{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,15]],"date-time":"2026-07-15T14:41:42Z","timestamp":1784126502794,"version":"3.55.0"},"reference-count":41,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"4","license":[{"start":{"date-parts":[[2025,4,1]],"date-time":"2025-04-01T00:00:00Z","timestamp":1743465600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2025,4,1]],"date-time":"2025-04-01T00:00:00Z","timestamp":1743465600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,4,1]],"date-time":"2025-04-01T00:00:00Z","timestamp":1743465600000},"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","award":["62025208"],"award-info":[{"award-number":["62025208"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62421002"],"award-info":[{"award-number":["62421002"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62302512"],"award-info":[{"award-number":["62302512"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Comput."],"published-print":{"date-parts":[[2025,4]]},"DOI":"10.1109\/tc.2024.3517748","type":"journal-article","created":{"date-parts":[[2024,12,16]],"date-time":"2024-12-16T19:16:19Z","timestamp":1734376579000},"page":"1196-1209","source":"Crossref","is-referenced-by-count":6,"title":["AutoPipe-H: A Heterogeneity-Aware Data-Paralleled Pipeline Approach on Commodity GPU Servers"],"prefix":"10.1109","volume":"74","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-0354-9467","authenticated-orcid":false,"given":"Weijie","family":"Liu","sequence":"first","affiliation":[{"name":"College of Computer, National University of Defence Technology, Changsha, Hunan, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6378-7002","authenticated-orcid":false,"given":"Kai","family":"Lu","sequence":"additional","affiliation":[{"name":"College of Computer, National University of Defence Technology, Changsha, Hunan, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3458-4732","authenticated-orcid":false,"given":"Zhiquan","family":"Lai","sequence":"additional","affiliation":[{"name":"College of Computer, National University of Defence Technology, Changsha, Hunan, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7419-1511","authenticated-orcid":false,"given":"Shengwei","family":"Li","sequence":"additional","affiliation":[{"name":"College of Computer, National University of Defence Technology, Changsha, Hunan, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0669-6892","authenticated-orcid":false,"given":"Keshi","family":"Ge","sequence":"additional","affiliation":[{"name":"College of Computer, National University of Defence Technology, Changsha, Hunan, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9743-2034","authenticated-orcid":false,"given":"Dongsheng","family":"Li","sequence":"additional","affiliation":[{"name":"College of Computer, National University of Defence Technology, Changsha, Hunan, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xicheng","family":"Lu","sequence":"additional","affiliation":[{"name":"College of Computer, National University of Defence Technology, Changsha, Hunan, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"key":"ref3","article-title":"Bert: Pre-training of deep bidirectional transformers for language understanding","author":"Devlin","year":"2018"},{"key":"ref4","article-title":"Language models are unsupervised multitask learners","author":"Radford","year":"2019"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.14778\/3415478.3415530"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1145\/3437801.3441593"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1145\/3472456.3472497"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.48550\/arxiv.1811.06965"},{"key":"ref9","first-page":"7937","article-title":"Memory-efficient pipeline-parallel DNN training","volume-title":"Int. Conf. Mach. Learn.","author":"Narayanan","year":"2021"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1145\/3341301.3359646"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/CLUSTER51413.2022.00043"},{"key":"ref12","first-page":"1","article-title":"Efficient large-scale language model training on GPU clusters using megatron-LM","volume-title":"Proc. Int. Conf. High Perform. Comput., Netw., Storage Anal.","author":"Narayanan","year":"2021"},{"key":"ref13","article-title":"Piper: Multidimensional planner for DNN parallelization","volume":"34","author":"Tarnawski","year":"2021","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1145\/3458817.3476145"},{"key":"ref15","article-title":"Adam: A method for stochastic optimization","author":"Kingma","year":"2014"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/TBDATA.2015.2472014"},{"key":"ref17","article-title":"Megatron-LM: Training multi-billion parameter language models using model parallelism","author":"Shoeybi","year":"2019"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1145\/3575693.3575703"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/EPEPS.2010.5642789"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/HOTI.2010.24"},{"key":"ref21","first-page":"2","article-title":"PCI express*(pcie*) 3.0 accelerator features","volume":"10","author":"Ajanovic","year":"2008","journal-title":"Intel Corporation"},{"key":"ref22","article-title":"Horovod: Fast and easy distributed deep learning in tensorflow","author":"Sergeev","year":"2018"},{"issue":"617\u2013632","key":"ref23","first-page":"102","article-title":"An introduction to the infiniband architecture","volume":"42","author":"Pfister","year":"2001","journal-title":"High Perform. Mass Storage Parallel I\/O"},{"key":"ref24","volume-title":"CUDA by Example: An Introduction to General-Purpose GPU Programming","author":"Sanders","year":"2010"},{"key":"ref25","article-title":"CUDNN: Efficient primitives for deep learning","author":"Chetlur","year":"2014"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2015.11"},{"key":"ref27","article-title":"Skip-thought vectors","volume":"28","author":"Kiros","year":"2015","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref28","article-title":"Openwebtext corpus","author":"Gokaslan","year":"2019"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.5555\/3454287.3455008"},{"key":"ref30","article-title":"Nccl 2.0","volume-title":"GPU Technol. Conf. (GTC)","volume":"2","author":"Jeaugey","year":"2017"},{"key":"ref31","article-title":"One weird trick for parallelizing convolutional neural networks","author":"Krizhevsky","year":"2014"},{"key":"ref32","article-title":"Automatic cross-replica sharding of weight update in data-parallel training","author":"Xu","year":"2020"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1145\/3458817.3476205"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1109\/SC41405.2020.00049"},{"key":"ref35","first-page":"1","article-title":"Hanayo: Harnessing wave-like pipeline parallelism for enhanced large model training efficiency","volume-title":"Proc. Int. Conf. High Perform. Comput., Netw., Storage Anal.","author":"Liu","year":"2023"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1109\/TPDS.2021.3132413"},{"key":"ref37","first-page":"2279","article-title":"Exploring hidden dimensions in parallelizing convolutional neural networks.","volume-title":"Proc. ICML","author":"Jia","year":"2018"},{"key":"ref38","article-title":"Learning generalizable device placement algorithms for distributed machine learning","volume":"32","author":"Venkatakrishnan","year":"2019","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref39","first-page":"1676","article-title":"Spotlight: Optimizing device placement for training deep neural networks","volume-title":"Int. Conf. Mach. Learn.","author":"Gao","year":"2018"},{"key":"ref40","first-page":"1","article-title":"Beyond data and model parallelism for deep neural networks.","volume-title":"Proc. Mach. Learn. Syst.","volume":"1","author":"Jia","year":"2019"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM48880.2022.9796787"}],"container-title":["IEEE Transactions on Computers"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/12\/10924434\/10803065.pdf?arnumber=10803065","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,3,13]],"date-time":"2025-03-13T04:53:08Z","timestamp":1741841588000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10803065\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,4]]},"references-count":41,"journal-issue":{"issue":"4"},"URL":"https:\/\/doi.org\/10.1109\/tc.2024.3517748","relation":{},"ISSN":["0018-9340","1557-9956","2326-3814"],"issn-type":[{"value":"0018-9340","type":"print"},{"value":"1557-9956","type":"electronic"},{"value":"2326-3814","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,4]]}}}