{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,6,18]],"date-time":"2025-06-18T04:19:14Z","timestamp":1750220354593,"version":"3.41.0"},"reference-count":92,"publisher":"Association for Computing Machinery (ACM)","issue":"2","license":[{"start":{"date-parts":[[2022,1,12]],"date-time":"2022-01-12T00:00:00Z","timestamp":1641945600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"CRISP"},{"name":"SRC-Global Research Collaboration"},{"DOI":"10.13039\/100000006","name":"Office of Naval Research","doi-asserted-by":"crossref","award":["N00014-21-1-2225"],"award-info":[{"award-number":["N00014-21-1-2225"]}],"id":[{"id":"10.13039\/100000006","id-type":"DOI","asserted-by":"crossref"}]},{"name":"NSF","award":["# 1527034, # 1730158, # 1826967, # 1911095, and # 2127780"],"award-info":[{"award-number":["# 1527034, # 1730158, # 1826967, # 1911095, and # 2127780"]}]}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["J. Emerg. Technol. Comput. Syst."],"published-print":{"date-parts":[[2022,4,30]]},"abstract":"<jats:p>\n            Stochastic computing (SC) reduces the complexity of computation by representing numbers with long streams of independent bits. However, increasing performance in SC comes with either an increase in area or a loss in accuracy. Processing in memory (PIM) computes data in-place while having high memory density and supporting bit-parallel operations with low energy consumption. In this article, we propose COSMO, an architecture for\n            <jats:underline>co<\/jats:underline>\n            mputing with\n            <jats:underline>s<\/jats:underline>\n            tochastic numbers in me\n            <jats:underline>mo<\/jats:underline>\n            ry, which enables SC in memory. The proposed architecture is general and can be used for a wide range of applications. It is a highly dense and parallel architecture that supports most SC encodings and operations in memory. It maximizes the performance and energy efficiency of SC by introducing several innovations: (i) in-memory parallel stochastic number generation, (ii) efficient implication-based logic in memory, (iii) novel memory bit line segmenting, (iv) a new memory-compatible SC addition operation, and (v) enabling flexible block allocation. To show the generality and efficiency of our stochastic architecture, we implement image processing, deep neural networks (DNNs), and hyperdimensional (HD) computing on the proposed hardware. Our evaluations show that running DNN inference on COSMO is 141\u00d7 faster and 80\u00d7 more energy efficient as compared to GPU.\n          <\/jats:p>","DOI":"10.1145\/3484731","type":"journal-article","created":{"date-parts":[[2022,1,12]],"date-time":"2022-01-12T14:06:38Z","timestamp":1641996398000},"page":"1-25","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["COSMO: Computing with Stochastic Numbers in Memory"],"prefix":"10.1145","volume":"18","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-5814-3934","authenticated-orcid":false,"given":"Saransh","family":"Gupta","sequence":"first","affiliation":[{"name":"University of California, San Diego, La Jolla, CA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5761-0622","authenticated-orcid":false,"given":"Mohsen","family":"Imani","sequence":"additional","affiliation":[{"name":"University of California, Irvine, CA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3417-0289","authenticated-orcid":false,"given":"Joonseop","family":"Sim","sequence":"additional","affiliation":[{"name":"University of California, San Diego, La Jolla, CA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1462-5143","authenticated-orcid":false,"given":"Andrew","family":"Huang","sequence":"additional","affiliation":[{"name":"University