{"id":"https://openalex.org/W4308075499","doi":"https://doi.org/10.1109/sips55645.2022.9919243","title":"Understanding and Optimizing INT4 Convolution for Accelerated DNN Inference on Tensor Cores","display_name":"Understanding and Optimizing INT4 Convolution for Accelerated DNN Inference on Tensor Cores","publication_year":2022,"publication_date":"2022-11-02","ids":{"openalex":"https://openalex.org/W4308075499","doi":"https://doi.org/10.1109/sips55645.2022.9919243"},"language":"en","primary_location":{"id":"doi:10.1109/sips55645.2022.9919243","is_oa":false,"landing_page_url":"https://doi.org/10.1109/sips55645.2022.9919243","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2022 IEEE Workshop on Signal Processing Systems (SiPS)","raw_type":"proceedings-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/A5073666957","display_name":"Junkyeong Choi","orcid":null},"institutions":[{"id":"https://openalex.org/I4575257","display_name":"Hanyang University","ror":"https://ror.org/046865y68","country_code":"KR","type":"education","lineage":["https://openalex.org/I4575257"]}],"countries":["KR"],"is_corresponding":true,"raw_author_name":"Junkyeong Choi","raw_affiliation_strings":["Hanyang University,Dept. of Electronic Engineering,South Korea","Nota Inc.,South Korea","Nota Inc., South Korea"],"affiliations":[{"raw_affiliation_string":"Hanyang University,Dept. of Electronic Engineering,South Korea","institution_ids":["https://openalex.org/I4575257"]},{"raw_affiliation_string":"Nota Inc.,South Korea","institution_ids":[]},{"raw_affiliation_string":"Nota Inc., South Korea","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5044511282","display_name":"Hyucksung Kwon","orcid":null},"institutions":[{"id":"https://openalex.org/I4575257","display_name":"Hanyang University","ror":"https://ror.org/046865y68","country_code":"KR","type":"education","lineage":["https://openalex.org/I4575257"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Hyucksung Kwon","raw_affiliation_strings":["Hanyang University,Dept. of Electronic Engineering,South Korea","Dept. of Electronic Engineering, Hanyang University, South Korea"],"affiliations":[{"raw_affiliation_string":"Hanyang University,Dept. of Electronic Engineering,South Korea","institution_ids":["https://openalex.org/I4575257"]},{"raw_affiliation_string":"Dept. of Electronic Engineering, Hanyang University, South Korea","institution_ids":["https://openalex.org/I4575257"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042246317","display_name":"Woongkyu Lee","orcid":"https://orcid.org/0000-0001-7106-9564"},"institutions":[{"id":"https://openalex.org/I4575257","display_name":"Hanyang University","ror":"https://ror.org/046865y68","country_code":"KR","type":"education","lineage":["https://openalex.org/I4575257"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Woongkyu Lee","raw_affiliation_strings":["Hanyang University,Dept. of Electronic Engineering,South Korea","Dept. of Electronic Engineering, Hanyang University, South Korea"],"affiliations":[{"raw_affiliation_string":"Hanyang University,Dept. of Electronic Engineering,South Korea","institution_ids":["https://openalex.org/I4575257"]},{"raw_affiliation_string":"Dept. of Electronic Engineering, Hanyang University, South Korea","institution_ids":["https://openalex.org/I4575257"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5068625256","display_name":"Jieun Lim","orcid":"https://orcid.org/0000-0002-8985-1957"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jieun Lim","raw_affiliation_strings":["Nota Inc.,South Korea","Nota Inc., South Korea"],"affiliations":[{"raw_affiliation_string":"Nota Inc.,South Korea","institution_ids":[]},{"raw_affiliation_string":"Nota Inc., South Korea","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101935712","display_name":"Jungwook Choi","orcid":"https://orcid.org/0000-0002-5691-4771"},"institutions":[{"id":"https://openalex.org/I4575257","display_name":"Hanyang University","ror":"https://ror.org/046865y68","country_code":"KR","type":"education","lineage":["https://openalex.org/I4575257"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jungwook Choi","raw_affiliation_strings":["Hanyang University,Dept. of Electronic Engineering,South Korea","Nota Inc.,South Korea","Dept. of Electronic Engineering, Hanyang University, South Korea"],"affiliations":[{"raw_affiliation_string":"Hanyang University,Dept. of Electronic Engineering,South Korea","institution_ids":["https://openalex.org/I4575257"]},{"raw_affiliation_string":"Nota