{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,10]],"date-time":"2026-07-10T02:26:34Z","timestamp":1783650394980,"version":"3.55.0"},"reference-count":43,"publisher":"Association for Computing Machinery (ACM)","issue":"1","license":[{"start":{"date-parts":[[2022,12,22]],"date-time":"2022-12-22T00:00:00Z","timestamp":1671667200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100000266","name":"U.K. Engineering and Physical Sciences Research Council","doi-asserted-by":"crossref","award":["EP\/N509796\/1"],"award-info":[{"award-number":["EP\/N509796\/1"]}],"id":[{"id":"10.13039\/501100000266","id-type":"DOI","asserted-by":"crossref"}]},{"name":"Royal Academy of Engineering\/The Leverhulme Trust Research Fellowship"}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Reconfigurable Technol. Syst."],"published-print":{"date-parts":[[2023,3,31]]},"abstract":"<jats:p>\n            <jats:bold>Long-Short Term Memory (LSTM)<\/jats:bold>\n            networks, and\n            <jats:bold>Recurrent Neural Networks (RNNs)<\/jats:bold>\n            in general, have demonstrated their suitability in many time series data applications, especially in\n            <jats:bold>Natural Language Processing (NLP)<\/jats:bold>\n            . Computationally, LSTMs introduce dependencies on previous outputs in each layer that complicate their computation and the design of custom computing architectures, compared to traditional feed-forward networks. Most neural network acceleration work has focused on optimising the core matrix-vector operations on highly capable FPGAs in server environments. Research that considers the embedded domain has often been unsuitable for streaming inference, relying heavily on batch processing to achieve high throughput. Moreover, many existing accelerator architectures have not focused on fully exploiting the underlying FPGA architecture, resulting in designs that achieve lower operating frequencies than the theoretical maximum. This paper presents a flexible overlay architecture for LSTMs on FPGA SoCs that is built around a streaming dataflow arrangement, uses DSP block capabilities directly, and is tailored to keep parameters within the architecture while moving input data serially to mitigate external memory access overheads. The architecture is designed as an overlay that can be configured to implement alternative models or update model parameters at runtime. It achieves higher operating frequency and demonstrates higher performance than other lightweight LSTM accelerators, as demonstrated in an FPGA SoC implementation.\n          <\/jats:p>","DOI":"10.1145\/3543069","type":"journal-article","created":{"date-parts":[[2022,6,21]],"date-time":"2022-06-21T09:16:12Z","timestamp":1655802972000},"page":"1-26","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":8,"title":["Streaming Overlay Architecture for Lightweight LSTM Computation on FPGA SoCs"],"prefix":"10.1145","volume":"16","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-1033-9951","authenticated-orcid":false,"given":"Lenos","family":"Ioannou","sequence":"first","affiliation":[{"name":"University of Warwick, Coventry, United Kingdom"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0568-5048","authenticated-orcid":false,"given":"Suhaib A.","family":"Fahmy","sequence":"additional","affiliation":[{"name":"King Abdullah University of Science and Technology (KAUST), Thuwal, Saudi Arabia"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2022,12,22]]},"reference":[{"key":"e_1_3_1_2_2","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2016.2587640"},{"key":"e_1_3_1_3_2","doi-asserted-by":"publisher","DOI":"10.1109\/FPL.2019.00043"},{"key":"e_1_3_1_4_2","volume-title":"Advances in Neural Information Processing Systems","author":"Hubara Itay","year":"2016","unstructured":"Itay Hubara, Matthieu Courbariaux, Daniel Soudry, Ran El-Yaniv, and Yoshua Bengio. 2016. Binarized neural networks. In Advances in Neural Information Processing Systems, Vol. 29. Curran Associates, Inc.https:\/\/proceedings.neurips.cc\/paper\/2016\/file\/d8330f857a17c53d217014ee776bfd50-Paper.pdf."