{"id":"https://openalex.org/W4319996342","doi":"https://doi.org/10.1109/jssc.2023.3234893","title":"A 95.6-TOPS/W Deep Learning Inference Accelerator With Per-Vector Scaled 4-bit Quantization in 5 nm","display_name":"A 95.6-TOPS/W Deep Learning Inference Accelerator With Per-Vector Scaled 4-bit Quantization in 5 nm","publication_year":2023,"publication_date":"2023-01-18","ids":{"openalex":"https://openalex.org/W4319996342","doi":"https://doi.org/10.1109/jssc.2023.3234893"},"language":"en","primary_location":{"id":"doi:10.1109/jssc.2023.3234893","is_oa":false,"landing_page_url":"https://doi.org/10.1109/jssc.2023.3234893","pdf_url":null,"source":{"id":"https://openalex.org/S83637746","display_name":"IEEE Journal of Solid-State Circuits","issn_l":"0018-9200","issn":["0018-9200","1558-173X"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Journal of Solid-State Circuits","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5001800455","display_name":"Ben Keller","orcid":"https://orcid.org/0000-0002-8117-1412"},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ben Keller","raw_affiliation_strings":["NVIDIA, Inc., Santa Clara, CA, USA"],"raw_orcid":"https://orcid.org/0000-0002-8117-1412","affiliations":[{"raw_affiliation_string":"NVIDIA, Inc., Santa Clara, CA, USA","institution_ids":["https://openalex.org/I4210127875"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5045219356","display_name":"Rangharajan Venkatesan","orcid":null},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Rangharajan Venkatesan","raw_affiliation_strings":["NVIDIA, Inc., Santa Clara, CA, USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NVIDIA, Inc., Santa Clara, CA, USA","institution_ids":["https://openalex.org/I4210127875"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5030894829","display_name":"Steve Dai","orcid":"https://orcid.org/0000-0002-5045-1964"},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Steve Dai","raw_affiliation_strings":["NVIDIA, Inc., Santa Clara, CA, USA"],"raw_orcid":"https://orcid.org/0000-0002-5045-1964","affiliations":[{"raw_affiliation_string":"NVIDIA, Inc., Santa Clara, CA, USA","institution_ids":["https://openalex.org/I4210127875"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102966476","display_name":"Stephen G. Tell","orcid":"https://orcid.org/0000-0002-2166-1452"},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Stephen G. Tell","raw_affiliation_strings":["NVIDIA, Inc., Durham, NC, USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NVIDIA, Inc., Durham, NC, USA","institution_ids":["https://openalex.org/I4210127875"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042467215","display_name":"Brian Zimmer","orcid":"https://orcid.org/0000-0001-9997-3141"},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Brian Zimmer","raw_affiliation_strings":["NVIDIA, Inc., Santa Clara, CA, USA"],"raw_orcid":"https://orcid.org/0000-0001-9997-3141","affiliations":[{"raw_affiliation_string":"NVIDIA, Inc., Santa Clara, CA, USA","institution_ids":["https://openalex.org/I4210127875"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5047303294","display_name":"Charbel Sakr","orcid":"https://orcid.org/0000-0001-5641-0541"},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Charbel Sakr","raw_affiliation_strings":["NVIDIA, Inc., Santa Clara, CA, USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NVIDIA, Inc., Santa Clara, CA, USA","institution_ids":["https://openalex.org/I4210127875"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5084342236","display_name":"William J. Dally","orcid":"https://orcid.org/0000-0003-4632-2876"},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"William J. Dally","raw_affiliation_strings":["NVIDIA, Inc., Santa Clara, CA, USA"],"raw_orcid":"https://orcid.org/0000-0003-4632-2876","affiliations":[{"raw_affiliation_string":"NVIDIA, Inc., Santa Clara, CA, USA","institution_ids":["https://openalex.org/I4210127875"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069026297","display_name":"C. Thomas Gray","orcid":"https://orcid.org/0000-0002-5137-5617"},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"C. Thomas Gray","raw_affiliation_strings":["NVIDIA, Inc., Durham, NC, USA"],"raw_orcid":"https://orcid.org/0000-0002-5137-5617","affiliations":[{"raw_affiliation_string":"NVIDIA, Inc., Durham, NC, USA","institution_ids":["https://openalex.org/I4210127875"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5010156116","display_name":"Brucek Khailany","orcid":"https://orcid.org/0000-0002-7584-3489"},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Brucek