{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,27]],"date-time":"2026-08-27T15:14:52Z","timestamp":1787843692414,"version":"build-2784847793"},"reference-count":43,"publisher":"Association for Computing Machinery (ACM)","issue":"4","license":[{"start":{"date-parts":[[2025,6,5]],"date-time":"2025-06-05T00:00:00Z","timestamp":1749081600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.acm.org\/publications\/policies\/copyright_policy#Background"}],"funder":[{"DOI":"10.13039\/501100003399","name":"Science and Technology Commission of Shanghai Municipality","doi-asserted-by":"crossref","award":["2021SHZDZX"],"award-info":[{"award-number":["2021SHZDZX"]}],"id":[{"id":"10.13039\/501100003399","id-type":"DOI","asserted-by":"crossref"}]},{"name":"National Key Research and Development Program of China","award":["2021ZD0202202"],"award-info":[{"award-number":["2021ZD0202202"]}]},{"DOI":"10.13039\/501100003399","name":"Science and Technology Commission of Shanghai Municipality","doi-asserted-by":"crossref","award":["2018SHZDZX01"],"award-info":[{"award-number":["2018SHZDZX01"]}],"id":[{"id":"10.13039\/501100003399","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Embed. Comput. Syst."],"published-print":{"date-parts":[[2025,7,31]]},"abstract":"<jats:p>\n            Deep neural networks (DNNs) have proved their great potential over various perceptual and cognitive tasks with the cost of ever-growing storage capacity and computation complexity. Sparse representations in neural networks have emerged as a compelling method to achieve substantial reductions in computational overhead and energy consumption. However, the introduction of sparsity presents challenges such as irregular memory accesses and wasted computation cycles. Traditional methods have attempted to address these challenges with varied success, unfortunately, often involving complex hardware designs or sacrificing sparsity\u2019s benefits. In this article, we propose a software-hardware codesign solution, comprising an offline workload balancing encoding algorithm toward arbitrary sparsity in DNNs and a dedicated Processing Element array\u2013based accelerator with a lightweight switch network. Extensive experiments are conducted to demonstrate the conclusion that our proposal is feasible to support various network structures and a wide range of sparsity ratios. With the encoding algorithm, a 1.16\u00d7\u20132.61\u00d7 acceleration is achieved compared to the baseline. The system-on-chip measures 7.9 mm\n            <jats:sup>2<\/jats:sup>\n            , achieving an energy efficiency of 0.7 TOPS\/W (dense), 2.1 TOPS\/W (at 75% sparsity), and 10.3 TOPS\/W (at 99.9% sparsity) at 0.9 V and 1,066-MHz clock frequency.\n          <\/jats:p>","DOI":"10.1145\/3725532","type":"journal-article","created":{"date-parts":[[2025,3,20]],"date-time":"2025-03-20T11:35:01Z","timestamp":1742470501000},"page":"1-25","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":5,"title":["A Workload-Balance-Aware Accelerator Enabling Dense-to-Arbitrary-Sparse Neural Networks"],"prefix":"10.1145","volume":"24","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-9961-7259","authenticated-orcid":false,"given":"Zihao","family":"Zhao","sequence":"first","affiliation":[{"name":"Fudan University, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9769-5480","authenticated-orcid":false,"given":"Yanhong","family":"Wang","sequence":"additional","affiliation":[{"name":"Fudan University, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-3151-1253","authenticated-orcid":false,"given":"Xu","family":"Jin","sequence":"additional","affiliation":[{"name":"Fudan University, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-4322-6729","authenticated-orcid":false,"given":"Haotian","family":"Zheng","sequence":"additional","affiliation":[{"name":"Fudan University, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3515-7764","authenticated-orcid":false,"given":"Maohua","family":"Nie","sequence":"additional","affiliation":[{"name":"Fudan University, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-3911-245X","authenticated-orcid":false,"given":"Longfei","family":"Gou","sequence":"additional","affiliation":[{"name":"Fudan University, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-5258-6896","authenticated-orcid":false,"given":"Junmin","family":"He","sequence":"additional","affiliation":[{"name":"Fudan University, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9545-1178","authenticated-orcid":false,"given":"Yongchuan","family":"Dong","sequence":"additional","affiliation":[{"name":"Fudan University, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6662-4316","authenticated-orcid":false,"given":"Qiaosha","family":"Zou","sequence":"additional","affiliation":[{"name":"Zhejiang Lab, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2155-303X","authenticated-orcid":false,"given":"Yiyun","family":"Zhang","sequence":"additional","affiliation":[{"name":"Fudan University, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3157-3464","authenticated-orcid":false,"given":"C.