{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,23]],"date-time":"2026-03-23T22:55:52Z","timestamp":1774306552045,"version":"3.50.1"},"reference-count":48,"publisher":"IOP Publishing","issue":"1","license":[{"start":{"date-parts":[[2022,2,18]],"date-time":"2022-02-18T00:00:00Z","timestamp":1645142400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2022,2,18]],"date-time":"2022-02-18T00:00:00Z","timestamp":1645142400000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/iopscience.iop.org\/info\/page\/text-and-data-mining"}],"content-domain":{"domain":["iopscience.iop.org"],"crossmark-restriction":false},"short-container-title":["Neuromorph. Comput. Eng."],"published-print":{"date-parts":[[2022,3,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>In-memory computing is a promising non-von Neumann approach to perform certain computational tasks efficiently within memory devices by exploiting their physical attributes. However, the computational accuracy achieved with this approach has been rather low, owing to significant inter-device variability and inhomogeneity across an array as well as intra-device variability and randomness from the analog memory devices. Bit slicing, a technique for constructing a high precision processor from several modules of lower precision, is a promising approach for overcoming this accuracy limitation. However, a systematic study to assess the precision ultimately achieved by bit slicing with analog in-memory computing has so far been lacking. In this work, we assess the computational error from bit slicing when performing in-memory matrix-vector multiplications. Using accurate models of phase-change memory crossbar arrays, we demonstrate that unlike in digital processors where bit slicing is used to extend the dynamic range of the number representation, bit slicing with in-memory computing should aim at minimizing the error from the analog matrix representation through averaging within a given dynamic range. The results are validated using a prototype phase-change memory chip and the impact on the neural network inference accuracy on CIFAR-10 and ImageNet benchmarks is evaluated.<\/jats:p>","DOI":"10.1088\/2634-4386\/ac4fb7","type":"journal-article","created":{"date-parts":[[2022,1,27]],"date-time":"2022-01-27T22:41:55Z","timestamp":1643323315000},"page":"014009","update-policy":"https:\/\/doi.org\/10.1088\/crossmark-policy","source":"Crossref","is-referenced-by-count":33,"title":["Precision of bit slicing with in-memory computing based on analog phase-change memory crossbars"],"prefix":"10.1088","volume":"2","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-1600-6151","authenticated-orcid":true,"given":"Manuel","family":"Le Gallo","sequence":"first","affiliation":[]},{"given":"S R","family":"Nandakumar","sequence":"additional","affiliation":[]},{"given":"Lazar","family":"Ciric","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4255-8622","authenticated-orcid":false,"given":"Irem","family":"Boybat","sequence":"additional","affiliation":[]},{"given":"Riduan","family":"Khaddam-Aljameh","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8413-5583","authenticated-orcid":true,"given":"Charles","family":"Mackin","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5603-5243","authenticated-orcid":true,"given":"Abu","family":"Sebastian","sequence":"additional","affiliation":[]}],"member":"266","published-online":{"date-parts":[[2022,2,18]]},"reference":[{"key":"nceac4fb7bib1","doi-asserted-by":"publisher","first-page":"260","DOI":"10.1109\/jproc.2018.2790840","article-title":"Neuro-inspired computing with emerging nonvolatile memorys","volume":"106","author":"Yu","year":"2018","journal-title":"Proc. IEEE"},{"key":"nceac4fb7bib2","doi-asserted-by":"publisher","first-page":"529","DOI":"10.1038\/s41565-020-0655-z","article-title":"Memory devices and applications for in-memory computing","volume":"15","author":"Sebastian","year":"2020","journal-title":"Nat. Nanotechnol."