{"id":"https://openalex.org/W4387789586","doi":"https://doi.org/10.1109/tcsi.2023.3312775","title":"An Integer-Only and Group-Vector Systolic Accelerator for Efficiently Mapping Vision Transformer on Edge","display_name":"An Integer-Only and Group-Vector Systolic Accelerator for Efficiently Mapping Vision Transformer on Edge","publication_year":2023,"publication_date":"2023-10-19","ids":{"openalex":"https://openalex.org/W4387789586","doi":"https://doi.org/10.1109/tcsi.2023.3312775"},"language":"en","primary_location":{"id":"doi:10.1109/tcsi.2023.3312775","doi":"https://doi.org/10.1109/tcsi.2023.3312775","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcsi.2023.3312775","pdf_url":null,"source":{"id":"https://openalex.org/S116977442","display_name":"IEEE Transactions on Circuits and Systems I Regular Papers","issn_l":"1549-8328","issn":["1549-8328","1558-0806"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["anvur-scientific","ccf-c","cwts-core","dongbi-b","jufo-2","norway-2","poland-140","russia-white-list-1","vabb-shw"],"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 Transactions on Circuits and Systems I: Regular Papers","raw_type":"journal-article"},"volume":null,"type":"article","is_traditional":true,"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/A5038495391","display_name":"Mingqiang Huang","orcid":"https://orcid.org/0000-0002-7794-3985","observed_orcids":["https://orcid.org/0000-0002-7794-3985"]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210145761","display_name":"Shenzhen Institutes of Advanced Technology","ror":"https://ror.org/04gh4er46","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210145761"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Mingqiang Huang","raw_affiliation_strings":["Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China"],"raw_orcid":"https://orcid.org/0000-0002-7794-3985","affiliations":[{"raw_affiliation_string":"Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210145761"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Junyi Luo","orcid":"https://orcid.org/0009-0000-0578-2206","observed_orcids":["https://orcid.org/0009-0000-0578-2206"]},"institutions":[{"id":"https://openalex.org/I3045169105","display_name":"Southern University of Science and Technology","ror":"https://ror.org/049tv2d57","country_code":"CN","type":"education","lineage":["https://openalex.org/I3045169105"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Junyi Luo","raw_affiliation_strings":["School of Microelectronics, Southern University of Science and Technology, Shenzhen, China"],"raw_orcid":"https://orcid.org/0009-0000-0578-2206","affiliations":[{"raw_affiliation_string":"School of Microelectronics, Southern University of Science and Technology, Shenzhen, China","institution_ids":["https://openalex.org/I3045169105"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101409855","display_name":"Chenchen Ding","orcid":"https://orcid.org/0009-0001-2026-1449","observed_orcids":["https://orcid.org/0009-0001-2026-1449"]},"institutions":[{"id":"https://openalex.org/I3045169105","display_name":"Southern University of Science and Technology","ror":"https://ror.org/049tv2d57","country_code":"CN","type":"education","lineage":["https://openalex.org/I3045169105"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chenchen Ding","raw_affiliation_strings":["School of Microelectronics, Southern University of Science and Technology, Shenzhen, China"],"raw_orcid":"https://orcid.org/0009-0001-2026-1449","affiliations":[{"raw_affiliation_string":"School of Microelectronics, Southern University of Science and Technology, Shenzhen, China","institution_ids":["https://openalex.org/I3045169105"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100733358","display_name":"Zikun Wei","orcid":"https://orcid.org/0009-0003-2684-5124","observed_orcids":["https://orcid.org/0009-0003-2684-5124"]},"institutions":[{"id":"https://openalex.org/I3045169105","display_name":"Southern University of Science and Technology","ror":"https://ror.org/049tv2d57","country_code":"CN","type":"education","lineage":["https://openalex.org/I3045169105"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zikun Wei","raw_affiliation_strings":["School of Microelectronics, Southern University of Science and Technology, Shenzhen, China"],"raw_orcid":"https://orcid.org/0009-0003-2684-5124","affiliations":[{"raw_affiliation_string":"School of Microelectronics, Southern University of Science and Technology, Shenzhen, China","institution_ids":["https://openalex.org/I3045169105"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5022383429","display_name":"Sixiao Huang","orcid":"https://orcid.org/0000-0002-2252-0678","observed_orcids":["https://orcid.org/0000-0002-2252-0678"]},"institutions":[{"id":"https://openalex.org/I3045169105","display_name":"Southern University of Science and