},{"key":"e_1_3_1_5_2","doi-asserted-by":"publisher","DOI":"10.1109\/FPL.2018.00024"},{"key":"e_1_3_1_6_2","first-page":"1135","volume-title":"International Conference on Neural Information Processing Systems (NIPS)","author":"Han Song","year":"2015","unstructured":"Song Han, Jeff Pool, John Tran, and William J. Dally. 2015. Learning both weights and connections for efficient neural networks. In International Conference on Neural Information Processing Systems (NIPS). 1135\u20131143."},{"key":"e_1_3_1_7_2","volume-title":"International Conference on Learning Representations (ICLR)","author":"Zhu Chenzhuo","year":"2017","unstructured":"Chenzhuo Zhu, Song Han, Huizi Mao, and William J. Dally. 2017. Trained ternary quantization. In International Conference on Learning Representations (ICLR)."},{"key":"e_1_3_1_8_2","doi-asserted-by":"publisher","DOI":"10.1109\/HCS49909.2020.9220415"},{"key":"e_1_3_1_9_2","doi-asserted-by":"publisher","DOI":"10.1109\/HPEC.2019.8916327"},{"issue":"3","key":"e_1_3_1_10_2","first-page":"56:1\u201356:39","article-title":"Toolflows for mapping convolutional neural networks on FPGAs: A survey and future directions","volume":"51","author":"Venieris Stylianos I.","year":"2018","unstructured":"Stylianos I. Venieris, Alexandros Kouris, and Christos-Savvas Bouganis. 2018. Toolflows for mapping convolutional neural networks on FPGAs: A survey and future directions. ACM Comput. Surv. 51, 3 (2018), 56:1\u201356:39.","journal-title":"ACM Comput. Surv."},{"key":"e_1_3_1_11_2","doi-asserted-by":"publisher","DOI":"10.23919\/FPL.2017.8056794"},{"key":"e_1_3_1_12_2","doi-asserted-by":"publisher","DOI":"10.1109\/FPL.2019.00061"},{"key":"e_1_3_1_13_2","first-page":"199","volume-title":"IEEE International Symposium on Field-Programmable Custom Computing Machines (FCCM)","author":"Nurvitadhi Eriko","year":"2019","unstructured":"Eriko Nurvitadhi, Dongup Kwon, Ali Jafari, Andrew Boutros, Jaewoong Sim, Phillip Tomson, Huseyin Sumbul, Gregory Chen, Phil Knag, Raghavan Kumar, Ram Krishnamurthy, Sergey Gribok, Bogdan Pasca, Martin Langhammer, Debbie Marr, and Aravind Dasu. 2019. Why compete when you can work together: FPGA-ASIC integration for persistent RNNs. In IEEE International Symposium on Field-Programmable Custom Computing Machines (FCCM). 199\u2013207."},{"key":"e_1_3_1_14_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA.2018.00012"},{"key":"e_1_3_1_15_2","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2019.2921977"},{"key":"e_1_3_1_16_2","unstructured":"Hey Siri: An On-device DNN-powered Voice Trigger for Apple\u2019s Personal Assistant. ([n. d.])."},{"key":"e_1_3_1_17_2","doi-asserted-by":"publisher","DOI":"10.23919\/MIPRO.2018.8400205"},{"key":"e_1_3_1_18_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.energy.2018.01.177"},{"key":"e_1_3_1_19_2","doi-asserted-by":"publisher","DOI":"10.1109\/PlatCon.2016.7456805"},{"key":"e_1_3_1_20_2","doi-asserted-by":"publisher","DOI":"10.1109\/SiPS.2016.48"},{"key":"e_1_3_1_21_2","doi-asserted-by":"publisher","DOI":"10.23919\/DATE.2017.7927210"},{"key":"e_1_3_1_22_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCD50377.2020.00086"},{"key":"e_1_3_1_23_2","doi-asserted-by":"publisher","DOI":"10.1007\/s11265-020-01549-8"},{"key":"e_1_3_1_24_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISCAS.2017.8050816"},{"key":"e_1_3_1_25_2","article-title":"Recurrent neural networks hardware implementation on FPGA","volume":"1511","author":"Chang Andre Xian Ming","year":"2015","unstructured":"Andre Xian Ming Chang, Berin Martini, and Eugenio Culurciello. 2015. Recurrent neural networks hardware implementation on FPGA. CoRR abs\/1511.05552 (2015). arXiv:1511.05552. http:\/\/arxiv.org\/abs\/1511.05552.","journal-title":"CoRR"},{"key":"e_1_3_1_26_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCD46524.2019.00014"},{"key":"e_1_3_1_27_2","doi-asserted-by":"publisher","DOI":"10.1109\/BigData47090.2019.9006030"},{"key":"e_1_3_1_28_2","doi-asserted-by":"publisher","DOI":"10.1145\/3366634"},{"key":"e_1_3_1_29_2","first-page":"11","volume-title":"ACM\/SIGDA International Symposium on Field-Programmable Gate Arrays (FPGA)","author":"Wang Shuo","year":"2018","unstructured":"Shuo Wang, Zhe Li, Caiwen Ding, Bo Yuan, Qinru Qiu, Yanzhi Wang, and Yun Liang. 2018. C-LSTM: Enabling efficient LSTM using structured compression techniques on FPGAs. In ACM\/SIGDA International Symposium on Field-Programmable Gate Arrays (FPGA). 