{"id":"https://openalex.org/W6959950620","doi":"https://doi.org/10.1109/tvlsi.2025.3587264","title":"Comma: A Communication-Minimized Model-Architecture Framework for Efficient Convolution Acceleration","display_name":"Comma: A Communication-Minimized Model-Architecture Framework for Efficient Convolution Acceleration","publication_year":2025,"publication_date":"2025-07-18","ids":{"openalex":"https://openalex.org/W6959950620","doi":"https://doi.org/10.1109/tvlsi.2025.3587264"},"language":"en","primary_location":{"id":"doi:10.1109/tvlsi.2025.3587264","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tvlsi.2025.3587264","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":null,"display_name":"Wendi Sun","orcid":"https://orcid.org/0009-0002-3198-2345"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wendi Sun","raw_affiliation_strings":["School of Microelectronics, University of Science and Technology of China, Hefei, Anhui, China"],"raw_orcid":"https://orcid.org/0009-0002-3198-2345","affiliations":[{"raw_affiliation_string":"School of Microelectronics, University of Science and Technology of China, Hefei, Anhui, China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Yifan Wang","orcid":"https://orcid.org/0009-0006-4017-8913"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yifan Wang","raw_affiliation_strings":["School of Microelectronics, University of Science and Technology of China, Hefei, Anhui, China"],"raw_orcid":"https://orcid.org/0009-0006-4017-8913","affiliations":[{"raw_affiliation_string":"School of Microelectronics, University of Science and Technology of China, Hefei, Anhui, China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Xu Cao","orcid":null},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xu Cao","raw_affiliation_strings":["School of Microelectronics, University of Science and Technology of China, Hefei, Anhui, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Microelectronics, University of Science and Technology of China, Hefei, Anhui, China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Yao Ge","orcid":"https://orcid.org/0009-0006-8230-4201"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yao Ge","raw_affiliation_strings":["School of Microelectronics, University of Science and Technology of China, Hefei, Anhui, China"],"raw_orcid":"https://orcid.org/0009-0006-8230-4201","affiliations":[{"raw_affiliation_string":"School of Microelectronics, University of Science and Technology of China, Hefei, Anhui, China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Song Chen","orcid":"https://orcid.org/0000-0003-0341-3428"},"institutions":[{"id":"https://openalex.org/I4210137491","display_name":"National Science Centre","ror":"https://ror.org/03ha2q922","country_code":"PL","type":"government","lineage":["https://openalex.org/I4210137491"]}],"countries":["PL"],"is_corresponding":false,"raw_author_name":"Song Chen","raw_affiliation_strings":["School of Microelectronics and the Institute of Artificial Intelligence, Hefei Comprehensive National Science Center, University of Science and Technology of China, Hefei, Anhui, China"],"raw_orcid":"https://orcid.org/0000-0003-0341-3428","affiliations":[{"raw_affiliation_string":"School of Microelectronics and the Institute of Artificial Intelligence, Hefei Comprehensive National Science Center, University of Science and Technology of China, Hefei, Anhui, China","institution_ids":["https://openalex.org/I4210137491"]}]},{"author_position":"last","author":{"id":null,"display_name":"Yi Kang","orcid":"https://orcid.org/0000-0002-5487-6855"},"institutions":[{"id":"https://openalex.org/I4210137491","display_name":"National Science Centre","ror":"https://ror.org/03ha2q922","country_code":"PL","type":"government","lineage":["https://openalex.org/I4210137491"]}],"countries":["PL"],"is_corresponding":false,"raw_author_name":"Yi Kang","raw_affiliation_strings":["School of Microelectronics and the Institute of Artificial Intelligence, Hefei Comprehensive National Science Center, University of Science and Technology of China, Hefei, Anhui, China"],"raw_orcid":"https://orcid.org/0000-0002-5487-6855","affiliations":[{"raw_affiliation_string":"School of Microelectronics and the Institute of Artificial Intelligence, Hefei Comprehensive National Science Center, University of Science and Technology of China, Hefei, Anhui, China","institution_ids":["https://openalex.org/I4210137491"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":6,"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":0.4573978,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"33","issue":"10","first_page":"2824","last_page":"2837"},"is_retracted":false,"is_paratext":false,"is_xpac":true,"primary_topic":{"id":"https://openalex.org/T12568","display_name":"Plant Taxonomy and Phylogenetics","score":0.600600004196167,"subfield":{"id":"https://openalex.org/subfields/1105","display_name":"Ecology, Evolution, Behavior and Systematics"},"field":{"id":"https://openalex.org/fields/11","display_name":"Agricultural and Biological