},{"key":"nceac4fb7bib3","first-page":"1","article-title":"Hermes core\u2014a 14\u00a0nm CMOS and PCM-based in-memory compute core using an array of 300 ps\/LSB linearized CCO-based ADCS and local digital processing","author":"Khaddam-Aljameh","year":"2021"},{"key":"nceac4fb7bib4","first-page":"236","article-title":"15.1 a programmable neural-network inference accelerator based on scalable in-memory computing","volume":"vol 64","author":"Jia","year":"2021"},{"key":"nceac4fb7bib5","doi-asserted-by":"publisher","first-page":"2473","DOI":"10.1038\/s41467-020-16108-9","article-title":"Accurate deep neural network inference using computational phase-change memory","volume":"11","author":"Joshi","year":"2020","journal-title":"Nat. Commun."},{"key":"nceac4fb7bib6","doi-asserted-by":"publisher","first-page":"641","DOI":"10.1038\/s41586-020-1942-4","article-title":"Fully hardware-implemented memristor convolutional neural network","volume":"577","author":"Yao","year":"2020","journal-title":"Nature"},{"key":"nceac4fb7bib7","doi-asserted-by":"publisher","first-page":"4304","DOI":"10.1109\/ted.2018.2865352","article-title":"Compressed sensing with approximate message passing using in-memory computing","volume":"65","author":"Le Gallo","year":"2018","journal-title":"IEEE Trans. Electron. Devices"},{"key":"nceac4fb7bib8","doi-asserted-by":"publisher","first-page":"246","DOI":"10.1038\/s41928-018-0054-8","article-title":"Mixed-precision in-memory computing","volume":"1","author":"Le Gallo","year":"2018","journal-title":"Nat. Electron."},{"key":"nceac4fb7bib9","doi-asserted-by":"crossref","DOI":"10.1109\/IEDM.2018.8614558","article-title":"8 bit precision in-memory multiplication with projected phase-change memory","author":"Giannopoulos","year":"2018"},{"key":"nceac4fb7bib10","first-page":"221","article-title":"Vector multiplications using memristive devices and applications thereof","author":"Zidan","year":"2020"},{"key":"nceac4fb7bib11","author":"Mick","year":"1980"},{"key":"nceac4fb7bib12","first-page":"1","article-title":"Memristive Boltzmann machine: a hardware accelerator for combinatorial optimization and deep learning","author":"Bojnordi","year":"2016"},{"key":"nceac4fb7bib13","first-page":"14","article-title":"ISAAC: a convolutional neural network accelerator with in situ analog arithmetic in crossbars","author":"Shafiee","year":"2016"},{"key":"nceac4fb7bib14","first-page":"27","article-title":"PRIME: a novel processing-in-memory architecture for neural network computation in ReRAM-based main memory","author":"Chi","year":"2016"},{"key":"nceac4fb7bib15","first-page":"1","article-title":"Binary neural network with 16 mb RRAM macro chip for classification and online training","author":"Yu","year":"2016"},{"key":"nceac4fb7bib16","first-page":"541","article-title":"Pipelayer: a pipelined ReRAM-based accelerator for deep learning","author":"Song","year":"2017"},{"key":"nceac4fb7bib17","first-page":"T174","article-title":"Achieving ideal accuracies in analog neuromorphic computing using periodic carry","author":"Agarwal","year":"2017"},{"key":"nceac4fb7bib18","doi-asserted-by":"publisher","first-page":"2514","DOI":"10.1038\/s41467-018-04933-y","article-title":"Neuromorphic computing with multi-memristive synapses","volume":"9","author":"Boybat","year":"2018","journal-title":"Nat. Commun."},{"key":"nceac4fb7bib19","first-page":"367","article-title":"Enabling scientific computing on memristive accelerators","author":"Feinberg","year":"2018"},{"key":"nceac4fb7bib20","doi-asserted-by":"publisher","first-page":"1900026","DOI":"10.1002\/aelm.201900026","article-title":"Weight programming in DNN analog hardware accelerators in the presence of NVM variability","volume":"5","author":"Mackin","year":"2019","journal-title":"Adv. Electron. Mater."},{"key":"nceac4fb7bib21","doi-asserted-by":"publisher","first-page":"411","DOI":"10.1038\/s41928-018-0100-6","article-title":"A general memristor-based partial differential equation solver","volume":"1","author":"Zidan","year":"2018","journal-title":"Nat. Electron."