{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,21]],"date-time":"2026-07-21T19:23:16Z","timestamp":1784661796820,"version":"3.55.0"},"publisher-location":"New York, NY, USA","reference-count":36,"publisher":"ACM","license":[{"start":{"date-parts":[[2024,8,4]],"date-time":"2024-08-04T00:00:00Z","timestamp":1722729600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.acm.org\/publications\/policies\/copyright_policy#Background"}],"funder":[{"DOI":"10.13039\/100000001","name":"NSF (National Science Foundation)","doi-asserted-by":"publisher","award":["CNS-2107147"],"award-info":[{"award-number":["CNS-2107147"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100006112","name":"Microsoft Research","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100006112","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100006920","name":"University of Pennsylvania","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100006920","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100007812","name":"University of Washington","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100007812","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2024,8,4]]},"DOI":"10.1145\/3651890.3672249","type":"proceedings-article","created":{"date-parts":[[2024,7,31]],"date-time":"2024-07-31T13:11:43Z","timestamp":1722431503000},"page":"16-37","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":30,"title":["Rethinking Machine Learning Collective Communication as a Multi-Commodity Flow Problem"],"prefix":"10.1145","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-8017-6652","authenticated-orcid":false,"given":"Xuting","family":"Liu","sequence":"first","affiliation":[{"name":"University of Pennsylvania, Philadelphia, United States of America"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2234-3250","authenticated-orcid":false,"given":"Behnaz","family":"Arzani","sequence":"additional","affiliation":[{"name":"Microsoft Research, Redmond, United States of America"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0006-2694-4685","authenticated-orcid":false,"given":"Siva Kesava Reddy","family":"Kakarla","sequence":"additional","affiliation":[{"name":"Microsoft Research, Redmond, United States of America"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0007-5794-0326","authenticated-orcid":false,"given":"Liangyu","family":"Zhao","sequence":"additional","affiliation":[{"name":"University of Washington, Seattle, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7683-208X","authenticated-orcid":false,"given":"Vincent","family":"Liu","sequence":"additional","affiliation":[{"name":"University of Pennsylvania, Philadelphia, United States of America"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6197-0818","authenticated-orcid":false,"given":"Miguel","family":"Castro","sequence":"additional","affiliation":[{"name":"OpenAI, San Francisco, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9494-6435","authenticated-orcid":false,"given":"Srikanth","family":"Kandula","sequence":"additional","affiliation":[{"name":"Microsoft, Redmond, United States of America"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0633-4004","authenticated-orcid":false,"given":"Luke","family":"Marshall","sequence":"additional","affiliation":[{"name":"Microsoft Research, Redmond, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2024,8,4]]},"reference":[{"key":"e_1_3_2_1_1_1","volume-title":"18th USENIX Symposium on Networked Systems Design and Implementation (NSDI 21)","author":"Abuzaid Firas","year":"2021","unstructured":"Firas Abuzaid, Srikanth Kandula, Behnaz Arzani, Ishai Menache, Matei Zaharia, and Peter Bailis. 2021. Contracting wide-area network topologies to solve flow problems quickly. In 18th USENIX Symposium on Networked Systems Design and Implementation (NSDI 21). 175--200."},{"key":"e_1_3_2_1_2_1","unstructured":"AMD 2023. AMD Instinct MI250 GPUs (slide 18). https:\/\/hc34.hotchips.org\/assets\/program\/conference\/day1\/GPU%20HPC\/HC2022.AMD.AlanSmith.v14.Final.20220820.pdf"},{"key":"e_1_3_2_1_3_1","unstructured":"Azure NDv2-series 2022. Azure NDv2-series. https:\/\/learn.microsoft.com\/en-us\/azure\/virtual-machines\/ndv2-series"},{"key":"e_1_3_2_1_4_1","volume-title":"Introduction to linear optimization","author":"Bertsimas Dimitris","unstructured":"Dimitris Bertsimas and John N Tsitsiklis. 