Technology","ror":"https://ror.org/049tv2d57","country_code":"CN","type":"education","lineage":["https://openalex.org/I3045169105"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Sixiao Huang","raw_affiliation_strings":["School of Microelectronics, Southern University of Science and Technology, Shenzhen, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Microelectronics, Southern University of Science and Technology, Shenzhen, China","institution_ids":["https://openalex.org/I3045169105"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5034853402","display_name":"Hao Yu","orcid":"https://orcid.org/0000-0002-2674-4118","observed_orcids":["https://orcid.org/0000-0002-2674-4118"]},"institutions":[{"id":"https://openalex.org/I3045169105","display_name":"Southern University of Science and Technology","ror":"https://ror.org/049tv2d57","country_code":"CN","type":"education","lineage":["https://openalex.org/I3045169105"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hao Yu","raw_affiliation_strings":["School of Microelectronics, Southern University of Science and Technology, Shenzhen, China"],"raw_orcid":"https://orcid.org/0000-0002-2674-4118","affiliations":[{"raw_affiliation_string":"School of Microelectronics, Southern University of Science and Technology, Shenzhen, China","institution_ids":["https://openalex.org/I3045169105"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":{"value":2345,"currency":"USD","value_usd":2345},"apc_paid":null,"fwci":6.1043,"has_fulltext":false,"cited_by_count":82,"citation_normalized_percentile":{"value":0.97905007,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":88,"max":100},"biblio":{"volume":"70","issue":"12","first_page":"5289","last_page":"5301"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9189000129699707,"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":0.9189000129699707,"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/T10904","display_name":"Embedded Systems Design Techniques","score":0.06210000067949295,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T13292","display_name":"Embedded Systems and FPGA Applications","score":0.00570000009611249,"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/vision-transformer","display_name":"vision transformer","score":0.8539999723434448},{"id":"https://openalex.org/keywords/edge-deployment","display_name":"edge deployment","score":0.7289999723434448},{"id":"https://openalex.org/keywords/integer-quantization","display_name":"integer quantization","score":0.7720000147819519},{"id":"https://openalex.org/keywords/algorithm-hardware-co-design","display_name":"algorithm-hardware co-design","score":0.8999999761581421},{"id":"https://openalex.org/keywords/fpga-accelerator","display_name":"FPGA accelerator","score":0.7710000276565552},{"id":"https://openalex.org/keywords/multi-head-self-attention","display_name":"multi-head self-attention","score":0.7429999709129333},{"id":"https://openalex.org/keywords/inference-latency","display_name":"inference latency","score":0.5109999775886536}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6909459829330444},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.5679550766944885},{"id":"https://openalex.org/C17349429","wikidata":"https://www.wikidata.org/wiki/Q1049914","display_name":"Matrix multiplication","level":3,"score":0.5540779232978821},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.5112738609313965},{"id":"https://openalex.org/C150741067","wikidata":"https://www.wikidata.org/wiki/Q2377218","display_name":"Systolic array","level":3,"score":0.4651094079017639},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.44854962825775146},{"id":"https://openalex.org/C2778456923","wikidata":"https://www.wikidata.org/wiki/Q5337692","display_name":"Edge computing","level":3,"score":0.4292772710323334},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.4286327362060547},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.41823646426200867},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.3928043842315674},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.298564076423645},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.13176584243774414},{"id":"https://openalex.org/C81860439","wikidata":"https://www.wikidata.org/wiki/Q251212","display_name":"Internet of Things","level":2,"score":0.10454005002975464},{"id":"https://openalex.org/C14580979","wikidata":"https://www.wikidata.org/wiki/Q876049","display_name":"Very-large-scale integration","level":2,"score":0.10280713438987732},{"id":"https://openalex.org/C62520636","wikidata":"https://www.wikidata.org/wiki/Q944","display_name":"Quantum mechanics","level":1,"score":0.0},{"id":"https://openalex.org/C79974875","wikidata":"https://www.wikidata.org/wiki/Q483639","display_name":"Cloud