Sciences"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},"topics":[{"id":"https://openalex.org/T12568","display_name":"Plant Taxonomy and Phylogenetics","score":0.600600004196167,"subfield":{"id":"https://openalex.org/subfields/1105","display_name":"Ecology, Evolution, Behavior and Systematics"},"field":{"id":"https://openalex.org/fields/11","display_name":"Agricultural and Biological Sciences"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},{"id":"https://openalex.org/T11229","display_name":"Genetics and Plant Breeding","score":0.0885000005364418,"subfield":{"id":"https://openalex.org/subfields/1110","display_name":"Plant Science"},"field":{"id":"https://openalex.org/fields/11","display_name":"Agricultural and Biological Sciences"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},{"id":"https://openalex.org/T10012","display_name":"Genetic diversity and population structure","score":0.08320000022649765,"subfield":{"id":"https://openalex.org/subfields/1311","display_name":"Genetics"},"field":{"id":"https://openalex.org/fields/13","display_name":"Biochemistry, Genetics and Molecular Biology"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/dataflow","display_name":"Dataflow","score":0.8465999960899353},{"id":"https://openalex.org/keywords/reduction","display_name":"Reduction (mathematics)","score":0.7519999742507935},{"id":"https://openalex.org/keywords/workflow","display_name":"Workflow","score":0.7387999892234802},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.6355000138282776},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.5530999898910522},{"id":"https://openalex.org/keywords/convolution","display_name":"Convolution (computer science)","score":0.517799973487854},{"id":"https://openalex.org/keywords/energy","display_name":"Energy (signal processing)","score":0.5171999931335449},{"id":"https://openalex.org/keywords/acceleration","display_name":"Acceleration","score":0.5019999742507935}],"concepts":[{"id":"https://openalex.org/C96324660","wikidata":"https://www.wikidata.org/wiki/Q205446","display_name":"Dataflow","level":2,"score":0.8465999960899353},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.7519999742507935},{"id":"https://openalex.org/C177212765","wikidata":"https://www.wikidata.org/wiki/Q627335","display_name":"Workflow","level":2,"score":0.7387999892234802},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7215999960899353},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.6355000138282776},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.5530999898910522},{"id":"https://openalex.org/C45347329","wikidata":"https://www.wikidata.org/wiki/Q5166604","display_name":"Convolution (computer science)","level":3,"score":0.517799973487854},{"id":"https://openalex.org/C186370098","wikidata":"https://www.wikidata.org/wiki/Q442787","display_name":"Energy (signal processing)","level":2,"score":0.5171999931335449},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.5019999742507935},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.49570000171661377},{"id":"https://openalex.org/C2742236","wikidata":"https://www.wikidata.org/wiki/Q924713","display_name":"Efficient energy use","level":2,"score":0.4909000098705292},{"id":"https://openalex.org/C206588197","wikidata":"https://www.wikidata.org/wiki/Q846574","display_name":"Reuse","level":2,"score":0.4507000148296356},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.43959999084472656},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.4196999967098236},{"id":"https://openalex.org/C2780165032","wikidata":"https://www.wikidata.org/wiki/Q16869822","display_name":"Energy consumption","level":2,"score":0.3952000141143799},{"id":"https://openalex.org/C47487241","wikidata":"https://www.wikidata.org/wiki/Q5227230","display_name":"Data access","level":2,"score":0.38609999418258667},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.3668000102043152},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.325300008058548},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.2915000021457672},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.2827000021934509},{"id":"https://openalex.org/C77553402","wikidata":"https://www.wikidata.org/wiki/Q13222579","display_name":"Upper and lower bounds","level":2,"score":0.28130000829696655},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.2696000039577484},{"id":"https://openalex.org/C489000","wikidata":"https://www.wikidata.org/wiki/Q747385","display_name":"Data flow diagram","level":2,"score":0.2628999948501587},{"id":"https://openalex.org/C64260653","wikidata":"https://www.wikidata.org/wiki/Q1194864","display_name":"Electronic design