographicVolumeNumber":"2002"}]},"relation_version_is_last":true,"weko_creator_id":"1"},"created":"2025-01-18T22:54:52.614444+00:00","updated":"2025-01-22T20:21:58.867659+00:00"}