{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,24]],"date-time":"2026-07-24T13:06:34Z","timestamp":1784898394359,"version":"3.55.0"},"reference-count":56,"publisher":"Association for Computing Machinery (ACM)","issue":"3","license":[{"start":{"date-parts":[[2026,7,24]],"date-time":"2026-07-24T00:00:00Z","timestamp":1784851200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/legalcode"}],"funder":[{"name":"Listen2Future is co-funded by the European Union","award":["101096884"],"award-info":[{"award-number":["101096884"]}]}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Archit. Code Optim."],"published-print":{"date-parts":[[2026,9,30]]},"abstract":"<jats:p>\n                    Custom instruction (CI) identification is a key technique for tailoring application-specific instruction-set processors (ASIPs) to demanding embedded workloads, yet classical two-stage flows that enumerate candidate instructions and then select among them do not scale as data-dependence graph (DDG) sizes and I\/O interface bounds grow. This article proposes\n                    <jats:sc>RL-GNN-CI<\/jats:sc>\n                    , a reinforcement-learning framework that operates directly on compiler-generated DDGs to identify profitable, convex CIs under joint I\/O and area constraints. A graph neural network encoder, shared across all candidate instructions, conditions node and graph embeddings on the evolving CI state, while a policy head incrementally grows and commits CI candidates on each graph. We instantiate this framework with proximal policy optimization (PPO) and deep Q-learning (DQN), training on synthetic DDGs that match real-workload statistics. A controlled ablation study over action masking, reward shaping, and gain normalization identifies strict masking with either a simplified or absolute-gain reward as the most effective configurations. On DDGs extracted from MiBench, MediaBench, and JM\u00a0H.264 kernels under 240 benchmark\u2013constraint scenarios and a common latency-and-area model, the best\n                    <jats:sc>RL-GNN-CI<\/jats:sc>\n                    variants exceed an integer-linear-programming baseline by 5\u20137% and a tabular SARSA baseline by 18\u201320% in speedup, while using about 18% less CI area on average.\n                  <\/jats:p>","DOI":"10.1145\/3828552","type":"journal-article","created":{"date-parts":[[2026,7,1]],"date-time":"2026-07-01T20:41:17Z","timestamp":1782938477000},"page":"1-25","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Reinforcement Learning on Data-Dependence Graphs for Custom Instruction Identification"],"prefix":"10.1145","volume":"23","author":[{"ORCID":"https:\/\/orcid.org\/0009-0003-8345-6723","authenticated-orcid":false,"given":"Eslam","family":"Hussein","sequence":"first","affiliation":[{"name":"Dresden University of Technology","place":["Dresden, Germany"]},{"name":"Infineon Technologies Dresden GmbH","place":["Dresden, Germany"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0294-8594","authenticated-orcid":false,"given":"Bernd","family":"Waschneck","sequence":"additional","affiliation":[{"name":"Infineon Technologies AG","place":["Neubiberg, Germany"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3502-0872","authenticated-orcid":false,"given":"Christian","family":"Mayr","sequence":"additional","affiliation":[{"name":"Dresden University of Technology","place":["Dresden, Germany"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2026,7,24]]},"reference":[{"key":"e_1_3_1_2_2","doi-asserted-by":"publisher","unstructured":"Kubilay Atasu G\u00fcnhan D\u00fcndar and Can \u00d6zturan. 