{"id":"https://openalex.org/W7162414818","doi":"https://doi.org/10.1109/tvlsi.2026.3693713","title":"Fused-STA: Automated Design Space Exploration of a Fused Systolic Tensor Array for Universal Deep Learning Acceleration","display_name":"Fused-STA: Automated Design Space Exploration of a Fused Systolic Tensor Array for Universal Deep Learning Acceleration","publication_year":2026,"publication_date":"2026-05-26","ids":{"openalex":"https://openalex.org/W7162414818","doi":"https://doi.org/10.1109/tvlsi.2026.3693713"},"language":null,"primary_location":{"id":"doi:10.1109/tvlsi.2026.3693713","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tvlsi.2026.3693713","pdf_url":null,"source":{"id":"https://openalex.org/S37538908","display_name":"IEEE Transactions on Very Large Scale Integration (VLSI) Systems","issn_l":"1063-8210","issn":["1063-8210","1557-9999"],"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 Transactions on Very Large Scale Integration (VLSI) Systems","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/A5000928832","display_name":"J D Lee","orcid":null},"institutions":[{"id":"https://openalex.org/I193352282","display_name":"Daegu Gyeongbuk Institute of Science and Technology","ror":"https://ror.org/03frjya69","country_code":"KR","type":"education","lineage":["https://openalex.org/I193352282"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jooyeon Lee","raw_affiliation_strings":["Daegu Gyeongbuk Institute of Science and Technology"],"raw_orcid":"https://orcid.org/0000-0002-3338-4982","affiliations":[{"raw_affiliation_string":"Daegu Gyeongbuk Institute of Science and Technology","institution_ids":["https://openalex.org/I193352282"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121541079","display_name":"Sangwoo Jung","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sangwoo Jung","raw_affiliation_strings":["Korea University"],"raw_orcid":"https://orcid.org/0009-0008-1150-9888","affiliations":[{"raw_affiliation_string":"Korea University","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5137013102","display_name":"Jaeha Kung","orcid":"https://orcid.org/0000-0002-2027-8531"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jaeha Kung","raw_affiliation_strings":["Korea University"],"raw_orcid":"https://orcid.org/0000-0002-2027-8531","affiliations":[{"raw_affiliation_string":"Korea University","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.57973174,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"34","issue":"8","first_page":"2355","last_page":"2368"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12303","display_name":"Tensor decomposition and applications","score":0.44920000433921814,"subfield":{"id":"https://openalex.org/subfields/2605","display_name":"Computational Mathematics"},"field":{"id":"https://openalex.org/fields/26","display_name":"Mathematics"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T12303","display_name":"Tensor decomposition and applications","score":0.44920000433921814,"subfield":{"id":"https://openalex.org/subfields/2605","display_name":"Computational Mathematics"},"field":{"id":"https://openalex.org/fields/26","display_name":"Mathematics"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11206","display_name":"Model Reduction and Neural Networks","score":0.08709999918937683,"subfield":{"id":"https://openalex.org/subfields/3109","display_name":"Statistical and Nonlinear Physics"},"field":{"id":"https://openalex.org/fields/31","display_name":"Physics and Astronomy"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10904","display_name":"Embedded Systems Design Techniques","score":0.05480000004172325,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/acceleration","display_name":"Acceleration","score":0.5670999884605408},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.4909000098705292},{"id":"https://openalex.org/keywords/nasa-deep-space-network","display_name":"NASA Deep Space