{"id":"https://openalex.org/W4386858078","doi":"https://doi.org/10.1109/tcad.2023.3317169","title":"GTCO: Graph and Tensor Co-Design for Transformer-Based Image Recognition on Tensor Cores","display_name":"GTCO: Graph and Tensor Co-Design for Transformer-Based Image Recognition on Tensor Cores","publication_year":2023,"publication_date":"2023-09-19","ids":{"openalex":"https://openalex.org/W4386858078","doi":"https://doi.org/10.1109/tcad.2023.3317169"},"language":"en","primary_location":{"id":"doi:10.1109/tcad.2023.3317169","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcad.2023.3317169","pdf_url":null,"source":{"id":"https://openalex.org/S100835903","display_name":"IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems","issn_l":"0278-0070","issn":["0278-0070","1937-4151"],"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 Computer-Aided Design of Integrated Circuits and 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/A5020529992","display_name":"Yang Bai","orcid":null},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":true,"raw_author_name":"Yang Bai","raw_affiliation_strings":["Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, SAR"],"raw_orcid":"https://orcid.org/0000-0002-5337-1783","affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, SAR","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5033561795","display_name":"Xufeng Yao","orcid":"https://orcid.org/0000-0002-7994-6290"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Xufeng Yao","raw_affiliation_strings":["Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, SAR"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, SAR","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100427492","display_name":"Qi Sun","orcid":"https://orcid.org/0000-0001-5153-6698"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Qi Sun","raw_affiliation_strings":["Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, SAR"],"raw_orcid":"https://orcid.org/0000-0001-5153-6698","affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, SAR","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5013503446","display_name":"Wenqian Zhao","orcid":"https://orcid.org/0000-0001-9501-9254"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Wenqian Zhao","raw_affiliation_strings":["Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, SAR"],"raw_orcid":"https://orcid.org/0000-0001-9501-9254","affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, SAR","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5071519571","display_name":"Shixin Chen","orcid":"https://orcid.org/0000-0002-9401-0482"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Shixin Chen","raw_affiliation_strings":["Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, SAR"],"raw_orcid":"https://orcid.org/0000-0002-9401-0482","affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, SAR","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100742342","display_name":"Zixiao Wang","orcid":"https://orcid.org/0000-0002-5351-0448"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Zixiao Wang","raw_affiliation_strings":["Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, SAR"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, SAR","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5051340429","display_name":"Bei Yu","orcid":"https://orcid.org/0000-0001-6406-4810"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Bei Yu","raw_affiliation_strings":["Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, SAR"],"raw_orcid":"https://orcid.org/0000-0001-6406-4810","affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, SAR","institution_ids":["https://openalex.org/I177725633"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5020529992"],"corresponding_institution_ids":["https://openalex.org/I177725633"],"apc_list":null,"apc_paid":null,"fwci":0.942,"has_fulltext":false,"cited_by_count":8,"citation_normalized_percentile":{"value":0.77433374,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":98},"biblio":{"volume":"43","issue":"2","first_page":"586","last_page":"599"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9976000189781189,"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":0.9976000189781189,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.996999979019165,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.972599983215332,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"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/computer-science","display_name":"Computer science","score":0.750967264175415},{"id":"https://openalex.org/keywords/tensor-contraction","display_name":"Tensor contraction","score":0.5344191193580627},{"id":"https://openalex.org/keywords/intrinsics","display_name":"Intrinsics","score":0.4913412630558014},{"id":"https://openalex.org/keywords/theoretical-computer-science","display_name":"Theoretical computer science","score":0.4735160171985626},{"id":"https://openalex.org/keywords/tensor","display_name":"Tensor (intrinsic definition)","score":0.4464235305786133},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.39818909764289856},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.35939162969589233},{"id":"https://openalex.org/keywords/computational-science","display_name":"Computational science","score":0.33608439564704895},{"id":"https://openalex.org/keywords/tensor-product","display_name":"Tensor product","score":0.163611501455307},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.13835906982421875}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.750967264175415},{"id":"https://openalex.org/C124007464","wikidata":"https://www.wikidata.org/wiki/Q428091","display_name":"Tensor contraction","level":3,"score":0.5344191193580627},{"id":"https://openalex.org/C2908650547","wikidata":"https://www.wikidata.org/wiki/Q20999234","display_name":"Intrinsics","level":2,"score":0.4913412630558014},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.4735160171985626},{"id":"https://openalex.org/C155281189","wikidata":"https://www.wikidata.org/wiki/Q3518150","display_name":"Tensor (intrinsic definition)","level":2,"score":0.4464235305786133},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.39818909764289856},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.35939162969589233},{"id":"https://openalex.org/C459310","wikidata":"https://www.wikidata.org/wiki/Q117801","display_name":"Computational science","level":1,"score":0.33608439564704895},{"id":"https://openalex.org/C51255310","wikidata":"https://www.wikidata.org/wiki/Q1163016","display_name":"Tensor product","level":2,"score":0.163611501455307},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.13835906982421875},{"id":"https://openalex.org/C202444582","wikidata":"https://www.wikidata.org/wiki/Q837863","display_name":"Pure mathematics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tcad.2023.3317169","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcad.2023.3317169","pdf_url":null,"source":{"id":"https://openalex.org/S100835903","display_name":"IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems","issn_l":"0278-0070","issn":["0278-0070","1937-4151"],"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 Computer-Aided Design of Integrated Circuits and Systems","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":45,"referenced_works":["https://openalex.org/W1989988531","https://openalex.org/W2155893237","https://openalex.org/W2183341477","https://openalex.org/W2194775991","https://openalex.org/W2212242068","https://openalex.org/W2295598076","https://openalex.org/W2792643794","https://openalex.org/W2798405286","https://openalex.org/W2896457183","https://openalex.org/W2899663614","https://openalex.org/W2945146780","https://openalex.org/W2962856739","https://openalex.org/W2965058486","https://openalex.org/W2981758446","https://openalex.org/W3033529678","https://openalex.org/W3091795552","https://openalex.org/W3096609285","https://openalex.org/W3159481202","https://openalex.org/W3170841864","https://openalex.org/W4200122815","https://openalex.org/W4206706211","https://openalex.org/W4281489207","https://openalex.org/W4283811336","https://openalex.org/W4287748555","https://openalex.org/W4293024135","https://openalex.org/W4302296459","https://openalex.org/W4319459088","https://openalex.org/W4382202854","https://openalex.org/W4384836452","https://openalex.org/W6637151318","https://openalex.org/W6637373629","https://openalex.org/W6713134421","https://openalex.org/W6739901393","https://openalex.org/W6749029207","https://openalex.org/W6751349269","https://openalex.org/W6752057402","https://openalex.org/W6755207826","https://openalex.org/W6755977528","https://openalex.org/W6779163297","https://openalex.org/W6779728309","https://openalex.org/W6780069040","https://openalex.org/W6780226713","https://openalex.org/W6785429063","https://openalex.org/W6797854001","https://openalex.org/W6846781997"],"related_works":["https://openalex.org/W2783190869","https://openalex.org/W2117014006","https://openalex.org/W2122338824","https://openalex.org/W4233815414","https://openalex.org/W2372170743","https://openalex.org/W1558545464","https://openalex.org/W3211620692","https://openalex.org/W2965967938","https://openalex.org/W4327930469","https://openalex.org/W3101543398"],"abstract_inverted_index":{"Deep":[0],"learning":[1,171],"frameworks":[2],"or":[3],"compilers":[4],"optimize":[5],"the":[6,32,67,72,92,152,168],"operators":[7,37],"in":[8,71],"computation":[9,118],"graph":[10],"using":[11,75],"fixed":[12],"templates":[13],"via":[14],"significant":[15],"engineering":[16],"efforts,":[17],"which":[18],"may":[19],"miss":[20],"potential":[21],"optimizations":[22],"such":[23],"as":[24],"operator":[25,68],"fusion.":[26],"Therefore,":[27],"automatically":[28],"implementing":[29],"and":[30,81,97,109,132],"optimizing":[31],"emerging":[33],"new":[34,87],"combinations":[35],"of":[36,44],"on":[38,63,143,174],"a":[39,52,76,83,110],"specific":[40],"hardware":[41,111,141],"accelerator":[42],"is":[43,124],"importance.":[45],"In":[46],"this":[47],"article,":[48],"we":[49],"introduce":[50],"GTCO,":[51],"tensor":[53,117,134],"compilation":[54],"system":[55],"designed":[56],"to":[57,119,158,167],"accelerate":[58],"transformer-based":[59,73],"vision":[60],"models\u2019":[61],"inference":[62],"GPUs.":[64],"GTCO":[65,150],"tackles":[66],"fusion":[69],"techniques":[70],"model":[74],"novel":[77],"dynamic":[78],"programming":[79],"algorithm":[80],"proposes":[82],"search":[84,107],"policy":[85],"with":[86,113,140,177],"sketch":[88],"generation":[89],"rules":[90],"for":[91],"fused":[93],"batch":[94],"matrix":[95],"multiplication":[96],"softmax":[98],"operators.":[99],"Tensor":[100,178],"programs":[101],"are":[102],"sampled":[103],"from":[104,116],"an":[105],"effective":[106],"space,":[108],"abstraction":[112],"hierarchical":[114],"mapping":[115],"domain-specific":[120],"accelerators":[121],"(Tensor":[122],"Cores)":[123],"formally":[125],"defined.":[126],"Finally,":[127],"our":[128],"framework":[129],"can":[130],"map":[131],"transform":[133],"expression":[135],"into":[136],"efficient":[137],"CUDA":[138],"kernels":[139],"intrinsics":[142],"GPU.":[144],"Our":[145],"experimental":[146],"results":[147],"demonstrate":[148],"that":[149],"improves":[151],"end-to-end":[153],"execution":[154],"performance":[155],"by":[156],"up":[157],"<inline-formula":[159],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[160],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">":[161],"<tex-math":[162],"notation=\"LaTeX\">$1.73\\times":[163],"$":[164],"</tex-math></inline-formula>":[165],"relative":[166],"cutting-edge":[169],"deep":[170],"library":[172],"TensorRT":[173],"NVIDIA":[175],"GPUs":[176],"Cores.":[179]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":4},{"year":2024,"cited_by_count":2},{"year":2023,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}