Inc.,South Korea","institution_ids":[]},{"raw_affiliation_string":"Dept. of Electronic Engineering, Hanyang University, South Korea","institution_ids":["https://openalex.org/I4575257"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5073666957"],"corresponding_institution_ids":["https://openalex.org/I4575257"],"apc_list":null,"apc_paid":null,"fwci":0.1007,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.39080117,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":94},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"6"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9995999932289124,"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":0.9995999932289124,"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":0.9991000294685364,"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/T12303","display_name":"Tensor decomposition and applications","score":0.988099992275238,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8012793064117432},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.6315348148345947},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.6215574741363525},{"id":"https://openalex.org/keywords/convolution","display_name":"Convolution (computer science)","score":0.6089890003204346},{"id":"https://openalex.org/keywords/tensor","display_name":"Tensor (intrinsic definition)","score":0.581641674041748},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.5625668168067932},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.5611344575881958},{"id":"https://openalex.org/keywords/reuse","display_name":"Reuse","score":0.5438907146453857},{"id":"https://openalex.org/keywords/matrix-multiplication","display_name":"Matrix multiplication","score":0.4793168306350708},{"id":"https://openalex.org/keywords/schedule","display_name":"Schedule","score":0.46759670972824097},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.4518706798553467},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.4424085319042206},{"id":"https://openalex.org/keywords/matrix","display_name":"Matrix (chemical analysis)","score":0.4343329071998596},{"id":"https://openalex.org/keywords/operand","display_name":"Operand","score":0.43321698904037476},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4229673147201538},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.37134477496147156},{"id":"https://openalex.org/keywords/computational-science","display_name":"Computational science","score":0.3711393177509308},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.35976141691207886},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.287442684173584},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.2173789143562317},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.09467646479606628}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8012793064117432},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.6315348148345947},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.6215574741363525},{"id":"https://openalex.org/C45347329","wikidata":"https://www.wikidata.org/wiki/Q5166604","display_name":"Convolution (computer science)","level":3,"score":0.6089890003204346},{"id":"https://openalex.org/C155281189","wikidata":"https://www.wikidata.org/wiki/Q3518150","display_name":"Tensor (intrinsic definition)","level":2,"score":0.581641674041748},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.5625668168067932},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.5611344575881958},{"id":"https://openalex.org/C206588197","wikidata":"https://www.wikidata.org/wiki/Q846574","display_name":"Reuse","level":2,"score":0.5438907146453857},{"id":"https://openalex.org/C17349429","wikidata":"https://www.wikidata.org/wiki/Q1049914","display_name":"Matrix multiplication","level":3,"score":0.4793168306350708},{"id":"https://openalex.org/C68387754","wikidata":"https://www.wikidata.org/wiki/Q7271585","display_name":"Schedule","level":2,"score":0.46759670972824097},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4518706798553467},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.4424085319042206},{"id":"https://openalex.org/C106487976","wikidata":"https://www.wikidata.org/wiki/Q685816","display_name":"Matrix (chemical analysis)","level":2,"score":0.4343329071998596},{"id":"https://openalex.org/C55526617","wikidata":"https://www.wikidata.org/wiki/Q719375","display_name":"Operand","level":2,"score":0.43321698904037476},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4229673147201538},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.37134477496147156},{"id":"https://openalex.org/C459310","wikidata":"https://www.wikidata.org/wiki/Q117801","display_name":"Computational