computing","level":2,"score":0.0},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.0},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.0},{"id":"https://openalex.org/C84114770","wikidata":"https://www.wikidata.org/wiki/Q46344","display_name":"Quantum","level":2,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tcsi.2023.3312775","doi":"https://doi.org/10.1109/tcsi.2023.3312775","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcsi.2023.3312775","pdf_url":null,"source":{"id":"https://openalex.org/S116977442","display_name":"IEEE Transactions on Circuits and Systems I Regular Papers","issn_l":"1549-8328","issn":["1549-8328","1558-0806"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["anvur-scientific","ccf-c","cwts-core","dongbi-b","jufo-2","norway-2","poland-140","russia-white-list-1","vabb-shw"],"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 Transactions on Circuits and Systems I: Regular Papers","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Industry, innovation and infrastructure","id":"https://metadata.un.org/sdg/9","score":0.6895999908447266}],"study_designs":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320336569","display_name":"Shenzhen Science and Technology Innovation Program","ror":null}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":29,"referenced_works":["https://openalex.org/W1686810756","https://openalex.org/W2048266589","https://openalex.org/W2094756095","https://openalex.org/W2194775991","https://openalex.org/W2606722458","https://openalex.org/W2626778328","https://openalex.org/W2896457183","https://openalex.org/W2963341956","https://openalex.org/W2965373594","https://openalex.org/W3030163527","https://openalex.org/W3047848469","https://openalex.org/W3115388607","https://openalex.org/W3116489684","https://openalex.org/W3118568258","https://openalex.org/W3121523901","https://openalex.org/W3130240120","https://openalex.org/W3135367836","https://openalex.org/W3162542754","https://openalex.org/W3199934250","https://openalex.org/W3206837665","https://openalex.org/W4280493494","https://openalex.org/W4282008392","https://openalex.org/W4285250061","https://openalex.org/W4293025835","https://openalex.org/W4308479898","https://openalex.org/W4312933868","https://openalex.org/W4360831786","https://openalex.org/W4390873361","https://openalex.org/W6766673545"],"related_works":["https://openalex.org/W7166801031","https://openalex.org/W4400811522","https://openalex.org/W4417260734","https://openalex.org/W7162073012","https://openalex.org/W7140462333","https://openalex.org/W4415939076","https://openalex.org/W4411544631","https://openalex.org/W4401357692","https://openalex.org/W4411798947","https://openalex.org/W7084126764"],"abstract_inverted_index":{"Transformer-like":[0],"network":[1],"has":[2,125],"shown":[3],"remarkable":[4],"high":[5,75,78,155],"performance":[6],"in":[7,20,30],"both":[8,73,96],"natural":[9],"language":[10],"processing":[11],"and":[12,24,63,77,99,108,140,146],"computer":[13],"vision.":[14],"However,":[15],"the":[16,25,45,54,74,91,100,121,164],"huge":[17],"computational":[18],"demands":[19],"non-linear":[21,57],"floating-point":[22],"arithmetic":[23],"irregular":[26,101],"memory":[27],"access":[28],"requirement":[29],"self-attention":[31,103],"mechanism":[32],"make":[33],"it":[34],"still":[35],"a":[36],"challenge":[37],"to":[38,71,88,118],"deploy":[39],"Transformer":[40,168],"on":[41,128],"edge.":[42],"To":[43],"address":[44],"above":[46],"issues,":[47],"we":[48,81],"propose":[49],"integer-only":[50],"quantization":[51],"scheme":[52],"for":[53,144],"simplification":[55],"of":[56,138],"operations":[58,94],"(such":[59],"as":[60,154,156],"LayerNorm,":[61],"Softmax":[62],"Gelu),":[64],"meanwhile":[65],"algorithm-hardware":[66],"co-design":[67],"strategy":[68,107,114],"is":[69],"applied":[70],"guarantee":[72],"accuracy":[76],"efficiency.":[79],"Besides,":[80],"construct":[82],"general-purpose":[83],"group":[84],"vector":[85],"systolic":[86],"array":[87],"efficiently":[89],"accelerate":[90],"matrix":[92],"multiplication":[93],"including":[95],"regular":[97],"matrix-multiplication/convolution":[98],"multi-head":[102],"mechanism.":[104],"Unified":[105],"data-package":[106],"flexible":[109],"on-/off-chip":[110],"data":[111],"storage":[112],"management":[113],"are":[115],"also":[116],"proposed":[117],"further":[119],"improve":[120],"performance.":[122],"The":[123,149],"design":[124],"been":[126],"deployed":[127],"Xilinx":[129],"ZCU102":[130],"FPGA":[131,167],"platform,":[132],"achieving":[133],"an":[134],"overall":[135],"inference":[136],"latency":[137],"4.077ms":[139],"11.15ms":[141],"per":[142],"image":[143],"ViT-tiny":[145],"ViT-s,":[147],"respectively.":[148],"average":[150],"throughput":[151],"can":[152],"reach":[153],"762.7":[157],"GOPs,":[158],"which":[159],"shows":[160],"significant":[161],"improvement":[162],"over":[163],"previous":[165],"state-of-the-art":[166],"accelerator.":[169]},"counts_by_year":[{"year":2026,"cited_by_count":26},{"year":2025,"cited_by_count":43},{"year":2024,"cited_by_count":12},{"year":2023,"cited_by_count":1}],"updated_date":"2026-11-08T12:12:28.789561","created_date":"2023-10-20T00:00:00"}