{"created":"2023-09-20T08:10:58.907251+00:00","id":6410,"links":{},"metadata":{"_buckets":{"deposit":"da6394b4-b62b-4fc2-b918-c1fa5a5d51e0"},"_deposit":{"created_by":7,"id":"6410","owners":[7],"pid":{"revision_id":0,"type":"depid","value":"6410"},"status":"published"},"_oai":{"id":"oai:nied-repo.bosai.go.jp:00006410","sets":[]},"author_link":[],"item_10001_biblio_info_7":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicIssueDates":{"bibliographicIssueDate":"2021-02-17","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"1","bibliographicPageEnd":"294","bibliographicPageStart":"283","bibliographicVolumeNumber":"14","bibliographic_titles":[{"bibliographic_title":"Journal of Information Processing","bibliographic_titleLang":"en"}]}]},"item_10001_description_5":{"attribute_name":"抄録","attribute_value_mlt":[{"subitem_description":"Exascale computers consume huge amounts of power and their variation over time makes system energy management important. Because of time lag in cooling-units operation, predictive control is desirable for effective power control. In this work, we report a state-of-the-art power prediction model. Conventional methods with topic model use the power of past job as a prediction based on the similarity of job information. The prediction, however, fails, if there is no correct data before. To resolve this, we developed a recurrent neural network model with variable network size, which detects features of power shape from its power history and enables precise prediction during job execution. By integrating these models into a single algorithm, the optimal model is automatically adopted for prediction according to the job status. We demonstrated high-precision prediction with an average relative error of 5.7% in K computer as compared to that of 20.1% by the conventional method.","subitem_description_language":"en","subitem_description_type":"Other"}]},"item_10001_publisher_8":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"Information Processing Society of Japan","subitem_publisher_language":"en"}]},"item_10001_relation_14":{"attribute_name":"DOI","attribute_value_mlt":[{"subitem_relation_type_id":{"subitem_relation_type_id_text":"10.2197/ipsjjip.29.283"}}]},"item_10001_source_id_9":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"1882-6652","subitem_source_identifier_type":"EISSN"}]},"item_creator":{"attribute_name":"著者","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Suzuki Shigeto","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"Hiraoka Michiko","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"Shiraishi Takashi","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"Kreshpa Enxhi","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"Yamamoto Takuji","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"Fukuda Hiroyuki","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"Matsui Shuji","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"Fujisaki Masahide","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"Uno Atsuya","creatorNameLang":"en"}]}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"eng"}]},"item_title":"Power Prediction for Sustainable HPC","item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"Power Prediction for Sustainable HPC","subitem_title_language":"en"}]},"item_type_id":"40001","owner":"7","path":["1670839190650"],"pubdate":{"attribute_name":"PubDate","attribute_value":"2023-09-20"},"publish_date":"2023-09-20","publish_status":"0","recid":"6410","relation_version_is_last":true,"title":["Power Prediction for Sustainable HPC"],"weko_creator_id":"7","weko_shared_id":-1},"updated":"2023-09-20T08:11:00.245856+00:00"}