Khailany","raw_affiliation_strings":["NVIDIA, Inc., Austin, TX, USA"],"raw_orcid":"https://orcid.org/0000-0002-7584-3489","affiliations":[{"raw_affiliation_string":"NVIDIA, Inc., Austin, TX, USA","institution_ids":["https://openalex.org/I4210127875"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I4210127875"],"apc_list":null,"apc_paid":null,"fwci":5.302,"has_fulltext":false,"cited_by_count":54,"citation_normalized_percentile":{"value":0.97092534,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":100},"biblio":{"volume":"58","issue":"4","first_page":"1129","last_page":"1141"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9997000098228455,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9997000098228455,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9990000128746033,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12808","display_name":"Ferroelectric and Negative Capacitance Devices","score":0.996999979019165,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6861088275909424},{"id":"https://openalex.org/keywords/tops","display_name":"TOPS","score":0.6479174494743347},{"id":"https://openalex.org/keywords/quantization","display_name":"Quantization (signal processing)","score":0.5844712853431702},{"id":"https://openalex.org/keywords/efficient-energy-use","display_name":"Efficient energy use","score":0.5423683524131775},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5394125580787659},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.5146216154098511},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.502694845199585},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.43479686975479126},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.42626795172691345},{"id":"https://openalex.org/keywords/hardware-acceleration","display_name":"Hardware acceleration","score":0.4258463382720947},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.4228243827819824},{"id":"https://openalex.org/keywords/register-file","display_name":"Register file","score":0.41467800736427307},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.411857545375824},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.33859699964523315},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.3374973237514496},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.3306873142719269},{"id":"https://openalex.org/keywords/voltage","display_name":"Voltage","score":0.2435062825679779},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.19614818692207336},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.15028122067451477},{"id":"https://openalex.org/keywords/electrical-engineering","display_name":"Electrical engineering","score":0.12648162245750427},{"id":"https://openalex.org/keywords/instruction-set","display_name":"Instruction set","score":0.11887416243553162},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.09562283754348755}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6861088275909424},{"id":"https://openalex.org/C2777675136","wikidata":"https://www.wikidata.org/wiki/Q835642","display_name":"TOPS","level":3,"score":0.6479174494743347},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.5844712853431702},{"id":"https://openalex.org/C2742236","wikidata":"https://www.wikidata.org/wiki/Q924713","display_name":"Efficient energy use","level":2,"score":0.5423683524131775},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5394125580787659},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.5146216154098511},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.502694845199585},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.43479686975479126},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.42626795172691345},{"id":"https://openalex.org/C13164978","wikidata":"https://www.wikidata.org/wiki/Q600158","display_name":"Hardware acceleration","level":3,"score":0.4258463382720947},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.4228243827819824},{"id":"https://openalex.org/C117280010","wikidata":"https://www.wikidata.org/wiki/Q180944","display_name":"Register file","level":3,"score":0.41467800736427307},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.411857545375824},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.33859699964523315},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.3374973237514496},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3306873142719269},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.2435062825679779},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.19614818692207336},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.15028122067451477},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.12648162245750427},{"id":"https://openalex.org/C202491316","wikidata":"https://www.wikidata.org/wiki/Q272683","display_name":"Instruction set","level":2,"score":0.11887416243553162},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.09562283754348755},{"id":"https://openalex.org/C159737794","wikidata":"https://www.wikidata.org/wiki/Q124274","display_name":"Azimuth","level":2,"score":0.0},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/jssc.2023.3234893","is_oa":false,"landing_page_url":"https://doi.org/10.1109/jssc.2023.3234893","pdf_url":null,"source":{"id":"https://openalex.org/S83637746","display_name":"IEEE Journal of Solid-State Circuits","issn_l":"0018-9200","issn":["0018-9200","1558-173X"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Journal of Solid-State