2005. An integer linear programming approach for identifying instruction-set extensions. In Proceedings of the 3rd IEEE\/ACM\/IFIP International Conference on Hardware\/software Codesign and System Synthesis. ACM New York NY USA 172\u2013177. DOI:10.1145\/1084834.1084880","DOI":"10.1145\/1084834.1084880"},{"key":"e_1_3_1_3_2","doi-asserted-by":"publisher","DOI":"10.1109\/TVLSI.2010.2090543"},{"key":"e_1_3_1_4_2","doi-asserted-by":"publisher","unstructured":"Kubilay Atasu Oskar Mencer Wayne Luk Can \u00d6zturan and G\u00fcnhan D\u00fcndar. 2008. Fast custom instruction identification by convex subgraph enumeration. In Proceedings of the 2008 International Conference on Application-Specific Systems Architectures and Processors. IEEE Piscataway NJ USA 1\u20136. DOI:10.1109\/ASAP.2008.4580145","DOI":"10.1109\/ASAP.2008.4580145"},{"key":"e_1_3_1_5_2","first-page":"449","volume-title":"Proceedings of the International Conference on Machine Learning","author":"Bellemare Marc G.","year":"2017","unstructured":"Marc G. Bellemare, Will Dabney, and R\u00e9mi Munos. 2017. A distributional perspective on reinforcement learning. In Proceedings of the International Conference on Machine Learning. PMLR, Sydney, Australia, 449\u2013458."},{"key":"e_1_3_1_6_2","first-page":"15","volume-title":"Proc. Int. Conf. on Learning Representations (ICLR) Workshop","author":"Bello Irwan","year":"2017","unstructured":"Irwan Bello, Hieu Pham, Quoc V. Le, Mohammad Norouzi, and Samy Bengio. 2017. Neural combinatorial optimization with reinforcement learning. In Proc. Int. Conf. on Learning Representations (ICLR) Workshop. OpenReview.net, Toulon, France, 15 pages. Retrieved from https:\/\/openreview.net\/forum?id=rJY3vK9eg"},{"key":"e_1_3_1_7_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejor.2020.07.063"},{"key":"e_1_3_1_8_2","doi-asserted-by":"publisher","unstructured":"Yoshua Bengio J\u00e9r\u00f4me Louradour Ronan Collobert and Jason Weston. 2009. Curriculum learning. In Proceedings of the 26th Annual International Conference on Machine Learning. ACM New York NY USA 41\u201348. DOI:10.1145\/1553374.1553380","DOI":"10.1145\/1553374.1553380"},{"key":"e_1_3_1_9_2","doi-asserted-by":"publisher","unstructured":"Paolo Bonzini and Laura Pozzi. 2006. Code transformation strategies for extensible embedded processors. In Proceedings of the 2006 International Conference on Compilers Architecture and Synthesis for Embedded Systems. ACM New York NY USA 242\u2013252. DOI:10.1145\/1176760.1176791","DOI":"10.1145\/1176760.1176791"},{"key":"e_1_3_1_10_2","doi-asserted-by":"publisher","DOI":"10.1109\/DATE.2007.364482"},{"key":"e_1_3_1_11_2","doi-asserted-by":"publisher","DOI":"10.1109\/TVLSI.2008.2001863"},{"key":"e_1_3_1_12_2","first-page":"26","volume-title":"Proc. Int. Conf. on Learning Representations (ICLR)","author":"Brody Shaked","year":"2022","unstructured":"Shaked Brody, Uri Alon, and Eran Yahav. 2022. How attentive are graph attention networks?. In Proc. Int. Conf. on Learning Representations (ICLR). OpenReview.net, Virtual, 26 pages. Retrieved from https:\/\/openreview.net\/forum?id=F72ximsx7C1"},{"key":"e_1_3_1_13_2","unstructured":"Cadence Design Systems. 2023. Tensilica Xtensa Controllers and Extensible Processors. Retrieved December 2025 from https:\/\/www.cadence.com\/en_US\/home\/tools\/silicon-solutions\/compute-ip\/tensilica-xtensa-controllers-and-extensible-processors.html"},{"key":"e_1_3_1_14_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA.2005.9"},{"key":"e_1_3_1_15_2","unstructured":"Codasip. 