{"id":"https://openalex.org/W3166036439","doi":"https://doi.org/10.1109/tpds.2021.3084813","title":"Optimizing Depthwise Separable Convolution Operations on GPUs","display_name":"Optimizing Depthwise Separable Convolution Operations on GPUs","publication_year":2021,"publication_date":"2021-05-28","ids":{"openalex":"https://openalex.org/W3166036439","doi":"https://doi.org/10.1109/tpds.2021.3084813","mag":"3166036439"},"language":"en","primary_location":{"id":"doi:10.1109/tpds.2021.3084813","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tpds.2021.3084813","pdf_url":null,"source":{"id":"https://openalex.org/S97130795","display_name":"IEEE Transactions on Parallel and Distributed Systems","issn_l":"1045-9219","issn":["1045-9219","1558-2183","2161-9883"],"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 Parallel and Distributed Systems","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://eprints.whiterose.ac.uk/id/eprint/174797/1/main.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5082302042","display_name":"Gangzhao Lu","orcid":null},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Gangzhao Lu","raw_affiliation_strings":["School of Cyberspace Science, Harbin Institute of Technology, Harbin, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Cyberspace Science, Harbin Institute of Technology, Harbin, China","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5085221201","display_name":"Weizhe Zhang","orcid":"https://orcid.org/0000-0003-4783-876X"},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Weizhe Zhang","raw_affiliation_strings":["School of Cyberspace Science, Harbin Institute of Technology, Harbin, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Cyberspace Science, Harbin Institute of Technology, Harbin, China","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100401045","display_name":"Zheng Wang","orcid":"https://orcid.org/0000-0001-6157-0662"},"institutions":[{"id":"https://openalex.org/I130828816","display_name":"University of Leeds","ror":"https://ror.org/024mrxd33","country_code":"GB","type":"education","lineage":["https://openalex.org/I130828816"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Zheng Wang","raw_affiliation_strings":["School of Computing, University of Leeds, Leeds, U.K"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computing, University of Leeds, Leeds, U.K","institution_ids":["https://openalex.org/I130828816"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":4.6167,"has_fulltext":true,"cited_by_count":81,"citation_normalized_percentile":{"value":0.96030152,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":93,"max":100},"biblio":{"volume":"33","issue":"1","first_page":"70","last_page":"87"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":1.0,"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":1.0,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.9983999729156494,"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"}},{"id":"https://openalex.org/T10812","display_name":"Human Pose and Action Recognition","score":0.9973999857902527,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.888237476348877},{"id":"https://openalex.org/keywords/convolution","display_name":"Convolution (computer science)","score":0.7022106647491455},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.6895030736923218},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.5257904529571533},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.4968469440937042},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.4760144352912903},{"id":"https://openalex.org/keywords/separable-space","display_name":"Separable space","score":0.4674454927444458},{"id":"https://openalex.org/keywords/computational-science","display_name":"Computational science","score":0.4602571427822113},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.45330461859703064},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.44700387120246887},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.37941253185272217},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.3737459182739258},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.22622475028038025},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.1744196116924286}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.888237476348877},{"id":"https://openalex.org/C45347329","wikidata":"https://www.wikidata.org/wiki/Q5166604","display_name":"Convolution (computer science)","level":3,"score":0.7022106647491455},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.6895030736923218},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.5257904529571533},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.4968469440937042},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4760144352912903},{"id":"https://openalex.org/C70710897","wikidata":"https://www.wikidata.org/wiki/Q680081","display_name":"Separable space","level":2,"score":0.4674454927444458},{"id":"https://openalex.org/C459310","wikidata":"https://www.wikidata.org/wiki/Q117801","display_name":"Computational