{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,5]],"date-time":"2026-06-05T06:02:56Z","timestamp":1780639376397,"version":"3.54.1"},"reference-count":49,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"7","license":[{"start":{"date-parts":[[2026,7,1]],"date-time":"2026-07-01T00:00:00Z","timestamp":1782864000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2026,7,1]],"date-time":"2026-07-01T00:00:00Z","timestamp":1782864000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,7,1]],"date-time":"2026-07-01T00:00:00Z","timestamp":1782864000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"National Key R&D Program of China","award":["2024YFB4505701"],"award-info":[{"award-number":["2024YFB4505701"]}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62572454"],"award-info":[{"award-number":["62572454"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Parallel Distrib. Syst."],"published-print":{"date-parts":[[2026,7]]},"DOI":"10.1109\/tpds.2026.3691363","type":"journal-article","created":{"date-parts":[[2026,5,8]],"date-time":"2026-05-08T19:41:43Z","timestamp":1778269303000},"page":"1691-1708","source":"Crossref","is-referenced-by-count":0,"title":["nnScaler-M: Constraint-Guided and Placement-Aware Parallelization Plan Generation for Deep Learning Training"],"prefix":"10.1109","volume":"37","author":[{"ORCID":"https:\/\/orcid.org\/0009-0003-2985-9361","authenticated-orcid":false,"given":"Jiaan","family":"Zhu","sequence":"first","affiliation":[{"name":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-2892-4115","authenticated-orcid":false,"given":"Qinghe","family":"Wang","sequence":"additional","affiliation":[{"name":"School of Artificial Intelligence, Anhui University, Hefei, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-9039-8944","authenticated-orcid":false,"given":"Long","family":"Zhao","sequence":"additional","affiliation":[{"name":"School of Artificial Intelligence, Anhui University, Hefei, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-4936-167X","authenticated-orcid":false,"given":"Chizheng","family":"Fang","sequence":"additional","affiliation":[{"name":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3273-7624","authenticated-orcid":false,"given":"Zewen","family":"Jin","sequence":"additional","affiliation":[{"name":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0007-6073-7011","authenticated-orcid":false,"given":"Youhui","family":"Bai","sequence":"additional","affiliation":[{"name":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8050-6196","authenticated-orcid":false,"given":"Zhiqi","family":"Lin","sequence":"additional","affiliation":[{"name":"ByteDance Inc., Hangzhou, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7064-6120","authenticated-orcid":false,"given":"Cheng","family":"Li","sequence":"additional","affiliation":[{"name":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","article-title":"ChatGPT introducing","year":"2022"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01170"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-021-03819-2"},{"key":"ref4","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","year":"2021"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.48550\/arxiv.1811.06965"},{"key":"ref6","first-page":"1","article-title":"Beyond data and model parallelism for deep neural networks","volume-title":"Proc. Mach. Learn. Syst.","author":"Jia","year":"2019"},{"key":"ref7","article-title":"Megatron-LM: Training multi-billion parameter language models using GPU model parallelism","author":"Shoeybi","year":"2019"},{"key":"ref8","first-page":"559","article-title":"ALPA: Automating inter-and intra-operator parallelism for distributed deep learning","volume-title":"Proc. 16th USENIX Symp. Operating Syst. Des. Implementation","author":"Zheng","year":"2022"},{"key":"ref9","first-page":"8026","article-title":"PyTorch: An imperative style, high-performance deep learning library","volume-title":"Proc. 33rd Int. Conf. Neural Inf. Process. Syst.","author":"Paszke","year":"2019"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1145\/3394486.3406703"},{"issue":"140","key":"ref11","first-page":"1","article-title":"Exploring the limits of transfer learning with a unified text-to-text transformer","volume":"21","author":"Raffel","year":"2020","journal-title":"J. Mach. Learn. Res."