1997. Introduction to linear optimization. Vol. 6. Athena Scientific Belmont, MA."},{"key":"e_1_3_2_1_5_1","doi-asserted-by":"publisher","DOI":"10.1145\/3437801.3441620"},{"key":"e_1_3_2_1_6_1","unstructured":"ChatGPT runs 10K Nvidia training GPUs with potential for thousands more 2023. ChatGPT runs 10K Nvidia training GPUs with potential for thousands more. https:\/\/www.fierceelectronics.com\/sensors\/chatgpt-runs-10k-nvidia-training-gpus-potential-thousands-more"},{"key":"e_1_3_2_1_7_1","doi-asserted-by":"publisher","DOI":"10.1145\/3575693.3575724"},{"key":"e_1_3_2_1_8_1","doi-asserted-by":"publisher","DOI":"10.1145\/3437963.3441727"},{"key":"e_1_3_2_1_9_1","volume-title":"Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805","author":"Devlin Jacob","year":"2018","unstructured":"Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2018. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805 (2018)."},{"key":"e_1_3_2_1_11_1","doi-asserted-by":"publisher","DOI":"10.1109\/INFCOM.1993.253246"},{"key":"e_1_3_2_1_12_1","doi-asserted-by":"publisher","DOI":"10.1109\/tssc.1968.300136"},{"key":"e_1_3_2_1_13_1","doi-asserted-by":"publisher","DOI":"10.1145\/2486001.2486012"},{"key":"e_1_3_2_1_14_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ipl.2010.02.001"},{"key":"e_1_3_2_1_15_1","doi-asserted-by":"publisher","DOI":"10.1145\/2534169.2486019"},{"key":"e_1_3_2_1_16_1","volume-title":"GPU Technology Conference (GTC)","volume":"2","author":"Jeaugey Sylvain","year":"2017","unstructured":"Sylvain Jeaugey. 2017. NCCL 2.0. In GPU Technology Conference (GTC), Vol. 2."},{"key":"e_1_3_2_1_17_1","doi-asserted-by":"publisher","DOI":"10.1145\/2619239.2626336"},{"key":"e_1_3_2_1_18_1","doi-asserted-by":"publisher","DOI":"10.1145\/3620666.3651362"},{"key":"e_1_3_2_1_19_1","volume-title":"19th USENIX Symposium on Networked Systems Design and Implementation (NSDI 22)","author":"Krishnaswamy Umesh","year":"2022","unstructured":"Umesh Krishnaswamy, Rachee Singh, Nikolaj Bj\u00f8rner, and Himanshu Raj. 2022. Decentralized cloud wide-area network traffic engineering with BLASTSHIELD. In 19th USENIX Symposium on Networked Systems Design and Implementation (NSDI 22). 325--338."},{"key":"e_1_3_2_1_20_1","volume-title":"20th USENIX Symposium on Networked Systems Design and Implementation (NSDI 23)","author":"Krishnaswamy Umesh","year":"2023","unstructured":"Umesh Krishnaswamy, Rachee Singh, Paul Mattes, Paul-Andre C Bissonnette, Nikolaj Bj\u00f8rner, Zahira Nasrin, Sonal Kothari, Prabhakar Reddy, John Abeln, Srikanth Kandula, et al. 2023. OneWAN is better than two: Unifying a split WAN architecture. In 20th USENIX Symposium on Networked Systems Design and Implementation (NSDI 23). 515--529."},{"key":"e_1_3_2_1_21_1","doi-asserted-by":"publisher","DOI":"10.1145\/2018436.2018446"},{"key":"e_1_3_2_1_22_1","volume-title":"20th USENIX Symposium on Networked Systems Design and Implementation (NSDI 23)","author":"Mahajan Kshiteej","year":"2023","unstructured":"Kshiteej Mahajan, Ching-Hsiang Chu, Srinivas Sridharan, and Aditya Akella. 2023. Better Together: Jointly Optimizing ML Collective Scheduling and Execution Planning using SYNDICATE. In 20th USENIX Symposium on Networked Systems Design and Implementation (NSDI 23). USENIX Association, Boston, MA, 809--824. https:\/\/www.usenix.org\/conference\/nsdi23\/presentation\/mahajan"},{"key":"e_1_3_2_1_23_1","unstructured":"MSCCL codebase 2023. MSCCL codebase. https:\/\/github.com\/microsoft\/msccl"},{"key":"e_1_3_2_1_24_1","doi-asserted-by":"publisher","DOI":"10.1145\/3563766.3564102"},{"key":"e_1_3_2_1_25_1","doi-asserted-by":"publisher","DOI":"10.1145\/3477132.3483588"},{"key":"e_1_3_2_1_26_1","unstructured":"NCCL Tests 2021. NCCL Tests. https:\/\/github.com\/NVIDIA\/nccl-tests\/blob\/d028efcf35101c6663ae8c5f33ad41bad00efb4d\/doc\/PERFORMANCE.md#algorithm-bandwidth"},{"key":"e_1_3_2_1_27_1","doi-asserted-by":"publisher","DOI":"10.1109\/INFCOM.1994.337651"},{"key":"e_1_3_2_1_28_1","unstructured":"Nvidia DGX System 2023. Nvidia DGX System. https:\/\/www.nvidia.com\/en-us\/data-center\/dgx-systems\/"},{"key":"e_1_3_2_1_29_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.cor.2003.12.007"},{"key":"e_1_3_2_1_30_1","volume-title":"Porto, Portugal 1","author":"Pedroso Joo Pedro","year":"2011","unstructured":"Joo Pedro Pedroso. 