-J. Richard","family":"Shi","sequence":"additional","affiliation":[{"name":"Fudan University, Shanghai, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2025,6,5]]},"reference":[{"key":"e_1_3_1_2_2","doi-asserted-by":"publisher","DOI":"10.1145\/3007787.3001138"},{"key":"e_1_3_1_3_2","volume-title":"Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems (NeurIPS\u201920)","author":"Larochelle Hugo","unstructured":"Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel M. Ziegler, Jeffrey Wu, Clemens Winter, Christopher Hesse, Mark Chen, Eric Sigler, Mateusz Litwin, Scott Gray, Benjamin Chess, Jack Clark, Christopher Berner, Sam McCandlish, Alec Radford, Ilya Sutskever, and Dario Amodei. 2020. Language models are few-shot learners. In Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems (NeurIPS\u201920), Hugo Larochelle, Marc\u2019Aurelio Ranzato, Raia Hadsell, Maria-Florina Balcan, and Hsuan-Tien Lin (Eds.). https:\/\/proceedings.neurips.cc\/paper\/2020\/hash\/1457c0d6bfcb4967418bfb8ac142f64a-Abstract.html"},{"key":"e_1_3_1_4_2","doi-asserted-by":"publisher","DOI":"10.1109\/JETCAS.2019.2910232"},{"key":"e_1_3_1_5_2","doi-asserted-by":"publisher","DOI":"10.1109\/JSSC.2016.2616357"},{"key":"e_1_3_1_6_2","doi-asserted-by":"publisher","DOI":"10.1145\/276675.276685"},{"key":"e_1_3_1_7_2","doi-asserted-by":"publisher","DOI":"10.5555\/2969239.2969366"},{"key":"e_1_3_1_8_2","doi-asserted-by":"publisher","DOI":"10.5555\/2969239.2969366"},{"key":"e_1_3_1_9_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"e_1_3_1_10_2","doi-asserted-by":"publisher","DOI":"10.1145\/3392717.3392751"},{"key":"e_1_3_1_11_2","first-page":"241:1\u2013241:124","article-title":"Sparsity in deep learning: Pruning and growth for efficient inference and training in neural networks","volume":"22","author":"Hoefler Torsten","year":"2021","unstructured":"Torsten Hoefler, Dan Alistarh, Tal Ben-Nun, Nikoli Dryden, and Alexandra Peste. 2021. Sparsity in deep learning: Pruning and growth for efficient inference and training in neural networks. J. Mach. Learn. Res. 22 (2021), 241:1\u2013241:124. http:\/\/jmlr.org\/papers\/v22\/21-0366.html","journal-title":"J. Mach. Learn. Res."},{"key":"e_1_3_1_12_2","doi-asserted-by":"publisher","DOI":"10.5555\/3546258.3546499"},{"key":"e_1_3_1_13_2","doi-asserted-by":"publisher","unstructured":"Peng Hu Xi Peng Hongyuan Zhu Mohamed Aly and Jie Lin. 2022. OPQ: Compressing deep neural networks with one-shot pruning-quantization. In Thirty-Fifth AAAI Conference on Artificial Intelligence AAAI 2021 Thirty-Third Conference on Innovative Applications of Artificial Intelligence IAAI 2021 The Eleventh Symposium on Educational Advances in Artificial Intelligence EAAI 2021 Virtual Event February 2-9 2021. AAAI Press 7780\u20137788. DOI:10.1609\/AAAI.V35I9.16950","DOI":"10.1609\/AAAI.V35I9.16950"},{"key":"e_1_3_1_14_2","doi-asserted-by":"publisher","DOI":"10.1145\/3079856.3080246"},{"key":"e_1_3_1_15_2","doi-asserted-by":"publisher","DOI":"10.1109\/JSSC.2020.2992900"},{"key":"e_1_3_1_16_2","doi-asserted-by":"publisher","DOI":"10.1109\/JSSC.2020.2992900"},{"key":"e_1_3_1_17_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCSI.2023.3258411"},{"key":"e_1_3_1_18_2","doi-asserted-by":"publisher","DOI":"10.1145\/3065386"},{"key":"e_1_3_1_19_2","doi-asserted-by":"publisher","DOI":"10.1145\/3297858.3304028"},{"key":"e_1_3_1_20_2","doi-asserted-by":"publisher","DOI":"10.1145\/3316781.3317879"},{"key":"e_1_3_1_21_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCSI.2024.3364093"},{"key":"e_1_3_1_22_2","first-page":"353","volume-title":"Proceedings of the Great Lakes Symposium on VLSI","author":"Liu Shiwei","year":"2021","unstructured":"Shiwei Liu, Zihao Zhao, Yanhong Wang, Qiaosha Zou, Yiyun Zhang, and C.