},{"key":"ref12","article-title":"The Llama 3 herd of models","author":"Grattafiori","year":"2024"},{"key":"ref13","article-title":"Project nnScaler","year":"2024"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1145\/3302424.3303953"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1145\/3437801.3441593"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1145\/3458817.3476145"},{"key":"ref18","article-title":"GSPMD: General and scalable parallelization for Ml computation graphs","author":"Xu","year":"2021"},{"key":"ref19","article-title":"Training deep nets with sublinear memory cost","author":"Chen","year":"2016"},{"key":"ref20","article-title":"Protein complex prediction with alphafold-multimer","author":"Evans","year":"2021","journal-title":"bioRxiv"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1145\/3341301.3359646"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.naacl-main.41"},{"key":"ref23","article-title":"Dynamic tensor rematerialization","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Kirisame","year":"2021"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1145\/3373376.3378530"},{"key":"ref25","first-page":"24829","article-title":"Piper: Multidimensional planner for DNN parallelization","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"34","author":"Tarnawski","year":"2021"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1145\/3627703.3629554"},{"key":"ref27","first-page":"563","article-title":"Metis: Fast automatic distributed training on heterogeneous $\\lbrace${GPUs$\\rbrace$}","volume-title":"Proc. USENIX Annu. Tech. Conf.","author":"Um","year":"2024"},{"key":"ref28","first-page":"463","article-title":"A unified architecture for accelerating distributed DNN training in heterogeneous GPU\/CPU clusters","volume-title":"Proc. 14th USENIX Symp. Operating Syst. Des. Implementation","author":"Jiang","year":"2020"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA57654.2024.00067"},{"key":"ref30","article-title":"TorchFX","year":"2020"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.2307\/j.ctvcm4g18.8"},{"key":"ref32","article-title":"Model hub\u2013natural language processing","year":"2022"},{"key":"ref33","article-title":"Fastfold: Reducing alphafold training time from 11 days to 67 hours","author":"Cheng","year":"2022"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1145\/3617232.3624864"},{"key":"ref35","article-title":"PyTorch Team","year":"2022"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1145\/3341301.3359642"},{"key":"ref37","article-title":"Distributed data parallelism","year":"2020"},{"key":"ref38","first-page":"341","article-title":"Reducing activation recomputation in large transformer models","volume-title":"Proc. Mach. Learn. Syst.","volume":"5","author":"Korthikanti","year":"2023"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1109\/sc41405.2020.00024"},{"key":"ref40","first-page":"6543","article-title":"TeraPipe: Token-level pipeline parallelism for training large-scale language models","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Li","year":"2021"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1145\/2987550.2987586"},{"key":"ref42","article-title":"Horovod: Fast and easy distributed deep learning in tensorflow","author":"Sergeev","year":"2018"},{"key":"ref43","article-title":"NVIDIA NVLINK","year":"2016"},{"key":"ref44","article-title":"NVIDIA collective communications library","year":"2021"},{"key":"ref45","article-title":"The falcon series of open language models","author":"Almazrouei","year":"2023"},{"key":"ref46","first-page":"551","article-title":"Zero-offload: Democratizing billion-scale model training","volume-title":"Proc. USENIX Annu. Tech. Conf.","author":"Ren","year":"2021"},{"key":"ref47","first-page":"11091","article-title":"Longrope: Extending LLM context window beyond 2 million tokens","volume-title":"Proc. 41st Int. Conf. Mach. Learn.","author":"Ding","year":"2024"},{"key":"ref48","first-page":"578","article-title":"TVM: An automated end-to-end optimizing compiler for deep learning","volume-title":"Proc. 13th USENIX Symp. Operating Syst. Des. Implementation","author":"Chen","year":"2018"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.52202\/068431-1189"}],"container-title":["IEEE Transactions on Parallel and Distributed Systems"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/71\/11520459\/11513008.pdf?arnumber=11513008","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,5]],"date-time":"2026-06-05T05:22:17Z","timestamp":1780636937000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11513008\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7]]},"references-count":49,"journal-issue":{"issue":"7"},"URL":"https:\/\/doi.org\/10.1109\/tpds.2026.3691363","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":[[2026,7]]}}}