science","level":1,"score":0.4602571427822113},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.45330461859703064},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.44700387120246887},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.37941253185272217},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.3737459182739258},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.22622475028038025},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.1744196116924286},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.0},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.0},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/tpds.2021.3084813","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tpds.2021.3084813","pdf_url":null,"source":{"id":"https://openalex.org/S97130795","display_name":"IEEE Transactions on Parallel and Distributed Systems","issn_l":"1045-9219","issn":["1045-9219","1558-2183","2161-9883"],"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 Parallel and Distributed Systems","raw_type":"journal-article"},{"id":"pmh:oai:eprints.whiterose.ac.uk:174797","is_oa":true,"landing_page_url":"https://orcid.org/0000-0001-6157-0662>","pdf_url":"https://eprints.whiterose.ac.uk/id/eprint/174797/1/main.pdf","source":{"id":"https://openalex.org/S4306400854","display_name":"White Rose Research Online (University of Leeds, The University of Sheffield, University of York)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I2800616092","host_organization_name":"White Rose University Consortium","host_organization_lineage":["https://openalex.org/I2800616092"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"}],"best_oa_location":{"id":"pmh:oai:eprints.whiterose.ac.uk:174797","is_oa":true,"landing_page_url":"https://orcid.org/0000-0001-6157-0662>","pdf_url":"https://eprints.whiterose.ac.uk/id/eprint/174797/1/main.pdf","source":{"id":"https://openalex.org/S4306400854","display_name":"White Rose Research Online (University of Leeds, The University of Sheffield, University of York)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I2800616092","host_organization_name":"White Rose University Consortium","host_organization_lineage":["https://openalex.org/I2800616092"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"sustainable_development_goals":[{"score":0.47999998927116394,"display_name":"Affordable and clean energy","id":"https://metadata.un.org/sdg/7"}],"awards":[{"id":"https://openalex.org/G1053123284","display_name":"\u5d4c\u5165\u5f0f\u7cfb\u7edf\u5b9e\u65f6\u4efb\u52a1\u8c03\u5ea6\u7b97\u6cd5\u4e0e\u53ef\u8c03\u5ea6\u6027\u5206\u6790\u7406\u8bba\u7814\u7a76","funder_award_id":"61672186","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G1626709283","display_name":null,"funder_award_id":"61672186","funder_id":"https://openalex.org/F4320335777","funder_display_name":"National Key Research and Development Program of China"},{"id":"https://openalex.org/G1719090559","display_name":null,"funder_award_id":"2019B010136001","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G4583476002","display_name":null,"funder_award_id":"61872294","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G6010123351","display_name":null,"funder_award_id":"JCYJ20190806143418198","funder_id":"https://openalex.org/F4320335777","funder_display_name":"National Key Research and Development Program of China"},{"id":"https://openalex.org/G7006078021","display_name":null,"funder_award_id":"JCYJ20190806143418198","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G7045393196","display_name":null,"funder_award_id":"2019B010136001","funder_id":"https://openalex.org/F4320335777","funder_display_name":"National Key Research and Development Program of China"},{"id":"https://openalex.org/G8910962047","display_name":null,"funder_award_id":"2017YFB0202901","funder_id":"https://openalex.org/F4320335777","funder_display_name":"National Key Research and Development Program of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320335777","display_name":"National Key Research and Development Program of China","ror":null}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W3166036439.pdf","grobid_xml":"https://content.openalex.org/works/W3166036439.grobid-xml"},"referenced_works_count":74,"referenced_works":["https://openalex.org/W1530262073","https://openalex.org/W1667652561","https://openalex.org/W1789336918","https://openalex.org/W1922123711","https://openalex.org/W1994869533","https://openalex.org/W2072912074","https://openalex.org/W2097117768","https://openalex.org/W2108598243","https://openalex.org/W2163687928","https://openalex.org/W2172654076","https://openalex.org/W2302330221","https://openalex.org/W2402144811","https://openalex.org/W2410272182","https://openalex.org/W2510697685","https://openalex.org/W2605739168","https://openalex.org/W2670151302","https://openalex.org/W2786374423","https://openalex.org/W2796649226","https://openalex.org/W2804777790","https://openalex.org/W2885184702","https://openalex.org/W2889647155","https://openalex.org/W2899366056","https://openalex.org/W2913790721","https://openalex.org/W2914086833","https://openalex.org/W2950649068","https://openalex.