{"updated":"2025-01-19T16:32:36.062957+00:00","links":{},"created":"2025-01-19T01:15:26.019124+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00214626","sets":["6504:10735:10736"]},"path":["10736"],"owner":"44499","recid":"214626","title":["畳み込みニューラルネットワークの行列計算におけるTensorコアを用いた並列処理"],"pubdate":{"attribute_name":"公開日","attribute_value":"2021-03-04"},"_buckets":{"deposit":"6dbeee52-61c2-446f-97f5-30e7ba294839"},"_deposit":{"id":"214626","pid":{"type":"depid","value":"214626","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"畳み込みニューラルネットワークの行列計算におけるTensorコアを用いた並列処理","author_link":["551738","551739"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"畳み込みニューラルネットワークの行列計算におけるTensorコアを用いた並列処理"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"コンピュータシステム","subitem_subject_scheme":"Other"}]},"item_type_id":"22","publish_date":"2021-03-04","item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"item_22_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"明大"},{"subitem_text_value":"明大"}]},"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/214626/files/IPSJ-Z83-1H-04.pdf","label":"IPSJ-Z83-1H-04.pdf"},"date":[{"dateType":"Available","dateValue":"2021-12-28"}],"format":"application/pdf","filename":"IPSJ-Z83-1H-04.pdf","filesize":[{"value":"153.6 kB"}],"mimetype":"application/pdf","accessrole":"open_date","version_id":"1c05d677-e642-4565-adca-83ffcb6a5301","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2021 by the Information Processing Society of Japan"}]},"item_22_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"大内, 佑一朗"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"吉田, 明正"}],"nameIdentifiers":[{}]}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_5794","resourcetype":"conference paper"}]},"item_22_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN00349328","subitem_source_identifier_type":"NCID"}]},"item_22_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"物体検出のディープラーニングに用いられる畳み込みニューラルネットワーク（CNN）では多くの行列計算が行われており，学習時間を短縮するためには行列計算の高速化が不可欠である．行列計算の高速化には，GPUの活用が広く行われているが，さらなる速度向上を実現するためには，単精度浮動小数点数FP32の代わりに，半精度浮動小数点数FP16を利用して，GPUの演算性能を高めることが挙げられる．加えて，Tensorコアを活用することにより，行列計算の高速化が可能になる．そこで本手法では，FP16およびTensorコアを利用するCUDAプログラムを開発しており，CNNの行列計算部分に適用した． NVIDIA Quadro RTX 6000上で行った性能評価の結果，提案手法の有効性が確認された．","subitem_description_type":"Other"}]},"item_22_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"26","bibliographic_titles":[{"bibliographic_title":"第83回全国大会講演論文集"}],"bibliographicPageStart":"25","bibliographicIssueDates":{"bibliographicIssueDate":"2021-03-04","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"1","bibliographicVolumeNumber":"2021"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":214626}