{"id":"https://openalex.org/W4308090217","doi":"https://doi.org/10.1109/hpec55821.2022.9926290","title":"Towards Fast GPU-based Sparse DNN Inference: A Hybrid Compute Model","display_name":"Towards Fast GPU-based Sparse DNN Inference: A Hybrid Compute Model","publication_year":2022,"publication_date":"2022-09-19","ids":{"openalex":"https://openalex.org/W4308090217","doi":"https://doi.org/10.1109/hpec55821.2022.9926290"},"language":"en","primary_location":{"id":"doi:10.1109/hpec55821.2022.9926290","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpec55821.2022.9926290","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2022 IEEE High Performance Extreme Computing Conference (HPEC)","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/A5100567290","display_name":"Shaoxian Xu","orcid":null},"institutions":[{"id":"https://openalex.org/I47720641","display_name":"Huazhong University of Science and Technology","ror":"https://ror.org/00p991c53","country_code":"CN","type":"education","lineage":["https://openalex.org/I47720641"]},{"id":"https://openalex.org/I4210123185","display_name":"Zhejiang Lab","ror":"https://ror.org/02m2h7991","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210123185"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Shaoxian Xu","raw_affiliation_strings":["National Engineering Research Center for Big Data Technology and System/Services Computing Technology, School of Computer Science and Technology, Huazhong University of Science and Technology,System Lab/Cluster and Grid Computing Lab,Wuhan,China,430074","Zhejiang Lab, Research Center for Graph Computing, Hangzhou, China"],"affiliations":[{"raw_affiliation_string":"National Engineering Research Center for Big Data Technology and System/Services Computing Technology, School of Computer Science and Technology, Huazhong University of Science and Technology,System Lab/Cluster and Grid Computing Lab,Wuhan,China,430074","institution_ids":["https://openalex.org/I47720641"]},{"raw_affiliation_string":"Zhejiang Lab, Research Center for Graph Computing, Hangzhou, China","institution_ids":["https://openalex.org/I4210123185"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104031837","display_name":"Minkang Wu","orcid":null},"institutions":[{"id":"https://openalex.org/I4210123185","display_name":"Zhejiang Lab","ror":"https://ror.org/02m2h7991","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210123185"]},{"id":"https://openalex.org/I47720641","display_name":"Huazhong University of Science and Technology","ror":"https://ror.org/00p991c53","country_code":"CN","type":"education","lineage":["https://openalex.org/I47720641"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Minkang Wu","raw_affiliation_strings":["National Engineering Research Center for Big Data Technology and System/Services Computing Technology, School of Computer Science and Technology, Huazhong University of Science and Technology,System Lab/Cluster and Grid Computing Lab,Wuhan,China,430074","Zhejiang Lab, Research Center for Graph Computing, Hangzhou, China"],"affiliations":[{"raw_affiliation_string":"National Engineering Research Center for Big Data Technology and System/Services Computing Technology, School of Computer Science and Technology, Huazhong University of Science and Technology,System Lab/Cluster and Grid Computing Lab,Wuhan,China,430074","institution_ids":["https://openalex.org/I47720641"]},{"raw_affiliation_string":"Zhejiang Lab, Research Center for Graph Computing, Hangzhou, China","institution_ids":["https://openalex.org/I4210123185"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103241222","display_name":"Long Zheng","orcid":"https://orcid.org/0000-0002-5812-7317"},"institutions":[{"id":"https://openalex.org/I4210123185","display_name":"Zhejiang Lab","ror":"https://ror.org/02m2h7991","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210123185"]},{"id":"https://openalex.org/I47720641","display_name":"Huazhong University of Science and Technology","ror":"https://ror.org/00p991c53","country_code":"CN","type":"education","lineage":["https://openalex.org/I47720641"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Long Zheng","raw_affiliation_strings":["National Engineering Research Center for Big Data Technology and System/Services Computing Technology, School of Computer Science and Technology, Huazhong University of Science and Technology,System Lab/Cluster and Grid Computing Lab,Wuhan,China,430074","Zhejiang Lab, Research Center for Graph Computing, Hangzhou, China"],"affiliations":[{"raw_affiliation_string":"National Engineering Research Center for Big Data Technology and System/Services Computing Technology, School of Computer Science and Technology, Huazhong University of Science and Technology,System Lab/Cluster and Grid Computing Lab,Wuhan,China,430074","institution_ids":["https://openalex.org/I47720641"]},{"raw_affiliation_string":"Zhejiang Lab, Research Center for Graph Computing, Hangzhou, China","institution_ids":["https://openalex.org/I4210123185"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5066761150","display_name":"Zhiyuan Shao","orcid":"https://orcid.org/0000-0003-2139-6465"},"institutions":[{"id":"https://openalex.org/I4210123185","display_name":"Zhejiang Lab","ror":"https://ror.org/02m2h7991","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210123185"]},{"id":"https://openalex.org/I47720641","display_name":"Huazhong University of Science and Technology","ror":"https://ror.org/00p991c53","country_code":"CN","type":"education","lineage":["https://openalex.org/I47720641"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhiyuan Shao","raw_affiliation_strings":["National Engineering Research Center for Big Data Technology and System/Services Computing Technology, School of Computer Science and Technology, Huazhong University of Science and Technology,System Lab/Cluster and Grid Computing Lab,Wuhan,China,430074","Zhejiang Lab, Research Center for Graph Computing, Hangzhou, China"],"affiliations":[{"raw_affiliation_string":"National Engineering Research Center for Big Data Technology and System/Services Computing Technology, School of Computer Science and