{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,11]],"date-time":"2026-06-11T16:17:01Z","timestamp":1781194621631,"version":"3.54.1"},"publisher-location":"New York, NY, USA","reference-count":40,"publisher":"ACM","funder":[{"name":"Tsinghua-Toyota Joint Research Fund"},{"name":"Open Research Fund Program of Beijing National Research Center for Information Science and Technology"},{"name":"Beijing Science and Technology Plan Project","award":["Z241100004824002"],"award-info":[{"award-number":["Z241100004824002"]}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["92373103, 62204271"],"award-info":[{"award-number":["92373103, 62204271"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2025,6,21]]},"DOI":"10.1145\/3695053.3731113","type":"proceedings-article","created":{"date-parts":[[2025,6,20]],"date-time":"2025-06-20T16:43:11Z","timestamp":1750437791000},"page":"1806-1818","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["MagiCache: A Virtual In-Cache Computing Engine"],"prefix":"10.1145","author":[{"ORCID":"https:\/\/orcid.org\/0009-0002-9136-1831","authenticated-orcid":false,"given":"Renhao","family":"Fan","sequence":"first","affiliation":[{"name":"Department of Computer Science and Technology, Tsinghua University, Beijing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6283-8405","authenticated-orcid":false,"given":"Yikai","family":"Cui","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Technology, Tsinghua University, Beijing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-6797-5228","authenticated-orcid":false,"given":"Weike","family":"Li","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Technology, Tsinghua University, Beijing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4006-8870","authenticated-orcid":false,"given":"Mingyu","family":"Wang","sequence":"additional","affiliation":[{"name":"School of Microelectronics Science and Technology, Sun Yat-Sen University, Guangzhou, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-1131-3024","authenticated-orcid":false,"given":"Zhaolin","family":"Li","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Technology, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2025,6,20]]},"reference":[{"key":"e_1_3_3_1_2_2","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA.2017.21"},{"key":"e_1_3_3_1_3_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISCAS45731.2020.9181068"},{"key":"e_1_3_3_1_4_2","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA56546.2023.10071074"},{"key":"e_1_3_3_1_5_2","doi-asserted-by":"publisher","DOI":"10.5555\/2371183"},{"key":"e_1_3_3_1_6_2","doi-asserted-by":"crossref","unstructured":"Nathan Binkert Bradford Beckmann Gabriel Black Steven\u00a0K Reinhardt Ali Saidi Arkaprava Basu Joel Hestness Derek\u00a0R Hower Tushar Krishna Somayeh Sardashti Rathijit Sen Korey Sewell Shoaib Muhammad Nilay Vaish Mark\u00a0D Hill and David\u00a0A Wood. 2011. The gem5 simulator. ACM SIGARCH computer architecture news 39 2 (2011) 1\u20137.","DOI":"10.1145\/2024716.2024718"},{"key":"e_1_3_3_1_7_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA45697.2020.00044"},{"key":"e_1_3_3_1_8_2","doi-asserted-by":"publisher","DOI":"10.1109\/IISWC.2009.5306797"},{"key":"e_1_3_3_1_9_2","first-page":"C132\u2013C133","volume-title":"2013 Symposium on VLSI Circuits","author":"Chen Wei","year":"2013","unstructured":"Wei Chen, Szu-Liang Chen, Siufu Chiu, Raghuraman Ganesan, Venkata Lukka, Wei\u00a0Wing Mar, and Stefan Rusu. 2013. A 22nm 2.5 MB slice on-die L3 cache for the next generation Xeon\u00ae processor. In 2013 Symposium on VLSI Circuits. IEEE, C132\u2013C133."