org/W2951810231","https://openalex.org/W2953384591","https://openalex.org/W2955425717","https://openalex.org/W2963069637","https://openalex.org/W2963163009","https://openalex.org/W2963340555","https://openalex.org/W2963367891","https://openalex.org/W2963887620","https://openalex.org/W2963947383","https://openalex.org/W2964017942","https://openalex.org/W2964157630","https://openalex.org/W2964301839","https://openalex.org/W2969585684","https://openalex.org/W2971383048","https://openalex.org/W2981751377","https://openalex.org/W2982083293","https://openalex.org/W2983655274","https://openalex.org/W2983670846","https://openalex.org/W2995429865","https://openalex.org/W2997471409","https://openalex.org/W2997999719","https://openalex.org/W2998957070","https://openalex.org/W3008378296","https://openalex.org/W3009472344","https://openalex.org/W3012249773","https://openalex.org/W3022446491","https://openalex.org/W3034326778","https://openalex.org/W3034339621","https://openalex.org/W3034681889","https://openalex.org/W3035421056","https://openalex.org/W3041406277","https://openalex.org/W3043216033","https://openalex.org/W3043303806","https://openalex.org/W3095498018","https://openalex.org/W4302296459","https://openalex.org/W6631660994","https://openalex.org/W6637151318","https://openalex.org/W6638020065","https://openalex.org/W6640442106","https://openalex.org/W6713134421","https://openalex.org/W6740373327","https://openalex.org/W6750448596","https://openalex.org/W6752390237","https://openalex.org/W6754777574","https://openalex.org/W6762718338","https://openalex.org/W6769580333","https://openalex.org/W6771298329","https://openalex.org/W6780635486","https://openalex.org/W6784701996"],"related_works":["https://openalex.org/W2009525028","https://openalex.org/W4321064619","https://openalex.org/W2530952058","https://openalex.org/W2582836483","https://openalex.org/W4299366318","https://openalex.org/W3158431807","https://openalex.org/W4297580547","https://openalex.org/W2951583185","https://openalex.org/W2531409750","https://openalex.org/W4308155352"],"abstract_inverted_index":{"The":[0],"depthwise":[1,30,86],"separable":[2,31,87],"convolution":[3,111],"is":[4,13,200],"commonly":[5],"seen":[6],"in":[7,71],"convolutional":[8],"neural":[9],"networks":[10],"(CNNs),":[11],"and":[12,60,106,124,153,173,181,189,243,247,251,259,263],"widely":[14],"used":[15],"to":[16,47,80,102,113,140,149,154,218],"reduce":[17,114],"the":[18,61,68,82,91,104,110,115,122,125,129,143,156,204,237,253],"computation":[19],"overhead":[20],"of":[21,29,45,64,84,109,117,128,241,257],"a":[22,42,72,135,229],"standard":[23],"multi-channel":[24],"2D":[25,130],"convolution.":[26,131],"Existing":[27],"implementations":[28],"convolutions":[32,88],"target":[33],"accelerating":[34],"model":[35,58,65,69],"training":[36,59,239],"with":[37,41],"large":[38,43],"batch":[39,231],"sizes":[40],"number":[44,116],"samples":[46,74],"be":[48],"processed":[49],"at":[50,75],"once.":[51,76],"Such":[52],"approaches":[53],"are":[54],"inadequate":[55],"for":[56,203],"small-batch-sized":[57],"typical":[62],"scenario":[63],"inference":[66,255],"where":[67],"takes":[70],"few":[73],"This":[77],"article":[78],"aims":[79],"bridge":[81],"gap":[83],"optimizing":[85],"by":[89,97,245,261],"targeting":[90],"GPU":[92,147,151,166,172,206],"architecture.":[93,207],"We":[94,160,193,224],"achieve":[95],"this":[96],"designing":[98],"two":[99,165,182],"novel":[100],"algorithms":[101],"improve":[103,150],"column":[105],"row":[107],"reuse":[108],"operation":[112],"memory":[118,157],"operations":[119],"performed":[120],"on":[121,164],"width":[123],"height":[126],"dimensions":[127],"Our":[132],"approach":[133,163,196,213,234],"employs":[134],"dynamic":[136],"tile":[137],"size":[138],"scheme":[139],"adaptively":[141],"distribute":[142],"computational":[144],"data":[145,183],"across":[146],"threads":[148],"utilization":[152],"hide":[155],"access":[158],"latency.":[159],"apply":[161],"our":[162,195,212,233],"platforms:":[167],"an":[168,174],"NVIDIA":[169,176,205],"RTX":[170],"2080Ti":[171],"embedded":[175],"Jetson":[177],"AGX":[178],"Xavier":[179],"GPU,":[180],"types:":[184],"32-bit":[185],"floating":[186],"point":[187],"(FP32)":[188],"8-bit":[190],"integer":[191],"(INT8).":[192],"compared":[194],"against":[197],"cuDNN":[198],"that":[199,211],"heavily":[201],"tuned":[202],"Experimental":[208],"results":[209],"show":[210,225],"delivers":[214],"over":[215,222],"2\u00d7":[216],"(up":[217],"3\u00d7)":[219],"performance":[220],"improvement":[221],"cuDNN.":[223],"that,":[226],"when":[227],"using":[228],"moderate":[230],"size,":[232],"averagely":[235],"reduces":[236,252],"end-to-end":[238,254],"time":[240,256],"MobileNet":[242,258],"EfficientNet":[244,260],"9.7":[246],"7.3":[248],"percent":[249,265],"respectively,":[250],"12.2":[262],"11.6":[264],"respectively.":[266]},"counts_by_year":[{"year":2026,"cited_by_count":7},{"year":2025,"cited_by_count":25},{"year":2024,"cited_by_count":23},{"year":2023,"cited_by_count":16},{"year":2022,"cited_by_count":8},{"year":2021,"cited_by_count":2}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}