{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,24]],"date-time":"2026-04-24T01:20:56Z","timestamp":1776993656778,"version":"3.51.4"},"reference-count":37,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"11","license":[{"start":{"date-parts":[[2022,11,1]],"date-time":"2022-11-01T00:00:00Z","timestamp":1667260800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2022,11,1]],"date-time":"2022-11-01T00:00:00Z","timestamp":1667260800000},"content-version":"am","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2022,11,1]],"date-time":"2022-11-01T00:00:00Z","timestamp":1667260800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2022,11,1]],"date-time":"2022-11-01T00:00:00Z","timestamp":1667260800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","award":["CCF-1763747"],"award-info":[{"award-number":["CCF-1763747"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","award":["CNS-1816887"],"award-info":[{"award-number":["CNS-1816887"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Parallel Distrib. Syst."],"published-print":{"date-parts":[[2022,11,1]]},"DOI":"10.1109\/tpds.2022.3151739","type":"journal-article","created":{"date-parts":[[2022,2,16]],"date-time":"2022-02-16T20:27:21Z","timestamp":1645043241000},"page":"2900-2912","source":"Crossref","is-referenced-by-count":7,"title":["Predicting Throughput of Distributed Stochastic Gradient Descent"],"prefix":"10.1109","volume":"33","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-8308-0231","authenticated-orcid":false,"given":"Zhuojin","family":"Li","sequence":"first","affiliation":[{"name":"Department of Computer Science, University of Southern California, Los Angeles, CA, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5110-203X","authenticated-orcid":false,"given":"Marco","family":"Paolieri","sequence":"additional","affiliation":[{"name":"Department of Computer Science, University of Southern California, Los Angeles, CA, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8353-5040","authenticated-orcid":false,"given":"Leana","family":"Golubchik","sequence":"additional","affiliation":[{"name":"Department of Computer Science, University of Southern California, Los Angeles, CA, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Sung-Han","family":"Lin","sequence":"additional","affiliation":[{"name":"Meta, Menlo Park, CA, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4506-795X","authenticated-orcid":false,"given":"Wumo","family":"Yan","sequence":"additional","affiliation":[{"name":"Department of Computer Science, University of Southern California, Los Angeles, CA, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.308"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2015.7298594"},{"key":"ref31","author":"snir","year":"1998","journal-title":"MPI—The Complete Reference"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/DASC\/PiCom\/DataCom\/CyberSciTec.2018.000-4"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1145\/3337821.3337873"},{"key":"ref36","first-page":"363","article-title":"Ernest: Efficient performance prediction for large-scale advanced analytics","author":"venkataraman","year":"2016","journal-title":"Proc 13th Usenix Conf Netw Syst Des Implementation"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1109\/ICDE.2017.160"},{"key":"ref34","first-page":"6105","article-title":"Efficientnet: Rethinking model scaling for convolutional neural networks","author":"tan","year":"2019","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref11","first-page":"1223","article-title":"More effective distributed ML via a stale synchronous parallel parameter server","author":"ho","year":"2013","journal-title":"Proc 26th Int Conf Adv Neural Inf Process Syst"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.243"},{"key":"ref13","article-title":"Highly scalable deep learning training system with mixed-precision: Training imagenet in four minutes","author":"jia","year":"2018"},{"key":"ref14","article-title":"Dissecting the NVidia turing T4 GPU via microbenchmarking","author":"jia","year":"2019"},{"key":"ref15","article-title":"Dissecting the NVIDIA volta GPU architecture via microbenchmarking","author":"jia","year":"2018"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1145\/3079856.3080246"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1038\/nature14539"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.14778\/3415478.3415530"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1145\/3358960.3379141"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/TCOM.1979.1094531"},{"key":"ref4","article-title":"PowerAI DDL","author":"cho","year":"2017"},{"key":"ref27","article-title":"Paleo: A performance model for deep neural networks","author":"qi","year":"2017","journal-title":"Proc 5th Int Conf Learn Representations"},{"key":"ref3","article-title":"Revisiting distributed synchronous SGD","author":"chen","year":"2016","journal-title":"CoRR"},{"key":"ref6","first-page":"1","article-title":"The design and operation of CloudLab","author":"duplyakin","year":"2019","journal-title":"Proc Conf USENIX Annu Tech Conf"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1145\/322186.322195"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1145\/2408776.2408794"},{"key":"ref8","article-title":"Accurate, large minibatch SGD: Training imagenet in 1 hour","author":"goyal","year":"2017"},{"key":"ref7","article-title":"Collective communications library","year":"2017"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1145\/3320060"},{"key":"ref9","article-title":"Omnivore: An optimizer for multi-device deep learning on CPUs and GPUs","author":"hadjis","year":"2016","journal-title":"CoRR"},{"key":"ref1","first-page":"265","article-title":"TensorFlow: A system for large-scale machine learning","author":"abadi","year":"2016","journal-title":"Proc 12th USENIX Symp Oper Syst Des Implementation"},{"key":"ref20","first-page":"3043","article-title":"Asynchronous decentralized parallel stochastic gradient descent","author":"lian","year":"2018","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref22","article-title":"Nvidia collective communication library","year":"2016","journal-title":"NCCL"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1109\/MASCOTS.2018.00037"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1016\/j.jpdc.2008.09.002"},{"key":"ref23","article-title":"PyTorch: An imperative style, high-performance deep learning library","author":"paszke","year":"2019"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1145\/3190508.3190517"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2019.2916550"}],"container-title":["IEEE Transactions on Parallel and Distributed Systems"],"original-title":[],"link":[{"URL":"https:\/\/ieeexplore.ieee.org\/ielam\/71\/9782123\/9714870-aam.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/71\/9782123\/09714870.pdf?arnumber=9714870","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,5,27]],"date-time":"2022-05-27T20:56:34Z","timestamp":1653684994000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/9714870\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,11,1]]},"references-count":37,"journal-issue":{"issue":"11"},"URL":"https:\/\/doi.org\/10.1109\/tpds.2022.3151739","relation":{},"ISSN":["1045-9219","1558-2183","2161-9883"],"issn-type":[{"value":"1045-9219","type":"print"},{"value":"1558-2183","type":"electronic"},{"value":"2161-9883","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,11,1]]}}}