of California, San Diego, La Jolla, CA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1390-7683","authenticated-orcid":false,"given":"Fan","family":"Wu","sequence":"additional","affiliation":[{"name":"University of California, San Diego, La Jolla, CA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1048-1285","authenticated-orcid":false,"given":"Jaeyoung","family":"Kang","sequence":"additional","affiliation":[{"name":"University of California, San Diego, La Jolla, CA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5947-9632","authenticated-orcid":false,"given":"Yeseong","family":"Kim","sequence":"additional","affiliation":[{"name":"Daegu Gyeongbuk Institue of Science and Technology, Republic of Korea"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6954-997X","authenticated-orcid":false,"given":"Tajana \u0160imuni\u0107","family":"Rosing","sequence":"additional","affiliation":[{"name":"University of California, San Diego, La Jolla, CA"}]}],"member":"320","published-online":{"date-parts":[[2022,1,12]]},"reference":[{"key":"e_1_3_1_2_2","unstructured":"2016. Pytorch. Retrieved 1 Dec. 2020 from https:\/\/github.com\/pytorch\/pytorch."},{"key":"e_1_3_1_3_2","unstructured":"2017. NVIDIA GTX 1080 ti specifications. Retrieved 1 Dec. 2020 from https:\/\/www.techpowerup.com\/gpu-specs\/geforce-gtx-1080-ti.c2877."},{"key":"e_1_3_1_4_2","unstructured":"2020. Edge TPU performance benchmarks. Retrieved 1 Dec. 2020 from https:\/\/coral.ai\/docs\/edgetpu\/benchmarks."},{"key":"e_1_3_1_5_2","doi-asserted-by":"publisher","DOI":"10.1145\/2872887.2750385"},{"key":"e_1_3_1_6_2","doi-asserted-by":"publisher","DOI":"10.1109\/comst.2015.2444095"},{"key":"e_1_3_1_7_2","doi-asserted-by":"publisher","DOI":"10.1145\/2465787.2465794"},{"key":"e_1_3_1_8_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2017.2778107"},{"key":"e_1_3_1_9_2","doi-asserted-by":"publisher","DOI":"10.1109\/COMST.2014.2320099"},{"key":"e_1_3_1_10_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-642-35395-6_30"},{"key":"e_1_3_1_11_2","doi-asserted-by":"publisher","DOI":"10.1109\/TVLSI.2017.2654298"},{"key":"e_1_3_1_12_2","doi-asserted-by":"publisher","DOI":"10.1109\/MM.2014.55"},{"key":"e_1_3_1_13_2","doi-asserted-by":"publisher","DOI":"10.1109\/VLSIT.2014.6894411"},{"key":"e_1_3_1_14_2","doi-asserted-by":"publisher","DOI":"10.1038\/nature08940"},{"key":"e_1_3_1_15_2","doi-asserted-by":"publisher","DOI":"10.1109\/12.954505"},{"key":"e_1_3_1_16_2","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2015.2413754"},{"key":"e_1_3_1_17_2","unstructured":"Liu Changyu. 2020. Retrieved from Alexnet-pytorch. https:\/\/pypi.org\/project\/alexnet-pytorch\/."},{"key":"e_1_3_1_18_2","unstructured":"Liu Changyu. 2020. Retrieved from Googlenet-pytorch. https:\/\/pypi.org\/project\/googlenet-pytorch\/."},{"key":"e_1_3_1_19_2","unstructured":"Liu Changyu. 2020. Retrieved from Resnet-pytorch. https:\/\/pypi.org\/project\/resnet-pytorch\/."},{"key":"e_1_3_1_20_2","unstructured":"Liu Changyu. 2020. Retrieved from VGG-PyTorch. https:\/\/pypi.org\/project\/vgg-pytorch\/."},{"key":"e_1_3_1_21_2","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO.2014.58"},{"key":"e_1_3_1_22_2","doi-asserted-by":"publisher","DOI":"10.1145\/3007787.3001140"},{"key":"e_1_3_1_23_2","volume-title":"ISOLET - Data Set","author":"Cole Ron","year":"1994","unstructured":"Ron Cole and Mark Fanty. 1994. ISOLET - Data Set. UCI Machine Learning Repository. Retrieved 1 Dec., 2020 from https:\/\/archive.ics.uci.edu\/ml\/datasets\/ISOLET."