{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,5]],"date-time":"2026-03-05T15:45:15Z","timestamp":1772725515080,"version":"3.50.1"},"reference-count":49,"publisher":"IEEE","license":[{"start":{"date-parts":[[2024,10,28]],"date-time":"2024-10-28T00:00:00Z","timestamp":1730073600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2024,10,28]],"date-time":"2024-10-28T00:00:00Z","timestamp":1730073600000},"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":[],"published-print":{"date-parts":[[2024,10,28]]},"DOI":"10.1109\/icnp61940.2024.10858570","type":"proceedings-article","created":{"date-parts":[[2025,2,4]],"date-time":"2025-02-04T18:29:45Z","timestamp":1738693785000},"page":"1-12","source":"Crossref","is-referenced-by-count":4,"title":["Rina: Enhancing Ring-Allreduce with in-Network Aggregation in Distributed Model Training"],"prefix":"10.1109","author":[{"given":"Zixuan","family":"Chen","sequence":"first","affiliation":[{"name":"Fudan University,School of Computer Science,Shanghai,China"}]},{"given":"Xuandong","family":"Liu","sequence":"additional","affiliation":[{"name":"Fudan University,School of Computer Science,Shanghai,China"}]},{"given":"Minglin","family":"Li","sequence":"additional","affiliation":[{"name":"Fudan University,School of Computer Science,Shanghai,China"}]},{"given":"Yinfan","family":"Hu","sequence":"additional","affiliation":[{"name":"Fudan University,School of Computer Science,Shanghai,China"}]},{"given":"Hao","family":"Mei","sequence":"additional","affiliation":[{"name":"Fudan University,School of Computer Science,Shanghai,China"}]},{"given":"Huifeng","family":"Xing","sequence":"additional","affiliation":[{"name":"Fudan University,School of Computer Science,Shanghai,China"}]},{"given":"Hao","family":"Wang","sequence":"additional","affiliation":[{"name":"Fudan University,School of Computer Science,Shanghai,China"}]},{"given":"Wanxin","family":"Shi","sequence":"additional","affiliation":[{"name":"Fudan University,School of Computer Science,Shanghai,China"}]},{"given":"Sen","family":"Liu","sequence":"additional","affiliation":[{"name":"Fudan University,School of Computer Science,Shanghai,China"}]},{"given":"Yang","family":"Xu","sequence":"additional","affiliation":[{"name":"Fudan University,School of Computer Science,Shanghai,China"}]}],"member":"263","reference":[{"key":"ref1","year":"2023","journal-title":"Gpt-4 technical report"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1613\/jair.1.12007"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00371"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1109\/IWQoS57198.2023.10188699"},{"key":"ref6","first-page":"785","article-title":"Scaling distributed machine learning with {In-Network} aggregation","author":"Sapio","year":"2021","journal-title":"18th USENIX Symposium on Networked Systems Design and Implementation (NSDI 21)"},{"key":"ref7","first-page":"741","article-title":"$\\{$ ATP $\\}$: In-network aggregation for multi-tenant learning","author":"Lao","year":"2021","journal-title":"18th USENIX Symposium on Networked Systems Design and Implementation (NSDI 21)"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM53939.2023.10228956"},{"key":"ref9","first-page":"829","article-title":"In-network aggregation for shared machine learning clusters","volume":"3","author":"Gebara","year":"2021","journal-title":"Proceedings of Machine Learning and Systems"},{"key":"ref10","article-title":"Measuring the effects of data parallelism on neural network training","author":"Shallue","year":"2018","journal-title":"arXiv preprint"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1093\/comjnl\/30.5.413"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1145\/3567955.3567959"},{"key":"ref13","first-page":"265","article-title":"On optimization methods for deep learning","author":"Le","year":"2011","journal-title":"Proceedings of the 28th international conference on international conference on machine learning"},{"key":"ref14","article-title":"Megatron-lm: Training multi-billion parameter language models using model parallelism","author":"Shoeybi","year":"2019","journal-title":"arXiv preprint"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.5555\/2685048.2685095"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1145\/79173.79181"},{"key":"ref17","article-title":"Asynchronous parallel stochastic gradient for nonconvex optimization","volume":"28","author":"Lian","year":"2015","journal-title":"Advances in neural information processing