2023. Codasip RISC-V Processors. Retrieved December 2025 from https:\/\/codasip.com\/products\/codasip-risc-v-processors\/"},{"key":"e_1_3_1_16_2","volume-title":"Statistical Power Analysis for the Behavioral Sciences (2nd. ed.)","author":"Cohen Jacob","year":"1988","unstructured":"Jacob Cohen. 1988. Statistical Power Analysis for the Behavioral Sciences (2nd. ed.). Lawrence Erlbaum Associates, Hillsdale, NJ, USA."},{"key":"e_1_3_1_17_2","doi-asserted-by":"publisher","unstructured":"Jason Cong Yiping Fan Guoling Han Ashok Jagannathan Glenn Reinman and Zhiru Zhang. 2005. Instruction set extension with shadow registers for configurable processors. In Proceedings of the 2005 ACM\/SIGDA 13th International Symposium on Field-programmable Gate Arrays. ACM New York NY USA 99\u2013106. DOI:10.1145\/1046192.1046206","DOI":"10.1145\/1046192.1046206"},{"key":"e_1_3_1_18_2","doi-asserted-by":"publisher","DOI":"10.1145\/968280.968307"},{"key":"e_1_3_1_19_2","first-page":"3061","volume-title":"Proceedings of the International Conference on Machine Learning.","author":"Fedus William","year":"2020","unstructured":"William Fedus, Prajit Ramachandran, Rishabh Agarwal, Yoshua Bengio, Hugo Larochelle, Mark Rowland, and Will Dabney. 2020. Revisiting fundamentals of experience replay. In Proceedings of the International Conference on Machine Learning.PMLR, Virtual, 3061\u20133071. Retrieved from https:\/\/proceedings.mlr.press\/v119\/fedus20a.html"},{"key":"e_1_3_1_20_2","first-page":"21","volume-title":"Proceedings of the Int. Conf. on Learning Representations","author":"Fortunato Meire","year":"2018","unstructured":"Meire Fortunato, Mohammad Gheshlaghi Azar, Bilal Piot, Jacob Menick, Ian Osband, Alex Graves, Vlad Mnih, Remi Munos, Demis Hassabis, Olivier Pietquin, et\u00a0al. 2018. Noisy networks for exploration. In Proceedings of the Int. Conf. on Learning Representations. OpenReview.net, Vancouver, Canada, 21 pages. Retrieved from https:\/\/openreview.net\/forum?id=rywHCPkAW"},{"key":"e_1_3_1_21_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.micpro.2009.02.010"},{"key":"e_1_3_1_22_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2014.2387375"},{"key":"e_1_3_1_23_2","doi-asserted-by":"publisher","DOI":"10.1109\/WWC.2001.990739"},{"key":"e_1_3_1_24_2","unstructured":"Andreas Hager-Clukas Philipp van Kempen and Stefan Wallentowitz. 2025. ARISE: Automating RISC-V Instruction Set Extension. arxiv:2508.07725. Retrieved from https:\/\/arxiv.org\/abs\/2508.07725"},{"key":"e_1_3_1_25_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.sysarc.2024.103080"},{"key":"e_1_3_1_26_2","unstructured":"IBM. 2023. IBM ILOG CPLEX Optimization Studio. Retrieved October 2025 from https:\/\/www.ibm.com\/products\/ilog-cplex-optimization-studio"},{"key":"e_1_3_1_27_2","doi-asserted-by":"publisher","DOI":"10.1145\/1146909.1146924"},{"key":"e_1_3_1_28_2","doi-asserted-by":"publisher","DOI":"10.2466\/11.IT.3.1"},{"key":"e_1_3_1_29_2","volume-title":"Proceedings of the Advances in Neural Information Processing Systems","volume":"30","author":"Khalil Elias","year":"2017","unstructured":"Elias Khalil, Hanjun Dai, Yuyu Zhang, Bistra Dilkina, and Le Song. 2017. Learning combinatorial optimization algorithms over graphs. In Proceedings of the Advances in Neural Information Processing Systems, I. Guyon, U. Von Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett (Eds.). Vol. 30, Curran Associates, Inc., Red Hook, NY, USA. Retrieved from https:\/\/proceedings.neurips.cc\/paper_files\/paper\/2017\/file\/d9896106ca98d3d05b8cbdf4fd8b13a1-Paper.pdf"},{"key":"e_1_3_1_30_2","first-page":"14","volume-title":"Proceedings of the Int. Conf. on Learning Representations","author":"Kipf Thomas N.","year":"2017","unstructured":"Thomas N. Kipf and Max Welling. 