-J. Richard Shi. 2021. Systolic-array deep-learning acceleration exploring pattern-indexed coordinate-assisted sparsity for real-time on-device speech processing. In Proceedings of the Great Lakes Symposium on VLSI. 353\u2013358."},{"key":"e_1_3_1_23_2","doi-asserted-by":"publisher","DOI":"10.1201\/b17320-16"},{"key":"e_1_3_1_24_2","doi-asserted-by":"publisher","DOI":"10.1145\/3466752.3480125"},{"key":"e_1_3_1_25_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPRW.2017.241"},{"key":"e_1_3_1_26_2","doi-asserted-by":"publisher","DOI":"10.1023\/A:1009953814988"},{"key":"e_1_3_1_27_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISSCC.2017.7870353"},{"key":"e_1_3_1_28_2","doi-asserted-by":"publisher","DOI":"10.1145\/3373376.3378534"},{"key":"e_1_3_1_29_2","unstructured":"Aaron van den Oord Sander Dieleman Heiga Zen Karen Simonyan Oriol Vinyals Alex Graves Nal Kalchbrenner Andrew Senior and Koray Kavukcuoglu. 2016. Wavenet: A generative model for raw audio. arXiv preprint arXiv:1609.03499 (2016)."},{"key":"e_1_3_1_30_2","doi-asserted-by":"publisher","DOI":"10.1145\/3079856.3080254"},{"key":"e_1_3_1_31_2","doi-asserted-by":"publisher","DOI":"10.23919\/DATE.2018.8341971"},{"key":"e_1_3_1_32_2","first-page":"13316","volume-title":"Advances in Neural Information Processing Systems","author":"Pool Jeff","year":"2021","unstructured":"Jeff Pool and Chong Yu. 2021. Channel permutations for N:M Sparsity. In Advances in Neural Information Processing Systems, M. Ranzato, A. Beygelzimer, Y. Dauphin, P. S. Liang, and J. Wortman Vaughan (Eds.), Vol. 34. Curran Associates, Inc., 13316\u201313327. https:\/\/proceedings.neurips.cc\/paper_files\/paper\/2021\/file\/6e8404c3b93a9527c8db241a1846599a-Paper.pdf"},{"key":"e_1_3_1_33_2","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA47549.2020.00015"},{"key":"e_1_3_1_34_2","volume-title":"Proceedings of the 3rd International Conference on Learning Representations (ICLR\u201915)","author":"Simonyan Karen","year":"2015","unstructured":"Karen Simonyan and Andrew Zisserman. 2015. Very deep convolutional networks for large-scale image recognition. In Proceedings of the 3rd International Conference on Learning Representations (ICLR\u201915), Yoshua Bengio and Yann LeCun (Eds.). http:\/\/arxiv.org\/abs\/1409.1556"},{"key":"e_1_3_1_35_2","doi-asserted-by":"publisher","DOI":"10.5555\/3327144.3327303"},{"key":"e_1_3_1_36_2","doi-asserted-by":"publisher","DOI":"10.1109\/MMSP.2018.8547084"},{"key":"e_1_3_1_37_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2019.8682804"},{"key":"e_1_3_1_38_2","doi-asserted-by":"publisher","DOI":"10.1109\/JSSC.2023.3269148"},{"key":"e_1_3_1_39_2","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA51647.2021.00018"},{"key":"e_1_3_1_40_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCSI.2022.3184175"},{"key":"e_1_3_1_41_2","doi-asserted-by":"publisher","unstructured":"Yanhong Wang Zihao Zhao Xu Jin Haotian Zheng Maohua Nie Qiaosha Zou and C-J Richard Shi. 2022. AutoMap: Automatic mapping of neural networks to deep learning accelerators for edge devices. IEEE Trans. Comput. Aided Des. Integr. Circuits Syst. 42 9 (2023) 2994\u20133006. DOI:10.1109\/TCAD.2022.3232070","DOI":"10.1109\/TCAD.2022.3232070"},{"key":"e_1_3_1_42_2","doi-asserted-by":"publisher","DOI":"10.1109\/JSSC.2019.2946771"},{"key":"e_1_3_1_43_2","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO.2016.7783723"},{"key":"e_1_3_1_44_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR46437.2021.00360"}],"container-title":["ACM Transactions on Embedded Computing Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3725532","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3725532","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,19]],"date-time":"2025-06-19T01:57:04Z","timestamp":1750298224000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3725532"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,6,5]]},"references-count":43,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2025,7,31]]}},"alternative-id":["10.1145\/3725532"],"URL":"https:\/\/doi.org\/10.1145\/3725532","relation":{},"ISSN":["1539-9087","1558-3465"],"issn-type":[{"value":"1539-9087","type":"print"},{"value":"1558-3465","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,6,5]]},"assertion":[{"value":"2024-08-10","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-03-03","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-06-05","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}