},{"key":"e_1_3_3_1_10_2","doi-asserted-by":"publisher","DOI":"10.1145\/3061639.3062326"},{"key":"e_1_3_3_1_11_2","doi-asserted-by":"crossref","unstructured":"Ping Chi Shuangchen Li Cong Xu Tao Zhang Jishen Zhao Yongpan Liu Yu Wang and Yuan Xie. 2016. Prime: A novel processing-in-memory architecture for neural network computation in reram-based main memory. ACM SIGARCH Computer Architecture News 44 3 (2016) 27\u201339.","DOI":"10.1145\/3007787.3001140"},{"key":"e_1_3_3_1_12_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA.2018.00040"},{"key":"e_1_3_3_1_13_2","doi-asserted-by":"crossref","unstructured":"Roger Espasa Federico Ardanaz Joel Emer Stephen Felix Julio Gago Roger Gramunt Isaac Hernandez Toni Juan Geoff Lowney Matthew Mattina and Andre Seznec. 2002. Tarantula: a vector extension to the alpha architecture. ACM SIGARCH Computer Architecture News 30 2 (2002) 281\u2013292.","DOI":"10.1145\/545214.545247"},{"key":"e_1_3_3_1_14_2","doi-asserted-by":"publisher","DOI":"10.1145\/3613424.3614268"},{"key":"e_1_3_3_1_15_2","doi-asserted-by":"crossref","unstructured":"Tim Finkbeiner Glen Hush Troy Larsen Perry Lea John Leidel and Troy Manning. 2017. In-memory intelligence. IEEE Micro 37 4 (2017) 30\u201338.","DOI":"10.1109\/MM.2017.3211117"},{"key":"e_1_3_3_1_16_2","doi-asserted-by":"publisher","DOI":"10.1145\/3307650.3322257"},{"key":"e_1_3_3_1_17_2","doi-asserted-by":"publisher","DOI":"10.1145\/3352460.3358260"},{"key":"e_1_3_3_1_18_2","doi-asserted-by":"publisher","DOI":"10.1145\/2966986.2980098"},{"key":"e_1_3_3_1_19_2","doi-asserted-by":"crossref","unstructured":"Min Huang Moty Mehalel Ramesh Arvapalli and Songnian He. 2013. An energy efficient 32-nm 20-mb shared on-die L3 cache for intel\u00ae xeon\u00ae processor E5 family. IEEE Journal of Solid-State Circuits 48 8 (2013) 1954\u20131962.","DOI":"10.1109\/JSSC.2013.2258815"},{"key":"e_1_3_3_1_20_2","doi-asserted-by":"publisher","DOI":"10.1145\/3307650.3322237"},{"key":"e_1_3_3_1_21_2","doi-asserted-by":"crossref","unstructured":"Supreet Jeloka Naveen\u00a0Bharathwaj Akesh Dennis Sylvester and David Blaauw. 2016. A 28 nm configurable memory (TCAM\/BCAM\/SRAM) using push-rule 6T bit cell enabling logic-in-memory. IEEE Journal of Solid-State Circuits 51 4 (2016) 1009\u20131021.","DOI":"10.1109\/JSSC.2016.2515510"},{"key":"e_1_3_3_1_22_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISCAS.2015.7169194"},{"key":"e_1_3_3_1_23_2","doi-asserted-by":"crossref","unstructured":"Yann LeCun Yoshua Bengio and Geoffrey Hinton. 2015. Deep learning. nature 521 7553 (2015) 436\u2013444.","DOI":"10.1038\/nature14539"},{"key":"e_1_3_3_1_24_2","doi-asserted-by":"publisher","DOI":"10.1145\/3240302.3240312"},{"key":"e_1_3_3_1_25_2","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO.2018.00062"},{"key":"e_1_3_3_1_26_2","doi-asserted-by":"publisher","DOI":"10.1145\/3123939.3123977"},{"key":"e_1_3_3_1_27_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA45697.2020.00073"},{"key":"e_1_3_3_1_28_2","unstructured":"Jason Lowe-Power Abdul\u00a0Mutaal Ahmad Ayaz Akram Mohammad Alian Rico Amslinger Matteo Andreozzi Adri\u00e0 Armejach Nils Asmussen Brad Beckmann Srikant Bharadwaj Gabe Black Gedare Bloom Bobby\u00a0R Bruce Daniel\u00a0Rodrigues Carvalho Jeronimo Castrillon Lizhong Chen Nicolas Derumigny Stephan Diestelhorst Wendy Elsasser Carlos Escuin Marjan Fariborz Amin Farmahini-Farahani Pouya Fotouhi Ryan Gambord Jayneel Gandhi Dibakar Gope Thomas Grass Anthony Gutierrez Bagus Hanindhito Andreas Hansson Swapnil Haria Austin Harris Timothy Hayes Adrian Herrera Matthew Horsnell Syed Ali\u00a0Raza Jafri Radhika Jagtap Hanhwi Jang Reiley Jeyapaul Timothy\u00a0M Jones Matthias Jung Subash Kannoth Hamidreza Khaleghzadeh Yuetsu Kodama Tushar Krishna Tommaso Marinelli Christian Menard Andrea Mondelli Miquel Moreto Tiago M\u00fcck Omar Naji Krishnendra Nathella Hoa Nguyen Nikos Nikoleris Lena\u00a0E Olson Marc Orr Binh Pham Pablo Prieto Trivikram Reddy Alec Roelke Mahyar Samani Andreas Sandberg Javier Setoain Boris Shingarov Matthew\u00a0D Sinclair Tuan Ta Rahul Thakur Giacomo Travaglini Michael Upton Nilay Vaish Ilias Vougioukas Willian Wang Zhengrong Wang Norbert Wehn Christian Weis David\u00a0A Wood Hongil Yoon and \u00c9der\u00a0F Zulian. 2020. The gem5 simulator: Version 20.0+. arXiv preprint arXiv:https:\/\/arXiv.org\/abs\/2007.03152 (2020)."