},{"key":"nceac4fb7bib22","first-page":"404","article-title":"29.1 a 40\u00a0nm 64\u00a0kb 56.67 TOPS\/W read-disturb-tolerant compute-in-memory\/digital RRAM macro with active-feedback-based read and in situ write verification","volume":"vol 64","author":"Yoon","year":"2021"},{"key":"nceac4fb7bib23","first-page":"245","article-title":"16.1 a 22\u00a0nm 4 mb 8b-precision ReRAM computing-in-memory macro with 11.91 to 195.7 TOPS\/W for tiny AI edge devices","volume":"vol 64","author":"Xue","year":"2021"},{"key":"nceac4fb7bib24","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1038\/s41586-018-0180-5","article-title":"Equivalent-accuracy accelerated neural-network training using analogue memory","volume":"558","author":"Ambrogio","year":"2018","journal-title":"Nature"},{"key":"nceac4fb7bib25","first-page":"1","article-title":"Swipe: enhancing robustness of ReRAM crossbars for in-memory computing","author":"Gonugondla","year":"2020"},{"key":"nceac4fb7bib26","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1007\/s11432-020-3198-9","article-title":"Array-level boosting method with spatial extended allocation to improve the accuracy of memristor based computing-in-memory chips","volume":"64","author":"Zhang","year":"2021","journal-title":"Sci. China Inf. Sci."},{"key":"nceac4fb7bib27","first-page":"1","article-title":"Conductance variations and their impact on the precision of in-memory computing with resistive switching memory (RRAM)","author":"Pedretti","year":"2021"},{"key":"nceac4fb7bib28","doi-asserted-by":"publisher","first-page":"774","DOI":"10.3389\/fnins.2018.00774","article-title":"Deep learning with spiking neurons: opportunities and challenges","volume":"12","author":"Pfeiffer","year":"2018","journal-title":"Front. Neurosci."},{"key":"nceac4fb7bib29","doi-asserted-by":"publisher","first-page":"97","DOI":"10.1109\/msp.2019.2933719","article-title":"Low-power neuromorphic hardware for signal processing applications: a review of architectural and system-level design approaches","volume":"36","author":"Rajendran","year":"2019","journal-title":"IEEE Signal Process. Mag."},{"key":"nceac4fb7bib30","doi-asserted-by":"publisher","DOI":"10.1063\/1.5042413","article-title":"Tutorial: brain-inspired computing using phase-change memory devices","volume":"124","author":"Sebastian","year":"2018","journal-title":"J. Appl. Phys."},{"key":"nceac4fb7bib31","doi-asserted-by":"publisher","first-page":"2213","DOI":"10.1109\/ted.2020.2979606","article-title":"A parallel multibit programing scheme with high precision for RRAM-based neuromorphic systems","volume":"67","author":"Chen","year":"2020","journal-title":"IEEE Trans. Electron Devices"},{"key":"nceac4fb7bib32","first-page":"1","article-title":"Mushroom-type phase change memory with projection liner: an array-level demonstration of conductance drift and noise mitigation","author":"Bruce","year":"2021"},{"key":"nceac4fb7bib33","first-page":"727","article-title":"Phase-change memory models for deep learning training and inference","author":"Nandakumar","year":"2019"},{"key":"nceac4fb7bib34","first-page":"1","article-title":"Phase change memory as synapse for ultra-dense neuromorphic systems: application to complex visual pattern extraction","author":"Suri","year":"2011"},{"key":"nceac4fb7bib35","doi-asserted-by":"crossref","DOI":"10.1109\/IEDM13553.2020.9371990","article-title":"Precision of synaptic weights programmed in phase-change memory devices for deep learning inference","author":"Nandakumar","year":"2020"},{"key":"nceac4fb7bib36","first-page":"1","article-title":"Reducing the impact of phase-change memory conductance drift on the inference of large-scale hardware neural networks","author":"Ambrogio","year":"2019"},{"key":"nceac4fb7bib37","doi-asserted-by":"publisher","first-page":"1","DOI":"10.3389\/fnins.2017.00538","article-title":"Training deep convolutional neural networks with resistive cross-point devices","volume":"11","author":"Gokmen","year":"2017","journal-title":"Front. Neurosci."},{"key":"nceac4fb7bib38","doi-asserted-by":"publisher","first-page":"312","DOI":"10.1109\/ted.2016.2630925","article-title":"3D memristor crossbars for analog and neuromorphic computing applications","volume":"64","author":"Adam","year":"2017","journal-title":"IEEE Trans. Electron Devices"},{"key":"nceac4fb7bib39","doi-asserted-by":"publisher","first-page":"674154","DOI":"10.3389\/fncom.2021.674154","article-title":"Accelerating inference of convolutional neural networks using in-memory computing","volume":"15","author":"Dazzi","year":"2021","journal-title":"Frontiers in Computational Neuroscience"},{"key":"nceac4fb7bib40","doi-asserted-by":"publisher","first-page":"657","DOI":"10.1007\/s10462-021-10060-w","article-title":"Reduction 93.7% time and power consumption using a memristor-based imprecise gradient update algorithm","volume":"55","author":"Li","year":"2021","journal-title":"Artif. Intell. Rev."