2017. Semi-supervised classification with graph convolutional networks. In Proceedings of the Int. Conf. on Learning Representations. OpenReview.net, Toulon, France, 14 pages. Retrieved from https:\/\/openreview.net\/forum?id=SJU4ayYgl"},{"key":"e_1_3_1_31_2","first-page":"17","volume-title":"Proceedings of the Int. Conf. on Learning Representations","author":"Kool Wouter","year":"2019","unstructured":"Wouter Kool, Herke van Hoof, and Max Welling. 2019. Attention, learn to solve routing problems!. In Proceedings of the Int. Conf. on Learning Representations. OpenReview.net, New Orleans, LA, USA, 17 pages. Retrieved from https:\/\/openreview.net\/forum?id=ByxBFsRqYm"},{"key":"e_1_3_1_32_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.sysarc.2010.04.004"},{"key":"e_1_3_1_33_2","first-page":"3053","volume-title":"Proceedings of the Int. Conf. on Machine Learning","author":"Liang Eric","year":"2018","unstructured":"Eric Liang, Richard Liaw, Robert Nishihara, Philipp Moritz, Roy Fox, Ken Goldberg, Joseph Gonzalez, Michael I. Jordan, and Ion Stoica. 2018. RLlib: Abstractions for distributed reinforcement learning. In Proceedings of the Int. Conf. on Machine Learning. PMLR, Stockholm, Sweden, 3053\u20133062. Retrieved from https:\/\/proceedings.mlr.press\/v80\/liang18b.html"},{"key":"e_1_3_1_34_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.cor.2021.105400"},{"key":"e_1_3_1_35_2","first-page":"9","volume-title":"Proceedings of the NIPS Deep Learning Workshop","author":"Mnih Volodymyr","year":"2013","unstructured":"Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Alex Graves, Ioannis Antonoglou, Daan Wierstra, and Martin Riedmiller. 2013. Playing Atari with deep reinforcement learning. In Proceedings of the NIPS Deep Learning Workshop. NIPS Foundation, Lake Tahoe, NV, USA, 9 pages."},{"key":"e_1_3_1_36_2","doi-asserted-by":"publisher","DOI":"10.1038\/nature14236"},{"key":"e_1_3_1_37_2","first-page":"278","volume-title":"Proceedings of the 16th Int. Conf. on Machine Learning.","author":"Ng Andrew Y.","year":"1999","unstructured":"Andrew Y. Ng, Daishi Harada, and Stuart Russell. 1999. Policy invariance under reward transformations: Theory and application to reward shaping. In Proceedings of the 16th Int. Conf. on Machine Learning.Morgan Kaufmann, San Francisco, CA, USA, 278\u2013287."},{"key":"e_1_3_1_38_2","doi-asserted-by":"publisher","DOI":"10.1109\/VLSID.2007.40"},{"key":"e_1_3_1_39_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2005.855950"},{"key":"e_1_3_1_40_2","doi-asserted-by":"publisher","DOI":"10.1145\/1086297.1086300"},{"key":"e_1_3_1_41_2","doi-asserted-by":"publisher","DOI":"10.1109\/TVLSI.2011.2173221"},{"key":"e_1_3_1_42_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCAD66269.2025.11240781"},{"key":"e_1_3_1_43_2","volume-title":"RISC-V Bit-Manipulation ISA Extensions (Zba, Zbb, Zbc, Zbs)","author":"International RISC-V","year":"2021","unstructured":"RISC-V International. 2021. RISC-V Bit-Manipulation ISA Extensions (Zba, Zbb, Zbc, Zbs). Technical Report. RISC-V International. Retrieved February 2026 from https:\/\/docs.riscv.org\/reference\/isa\/unpriv\/b-st-ext.html"},{"key":"e_1_3_1_44_2","volume-title":"RISC-V Cryptography Extensions Volume I: Scalar and Entropy Source Instructions","author":"International RISC-V","year":"2023","unstructured":"RISC-V International. 2023. RISC-V Cryptography Extensions Volume I: Scalar and Entropy Source Instructions. Technical Report. RISC-V International. Retrieved February 2026 from https:\/\/docs.riscv.org\/reference\/isa\/unpriv\/scalar-crypto.html"},{"key":"e_1_3_1_45_2","unstructured":"John Schulman Filip Wolski Prafulla Dhariwal Alec Radford and Oleg Klimov. 