{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,29]],"date-time":"2026-04-29T11:09:17Z","timestamp":1777460957136,"version":"3.51.4"},"publisher-location":"New York, NY, USA","reference-count":49,"publisher":"ACM","content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2026,4,27]]},"DOI":"10.1145\/3805621.3807615","type":"proceedings-article","created":{"date-parts":[[2026,4,28]],"date-time":"2026-04-28T13:08:45Z","timestamp":1777381725000},"page":"407-416","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["On Harnessing Idle Compute at the Edge for Foundation Model Training"],"prefix":"10.1145","author":[{"ORCID":"https:\/\/orcid.org\/0009-0007-4367-9198","authenticated-orcid":false,"given":"Leyang","family":"Xue","sequence":"first","affiliation":[{"name":"University of Edinburgh, Edinburgh, United Kingdom"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-5593-8752","authenticated-orcid":false,"given":"Meghana","family":"Madhyastha","sequence":"additional","affiliation":[{"name":"Johns Hopkins University, Baltimore, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2360-7019","authenticated-orcid":false,"given":"Myungjin","family":"Lee","sequence":"additional","affiliation":[{"name":"Cisco Research, Bellevue, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8100-506X","authenticated-orcid":false,"given":"Amos","family":"Storkey","sequence":"additional","affiliation":[{"name":"The University of Edinburgh, Edinburgh, United Kingdom"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2924-1997","authenticated-orcid":false,"given":"Randal","family":"Burns","sequence":"additional","affiliation":[{"name":"Johns Hopkins University, Baltimore, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6946-5143","authenticated-orcid":false,"given":"Mahesh","family":"Marina","sequence":"additional","affiliation":[{"name":"University of Edinburgh, Edinburgh, United Kingdom"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"320","published-online":{"date-parts":[[2026,4,28]]},"reference":[{"key":"e_1_3_2_1_1_1","unstructured":"Backlinko. 2023. Smartphone Usage Statistics. https:\/\/backlinko.com\/smartphone-usage-statistics. Accessed: 2024-07-28."},{"key":"e_1_3_2_1_2_1","article-title":"Compute-Efficient Deep Learning: Algorithmic Trends and Opportunities","volume":"24","author":"Bartoldson Brian R.","year":"2023","unstructured":"Brian R. Bartoldson, Bhavya Kailkhura, and Davis W. Blalock. 2023. Compute-Efficient Deep Learning: Algorithmic Trends and Opportunities. J. Mach. Learn. Res. 24 (2023), 122:1\u2013122:77.","journal-title":"J. Mach. Learn. Res."},{"key":"e_1_3_2_1_3_1","volume-title":"Oakestra: A Lightweight Hierarchical Orchestration Framework for Edge Computing","author":"Bartolomeo Giovanni","year":"2023","unstructured":"Giovanni Bartolomeo, Mehdi Yosofie, Simon B\u00e4urle, Oliver Haluszczynski, Nitinder Mohan, and J\u00f6rg Ott. 2023. Oakestra: A Lightweight Hierarchical Orchestration Framework for Edge Computing. In USENIX ATC. USENIX Association, 215\u2013231."},{"key":"e_1_3_2_1_4_1","unstructured":"Rishi Bommasani Drew A. Hudson Ehsan Adeli et al. 2021. On the Opportunities and Risks of Foundation Models. arXiv:2108.07258"},{"key":"e_1_3_2_1_5_1","unstructured":"BT. 2024. Broadband deals. https:\/\/www.bt.com\/broadband\/deals."