11\u201320."},{"key":"e_1_3_1_30_2","article-title":"BRDS: An FPGA-based LSTM accelerator with row-balanced dual-ratio sparsification","volume":"2101","author":"Ghasemzadeh Seyed Abolfazl","year":"2021","unstructured":"Seyed Abolfazl Ghasemzadeh, Erfan Bank Tavakoli, Mehdi Kamal, Ali Afzali-Kusha, and Massoud Pedram. 2021. BRDS: An FPGA-based LSTM accelerator with row-balanced dual-ratio sparsification. CoRR abs\/2101.02667 (2021). arXiv:2101.02667. https:\/\/arxiv.org\/abs\/2101.02667.","journal-title":"CoRR"},{"key":"e_1_3_1_31_2","first-page":"1628","volume-title":"Design, Automation and Test in Europe Conference and Exhibition (DATE)","author":"Jain Abhishek K.","year":"2016","unstructured":"Abhishek K. Jain, Douglas L. Maskell, and Suhaib A. Fahmy. 2016. Throughput oriented FPGA overlays using DSP blocks. In Design, Automation and Test in Europe Conference and Exhibition (DATE). 1628\u20131633."},{"key":"e_1_3_1_32_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICFPT47387.2019.00066"},{"key":"e_1_3_1_33_2","doi-asserted-by":"publisher","DOI":"10.1109\/TVLSI.2020.2987202"},{"key":"e_1_3_1_34_2","volume-title":"UG579 UltraScale Architecture DSP Slice-User Guide","year":"2020","unstructured":"Xilinx. 2020. UG579 UltraScale Architecture DSP Slice-User Guide."},{"key":"e_1_3_1_35_2","volume-title":"7 Series DSP48E1 Slice User Guide (UG479)","year":"2018","unstructured":"2018. 7 Series DSP48E1 Slice User Guide (UG479)."},{"key":"e_1_3_1_36_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2015.2474363"},{"key":"e_1_3_1_37_2","first-page":"125","volume-title":"International Conference on Informatics, Cybernetics, and Computer Engineering (ICCE)","author":"Zhen-zhen Xie","year":"2012","unstructured":"Xie Zhen-zhen and Zhang Su-yu. 2012. A non-linear approximation of the sigmoid function based FPGA. In International Conference on Informatics, Cybernetics, and Computer Engineering (ICCE). 125\u2013132."},{"key":"e_1_3_1_38_2","unstructured":"Mart\u00edn Abadi et\u00a0al. 2015. TensorFlow: Large-Scale Machine Learning on Heterogeneous Systems. (2015). https:\/\/www.tensorflow.org\/."},{"key":"e_1_3_1_39_2","doi-asserted-by":"publisher","DOI":"10.1109\/FPL.2018.00015"},{"key":"e_1_3_1_40_2","doi-asserted-by":"publisher","DOI":"10.5555\/3203489"},{"key":"e_1_3_1_41_2","volume-title":"Zynq-7000 SoC (Z-7007S, Z-7012S, Z-7014S, Z-7010, Z-7015, and Z-7020) (DS187)","year":"2020","unstructured":"2020. Zynq-7000 SoC (Z-7007S, Z-7012S, Z-7014S, Z-7010, Z-7015, and Z-7020) (DS187)."},{"key":"e_1_3_1_42_2","volume-title":"Zynq-7000 SoC (Z-7030, Z-7035, Z-7045, and Z-7100): DC and AC Switching Characteristics (DS191)","year":"2018","unstructured":"2018. Zynq-7000 SoC (Z-7030, Z-7035, Z-7045, and Z-7100): DC and AC Switching Characteristics (DS191)."},{"key":"e_1_3_1_43_2","volume-title":"Virtex-7 T and XT FPGAs Data Sheet: DC and AC Switching Characteristics (DS183)","year":"2021","unstructured":"2021. Virtex-7 T and XT FPGAs Data Sheet: DC and AC Switching Characteristics (DS183)."},{"key":"e_1_3_1_44_2","volume-title":"Zynq Ultrascale+ MPSoC Data Sheet: DC and AC Switching Characteristics (DS925)","year":"2021","unstructured":"2021. Zynq Ultrascale+ MPSoC Data Sheet: DC and AC Switching Characteristics (DS925)."}],"container-title":["ACM Transactions on Reconfigurable Technology and Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3543069","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3543069","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,17]],"date-time":"2025-06-17T17:49:32Z","timestamp":1750182572000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3543069"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,12,22]]},"references-count":43,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2023,3,31]]}},"alternative-id":["10.1145\/3543069"],"URL":"https:\/\/doi.org\/10.1145\/3543069","relation":{},"ISSN":["1936-7406","1936-7414"],"issn-type":[{"value":"1936-7406","type":"print"},{"value":"1936-7414","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,12,22]]},"assertion":[{"value":"2021-10-31","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2022-05-26","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2022-12-22","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}