},{"key":"nceac4fb7bib41","first-page":"7675","article-title":"Training deep neural networks with 8 bit floating point numbers","author":"Wang","year":"2018"},{"key":"nceac4fb7bib42","first-page":"1796","article-title":"Ultra-low precision 4 bit training of deep neural networks","volume":"vol 33","author":"Sun","year":"2020"},{"key":"nceac4fb7bib43","doi-asserted-by":"publisher","first-page":"61","DOI":"10.1038\/nature14441","article-title":"Training and operation of an integrated neuromorphic network based on metal-oxide memristors","volume":"521","author":"Prezioso","year":"2015","journal-title":"Nature"},{"key":"nceac4fb7bib44","doi-asserted-by":"publisher","first-page":"15199","DOI":"10.1038\/ncomms15199","article-title":"Face classification using electronic synapses","volume":"8","author":"Yao","year":"2018","journal-title":"Nat. Commun."},{"key":"nceac4fb7bib45","doi-asserted-by":"publisher","first-page":"2385","DOI":"10.1038\/s41467-018-04484-2","article-title":"Efficient and self-adaptive in situ learning in multilayer memristor neural networks","volume":"9","author":"Li","year":"2018","journal-title":"Nat. Commun."},{"key":"nceac4fb7bib46","doi-asserted-by":"publisher","first-page":"434","DOI":"10.1038\/s42256-019-0089-1","article-title":"In situ training of feed-forward and recurrent convolutional memristor networks","volume":"1","author":"Wang","year":"2019","journal-title":"Nat. Mach. Intell."},{"key":"nceac4fb7bib47","doi-asserted-by":"publisher","first-page":"406","DOI":"10.3389\/fnins.2020.00406","article-title":"Mixed-precision deep learning based on computational memory","volume":"14","author":"Nandakumar","year":"2020","journal-title":"Front. Neurosci."},{"key":"nceac4fb7bib48","first-page":"1","article-title":"Fully on-chip MAC at 14\u00a0nm enabled by accurate row-wise programming of PCM-based weights and parallel vector-transport in duration-format","author":"Narayanan","year":"2021"}],"container-title":["Neuromorphic Computing and Engineering"],"original-title":[],"link":[{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac4fb7","content-type":"text\/html","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac4fb7\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac4fb7","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac4fb7\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac4fb7\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac4fb7\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac4fb7\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"similarity-checking"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac4fb7\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,2,18]],"date-time":"2022-02-18T15:39:15Z","timestamp":1645198755000},"score":1,"resource":{"primary":{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac4fb7"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,2,18]]},"references-count":48,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2022,2,18]]},"published-print":{"date-parts":[[2022,3,1]]}},"URL":"https:\/\/doi.org\/10.1088\/2634-4386\/ac4fb7","relation":{},"ISSN":["2634-4386"],"issn-type":[{"value":"2634-4386","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,2,18]]},"assertion":[{"value":"Precision of bit slicing with in-memory computing based on analog phase-change memory crossbars","name":"article_title","label":"Article Title"},{"value":"Neuromorphic Computing and Engineering","name":"journal_title","label":"Journal Title"},{"value":"paper","name":"article_type","label":"Article Type"},{"value":"\u00a9 2022 The Author(s). Published by IOP Publishing Ltd","name":"copyright_information","label":"Copyright Information"},{"value":"2021-09-30","name":"date_received","label":"Date Received","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2022-01-27","name":"date_accepted","label":"Date Accepted","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2022-02-18","name":"date_epub","label":"Online publication date","group":{"name":"publication_dates","label":"Publication dates"}}]}}