{"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00204962","sets":["6504","6504:10247","6504:10247:10257"]},"path":["6504","10247","10257"],"owner":"6748","recid":"204962","title":["組込み向けGPUを用いた畳み込み演算の高速化に関する検討"],"pubdate":{"attribute_name":"公開日","attribute_value":"2020-02-20"},"_buckets":{"deposit":"1f530376-8e34-4d6e-a9ba-7e19b3b86a40"},"_deposit":{"id":"204962","pid":{"type":"depid","value":"204962","revision_id":0},"owners":[6748],"status":"published","created_by":6748},"item_title":"組込み向けGPUを用いた畳み込み演算の高速化に関する検討","author_link":["508370","508373","508371","508372"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"組込み向けGPUを用いた畳み込み演算の高速化に関する検討"}]},"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":"2020-02-20","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":"三菱"},{"subitem_text_value":"三菱"},{"subitem_text_value":"三菱"},{"subitem_text_value":"三菱"}]},"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/204962/files/IPSJ-Z82-2A-02.pdf","label":"IPSJ-Z82-2A-02.pdf"},"date":[{"dateType":"Available","dateValue":"2020-06-19"}],"format":"application/pdf","filename":"IPSJ-Z82-2A-02.pdf","filesize":[{"value":"659.5 kB"}],"mimetype":"application/pdf","accessrole":"open_date","version_id":"1709b4bd-b323-4fd2-abc2-3e6c3e295732","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2020 by the Information Processing Society of Japan"}]},"item_22_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"立見, 駿介"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"山本, 亮"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"岡田, 尚也"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"小川, 吉大"}],"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":"深層学習のモデルの1つである畳み込みニューラルネットワーク（CNN）は、物体検出、物体認識、音声認識など多様な応用において有望な手法であり、組込みにおいても需要がある。CNNは畳み込み演算のために計算量が非常に大きいため、リアルタイム処理を実現するためにGPU活用が検討されており、様々な深層学習向けOSSを用いることでGPU利用による高速化が可能である。しかしながら、OSSを用いる場合、実装先GPUに制限があることや、実装コードの品質保証が難しいことが問題となる。そこで、本発表では、組込み向けGPUに対して畳み込み演算のフルスクラッチ実装を行い、高速化を検討する。","subitem_description_type":"Other"}]},"item_22_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"4","bibliographic_titles":[{"bibliographic_title":"第82回全国大会講演論文集"}],"bibliographicPageStart":"3","bibliographicIssueDates":{"bibliographicIssueDate":"2020-02-20","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"1","bibliographicVolumeNumber":"2020"}]},"relation_version_is_last":true,"weko_creator_id":"6748"},"id":204962,"updated":"2025-06-05T06:25:15.158726+00:00","links":{},"created":"2025-01-19T01:07:08.849389+00:00"}