{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,7]],"date-time":"2026-03-07T07:27:26Z","timestamp":1772868446746,"version":"3.50.1"},"reference-count":36,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"12","license":[{"start":{"date-parts":[[2025,12,1]],"date-time":"2025-12-01T00:00:00Z","timestamp":1764547200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2025,12,1]],"date-time":"2025-12-01T00:00:00Z","timestamp":1764547200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,12,1]],"date-time":"2025-12-01T00:00:00Z","timestamp":1764547200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"Guangxi Key Research and Development Project","award":["2024AB02018"],"award-info":[{"award-number":["2024AB02018"]}]},{"name":"Shandong Provincial Natural Science Foundation Project","award":["ZR2024LZH012"],"award-info":[{"award-number":["ZR2024LZH012"]}]},{"name":"Major Key Project of PCL","award":["PCL2025AS11"],"award-info":[{"award-number":["PCL2025AS11"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Parallel Distrib. Syst."],"published-print":{"date-parts":[[2025,12]]},"DOI":"10.1109\/tpds.2025.3626153","type":"journal-article","created":{"date-parts":[[2025,10,28]],"date-time":"2025-10-28T17:38:05Z","timestamp":1761673085000},"page":"2789-2800","source":"Crossref","is-referenced-by-count":1,"title":["MCG-Sched: Multi-Cluster GPU Scheduling for Resource Fragmentation Reduction and Load Balancing"],"prefix":"10.1109","volume":"36","author":[{"given":"Haijie","family":"Wu","sequence":"first","affiliation":[{"name":"School of Computer Science and Engineering, South China University of Technology, Guangzhou, China"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7346-0823","authenticated-orcid":false,"given":"Xinhua","family":"Wang","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, South China University of Technology, Guangzhou, China"}]},{"given":"Xiaoxuan","family":"Luo","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, South China University of Technology, Guangzhou, China"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3761-843X","authenticated-orcid":false,"given":"Wangbo","family":"Shen","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, South China University of Technology, Guangzhou, China"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6876-1795","authenticated-orcid":false,"given":"Weiwei","family":"Lin","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, South China University of Technology, Guangzhou, China"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1007\/s10462-023-10631-z"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1016\/j.cosrev.2024.100645"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1016\/j.eswa.2024.123740"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1145\/3597503.3639232"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1145\/3627703.3629578"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/TPDS.2021.3136245"},{"key":"ref7","first-page":"995","article-title":"Beware of fragmentation: Scheduling GPU-sharing workloads with fragmentation gradient descent","volume-title":"Proc. USENIX Annu. Tech. Conf.","author":"Weng"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1145\/3575693.3575705"},{"key":"ref9","first-page":"69","article-title":"Transparent gpu sharing in container clouds for deep learning workloads","volume-title":"Proc. 20th USENIX Symp. Netw. Syst. Des. Implementation","author":"Wu"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1145\/3710848.3710863"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1145\/3552326.3587445"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1145\/3575693.3575721"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1145\/3698038.3698532"},{"key":"ref14","first-page":"579","article-title":"Looking beyond gpus for dnn scheduling on multi-tenant clusters","volume-title":"Proc. 16th USENIX Symp. Operating Syst. Des. Implementation","author":"Mohan"},{"key":"ref15","first-page":"1","article-title":"Characterization and prediction of deep learning workloads in large-scale GPU datacenters","volume-title":"Proc. Int. Conf. High Perform. Comput., Netw., Storage Anal.","author":"Hu"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1016\/j.parco.2024.103082"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1145\/3651890.3672239"},{"key":"ref18","first-page":"563","article-title":"Mast: Global scheduling of ML training across GEO-distributed datacenters at hyperscale","volume-title":"Proc. 18th USENIX Symp. Operating Syst. Des. Implementation","author":"Choudhury"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/COMST.2023.3329027"},{"key":"ref20","article-title":"Improving training time and GPU utilization in GEO-distributed language model training","author":"Gandhi","year":"2024"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.26599\/BDMA.2025.9020031"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1016\/j.future.2024.107561"},{"key":"ref23","first-page":"1285","article-title":"Load is not what you should balance: Introducing prequal","volume-title":"Proc. 21st USENIX Symp. Netw. Syst. Des. Implementation","author":"Wydrowski"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1016\/j.eswa.2023.121450"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1007\/s10462-024-10925-w"},{"key":"ref26","first-page":"504","volume-title":"Proc. 53 rd Int. Conf. Parallel Process.","author":"Zhang"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1145\/3605573.3605638"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1145\/3711927"},{"key":"ref29","article-title":"Topology-aware preemptive scheduling for co-located LLM workloads","author":"Zhang","year":"2024"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/TPDS.2021.3138825"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1109\/TNET.2022.3202529"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1109\/TNET.2022.3190797"},{"key":"ref33","article-title":"Singularity: Planet-scale, preemptive and elastic scheduling of ai workloads","author":"Shukla","year":"2022"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1109\/TSC.2024.3442544"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1109\/TSC.2020.3044104"},{"key":"ref36","first-page":"947","article-title":"Analysis of large-scale multi-tenant GPU clusters for DNN training workloads","volume-title":"Proc. 2019 USENIX Annu. Tech. Conf.","author":"Jeon"}],"container-title":["IEEE Transactions on Parallel and Distributed Systems"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/71\/11205816\/11218789.pdf?arnumber=11218789","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,11,20]],"date-time":"2025-11-20T06:05:09Z","timestamp":1763618709000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11218789\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12]]},"references-count":36,"journal-issue":{"issue":"12"},"URL":"https:\/\/doi.org\/10.1109\/tpds.2025.3626153","relation":{},"ISSN":["1045-9219","1558-2183","2161-9883"],"issn-type":[{"value":"1045-9219","type":"print"},{"value":"1558-2183","type":"electronic"},{"value":"2161-9883","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,12]]}}}