},{"key":"e_1_3_2_1_6_1","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2019.2921977"},{"key":"e_1_3_2_1_7_1","volume-title":"ATP: Adaptive Tensor Parallelism for Foundation Models. arXiv:2301.08658","author":"Cheng Shenggan","year":"2023","unstructured":"Shenggan Cheng, Ziming Liu, Jiangsu Du, and Yang You. 2023. ATP: Adaptive Tensor Parallelism for Foundation Models. arXiv:2301.08658"},{"key":"e_1_3_2_1_8_1","volume-title":"Denis Mazur, Ilia Kobelev, Yacine Jernite, Thomas Wolf, and Gennady Pekhimenko.","author":"Diskin Michael","year":"2021","unstructured":"Michael Diskin, Alexey Bukhtiyarov, Max Ryabinin, Lucile Saulnier, Quentin Lhoest, Anton Sinitsin, Dmitry Popov, Dmitry V. Pyrkin, Maxim Kashirin, Alexander Borzunov, Albert Villanova del Moral, Denis Mazur, Ilia Kobelev, Yacine Jernite, Thomas Wolf, and Gennady Pekhimenko. 2021. Distributed Deep Learning In Open Collaborations. In NeurIPS. 7879\u20137897."},{"key":"e_1_3_2_1_9_1","unstructured":"GitHub. 2021. GitHub Copilot \u00b7 Your AI pair programmer. https:\/\/github.com\/features\/copilot. Accessed: 2024-05-17."},{"key":"e_1_3_2_1_10_1","unstructured":"Google. 2024. gRPC - An RPC library and framework. https:\/\/github.com\/grpc\/grpc. Accessed: 2024-05-17."},{"key":"e_1_3_2_1_11_1","unstructured":"Gurobi Optimization LLC. 2024. Gurobi Optimizer Reference Manual. https:\/\/www.gurobi.com"},{"key":"e_1_3_2_1_12_1","volume-title":"EDDL: A Distributed Deep Learning System for Resource-limited Edge Computing Environment","author":"Hao Pengzhan","year":"2021","unstructured":"Pengzhan Hao and Yifan Zhang. 2021. EDDL: A Distributed Deep Learning System for Resource-limited Edge Computing Environment. In SEC. IEEE, 1\u201313."},{"key":"e_1_3_2_1_13_1","volume-title":"HyoukJoong Lee, Jiquan Ngiam, Quoc V. Le, Yonghui Wu, and Zhifeng Chen.","author":"Huang Yanping","year":"2019","unstructured":"Yanping Huang, Youlong Cheng, Ankur Bapna, Orhan Firat, Dehao Chen, Mia Xu Chen, HyoukJoong Lee, Jiquan Ngiam, Quoc V. Le, Yonghui Wu, and Zhifeng Chen. 2019. GPipe: Efficient Training of Giant Neural Networks using Pipeline Parallelism. In NeurIPS. 103\u2013112."},{"key":"e_1_3_2_1_14_1","volume-title":"AI Benchmark: Running Deep Neural Networks on Android Smartphones. In ECCV Workshops (5) (Lecture Notes in Computer Science","volume":"314","author":"Ignatov Andrey","year":"2018","unstructured":"Andrey Ignatov, Radu Timofte, William Chou, Ke Wang, Max Wu, Tim Hartley, and Luc Van Gool. 2018. AI Benchmark: Running Deep Neural Networks on Android Smartphones. In ECCV Workshops (5) (Lecture Notes in Computer Science, Vol. 11133). Springer, 288\u2013314."},{"key":"e_1_3_2_1_15_1","unstructured":"Jared Kaplan Sam McCandlish Tom Henighan Tom B. Brown Benjamin Chess Rewon Child Scott Gray Alec Radford Jeffrey Wu and Dario Amodei. 2020. Scaling Laws for Neural Language Models. arXiv:abs\/2001.08361"},{"key":"e_1_3_2_1_16_1","unstructured":"Rupesh Khendry. 2023. The era of generative AI: Driving transformation in capital markets. https:\/\/www.microsoft.com\/en-us\/industry\/blog\/financial-services\/2023\/07\/10\/the-era-of-generative-ai-driving-transformation-in-capital-markets\/. Accessed: 2024-05-17."},{"key":"e_1_3_2_1_17_1","volume-title":"Jiachen Liu, Xiangfeng Zhu, Harsha V. Madhyastha, and Mosharaf Chowdhury.","author":"Lai Fan","year":"2022","unstructured":"Fan Lai, Yinwei Dai, Sanjay Sri Vallabh Singapuram, Jiachen Liu, Xiangfeng Zhu, Harsha V. Madhyastha, and Mosharaf Chowdhury. 2022. FedScale: Benchmarking Model and System Performance of Federated Learning at Scale. In ICML (Proceedings of Machine Learning Research, Vol. 162). PMLR, 11814\u201311827."