2017. Proximal policy optimization algorithms. arxiv:1707.06347. Retrieved from https:\/\/arxiv.org\/abs\/1707.06347"},{"key":"e_1_3_1_46_2","unstructured":"Synopsys Inc.2023. Synopsys ARC Processors for Embedded Applications. Retrieved December 2025 from https:\/\/www.synopsys.com\/designware-ip\/processor-solutions\/arc-processors.html"},{"key":"e_1_3_1_47_2","doi-asserted-by":"publisher","DOI":"10.1109\/NorCAS66540.2025.11231278"},{"key":"e_1_3_1_48_2","first-page":"12","volume-title":"Proceedings of the Int. Conf. on Learning Representations","author":"Veli\u010dkovi\u0107 Petar","year":"2018","unstructured":"Petar Veli\u010dkovi\u0107, Guillem Cucurull, Arantxa Casanova, Adriana Romero, Pietro Li\u00f2, and Yoshua Bengio. 2018. Graph attention networks. In Proceedings of the Int. Conf. on Learning Representations. OpenReview.net, Vancouver, Canada, 12 pages. Retrieved from https:\/\/openreview.net\/forum?id=rJXMpikCZ"},{"key":"e_1_3_1_49_2","unstructured":"Minjie Wang Da Zheng Zihao Ye Quan Gan Mufei Li Xiang Song Jinjing Zhou Chao Ma Lingfan Yu Yu Gai et\u00a0al. 2020. Deep Graph Library: A Graph-Centric Highly-Performant Package for Graph Neural Networks. arxiv:1909.01315 [cs.LG]."},{"key":"e_1_3_1_50_2","doi-asserted-by":"publisher","DOI":"10.1109\/TAI.2023.3308099"},{"key":"e_1_3_1_51_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.micpro.2016.05.001"},{"key":"e_1_3_1_52_2","doi-asserted-by":"publisher","DOI":"10.1109\/DASIP.2015.7367250"},{"key":"e_1_3_1_53_2","first-page":"1995","volume-title":"Proceedings of the International Conference on Machine Learning","author":"Wang Ziyu","year":"2016","unstructured":"Ziyu Wang, Tom Schaul, Matteo Hessel, Hado van Hasselt, Marc Lanctot, and Nando de Freitas. 2016. Dueling network architectures for deep reinforcement learning. In Proceedings of the International Conference on Machine Learning. PMLR, New York, NY, USA, 1995\u20132003. Retrieved from https:\/\/proceedings.mlr.press\/v48\/wangf16.html"},{"key":"e_1_3_1_54_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCSVT.2003.815165"},{"key":"e_1_3_1_55_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.micpro.2014.09.001"},{"key":"e_1_3_1_56_2","first-page":"17","volume-title":"Proc. Int. Conf. on Learning Representations","author":"Xu Keyulu","year":"2019","unstructured":"Keyulu Xu, Weihua Hu, Jure Leskovec, and Stefanie Jegelka. 2019. How powerful are graph neural networks?. In Proc. Int. Conf. on Learning Representations. OpenReview.net, New Orleans, LA, USA, 17 pages. Retrieved from https:\/\/openreview.net\/forum?id=ryGs6iA5Km"},{"key":"e_1_3_1_57_2","doi-asserted-by":"publisher","DOI":"10.1145\/1023833.1023844"}],"container-title":["ACM Transactions on Architecture and Code Optimization"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3828552","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,24]],"date-time":"2026-07-24T12:57:19Z","timestamp":1784897839000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3828552"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,24]]},"references-count":56,"journal-issue":{"issue":"3","published-print":{"date-parts":[[2026,9,30]]}},"alternative-id":["10.1145\/3828552"],"URL":"https:\/\/doi.org\/10.1145\/3828552","relation":{},"ISSN":["1544-3566","1544-3973"],"issn-type":[{"value":"1544-3566","type":"print"},{"value":"1544-3973","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,24]]},"assertion":[{"value":"2025-12-11","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-06-21","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-07-24","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}