systems"},{"key":"ref18","first-page":"336","article-title":"Mlperf training benchmark","volume":"2","author":"Mattson","year":"2020","journal-title":"Proceedings of Machine Learning and Systems"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1142\/6177"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1145\/3411029.3411037"},{"key":"ref21","volume-title":"Nvidia collective communications library (ncel)","year":"2024"},{"key":"ref22","volume-title":"Open mpi: Open source high performance computing","author":"MPI","year":"2024"},{"key":"ref23","first-page":"929","article-title":"Check-n-run: a checkpointing system for training deep learning recommendation models","author":"Eisenman","year":"2022","journal-title":"19th USENIX Symposium on Networked Systems Design and Implementation (NSDI 22)"},{"key":"ref24","article-title":"A study of checkpointing in large scale training of deep neural networks","author":"Rojas","year":"2020","journal-title":"arXiv preprint"},{"key":"ref25","article-title":"Highly scalable deep learning training system with mixed-precision: Training imagenet in four minutes","author":"Jia","year":"2018","journal-title":"arXiv preprint"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref27","author":"Krizhevsky","year":"2009","journal-title":"Learning multiple layers of features from tiny images"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1145\/1402946.1402967"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA.2008.19"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM.2019.8737587"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1145\/3605573.3605650"},{"key":"ref32","article-title":"Intel \u00ae tofino ${ }^{\\text{TM }}$ programmable ethernet switch asic"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1145\/3575693.3575708"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1145\/2829988.2787484"},{"key":"ref35","volume-title":"Raw ethernet programming: Basic introduction - code example"},{"key":"ref36","volume-title":"Tcp segmentation offload"},{"key":"ref37","first-page":"2011","journal-title":"ns-3 network simulator"},{"key":"ref38","volume-title":"Cuda, release: 11.2","year":"2020"},{"key":"ref39","article-title":"Very deep convolutional networks for large-scale image recognition","author":"Simonyan","year":"2014","journal-title":"arXiv preprint"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1109\/ICIVC.2017.7984661"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"ref42","article-title":"Bert: Pre-training of deep bidirectional transformers for language understanding","author":"Devlin","year":"2018","journal-title":"arXiv preprint"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D16-1264"},{"key":"ref44","first-page":"1737","article-title":"Deep learning with limited numerical precision","author":"Gupta","year":"2015","journal-title":"International conference on machine learning. PMLR"},{"key":"ref45","article-title":"Qsgd: Communication-efficient sgd via gradient quantization and encoding","volume":"30","author":"Alistarh","year":"2017","journal-title":"Advances in neural information processing systems"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2015-354"},{"key":"ref47","article-title":"Deep gradient compression: Reducing the communication bandwidth for distributed training","author":"Lin","year":"2017","journal-title":"arXiv preprint"},{"key":"ref48","article-title":"More effective distributed ml via a stale synchronous parallel parameter server","volume":"26","author":"Ho","year":"2013","journal-title":"Advances in neural information processing systems"},{"key":"ref49","article-title":"Local sgd converges fast and communicates little","author":"Stich","year":"2018","journal-title":"arXiv preprint"}],"event":{"name":"2024 IEEE 32nd International Conference on Network Protocols (ICNP)","location":"Charleroi, Belgium","start":{"date-parts":[[2024,10,28]]},"end":{"date-parts":[[2024,10,31]]}},"container-title":["2024 IEEE 32nd International Conference on Network Protocols (ICNP)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/10858485\/10858498\/10858570.pdf?arnumber=10858570","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,2,5]],"date-time":"2025-02-05T05:59:00Z","timestamp":1738735140000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10858570\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,10,28]]},"references-count":49,"URL":"https:\/\/doi.org\/10.1109\/icnp61940.2024.10858570","relation":{},"subject":[],"published":{"date-parts":[[2024,10,28]]}}}