},{"key":"e_1_3_2_1_18_1","volume-title":"AMP: Automatically Finding Model Parallel Strategies with Heterogeneity Awareness. In NeurIPS.","author":"Li Dacheng","year":"2022","unstructured":"Dacheng Li, Hongyi Wang, Eric P. Xing, and Hao Zhang. 2022. AMP: Automatically Finding Model Parallel Strategies with Heterogeneity Awareness. In NeurIPS."},{"key":"e_1_3_2_1_19_1","volume-title":"Alexander J. Smola, Amr Ahmed, Vanja Josifovski, James Long, Eugene J. Shekita, and Bor-Yiing Su.","author":"Li Mu","year":"2014","unstructured":"Mu Li, David G. Andersen, Jun Woo Park, Alexander J. Smola, Amr Ahmed, Vanja Josifovski, James Long, Eugene J. Shekita, and Bor-Yiing Su. 2014. Scaling Distributed Machine Learning with the Parameter Server. In OSDI. USENIX Association, 583\u2013598."},{"key":"e_1_3_2_1_20_1","doi-asserted-by":"crossref","unstructured":"Xiangyu Li Yuanchun Li Yuanzhe Li Ting Cao and Yunxin Liu. 2024. FlexNN: Efficient and Adaptive DNN Inference on Memory-Constrained Edge Devices. In MobiCom. ACM 709\u2013723.","DOI":"10.1145\/3636534.3649391"},{"key":"e_1_3_2_1_21_1","volume-title":"Mario: Near Zero-cost Activation Checkpointing in Pipeline Parallelism. In PPoPP. ACM, 197\u2013211.","author":"Liu Weijian","year":"2025","unstructured":"Weijian Liu, Mingzhen Li, Guangming Tan, and Weile Jia. 2025. Mario: Near Zero-cost Activation Checkpointing in Pipeline Parallelism. In PPoPP. ACM, 197\u2013211."},{"key":"e_1_3_2_1_22_1","doi-asserted-by":"publisher","DOI":"10.14778\/3598581.3598604"},{"key":"e_1_3_2_1_23_1","doi-asserted-by":"crossref","unstructured":"Deepak Narayanan Aaron Harlap Amar Phanishayee Vivek Seshadri Nikhil R. Devanur Gregory R. Ganger Phillip B. Gibbons and Matei Zaharia. 2019. PipeDream: generalized pipeline parallelism for DNN training. In SOSP. ACM 1\u201315.","DOI":"10.1145\/3341301.3359646"},{"key":"e_1_3_2_1_24_1","unstructured":"OpenAI. 2023. GPT-4 Technical Report. arXiv:2303.08774"},{"key":"e_1_3_2_1_25_1","volume-title":"LuoShen: A Hyper-Converged Programmable Gateway for Multi-Tenant Multi-Service Edge Clouds","author":"Pan Tian","unstructured":"Tian Pan, Kun Liu, Xionglie Wei, Yisong Qiao, Jun Hu, Zhiguo Li, Jun Liang, Tiesheng Cheng, Wenqiang Su, Jie Lu, Yuke Hong, Zhengzhong Wang, Zhi Xu, Chongjing Dai, Peiqiao Wang, Xuetao Jia, Jianyuan Lu, Enge Song, Jun Zeng, Biao Lyu, Ennan Zhai, Jiao Zhang, Tao Huang, Dennis Cai, and Shunmin Zhu. 2024. LuoShen: A Hyper-Converged Programmable Gateway for Multi-Tenant Multi-Service Edge Clouds. In NSDI. USENIX Association, 877\u2013892."},{"key":"e_1_3_2_1_26_1","volume-title":"HetPipe: Enabling Large DNN Training on (Whimpy) Heterogeneous GPU Clusters through Integration of Pipelined Model Parallelism and Data Parallelism","author":"Park Jay H.","unstructured":"Jay H. Park, Gyeongchan Yun, Chang M. Yi, Nguyen T. Nguyen, Seungmin Lee, Jaesik Choi, Sam H. Noh, and Young-ri Choi. 2020. HetPipe: Enabling Large DNN Training on (Whimpy) Heterogeneous GPU Clusters through Integration of Pipelined Model Parallelism and Data Parallelism. In USENIX ATC. USENIX Association, 307\u2013321."