{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,12,3]],"date-time":"2025-12-03T18:14:04Z","timestamp":1764785644106,"version":"3.44.0"},"reference-count":45,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"3","license":[{"start":{"date-parts":[[2025,7,1]],"date-time":"2025-07-01T00:00:00Z","timestamp":1751328000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2025,7,1]],"date-time":"2025-07-01T00:00:00Z","timestamp":1751328000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,7,1]],"date-time":"2025-07-01T00:00:00Z","timestamp":1751328000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"National Key Research and Development Project","award":["2022YFB2901604"],"award-info":[{"award-number":["2022YFB2901604"]}]},{"name":"Zhejiang Provincial Natural Science Foundation of China","award":["LZ22F010008"],"award-info":[{"award-number":["LZ22F010008"]}]},{"DOI":"10.13039\/501100004543","name":"China Scholarship Council","doi-asserted-by":"publisher","award":["202406320382"],"award-info":[{"award-number":["202406320382"]}],"id":[{"id":"10.13039\/501100004543","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Emerg. Topics Comput."],"published-print":{"date-parts":[[2025,7]]},"DOI":"10.1109\/tetc.2025.3573522","type":"journal-article","created":{"date-parts":[[2025,6,2]],"date-time":"2025-06-02T14:02:29Z","timestamp":1748872949000},"page":"1170-1184","source":"Crossref","is-referenced-by-count":1,"title":["PipeDAP: An Efficient Communication Framework for Scheduling Decoupled All-Reduce Primitives in Distributed DNN Training"],"prefix":"10.1109","volume":"13","author":[{"ORCID":"https:\/\/orcid.org\/0009-0003-0599-8383","authenticated-orcid":false,"given":"Yunqi","family":"Gao","sequence":"first","affiliation":[{"name":"Zhejiang University, Hangzhou, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0594-8998","authenticated-orcid":false,"given":"Bing","family":"Hu","sequence":"additional","affiliation":[{"name":"Zhejiang University, Hangzhou, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9948-9165","authenticated-orcid":false,"given":"Mahdi Boloursaz","family":"Mashhadi","sequence":"additional","affiliation":[{"name":"5GIC & 6GIC, Institute for Communication Systems (ICS), University of Surrey, Guildford, U.K."}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2153-9075","authenticated-orcid":false,"given":"Wei","family":"Wang","sequence":"additional","affiliation":[{"name":"Zhejiang University, Hangzhou, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6062-8639","authenticated-orcid":false,"given":"Rahim","family":"Tafazolli","sequence":"additional","affiliation":[{"name":"5GIC & 6GIC, Institute for Communication Systems (ICS), University of Surrey, Guildford, U.K."}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8941-8080","authenticated-orcid":false,"given":"M\u00e9rouane","family":"Debbah","sequence":"additional","affiliation":[{"name":"KU 6G Research Center, Department of Computer and Information Engineering, Khalifa University, Abu Dhabi, UAE"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/MCOM.001.2300364"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.24963\/ijcai.2024\/883"},{"article-title":"Accurate, large minibatch SGD: Training ImageNet in 1 hour","year":"2017","author":"Goyal","key":"ref3"},{"article-title":"Language models are few-shot learners","year":"2020","author":"Mann","key":"ref4"},{"article-title":"Megatron-LM: Training multi-billion parameter language models using model parallelism","year":"2019","author":"Shoeybi","key":"ref5"},{"key":"ref6","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"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM42981.2021.9488803"},{"key":"ref8","first-page":"8024","article-title":"PyTorch: An imperative style, high-performance deep learning library","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Paszke"},{"key":"ref9","first-page":"265","article-title":"TensorFlow: A system for large-scale machine learning","volume-title":"Proc. USENIX Conf. Operating Syst. Des. Implementation","author":"Abadi"},{"article-title":"MXNet: A flexible and efficient machine learning library for heterogeneous distributed systems","year":"2015","author":"Chen","key":"ref10"},{"article-title":"Highly scalable deep learning training system with mixed-precision: Training ImageNet in four minutes","year":"2018","author":"Jia","key":"ref11"},{"key":"ref12","first-page":"181","article-title":"Poseidon: An efficient communication architecture for distributed deep learning