Circuits","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Affordable and clean energy","id":"https://metadata.un.org/sdg/7","score":0.9100000262260437}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":62,"referenced_works":["https://openalex.org/W2108598243","https://openalex.org/W2194775991","https://openalex.org/W2318354022","https://openalex.org/W2593172471","https://openalex.org/W2604319603","https://openalex.org/W2618530766","https://openalex.org/W2625457103","https://openalex.org/W2890472095","https://openalex.org/W2915716523","https://openalex.org/W2925491732","https://openalex.org/W2963748441","https://openalex.org/W2969812992","https://openalex.org/W2979874885","https://openalex.org/W2980104813","https://openalex.org/W2997929983","https://openalex.org/W2999356938","https://openalex.org/W3015391117","https://openalex.org/W3017024317","https://openalex.org/W3035083896","https://openalex.org/W3109585842","https://openalex.org/W3134750800","https://openalex.org/W3159727696","https://openalex.org/W3172967059","https://openalex.org/W3177265267","https://openalex.org/W3180037928","https://openalex.org/W3183406752","https://openalex.org/W3185238080","https://openalex.org/W3189877953","https://openalex.org/W3196923642","https://openalex.org/W3207622241","https://openalex.org/W3211525823","https://openalex.org/W4220686330","https://openalex.org/W4220764408","https://openalex.org/W4220837341","https://openalex.org/W4220865834","https://openalex.org/W4220922172","https://openalex.org/W4221086307","https://openalex.org/W4221131761","https://openalex.org/W4224267386","https://openalex.org/W4286571774","https://openalex.org/W4286571786","https://openalex.org/W4286571858","https://openalex.org/W4293025109","https://openalex.org/W4312772368","https://openalex.org/W4385245566","https://openalex.org/W6631190155","https://openalex.org/W6638632666","https://openalex.org/W6690026940","https://openalex.org/W6720242923","https://openalex.org/W6737454590","https://openalex.org/W6742116046","https://openalex.org/W6748224102","https://openalex.org/W6755207826","https://openalex.org/W6762718338","https://openalex.org/W6766978945","https://openalex.org/W6767997687","https://openalex.org/W6768851824","https://openalex.org/W6776320331","https://openalex.org/W6788135285","https://openalex.org/W6790521546","https://openalex.org/W6793164127","https://openalex.org/W6838537176"],"related_works":["https://openalex.org/W3183118997","https://openalex.org/W3214410901","https://openalex.org/W3204400881","https://openalex.org/W3204296682","https://openalex.org/W2917767146","https://openalex.org/W4319952061","https://openalex.org/W4280636456","https://openalex.org/W4388913998","https://openalex.org/W4310584535","https://openalex.org/W4295935044"],"abstract_inverted_index":{"The":[0,62],"energy":[1,25,93],"efficiency":[2],"of":[3,29,60,84,127],"deep":[4],"neural":[5],"network":[6],"(DNN)":[7],"inference":[8,16],"can":[9],"be":[10],"improved":[11],"with":[12,87,121,137,149],"custom":[13],"accelerators.":[14],"DNN":[15,54,169],"accelerators":[17],"often":[18],"employ":[19],"specialized":[20],"hardware":[21],"techniques":[22,31],"to":[23,80,99,159],"improve":[24],"efficiency,":[26],"but":[27],"many":[28],"these":[30],"result":[32],"in":[33],"catastrophic":[34],"accuracy":[35,89,140,151],"loss":[36,90,141,152],"on":[37,116,142,153],"transformer-based":[38],"DNNs,":[39],"which":[40,70],"have":[41],"become":[42],"ubiquitous":[43],"for":[44,57,76,167],"natural":[45],"language":[46],"processing":[47],"(NLP)":[48],"tasks.":[49],"This":[50],"article":[51],"presents":[52],"a":[53,96,117,124,163],"accelerator":[55,64,131,165],"designed":[56],"efficient":[58],"execution":[59],"transformers.":[61],"proposed":[63],"implements":[65],"per-vector":[66],"scaled":[67],"quantization":[68],"(VSQ),":[69],"employs":[71],"an":[72],"independent":[73],"scale":[74],"factor":[75],"each":[77],"64-element":[78],"vector":[79],"enable":[81],"the":[82,102,130],"use":[83],"4-bit":[85,118],"arithmetic":[86],"little":[88],"and":[91,144],"low":[92],"overhead.":[94],"Using":[95],"multilevel":[97],"dataflow":[98],"maximize":[100],"reuse,":[101],"5-nm":[103],"prototype":[104],"achieves":[105,132],"95.6":[106],"tera-operations":[107],"per":[108,110],"second":[109],"Watt":[111],"(TOPS/W)":[112],"at":[113],"0.46":[114],"V":[115],"benchmarking":[119],"layer":[120],"VSQ.":[122],"At":[123],"nominal":[125],"voltage":[126],"0.67":[128],"V,":[129],"1734":[133],"inferences/s/W":[134,146],"(38.7":[135],"TOPS/W)":[136,148],"only":[138],"0.7%":[139],"BERT-Base":[143],"4714":[145],"(38.6":[147],"0.15%":[150],"ResNet-50":[154],"by":[155],"using":[156],"quantization-aware":[157],"fine-tuning":[158],"recover":[160],"accuracy,":[161],"demonstrating":[162],"practical":[164],"design":[166],"energy-efficient":[168],"inference.":[170]},"counts_by_year":[{"year":2026,"cited_by_count":8},{"year":2025,"cited_by_count":26},{"year":2024,"cited_by_count":13},{"year":2023,"cited_by_count":7}],"updated_date":"2026-07-16T13:24:37.021932","created_date":"2025-10-10T00:00:00"}