Network","score":0.4406000077724457},{"id":"https://openalex.org/keywords/tensor","display_name":"Tensor (intrinsic definition)","score":0.42149999737739563},{"id":"https://openalex.org/keywords/deep-space-exploration","display_name":"Deep space exploration","score":0.41769999265670776},{"id":"https://openalex.org/keywords/space-exploration","display_name":"Space exploration","score":0.4059999883174896},{"id":"https://openalex.org/keywords/space","display_name":"Space (punctuation)","score":0.3898000121116638}],"concepts":[{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.5670999884605408},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5548999905586243},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.4909000098705292},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4514000117778778},{"id":"https://openalex.org/C187107819","wikidata":"https://www.wikidata.org/wiki/Q835696","display_name":"NASA Deep Space Network","level":3,"score":0.4406000077724457},{"id":"https://openalex.org/C155281189","wikidata":"https://www.wikidata.org/wiki/Q3518150","display_name":"Tensor (intrinsic definition)","level":2,"score":0.42149999737739563},{"id":"https://openalex.org/C61776360","wikidata":"https://www.wikidata.org/wiki/Q5250441","display_name":"Deep space exploration","level":4,"score":0.41769999265670776},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.40869998931884766},{"id":"https://openalex.org/C104060986","wikidata":"https://www.wikidata.org/wiki/Q180046","display_name":"Space exploration","level":2,"score":0.4059999883174896},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.40459999442100525},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.3898000121116638},{"id":"https://openalex.org/C150741067","wikidata":"https://www.wikidata.org/wiki/Q2377218","display_name":"Systolic array","level":3,"score":0.33079999685287476},{"id":"https://openalex.org/C2986737658","wikidata":"https://www.wikidata.org/wiki/Q30103009","display_name":"Tensor decomposition","level":3,"score":0.29760000109672546},{"id":"https://openalex.org/C104267543","wikidata":"https://www.wikidata.org/wiki/Q208163","display_name":"Signal processing","level":3,"score":0.2721000015735626},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2671000063419342},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.26510000228881836},{"id":"https://openalex.org/C2776221188","wikidata":"https://www.wikidata.org/wiki/Q21072556","display_name":"Design space exploration","level":2,"score":0.26179999113082886},{"id":"https://openalex.org/C127313418","wikidata":"https://www.wikidata.org/wiki/Q1069","display_name":"Geology","level":0,"score":0.2590999901294708}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tvlsi.2026.3693713","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tvlsi.2026.3693713","pdf_url":null,"source":{"id":"https://openalex.org/S37538908","display_name":"IEEE Transactions on Very Large Scale Integration (VLSI) Systems","issn_l":"1063-8210","issn":["1063-8210","1557-9999"],"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 Transactions on Very Large Scale Integration (VLSI) Systems","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/7","score":0.571359395980835,"display_name":"Affordable and clean energy"}],"awards":[{"id":"https://openalex.org/G6136965849","display_name":null,"funder_award_id":"RS-2023-NR077140","funder_id":"https://openalex.org/F4320328359","funder_display_name":"Ministry of Science and ICT, South Korea"}],"funders":[{"id":"https://openalex.org/F4320322120","display_name":"National Research Foundation of Korea","ror":"https://ror.org/013aysd81"},{"id":"https://openalex.org/F4320328359","display_name":"Ministry of Science and ICT, South