{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,2,21]],"date-time":"2025-02-21T01:42:17Z","timestamp":1740102137810,"version":"3.37.3"},"reference-count":36,"publisher":"IEEE","license":[{"start":{"date-parts":[[2023,5,17]],"date-time":"2023-05-17T00:00:00Z","timestamp":1684281600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2023,5,17]],"date-time":"2023-05-17T00:00:00Z","timestamp":1684281600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100015749","name":"Communication University of China","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100015749","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2023,5,17]]},"DOI":"10.1109\/infocom53939.2023.10228922","type":"proceedings-article","created":{"date-parts":[[2023,8,29]],"date-time":"2023-08-29T17:40:43Z","timestamp":1693330843000},"page":"1-10","source":"Crossref","is-referenced-by-count":1,"title":["Libra: Contention-Aware GPU Thread Allocation for Data Parallel Training in High Speed Networks"],"prefix":"10.1109","author":[{"given":"Yunzhuo","family":"Liu","sequence":"first","affiliation":[{"name":"Shanghai Jiao Tong University"}]},{"given":"Bo","family":"Jiang","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University"}]},{"given":"Shizhen","family":"Zhao","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University"}]},{"given":"Tao","family":"Lin","sequence":"additional","affiliation":[{"name":"Communication University of China"}]},{"given":"Xinbing","family":"Wang","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University"}]},{"given":"Chenghu","family":"Zhou","sequence":"additional","affiliation":[{"name":"Chinese Academy of Sciences"}]}],"member":"263","reference":[{"doi-asserted-by":"publisher","key":"ref13","DOI":"10.1109\/INFOCOM41043.2020.9155446"},{"doi-asserted-by":"publisher","key":"ref35","DOI":"10.1109\/ICCV.2015.11"},{"doi-asserted-by":"publisher","key":"ref12","DOI":"10.1109\/INFOCOM41043.2020.9155269"},{"doi-asserted-by":"publisher","key":"ref34","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"ref15","first-page":"132","article-title":"Priority-based parameter propagation for distributed dnn training","author":"jayarajan","year":"2019","journal-title":"MLSys"},{"doi-asserted-by":"publisher","key":"ref14","DOI":"10.1145\/3341301.3359642"},{"key":"ref36","first-page":"1027","article-title":"Accelerating collective communication in data parallel training across deep learning frameworks","author":"romero","year":"2022","journal-title":"NSDI"},{"key":"ref31","article-title":"BERT: Pre-training of deep bidirectional transformers for language understanding","author":"devlin","year":"2018","journal-title":"NAACL-HLT"},{"key":"ref30","article-title":"Very deep convolutional networks for large-scale image recognition","author":"simonyan","year":"2015","journal-title":"ICLRE"},{"doi-asserted-by":"publisher","key":"ref11","DOI":"10.1109\/INFOCOM42981.2021.9488803"},{"year":"0","article-title":"NVIDIA TESLA V100","key":"ref33"},{"doi-asserted-by":"publisher","key":"ref10","DOI":"10.1109\/TPDS.2021.3052862"},{"year":"0","article-title":"GEFORCE GTX 1080 Ti","key":"ref32"},{"doi-asserted-by":"publisher","key":"ref2","DOI":"10.1177\/1094342005051521"},{"year":"2017","article-title":"baidu-allreduce","key":"ref1"},{"doi-asserted-by":"publisher","key":"ref17","DOI":"10.1109\/INFOCOM48880.2022.9796820"},{"key":"ref16","first-page":"418","article-title":"Tictac: Accelerating distributed deep learning with communication scheduling","author":"hashemi","year":"2019","journal-title":"MLSys"},{"key":"ref19","article-title":"Gradient sparsification for communication-efficient distributed optimization","volume":"31","author":"wangni","year":"2018","journal-title":"NIPS"},{"key":"ref18","article-title":"Double quantization for communication-efficient distributed optimization","volume":"32","author":"yu","year":"2019","journal-title":"NIPS"},{"year":"0","article-title":"oneapi collective communications library (oneccl)","key":"ref24"},{"year":"0","article-title":"Nvidia collective communications library (NCCL)","key":"ref23"},{"doi-asserted-by":"publisher","key":"ref26","DOI":"10.1145\/3453953.3453972"},{"doi-asserted-by":"publisher","key":"ref25","DOI":"10.1109\/ISCA52012.2021.00049"},{"key":"ref20","article-title":"Powersgd: Practical low-rank gradient compression for distributed optimization","volume":"32","author":"vogels","year":"2019","journal-title":"NIPS"},{"doi-asserted-by":"publisher","key":"ref22","DOI":"10.1145\/3126908.3126912"},{"year":"2018","author":"sergeev","article-title":"Horovod: fast and easy distributed deep learning in Tensorflow","key":"ref21"},{"doi-asserted-by":"publisher","key":"ref28","DOI":"10.1109\/CVPR.2017.195"},{"doi-asserted-by":"publisher","key":"ref27","DOI":"10.1109\/RTSS.2017.00017"},{"doi-asserted-by":"publisher","key":"ref29","DOI":"10.1109\/CVPR.2016.90"},{"doi-asserted-by":"publisher","key":"ref8","DOI":"10.14778\/3415478.3415530"},{"doi-asserted-by":"publisher","key":"ref7","DOI":"10.1145\/3302424.3303957"},{"key":"ref9","first-page":"181","article-title":"Poseidon: An efficient communication architecture for distributed deep learning on GPU clusters","author":"zhang","year":"2017","journal-title":"ATC"},{"doi-asserted-by":"publisher","key":"ref4","DOI":"10.1109\/MNET.011.2000530"},{"doi-asserted-by":"publisher","key":"ref3","DOI":"10.1016\/j.jpdc.2008.09.002"},{"doi-asserted-by":"publisher","key":"ref6","DOI":"10.1145\/3267809.3267840"},{"doi-asserted-by":"publisher","key":"ref5","DOI":"10.1145\/2901318.2901323"}],"event":{"name":"IEEE INFOCOM 2023 - IEEE Conference on Computer Communications","start":{"date-parts":[[2023,5,17]]},"location":"New York City, NY, USA","end":{"date-parts":[[2023,5,20]]}},"container-title":["IEEE INFOCOM 2023 - IEEE Conference on Computer Communications"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/10228851\/10228852\/10228922.pdf?arnumber=10228922","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,9,25]],"date-time":"2023-09-25T17:53:30Z","timestamp":1695664410000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10228922\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,5,17]]},"references-count":36,"URL":"https:\/\/doi.org\/10.1109\/infocom53939.2023.10228922","relation":{},"subject":[],"published":{"date-parts":[[2023,5,17]]}}}