science","level":1,"score":0.3711393177509308},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.35976141691207886},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.287442684173584},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.2173789143562317},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.09467646479606628},{"id":"https://openalex.org/C84114770","wikidata":"https://www.wikidata.org/wiki/Q46344","display_name":"Quantum","level":2,"score":0.0},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0},{"id":"https://openalex.org/C18903297","wikidata":"https://www.wikidata.org/wiki/Q7150","display_name":"Ecology","level":1,"score":0.0},{"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/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0},{"id":"https://openalex.org/C192562407","wikidata":"https://www.wikidata.org/wiki/Q228736","display_name":"Materials science","level":0,"score":0.0},{"id":"https://openalex.org/C159985019","wikidata":"https://www.wikidata.org/wiki/Q181790","display_name":"Composite material","level":1,"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/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/sips55645.2022.9919243","is_oa":false,"landing_page_url":"https://doi.org/10.1109/sips55645.2022.9919243","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2022 IEEE Workshop on Signal Processing Systems (SiPS)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320321681","display_name":"Ministry of Trade, Industry and Energy","ror":"https://ror.org/008nkqk13"},{"id":"https://openalex.org/F4320335489","display_name":"Institute for Information and Communications Technology Promotion","ror":"https://ror.org/01g0hqq23"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":16,"referenced_works":["https://openalex.org/W2194775991","https://openalex.org/W2618530766","https://openalex.org/W2791673912","https://openalex.org/W2804500013","https://openalex.org/W2895162358","https://openalex.org/W2980186997","https://openalex.org/W3036692157","https://openalex.org/W3043303806","https://openalex.org/W3101543398","https://openalex.org/W3102337469","https://openalex.org/W3205226039","https://openalex.org/W6751349269","https://openalex.org/W6752057402","https://openalex.org/W6779728309","https://openalex.org/W6779898197","https://openalex.org/W6786271619"],"related_works":["https://openalex.org/W2058965144","https://openalex.org/W2164382479","https://openalex.org/W2556282987","https://openalex.org/W1533508804","https://openalex.org/W2146343568","https://openalex.org/W98480971","https://openalex.org/W2891987081","https://openalex.org/W3157543420","https://openalex.org/W4224278821","https://openalex.org/W1535437188"],"abstract_inverted_index":{"Convolution":[0],"is":[1,23,41],"one":[2],"of":[3,7,91,99,122,137],"the":[4,48,70,89,135,138],"fundamental":[5],"operations":[6,121],"deep":[8],"neural":[9,124],"networks":[10,125],"with":[11,30],"demanding":[12],"matrix":[13,26,57],"computation.":[14],"In":[15],"a":[16,24,56],"graphic":[17],"processing":[18,27],"unit":[19],"(GPU),":[20],"Tensor":[21,83,131],"Core":[22,132],"specialized":[25],"hardware":[28],"equipped":[29],"reduced-precision":[31,49],"warp":[32],"matrix-multiply-accumulate":[33],"(WMMA)":[34],"instructions":[35],"to":[36,43,77,126,134],"increase":[37],"throughput.":[38],"However,":[39],"it":[40],"challenging":[42],"achieve":[44],"optimal":[45],"performance":[46],"since":[47],"WMMA":[50,80,114],"requires":[51],"many":[52],"elements":[53],"grouped":[54],"as":[55],"operand,":[58],"seriously":[59],"limiting":[60],"data":[61,104,109],"reuse":[62,90],"and":[63,66,103],"imposing":[64],"packing":[65,95],"layout":[67,105],"overhead":[68,98],"on":[69,82,119,130],"schedule.":[71],"This":[72],"work":[73],"proposes":[74],"three":[75],"techniques":[76,116],"enhance":[78],"INT4":[79,101,113],"utilization":[81],"Cores:":[84],"duplicate-aware":[85],"load":[86],"for":[87,96,107],"increasing":[88],"convolution":[92,120],"input,":[93],"register-level":[94],"alleviating":[97],"handling":[100],"data,":[102],"optimization":[106,115],"coalesced":[108],"transfer.":[110],"The":[111],"proposed":[112],"are":[117],"evaluated":[118],"popular":[123],"demonstrate":[127],"substantial":[128],"speedup":[129],"compared":[133],"state":[136],"art.":[139]},"counts_by_year":[{"year":2024,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}