automation","level":2,"score":0.2624000012874603},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.2605000138282776}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tvlsi.2025.3587264","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tvlsi.2025.3587264","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":[],"awards":[{"id":"https://openalex.org/G1085865041","display_name":null,"funder_award_id":"92473114","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G8638999836","display_name":null,"funder_award_id":"XDB0660000","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":31,"referenced_works":["https://openalex.org/W2015861736","https://openalex.org/W2017369466","https://openalex.org/W2152839228","https://openalex.org/W2155893237","https://openalex.org/W2289252105","https://openalex.org/W2295680811","https://openalex.org/W2442974303","https://openalex.org/W2605487586","https://openalex.org/W2606722458","https://openalex.org/W2612076670","https://openalex.org/W2618530766","https://openalex.org/W2725159389","https://openalex.org/W2773339846","https://openalex.org/W2795915628","https://openalex.org/W2798317693","https://openalex.org/W2807091852","https://openalex.org/W3010389911","https://openalex.org/W3017059173","https://openalex.org/W3136346557","https://openalex.org/W3158233068","https://openalex.org/W3185974615","https://openalex.org/W3214917546","https://openalex.org/W4210588783","https://openalex.org/W4214493665","https://openalex.org/W4220972538","https://openalex.org/W4240168186","https://openalex.org/W4247470470","https://openalex.org/W4320712936","https://openalex.org/W4389474030","https://openalex.org/W4399487612","https://openalex.org/W4400579220"],"related_works":[],"abstract_inverted_index":{"Convolutional":[0],"neural":[1],"networks":[2],"(CNNs)":[3],"are":[4,118],"still":[5],"indispensable":[6],"nowadays,":[7],"but":[8],"they":[9],"bring":[10],"a":[11,31,158],"dominantly":[12],"large":[13],"overhead":[14],"of":[15,95,107,170],"data":[16,61],"communication.":[17],"Current":[18],"works":[19],"mainly":[20],"focus":[21],"on":[22,59],"prior":[23],"off-chip":[24,43,53,83,128],"or":[25],"intuitive/heuristic":[26],"access":[27,93,129,269],"optimization":[28],"and":[29,42,64,100,116,124,127,150,179,189,211],"lack":[30],"theoretically":[32],"holistic":[33,159],"analysis.":[34],"Therefore,":[35],"this":[36],"article":[37],"proposes":[38],"an":[39,151],"overall":[40,125],"on-":[41,126],"communication-minimized":[44],"CNN":[45,155],"acceleration":[46],"framework.":[47],"First,":[48],"we":[49,85,145,259],"derive":[50],"the":[51,91,103,121,135,168,253,273],"layer-wise":[52],"communication":[54,114],"analytic":[55],"expressions":[56],"(AEs)":[57],"based":[58],"different":[60],"reuse":[62],"strategies":[63],"give":[65],"their":[66],"lower":[67,72],"bound":[68],"(LB),":[69],"reaching":[70],"6%\u201318%":[71],"than":[73],"that":[74],"proposed":[75,165],"in":[76,176],"work":[77,96,108,232],"<italic":[78,97,109,161,208,233,240],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[79,98,110,162,182,196,209,213,222,234,241,263],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">CLB</i>.":[80],"For":[81,252],"practical":[82],"access,":[84],"achieve":[86],"22%\u201335%":[87],"reduction":[88,193,219,247],"compared":[89,206,230,271],"with":[90,102,207,231,248,272],"adaptive":[92],"strategy":[94],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">Sense</i>":[99],"11%":[101],"dataflow":[104,149],"selection":[105],"approach":[106],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">Versa</i>.":[111],"Second,":[112],"on-chip":[113,203],"AEs":[115],"LB":[117,138],"derived":[119],"within":[120],"tile":[122],"range,":[123],"analysis":[130],"is":[131,164,174],"conducted":[132],"to":[133,166],"obtain":[134],"unified":[136],"energy":[137,192,246,270],"(ELB).":[139],"Guided":[140],"by":[141],"our":[142],"theoretical":[143],"model,":[144],"design":[146],"window-primitive":[147],"(WP)":[148],"energy-efficient":[152],"systolic-cross-line":[153],"(SCL)":[154],"accelerator.":[156],"Finally,":[157],"framework,":[160],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">Comma</i>,":[163],"streamline":[167],"workflow":[169],"model-architecture":[171],"co-design.":[172],"SCL":[173],"implemented":[175],"65-nm":[177],"technology":[178],"achieves":[180,244],"<inline-formula":[181,195,262],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">":[183,197,214,223,264],"<tex-math":[184,198,215,224,265],"notation=\"LaTeX\">$4.2\\times":[185],"$":[186,200,217,226,267],"</tex-math></inline-formula>":[187,201,218,227,268],"pJ/Multiply":[188],"ACcumulation":[190],"(MAC)":[191],"at":[194,220],"notation=\"LaTeX\">$1.4\\times":[199],"less":[202,228],"memory":[204,255],"area":[205,229,250],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">Eyeriss</i>,":[210],"1.4~<inline-formula":[212],"notation=\"LaTeX\">$1.9\\times":[216],"3~<inline-formula":[221],"notation=\"LaTeX\">$4\\times":[225],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">CLB</i>;":[235],"as":[236],"for":[237],"systolic":[238,274],"architecture,":[239],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">Sense</i>,":[242],"it":[243],"29%\u201344%":[245],"4%":[249],"overhead.":[251],"near":[254],"processing":[256],"(NMP)":[257],"situation,":[258],"reduce":[260],"around":[261],"notation=\"LaTeX\">$2\\times":[266],"NMP":[275],"architecture.":[276]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}