},{"key":"e_1_3_1_24_2","doi-asserted-by":"publisher","DOI":"10.1145\/3296957.3173171"},{"key":"e_1_3_1_25_2","doi-asserted-by":"publisher","DOI":"10.1145\/1465482.1465505"},{"key":"e_1_3_1_26_2","doi-asserted-by":"publisher","DOI":"10.1145\/313817.313860"},{"key":"e_1_3_1_27_2","doi-asserted-by":"publisher","DOI":"10.1109\/2.375174"},{"key":"e_1_3_1_28_2","doi-asserted-by":"publisher","DOI":"10.5555\/578271"},{"key":"e_1_3_1_29_2","doi-asserted-by":"publisher","DOI":"10.1145\/3240765.3240811"},{"key":"e_1_3_1_30_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISCAS.2018.8351561"},{"key":"e_1_3_1_31_2","doi-asserted-by":"publisher","DOI":"10.1109\/ETS.2013.6569370"},{"key":"e_1_3_1_32_2","doi-asserted-by":"publisher","DOI":"10.1145\/2744769.2747932"},{"key":"e_1_3_1_33_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2015.123"},{"key":"e_1_3_1_34_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"e_1_3_1_35_2","first-page":"172","volume-title":"Proceedings of the 2011 Symposium on VLSI Technology (VLSIT)","author":"Hwang Sang-Min","year":"2011","unstructured":"Sang-Min Hwang, Srinivasa Banna, Cathy Tang, Sunil Bhardwaj, Mayank Gupta, Tim Thurgate, David Kim, Jungtae Kwon, Joong-Sik Kim, Seung-Hwan Lee, and J. Y. Lee. 2011. Offset buried metal gate vertical floating body memory technology with excellent retention time for DRAM application. In Proceedings of the 2011 Symposium on VLSI Technology (VLSIT). IEEE, 172\u2013173."},{"key":"e_1_3_1_36_2","unstructured":"Mohsen Imani. 2018. HD-IDHV. Retrieved 1 Dec. 2020 from https:\/\/Github.Com\/Moimani\/Hd-Idhv."},{"key":"e_1_3_1_37_2","volume-title":"Proceedings of the 2019 Design, Automation & Test in Europe Conference & Exhibition (DATE)","author":"Rosing M. Imani, J. Messerly, F. Wu, W. Pi, T.","year":"2019","unstructured":"M. Imani, J. Messerly, F. Wu, W. Pi, T. Rosing. 2019. A binary learning framework for hyperdimensional computing. In Proceedings of the 2019 Design, Automation & Test in Europe Conference & Exhibition (DATE)."},{"key":"e_1_3_1_38_2","doi-asserted-by":"publisher","DOI":"10.1145\/3061639.3062337"},{"key":"e_1_3_1_39_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICRC.2017.8123650"},{"issue":"2","key":"e_1_3_1_40_2","article-title":"Memristive logic-in-memory integrated circuits for energy-efficient flexible electronics","volume":"28","author":"Jang Byung Chul","year":"2018","unstructured":"Byung Chul Jang, Yunyong Nam, Beom Jun Koo, Junhwan Choi, Sung Gap Im, Sang-Hee Ko Park, and Sung-Yool Choi. 2018. Memristive logic-in-memory integrated circuits for energy-efficient flexible electronics. Advanced Functional Materials 28, 2 (2018), 1704725.","journal-title":"Advanced Functional Materials"},{"key":"e_1_3_1_41_2","doi-asserted-by":"publisher","DOI":"10.1021\/nl803669s"},{"key":"e_1_3_1_42_2","doi-asserted-by":"publisher","DOI":"10.1007\/s12559-009-9009-8"},{"key":"e_1_3_1_43_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.eswa.2016.04.018"},{"key":"e_1_3_1_44_2","doi-asserted-by":"publisher","DOI":"10.1145\/2897937.2898011"},{"key":"e_1_3_1_45_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISOCC.2015.7401667"},{"key":"e_1_3_1_46_2","doi-asserted-by":"publisher","DOI":"10.5555\/3199700.3199704"},{"key":"e_1_3_1_47_2","doi-asserted-by":"publisher","DOI":"10.1109\/TNANO.2014.2300342"},{"key":"e_1_3_1_48_2","doi-asserted-by":"publisher","DOI":"10.5555\/2999134.2999257"},{"key":"e_1_3_1_49_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCSII.2014.2357292"},{"key":"e_1_3_1_50_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCSII.2015.2433536"},{"issue":"10","key":"e_1_3_1_51_2","first-page":"2054","article-title":"Memristor-based material implication (IMPLY) logic: Design principles and methodologies","volume":"22","author":"Kvatinsky Shahar","year":"2014","unstructured":"Shahar Kvatinsky, Guy Satat, Nimrod Wald, Eby G. Friedman, Avinoam Kolodny, and Uri C. Weiser. 