Korea","ror":"https://ror.org/01wpjm123"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":32,"referenced_works":["https://openalex.org/W1983394510","https://openalex.org/W2010202670","https://openalex.org/W2067523571","https://openalex.org/W2194775991","https://openalex.org/W2606722458","https://openalex.org/W2612695082","https://openalex.org/W2790925711","https://openalex.org/W2791673912","https://openalex.org/W2896457183","https://openalex.org/W2906043559","https://openalex.org/W2940862705","https://openalex.org/W2945146780","https://openalex.org/W2963163009","https://openalex.org/W2980104813","https://openalex.org/W2981207549","https://openalex.org/W2997929983","https://openalex.org/W3017521908","https://openalex.org/W3036878841","https://openalex.org/W3097528158","https://openalex.org/W3138516171","https://openalex.org/W3184376546","https://openalex.org/W4285335127","https://openalex.org/W4286300805","https://openalex.org/W4291653336","https://openalex.org/W4310362310","https://openalex.org/W4311887664","https://openalex.org/W4312933868","https://openalex.org/W4393042652","https://openalex.org/W4400232637","https://openalex.org/W4410583693","https://openalex.org/W4412987431","https://openalex.org/W7133205862"],"related_works":[],"abstract_inverted_index":{"Systolic":[0],"array":[1,28,81,89],"architectures":[2],"have":[3],"become":[4],"the":[5,23,187],"dominant":[6],"solution":[7],"for":[8,78,193],"accelerating":[9],"deep":[10],"neural":[11,126],"network":[12],"(DNN)":[13],"computations,":[14],"yet":[15],"designing":[16],"optimal":[17,70],"configurations":[18],"remains":[19],"challenging":[20],"due":[21],"to":[22,66,129],"vast":[24],"design":[25,43,83],"space":[26,44,84],"spanning":[27],"dimensions,":[29],"dataflow":[30,96],"strategies,":[31],"SRAM":[32,47,91],"allocations,":[33],"and":[34,95,110,174,209],"tiling":[35,50,93],"sizes.":[36],"Existing":[37],"tool-aided":[38],"optimization":[39,115],"frameworks":[40],"constrain":[41],"this":[42],"by":[45],"treating":[46],"sizes":[48,51],"or":[49],"as":[52],"fixed":[53],"input":[54],"parameters,":[55],"while":[56,149],"focusing":[57],"only":[58],"on":[59,156,160,207],"conventional":[60],"systolic":[61,79],"arrays,":[62],"limiting":[63],"their":[64],"ability":[65],"discover":[67],"a":[68,75,101,166,178,194,201],"truly":[69],"design.":[71],"This":[72],"article":[73],"presents":[74],"comprehensive":[76],"framework":[77],"tensor":[80,202],"(STA)":[82],"exploration":[85],"that":[86,135,169],"jointly":[87],"optimizes":[88],"configurations,":[90],"sizes,":[92,94],"types.":[97],"We":[98],"introduce":[99],"TensorSim,":[100],"cycle-accurate":[102],"simulator":[103],"with":[104,118],"an":[105,112],"ML-based":[106],"synthesis":[107],"prediction":[108],"model,":[109],"TensorOptimizer,":[111],"automated":[113],"multiobjective":[114],"framework.":[116],"Evaluation":[117],"TensorOptimizer":[119],"across":[120],"12":[121],"representative":[122],"DNNs\u2014from":[123],"edge-level":[124,208],"convolutional":[125],"networks":[127],"(CNNs)":[128],"billion-parameter":[130],"large":[131],"language":[132],"models":[133],"(LLMs)\u2014reveals":[134],"output":[136],"stationary":[137,142],"(OS)":[138],"dataflows":[139,144],"outperform":[140],"weight":[141],"(WS)":[143],"by$1.53\\times":[145],"$on":[146],"edge":[147],"workloads,":[148],"WS":[150,175],"achieves$2.13\\times":[151],"$higher":[152],"efficiency":[153],"over":[154,184,200],"OS":[155,173],"server-scale":[157],"computations.":[158],"Based":[159],"these":[161],"insights,":[162],"we":[163],"propose":[164],"Fused-STA,":[165],"reconfigurable":[167],"architecture":[168],"dynamically":[170],"switches":[171],"between":[172],"modes":[176],"within":[177],"unified":[179],"hardware":[180],"substrate.":[181],"Fused-STA":[182],"achieves":[183],"91%":[185],"of":[186],"oracle":[188],"single-dataflow":[189],"performance,":[190],"specifically":[191],"designed":[192],"single":[195],"network,":[196],"providing$2.34\\times":[197],"$and$1.92\\times":[198],"$improvements":[199],"processing":[203],"unit":[204],"(TPU)-like":[205],"baseline":[206],"server-level":[210],"benchmarks,":[211],"respectively.":[212]},"counts_by_year":[],"updated_date":"2026-08-13T07:04:57.449891","created_date":"2026-05-27T00:00:00"}