},{"key":"e_1_3_2_1_27_1","doi-asserted-by":"publisher","DOI":"10.1145\/3624719"},{"key":"e_1_3_2_1_28_1","doi-asserted-by":"crossref","unstructured":"Jeff Rasley Samyam Rajbhandari Olatunji Ruwase and Yuxiong He. 2020. DeepSpeed: System Optimizations Enable Training Deep Learning Models with Over 100 Billion Parameters. In KDD. ACM 3505\u20133506.","DOI":"10.1145\/3394486.3406703"},{"key":"e_1_3_2_1_29_1","volume-title":"ZeRO-Offload: Democratizing Billion-Scale Model Training. In USENIX Annual Technical Conference. USENIX Association, 551\u2013564","author":"Ren Jie","year":"2021","unstructured":"Jie Ren, Samyam Rajbhandari, Reza Yazdani Aminabadi, Olatunji Ruwase, Shuangyan Yang, Minjia Zhang, Dong Li, and Yuxiong He. 2021. ZeRO-Offload: Democratizing Billion-Scale Model Training. In USENIX Annual Technical Conference. USENIX Association, 551\u2013564."},{"key":"e_1_3_2_1_30_1","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01042"},{"key":"e_1_3_2_1_31_1","volume-title":"ICML (Proceedings of Machine Learning Research","volume":"29440","author":"Ryabinin Max","year":"2023","unstructured":"Max Ryabinin, Tim Dettmers, Michael Diskin, and Alexander Borzunov. 2023. SWARM Parallelism: Training Large Models Can Be Surprisingly Communication-Efficient. In ICML (Proceedings of Machine Learning Research, Vol. 202). PMLR, 29416\u201329440."},{"key":"e_1_3_2_1_32_1","unstructured":"Max Ryabinin Eduard Gorbunov Vsevolod Plokhotnyuk and Gennady Pekhimenko. 2021. Moshpit SGD: Communication-Efficient Decentralized Training on Heterogeneous Unreliable Devices. In NeurIPS. 18195\u201318211."},{"key":"e_1_3_2_1_33_1","unstructured":"Max Ryabinin and Anton Gusev. 2020. Towards Crowdsourced Training of Large Neural Networks using Decentralized Mixture-of-Experts. In NeurIPS."},{"key":"e_1_3_2_1_34_1","volume-title":"Lane","author":"Sani Lorenzo","year":"2025","unstructured":"Lorenzo Sani, Alex Iacob, Zeyu Cao, Royson Lee, Bill Marino, Yan Gao, Dongqi Cai, Zexi Li, Wanru Zhao, Xinchi Qiu, and Nicholas D. Lane. 2025. Photon: Federated LLM Pre-Training. In MLSys. mlsys.org."},{"key":"e_1_3_2_1_35_1","unstructured":"Mohammad Shoeybi Mostofa Patwary Raul Puri Patrick LeGresley Jared Casper and Bryan Catanzaro. 2019. Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism. arXiv:1909.08053"},{"key":"e_1_3_2_1_36_1","unstructured":"Craig S. Smith. 2023. What Large Models Cost You - There Is No Free AI Lunch. https:\/\/www.forbes.com\/sites\/craigsmith\/2023\/09\/08\/what-large-models-cost-you\u2013there-is-no-free-ai-lunch\/?sh=2b6d10724af7."},{"key":"e_1_3_2_1_37_1","unstructured":"SPEEDTEST. 2025. Speed Test Global Index. https:\/\/www.speedtest. net\/global-index\/united-states. Accessed: 2025-01-27."},{"key":"e_1_3_2_1_38_1","unstructured":"Zhenheng Tang Yuxin Wang Xin He Longteng Zhang Xinglin Pan Qiang Wang Rongfei Zeng Kaiyong Zhao Shaohuai Shi Bingsheng He and Xiaowen Chu. 2023. FusionAI: Decentralized Training and Deploying LLMs with Massive Consumer-Level GPUs. arXiv:2309.01172"},{"key":"e_1_3_2_1_39_1","volume-title":"Seyit Camtepe, and Lichao Sun.","author":"Thapa Chandra","year":"2022","unstructured":"Chandra Thapa, Mahawaga Arachchige Pathum Chamikara, Seyit Camtepe, and Lichao Sun. 2022. SplitFed: When Federated Learning Meets Split Learning. In AAAI. AAAI Press, 8485\u20138493."},{"key":"e_1_3_2_1_40_1","volume-title":"Bamboo: Making Preemptible Instances Resilient for Affordable Training of Large DNNs","author":"Thorpe John","year":"2023","unstructured":"John Thorpe, Pengzhan Zhao, Jonathan Eyolfson, Yifan Qiao, Zhihao Jia, Minjia Zhang, Ravi Netravali, and Guoqing Harry Xu. 2023. Bamboo: Making Preemptible Instances Resilient for Affordable Training of Large DNNs. In NSDI. USENIX Association, 497\u2013513."