},{"key":"e_1_3_3_1_29_2","volume-title":"Advanced compiler design implementation","author":"Muchnick Steven","year":"1997","unstructured":"Steven Muchnick. 1997. Advanced compiler design implementation. Morgan kaufmann."},{"key":"e_1_3_3_1_30_2","unstructured":"NVIDIA. 2018. Parallel Thread Execution ISA. https:\/\/docs.nvidia.com\/cuda\/parallel-thread-execution\/index.html"},{"key":"e_1_3_3_1_31_2","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA47549.2020.00034"},{"key":"e_1_3_3_1_32_2","doi-asserted-by":"crossref","unstructured":"Crist\u00f3bal Ram\u00edrez C\u00e9sar\u00a0Alejandro Hern\u00e1ndez Oscar Palomar Osman Unsal Marco\u00a0Antonio Ram\u00edrez and Adri\u00e1n Cristal. 2020. A risc-v simulator and benchmark suite for designing and evaluating vector architectures. ACM Transactions on Architecture and Code Optimization (TACO) 17 4 (2020) 1\u201330.","DOI":"10.1145\/3422667"},{"key":"e_1_3_3_1_33_2","unstructured":"RISC-V Foundation. 2021. RISC-V Vector Extension 1.0. https:\/\/github.com\/riscv\/riscv-v-spec\/releases\/tag\/v1.0"},{"key":"e_1_3_3_1_34_2","doi-asserted-by":"publisher","DOI":"10.1145\/3123939.3124544"},{"key":"e_1_3_3_1_35_2","doi-asserted-by":"crossref","unstructured":"Ali Shafiee Anirban Nag Naveen Muralimanohar Rajeev Balasubramonian John\u00a0Paul Strachan Miao Hu R\u00a0Stanley Williams and Vivek Srikumar. 2016. ISAAC: A convolutional neural network accelerator with in-situ analog arithmetic in crossbars. ACM SIGARCH Computer Architecture News 44 3 (2016) 14\u201326.","DOI":"10.1145\/3007787.3001139"},{"key":"e_1_3_3_1_36_2","doi-asserted-by":"crossref","unstructured":"William\u00a0Andrew Simon Yasir\u00a0Mahmood Qureshi Marco Rios Alexandre Levisse Marina Zapater and David Atienza. 2020. BLADE: An in-cache computing architecture for edge devices. IEEE Trans. Comput. 69 9 (2020) 1349\u20131363.","DOI":"10.1109\/TC.2020.2972528"},{"key":"e_1_3_3_1_37_2","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA.2017.55"},{"key":"e_1_3_3_1_38_2","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA.2018.00052"},{"key":"e_1_3_3_1_39_2","doi-asserted-by":"crossref","unstructured":"Nigel Stephens Stuart Biles Matthias Boettcher Jacob Eapen Mbou Eyole Giacomo Gabrielli Matt Horsnell Grigorios Magklis Alejandro Martinez Nathanael Premillieu Alastair Reid Alejandro Rico Pau Walker and ARM. 2017. The ARM scalable vector extension. IEEE micro 37 2 (2017) 26\u201339.","DOI":"10.1109\/MM.2017.35"},{"key":"e_1_3_3_1_40_2","doi-asserted-by":"crossref","unstructured":"Jingcheng Wang Xiaowei Wang Charles Eckert Arun Subramaniyan Reetuparna Das David Blaauw and Dennis Sylvester. 2019. A 28-nm compute SRAM with bit-serial logic\/arithmetic operations for programmable in-memory vector computing. IEEE Journal of Solid-State Circuits 55 1 (2019) 76\u201386.","DOI":"10.1109\/JSSC.2019.2939682"},{"key":"e_1_3_3_1_41_2","doi-asserted-by":"publisher","DOI":"10.1145\/3316781.3317900"}],"event":{"name":"ISCA '25: Proceedings of the 52nd Annual International Symposium on Computer Architecture","location":"Tokyo Japan","acronym":"SIGARCH '25","sponsor":["SIGARCH ACM Special Interest Group on Computer Architecture"]},"container-title":["Proceedings of the 52nd Annual International Symposium on Computer Architecture"],"original-title":[],"link":[{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3695053.3731113","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,21]],"date-time":"2025-06-21T11:09:17Z","timestamp":1750504157000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3695053.3731113"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,6,20]]},"references-count":40,"alternative-id":["10.1145\/3695053.3731113","10.1145\/3695053"],"URL":"https:\/\/doi.org\/10.1145\/3695053.3731113","relation":{},"subject":[],"published":{"date-parts":[[2025,6,20]]},"assertion":[{"value":"2025-06-20","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}