2014. Memristor-based material implication (IMPLY) logic: Design principles and methodologies. Tvlsi 22, 10 (2014), 2054\u20132066.","journal-title":"Tvlsi"},{"key":"e_1_3_1_52_2","unstructured":"Yann LeCun. 1998. The MNIST database of handwritten digits. Retrieved 1 Dec. 2020 from http:\/\/yann.Lecun.Com\/exdb\/mnist\/."},{"key":"e_1_3_1_53_2","first-page":"53","volume-title":"Proceedings of the International Conference on Artificial Neural Networks","volume":"60","author":"LeCun Yann","year":"1995","unstructured":"Yann LeCun, L. D. Jackel, Leon Bottou, A. Brunot, Corinna Cortes, J. S. Denker, Harris Drucker, I. Guyon, U. A. Muller, Eduard Sackinger, and P. Simard. 1995. Comparison of learning algorithms for handwritten digit recognition. In Proceedings of the International Conference on Artificial Neural Networks, Vol. 60. Perth, 53\u201360."},{"key":"e_1_3_1_54_2","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA.2013.6522354"},{"key":"e_1_3_1_55_2","doi-asserted-by":"publisher","DOI":"10.5555\/3130379.3130383"},{"key":"e_1_3_1_56_2","doi-asserted-by":"publisher","DOI":"10.1109\/NANOARCH.2009.5226356"},{"key":"e_1_3_1_57_2","doi-asserted-by":"publisher","DOI":"10.1145\/2847263.2847340"},{"key":"e_1_3_1_58_2","volume-title":"Caltech101 Library","author":"Li Fei-Fei","year":"2003","unstructured":"Fei-Fei Li, Marco Andreetto, Marc Aurelio Ranzato, and Pietro Perona. 2003. Caltech101 Library. Retrieved 1 Dec., 2020 from http:\/\/www.vision.caltech.edu\/Image_Datasets\/Caltech101."},{"key":"e_1_3_1_59_2","doi-asserted-by":"publisher","DOI":"10.1145\/2429384.2429483"},{"key":"e_1_3_1_60_2","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO.2018.00062"},{"key":"e_1_3_1_61_2","doi-asserted-by":"publisher","DOI":"10.1145\/2897937.2898064"},{"key":"e_1_3_1_62_2","first-page":"1","volume-title":"Proceedings of the Workshop on Near-Data Processing (WoNDP)","author":"Loh Gabriel H.","year":"2013","unstructured":"Gabriel H. Loh, Nuwan Jayasena, M. Oskin, Mark Nutter, David Roberts, Mitesh Meswani, Dong Ping Zhang, and Mike Ignatowski. 2013. A processing in memory taxonomy and a case for studying fixed-function pim. In Proceedings of the Workshop on Near-Data Processing (WoNDP). 1\u20134."},{"key":"e_1_3_1_63_2","volume-title":"UCI Machine Learning Repository","author":"Meidan Yair","year":"2018","unstructured":"Yair Meidan, Michael Bohadana, Yael Mathov, Yisroel Mirsky, Dominik Breitenbacher, Asaf Shabtai, and Yuval Elovici. 2018. UCI Machine Learning Repository. UCI Machine Learning Repository. Retrieved 1 Dec., 2020 from https:\/\/archive.ics.uci.edu\/ml\/datasets\/detection_of_IoT_botnet_attacks_N_BaIoT."},{"key":"e_1_3_1_64_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-04223-7_1"},{"key":"e_1_3_1_65_2","doi-asserted-by":"publisher","DOI":"10.1109\/MM.2018.053631140"},{"key":"e_1_3_1_66_2","article-title":"GPU-based deep learning inference: A performance and power analysis","year":"2015","unstructured":"Nvidia. 2015. GPU-based deep learning inference: A performance and power analysis. Nvidia Whitepaper.","journal-title":"Nvidia Whitepaper"},{"key":"e_1_3_1_67_2","volume-title":"Xilinx\/brevitas","author":"Pappalardo Alessandro","year":"2019","unstructured":"Alessandro Pappalardo. 