},{"key":"e_1_3_2_1_41_1","unstructured":"Hugo Touvron Louis Martin Kevin Stone and et al. 2023. Llama 2: Open Foundation and Fine-Tuned Chat Models. arXiv:2307.09288"},{"key":"e_1_3_2_1_42_1","volume-title":"Pietzuch","author":"Wagenl\u00e4nder Marcel","year":"2024","unstructured":"Marcel Wagenl\u00e4nder, Guo Li, Bo Zhao, Luo Mai, and Peter R. Pietzuch. 2024. Tenplex: Dynamic Parallelism for Deep Learning using Parallelizable Tensor Collections. In SOSP. ACM, 195\u2013210."},{"key":"e_1_3_2_1_43_1","doi-asserted-by":"crossref","unstructured":"Duo Wu Xianda Wang Yaqi Qiao Zhi Wang Junchen Jiang Shuguang Cui and Fangxin Wang. 2024. NetLLM: Adapting Large Language Models for Networking. In SIGCOMM. ACM 661\u2013678.","DOI":"10.1145\/3651890.3672268"},{"key":"e_1_3_2_1_44_1","doi-asserted-by":"publisher","DOI":"10.1145\/3757892.3757895"},{"key":"e_1_3_2_1_45_1","volume-title":"Marina","author":"Xue Leyang","year":"2025","unstructured":"Leyang Xue, Meghana Madhyastha, Myungjin Lee, Amos Storkey, Randal Burns, and Mahesh K. Marina. 2025. On Harnessing Idle Compute at the Edge for Foundation Model Training. arXiv:2512.22142"},{"key":"e_1_3_2_1_46_1","volume-title":"Asteroid: Resource-Efficient Hybrid Pipeline Parallelism for Collaborative DNN Training on Heterogeneous Edge Devices. In MobiCom. ACM, 312\u2013326.","author":"Ye Shengyuan","year":"2024","unstructured":"Shengyuan Ye, Liekang Zeng, Xiaowen Chu, Guoliang Xing, and Xu Chen. 2024. Asteroid: Resource-Efficient Hybrid Pipeline Parallelism for Collaborative DNN Training on Heterogeneous Edge Devices. In MobiCom. ACM, 312\u2013326."},{"key":"e_1_3_2_1_47_1","unstructured":"Binhang Yuan Yongjun He Jared Davis Tianyi Zhang Tri Dao Beidi Chen Percy Liang Christopher R\u00e9 and Ce Zhang. 2022. Decentralized Training of Foundation Models in Heterogeneous Environments. In NeurIPS."},{"key":"e_1_3_2_1_48_1","volume-title":"OPT: Open Pre-trained Transformer Language Models. arXiv:2205.01068","author":"Zhang Susan","year":"2022","unstructured":"Susan Zhang, Stephen Roller, Naman Goyal, et al. 2022. OPT: Open Pre-trained Transformer Language Models. arXiv:2205.01068"},{"key":"e_1_3_2_1_49_1","volume-title":"Alpa: Automating Inter- and Intra-Operator Parallelism for Distributed Deep Learning","author":"Zheng Lianmin","year":"2022","unstructured":"Lianmin Zheng, Zhuohan Li, Hao Zhang, Yonghao Zhuang, Zhifeng Chen, Yanping Huang, Yida Wang, Yuanzhong Xu, Danyang Zhuo, Eric P. Xing, Joseph E. Gonzalez, and Ion Stoica. 2022. Alpa: Automating Inter- and Intra-Operator Parallelism for Distributed Deep Learning. In OSDI. USENIX Association, 559\u2013578."}],"event":{"name":"EuroSys '26: 21st European Conference on Computer Systems","location":"Edinburgh Scotland Uk","acronym":"EuroMLSys '26","sponsor":["SIGOPS ACM Special Interest Group on Operating Systems"]},"container-title":["Proceedings of the Sixth European Workshop on Machine Learning and Systems"],"original-title":[],"link":[{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3805621.3807615","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,28]],"date-time":"2026-04-28T13:13:26Z","timestamp":1777382006000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3805621.3807615"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4,27]]},"references-count":49,"alternative-id":["10.1145\/3805621.3807615","10.1145\/3805621"],"URL":"https:\/\/doi.org\/10.1145\/3805621.3807615","relation":{},"subject":[],"published":{"date-parts":[[2026,4,27]]},"assertion":[{"value":"2026-04-28","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}