2011. Optimization with gurobi and python. INESC Porto and Universidade do Porto,, Porto, Portugal 1 (2011)."},{"key":"e_1_3_2_1_31_1","doi-asserted-by":"publisher","DOI":"10.1145\/3470496.3527382"},{"key":"e_1_3_2_1_32_1","unstructured":"ROCm 2024. AMD ROCm Software. https:\/\/www.amd.com\/en\/products\/software\/rocm.html"},{"key":"e_1_3_2_1_33_1","volume-title":"20th USENIX Symposium on Networked Systems Design and Implementation (NSDI 23)","author":"Shah Aashaka","year":"2023","unstructured":"Aashaka Shah, Vijay Chidambaram, Meghan Cowan, Saeed Maleki, Madan Musuvathi, Todd Mytkowicz, Jacob Nelson, Olli Saarikivi, and Rachee Singh. 2023. TACCL: Guiding Collective Algorithm Synthesis using Communication Sketches. In 20th USENIX Symposium on Networked Systems Design and Implementation (NSDI 23). USENIX Association, Boston, MA, 593--612. https:\/\/www.usenix.org\/conference\/nsdi23\/presentation\/shah"},{"key":"e_1_3_2_1_34_1","volume-title":"Guiding Collective Algorithm Synthesis using Communication Sketches","author":"TACCL","year":"2023","unstructured":"TACCL: Guiding Collective Algorithm Synthesis using Communication Sketches 2023. TACCL codebase. https:\/\/github.com\/microsoft\/taccl"},{"key":"e_1_3_2_1_35_1","first-page":"172","article-title":"Blink: Fast and generic collectives for distributed ML","volume":"2","author":"Wang Guanhua","year":"2020","unstructured":"Guanhua Wang, Shivaram Venkataraman, Amar Phanishayee, Nikhil Devanur, Jorgen Thelin, and Ion Stoica. 2020. Blink: Fast and generic collectives for distributed ML. Proceedings of Machine Learning and Systems 2 (2020), 172--186.","journal-title":"Proceedings of Machine Learning and Systems"},{"key":"e_1_3_2_1_36_1","volume-title":"TopoOpt: Co-optimizing Network Topology and Parallelization Strategy for Distributed Training Jobs. In 20th USENIX Symposium on Networked Systems Design and Implementation (NSDI 23)","author":"Wang Weiyang","year":"2023","unstructured":"Weiyang Wang, Moein Khazraee, Zhizhen Zhong, Manya Ghobadi, Zhihao Jia, Dheevatsa Mudigere, Ying Zhang, and Anthony Kewitsch. 2023. TopoOpt: Co-optimizing Network Topology and Parallelization Strategy for Distributed Training Jobs. In 20th USENIX Symposium on Networked Systems Design and Implementation (NSDI 23). USENIX Association, Boston, MA, 739--767. https:\/\/www.usenix.org\/conference\/nsdi23\/presentation\/wang-weiyang"},{"key":"e_1_3_2_1_37_1","doi-asserted-by":"publisher","unstructured":"Liangyu Zhao Siddharth Pal Tapan Chugh Weiyang Wang Prithwish Basu Joud Khoury and Arvind Krishnamurthy. 2022. Optimal Direct-Connect Topologies for Collective Communications. 10.48550\/ARXIV.2202.03356","DOI":"10.48550\/ARXIV.2202.03356"}],"event":{"name":"ACM SIGCOMM '24: ACM SIGCOMM 2024 Conference","location":"Sydney NSW Australia","acronym":"ACM SIGCOMM '24","sponsor":["SIGCOMM ACM Special Interest Group on Data Communication"]},"container-title":["Proceedings of the ACM SIGCOMM 2024 Conference"],"original-title":[],"link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3651890.3672249","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/abs\/10.1145\/3651890.3672249","content-type":"text\/html","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3651890.3672249","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3651890.3672249","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,18]],"date-time":"2025-06-18T17:49:12Z","timestamp":1750268952000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3651890.3672249"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,8,4]]},"references-count":36,"alternative-id":["10.1145\/3651890.3672249","10.1145\/3651890"],"URL":"https:\/\/doi.org\/10.1145\/3651890.3672249","relation":{},"subject":[],"published":{"date-parts":[[2024,8,4]]},"assertion":[{"value":"2024-08-04","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}