{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,18]],"date-time":"2026-03-18T13:55:04Z","timestamp":1773842104096,"version":"3.50.1"},"reference-count":12,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"5","license":[{"start":{"date-parts":[[2021,9,1]],"date-time":"2021-09-01T00:00:00Z","timestamp":1630454400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2021,9,1]],"date-time":"2021-09-01T00:00:00Z","timestamp":1630454400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2021,9,1]],"date-time":"2021-09-01T00:00:00Z","timestamp":1630454400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Micro"],"published-print":{"date-parts":[[2021,9,1]]},"DOI":"10.1109\/mm.2021.3097287","type":"journal-article","created":{"date-parts":[[2021,8,24]],"date-time":"2021-08-24T19:17:34Z","timestamp":1629832654000},"page":"101-112","source":"Crossref","is-referenced-by-count":12,"title":["Datacenter-Scale Analysis and Optimization of GPU Machine Learning Workloads"],"prefix":"10.1109","volume":"41","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-8377-9691","authenticated-orcid":false,"given":"Lukasz","family":"Wesolowski","sequence":"first","affiliation":[{"name":"Facebook, Inc., Menlo Park, CA, USA"}]},{"given":"Bilge","family":"Acun","sequence":"additional","affiliation":[{"name":"Facebook, Inc., Menlo Park, CA, USA"}]},{"given":"Valentin","family":"Andrei","sequence":"additional","affiliation":[{"name":"Facebook, Inc., Menlo Park, CA, USA"}]},{"given":"Adnan","family":"Aziz","sequence":"additional","affiliation":[{"name":"Facebook, Inc., Menlo Park, CA, USA"}]},{"given":"Gisle","family":"Dankel","sequence":"additional","affiliation":[{"name":"Facebook, Inc., Menlo Park, CA, USA"}]},{"given":"Christopher","family":"Gregg","sequence":"additional","affiliation":[{"name":"Stanford University, Stanford, CA, USA"}]},{"given":"Xiaoqiao","family":"Meng","sequence":"additional","affiliation":[{"name":"Facebook, Inc., Menlo Park, CA, USA"}]},{"given":"Cyril","family":"Meurillon","sequence":"additional","affiliation":[{"name":"Facebook, Inc., Menlo Park, CA, USA"}]},{"given":"Denis","family":"Sheahan","sequence":"additional","affiliation":[{"name":"Facebook, Inc., Menlo Park, CA, USA"}]},{"given":"Lei","family":"Tian","sequence":"additional","affiliation":[{"name":"Facebook, Inc., Menlo Park, CA, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4237-7493","authenticated-orcid":false,"given":"Janet","family":"Yang","sequence":"additional","affiliation":[{"name":"Carnegie Mellon University, Pittsburgh, PA, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7001-6647","authenticated-orcid":false,"given":"Peifeng","family":"Yu","sequence":"additional","affiliation":[{"name":"University of Michigan, Ann Arbor, MI, USA"}]},{"given":"Kim","family":"Hazelwood","sequence":"additional","affiliation":[{"name":"Facebook, Inc., Menlo Park, CA, USA"}]}],"member":"263","reference":[{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1145\/3079856.3080246"},{"key":"ref3","article-title":"Using cloud TPU tools","author":"google","year":"2020"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1109\/MM.2017.37"},{"key":"ref6","article-title":"Monitoring GPU utilization with amazon cloudwatch","author":"xu","year":"2017"},{"key":"ref11","first-page":"8024?8035","article-title":"Pytorch: An imperative style, high-performance deep learning library","author":"paszke","year":"2019"},{"key":"ref5","article-title":"GPU monitor","author":"salvaris","year":"2018"},{"key":"ref12","article-title":"Kineto profiling library","author":"facebook","year":"2020"},{"key":"ref8","article-title":"Cuda profiling tools interface (CUPTI)","year":"2020"},{"key":"ref7","article-title":"Monitor GPUs with cloudwatch","author":"amazon","year":"2020"},{"key":"ref2","first-page":"947?960","article-title":"Analysis of large-scale multi-tenant GPU clusters for DNN training workloads","author":"jeon","year":"2019","journal-title":"Proc USENIX Conf USENIX Annu Tech Conf"},{"key":"ref9","article-title":"Introducing big basin: Our next-generation AI hardware","author":"lee","year":"2017"},{"key":"ref1","article-title":"Workload analysis of blue waters","author":"jones","year":"2017"}],"container-title":["IEEE Micro"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/40\/9536946\/09484727.pdf?arnumber=9484727","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,5,10]],"date-time":"2022-05-10T14:50:47Z","timestamp":1652194247000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/9484727\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,9,1]]},"references-count":12,"journal-issue":{"issue":"5"},"URL":"https:\/\/doi.org\/10.1109\/mm.2021.3097287","relation":{},"ISSN":["0272-1732","1937-4143"],"issn-type":[{"value":"0272-1732","type":"print"},{"value":"1937-4143","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,9,1]]}}}