on GPU clusters","volume-title":"Proc. USENIX Annu. Tech. Conf.","author":"Zhang"},{"key":"ref13","first-page":"1709","article-title":"QSGD: Communication-efficient SGD via gradient quantization and encoding","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Alistarh"},{"key":"ref14","first-page":"1509","article-title":"TernGrad: Ternary gradients to reduce communication in distributed deep learning","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Wen"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1109\/ICDCS.2019.00220"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM.2019.8737587"},{"key":"ref17","first-page":"1223","article-title":"More effective distributed ML via a stale synchronous parallel parameter server","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Ho"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1145\/3341301.3359646"},{"article-title":"Horovod: Fast and easy distributed deep learning in TensorFlow","year":"2018","author":"Sergeev","key":"ref19"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM.2019.8737367"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1145\/3341301.3359642"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/ICDCS57875.2023.00054"},{"key":"ref23","first-page":"132","article-title":"Priority-based parameter propagation for distributed DNN training","volume-title":"Proc. 2nd Conf. Mach. Learn. Syst.","author":"Jayarajan"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM41043.2020.9155446"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1109\/TPDS.2023.3331372"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.5555\/2685048.2685095"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM48880.2022.9796820"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"ref29","first-page":"142","article-title":"Learning word vectors for sentiment analysis","volume-title":"Proc. Annu. Meeting Assoc. Comput. Linguistics: Hum. Lang. Technol.","author":"Maas"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"article-title":"Very deep convolutional networks for large-scale image recognition","year":"2014","author":"Simonyan","key":"ref31"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1810.04805"},{"issue":"8","key":"ref33","article-title":"Language models are unsupervised multitask learners","volume":"1","author":"Radford","year":"2019","journal-title":"OpenAI Blog"},{"article-title":"DeFT: Mitigating data dependencies for flexible communication scheduling in distributed training","year":"2025","author":"Meng","key":"ref34"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1109\/ICDCS51616.2021.00010"},{"key":"ref36","first-page":"629","article-title":"Gaia: Geo-distributed machine learning approaching LAN speeds","volume-title":"Proc. 14th USENIX Conf. Netw. Syst. Des. Implementation","author":"Hsieh"},{"key":"ref37","first-page":"172","article-title":"Blink: Fast and generic collectives for distributed ML","volume-title":"Proc. Conf. Mach. Learn. Syst.","author":"Wang"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1145\/2966884.2966912"},{"key":"ref39","first-page":"1","article-title":"Blueconnect: Novel hierarchical all-reduce on multi-tired network for deep learning","volume-title":"Proc. Conf. Syst. Mach. Learn.","author":"Cho"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1145\/3018743.3018769"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM.2019.8737595"},{"article-title":"Image classification at supercomputer scale","year":"2018","author":"Ying","key":"ref42"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1145\/3392717.3392771"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1145\/3503222.3507778"},{"article-title":"FLUX: Fast software-based communication overlap on GPUs through kernel fusion","year":"2024","author":"Chang","key":"ref45"}],"container-title":["IEEE Transactions on Emerging Topics in Computing"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/6245516\/11159571\/11021340.pdf?arnumber=11021340","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,12]],"date-time":"2025-09-12T17:31:17Z","timestamp":1757698277000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11021340\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,7]]},"references-count":45,"journal-issue":{"issue":"3"},"URL":"https:\/\/doi.org\/10.1109\/tetc.2025.3573522","relation":{},"ISSN":["2168-6750","2376-4562"],"issn-type":[{"type":"electronic","value":"2168-6750"},{"type":"electronic","value":"2376-4562"}],"subject":[],"published":{"date-parts":[[2025,7]]}}}