2019. Xilinx\/brevitas. Github.com. DOI:https:\/\/doi.org\/10.5281\/zenodo.3333552"},{"key":"e_1_3_1_68_2","doi-asserted-by":"publisher","DOI":"10.5555\/784104.784178"},{"key":"e_1_3_1_69_2","article-title":"Computing arithmetic functions using stochastic logic by series expansion","author":"Parhi Keshab","year":"2016","unstructured":"Keshab Parhi and Yin Liu. 2016. Computing arithmetic functions using stochastic logic by series expansion. IEEE Transactions on Emerging Topics in Computing 7, 1 (2016), 44\u201359.","journal-title":"IEEE Transactions on Emerging Topics in Computing"},{"key":"e_1_3_1_70_2","doi-asserted-by":"publisher","DOI":"10.1109\/TC.2010.202"},{"key":"e_1_3_1_71_2","doi-asserted-by":"publisher","DOI":"10.1145\/3195970.3195998"},{"key":"e_1_3_1_72_2","doi-asserted-by":"publisher","DOI":"10.1186\/s13634-016-0355-x"},{"key":"e_1_3_1_73_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCSI.2017.2705051"},{"key":"e_1_3_1_74_2","doi-asserted-by":"publisher","DOI":"10.1145\/2934583.2934624"},{"key":"e_1_3_1_75_2","doi-asserted-by":"publisher","DOI":"10.1007\/11945918_17"},{"key":"e_1_3_1_76_2","doi-asserted-by":"publisher","DOI":"10.1145\/3093337.3037746"},{"key":"e_1_3_1_77_2","doi-asserted-by":"publisher","DOI":"10.1109\/MCOM.2017.1600435CM"},{"key":"e_1_3_1_78_2","doi-asserted-by":"publisher","DOI":"10.1145\/3007787.3001139"},{"key":"e_1_3_1_79_2","doi-asserted-by":"publisher","DOI":"10.1109\/DCAS51144.2020.9330667"},{"key":"e_1_3_1_80_2","unstructured":"James M. Sibigtroth George L. Espinor and Bruce L. Morton. 2006. Memory bit line segment isolation. US Patent 7 042 765 ."},{"issue":"1","key":"e_1_3_1_81_2","first-page":"64","article-title":"A complementary resistive switch-based crossbar array adder","volume":"5","author":"Siemon Anne","year":"2015","unstructured":"Anne Siemon, Stephan Menzel, Rainer Waser, and Eike Linn. 2015. A complementary resistive switch-based crossbar array adder. Jetcas 5, 1 (2015), 64\u201374.","journal-title":"Jetcas"},{"key":"e_1_3_1_82_2","doi-asserted-by":"publisher","DOI":"10.1109\/ASPDAC.2017.7858405"},{"key":"e_1_3_1_83_2","unstructured":"Karen Simonyan and Andrew Zisserman. 2014. Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556."},{"key":"e_1_3_1_84_2","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA.2017.55"},{"key":"e_1_3_1_85_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.vlsi.2017.02.002"},{"key":"e_1_3_1_86_2","doi-asserted-by":"publisher","DOI":"10.1109\/TNANO.2016.2570248"},{"key":"e_1_3_1_87_2","doi-asserted-by":"publisher","DOI":"10.1109\/NVMSA.2017.8064475"},{"key":"e_1_3_1_88_2","doi-asserted-by":"publisher","DOI":"10.1109\/DSD.2014.75"},{"key":"e_1_3_1_89_2","doi-asserted-by":"publisher","DOI":"10.1109\/TED.2011.2160265"},{"key":"e_1_3_1_90_2","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA.2015.7056056"},{"key":"e_1_3_1_91_2","doi-asserted-by":"publisher","DOI":"10.5555\/2691365.2691450"},{"key":"e_1_3_1_92_2","doi-asserted-by":"publisher","DOI":"10.1145\/3195970.3196113"},{"key":"e_1_3_1_93_2","doi-asserted-by":"publisher","DOI":"10.1049\/cje.2016.11.016"}],"container-title":["ACM Journal on Emerging Technologies in Computing Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3484731","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3484731","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3484731","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,17]],"date-time":"2025-06-17T20:17:15Z","timestamp":1750191435000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3484731"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,1,12]]},"references-count":92,"journal-issue":{"issue":"2","published-print":{"date-parts":[[2022,4,30]]}},"alternative-id":["10.1145\/3484731"],"URL":"https:\/\/doi.org\/10.1145\/3484731","relation":{},"ISSN":["1550-4832","1550-4840"],"issn-type":[{"type":"print","value":"1550-4832"},{"type":"electronic","value":"1550-4840"}],"subject":[],"published":{"date-parts":[[2022,1,12]]},"assertion":[{"value":"2020-08-01","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2021-08-01","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2022-01-12","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}