Technology, Huazhong University of Science and Technology,System Lab/Cluster and Grid Computing Lab,Wuhan,China,430074","institution_ids":["https://openalex.org/I47720641"]},{"raw_affiliation_string":"Zhejiang Lab, Research Center for Graph Computing, Hangzhou, China","institution_ids":["https://openalex.org/I4210123185"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5114546959","display_name":"Xiangyu Ye","orcid":"https://orcid.org/0009-0008-1285-5176"},"institutions":[{"id":"https://openalex.org/I4210123185","display_name":"Zhejiang Lab","ror":"https://ror.org/02m2h7991","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210123185"]},{"id":"https://openalex.org/I47720641","display_name":"Huazhong University of Science and Technology","ror":"https://ror.org/00p991c53","country_code":"CN","type":"education","lineage":["https://openalex.org/I47720641"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiangyu Ye","raw_affiliation_strings":["National Engineering Research Center for Big Data Technology and System/Services Computing Technology, School of Computer Science and Technology, Huazhong University of Science and Technology,System Lab/Cluster and Grid Computing Lab,Wuhan,China,430074","Zhejiang Lab, Research Center for Graph Computing, Hangzhou, China"],"affiliations":[{"raw_affiliation_string":"National Engineering Research Center for Big Data Technology and System/Services Computing Technology, School of Computer Science and Technology, Huazhong University of Science and Technology,System Lab/Cluster and Grid Computing Lab,Wuhan,China,430074","institution_ids":["https://openalex.org/I47720641"]},{"raw_affiliation_string":"Zhejiang Lab, Research Center for Graph Computing, Hangzhou, China","institution_ids":["https://openalex.org/I4210123185"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5022398389","display_name":"Xiaofei Liao","orcid":"https://orcid.org/0000-0001-6302-813X"},"institutions":[{"id":"https://openalex.org/I47720641","display_name":"Huazhong University of Science and Technology","ror":"https://ror.org/00p991c53","country_code":"CN","type":"education","lineage":["https://openalex.org/I47720641"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiaofei Liao","raw_affiliation_strings":["National Engineering Research Center for Big Data Technology and System/Services Computing Technology, School of Computer Science and Technology, Huazhong University of Science and Technology,System Lab/Cluster and Grid Computing Lab,Wuhan,China,430074"],"affiliations":[{"raw_affiliation_string":"National Engineering Research Center for Big Data Technology and System/Services Computing Technology, School of Computer Science and Technology, Huazhong University of Science and Technology,System Lab/Cluster and Grid Computing Lab,Wuhan,China,430074","institution_ids":["https://openalex.org/I47720641"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5022262922","display_name":"Hai Jin","orcid":"https://orcid.org/0000-0002-3934-7605"},"institutions":[{"id":"https://openalex.org/I47720641","display_name":"Huazhong University of Science and Technology","ror":"https://ror.org/00p991c53","country_code":"CN","type":"education","lineage":["https://openalex.org/I47720641"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hai Jin","raw_affiliation_strings":["National Engineering Research Center for Big Data Technology and System/Services Computing Technology, School of Computer Science and Technology, Huazhong University of Science and Technology,System Lab/Cluster and Grid Computing Lab,Wuhan,China,430074"],"affiliations":[{"raw_affiliation_string":"National Engineering Research Center for Big Data Technology and System/Services Computing Technology, School of Computer Science and Technology, Huazhong University of Science and Technology,System Lab/Cluster and Grid Computing Lab,Wuhan,China,430074","institution_ids":["https://openalex.org/I47720641"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5100567290"],"corresponding_institution_ids":["https://openalex.org/I4210123185","https://openalex.org/I47720641"],"apc_list":null,"apc_paid":null,"fwci":0.5033,"has_fulltext":false,"cited_by_count":5,"citation_normalized_percentile":{"value":0.64071265,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":94,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"7"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9991000294685364,"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":0.9991000294685364,"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/T11273","display_name":"Advanced Graph Neural Networks","score":0.9983000159263611,"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/T11948","display_name":"Machine Learning in Materials Science","score":0.9979000091552734,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials 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":0.8406699895858765},{"id":"https://openalex.org/keywords/sparse-matrix","display_name":"Sparse matrix","score":0.6881136298179626},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.6792598962783813},{"id":"https://openalex.org/keywords/memory-bandwidth","display_name":"Memory bandwidth","score":0.578842282295227},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5199128985404968},{"id":"https://openalex.org/keywords/block-matrix","display_name":"Block matrix","score":0.49528399109840393},{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.4937673509120941},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.4814262092113495},{"id":"https://openalex.org/keywords/memory-footprint","display_name":"Memory footprint","score":0.45716750621795654},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.44693076610565186},{"id":"https://openalex.org/keywords/supercomputer","display_name":"Supercomputer","score":0.43758291006088257},{"id":"https://openalex.org/keywords/theoretical-computer-science","display_name":"Theoretical computer science","score":0.35048627853393555},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.34939008951187134},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.2169581651687622}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8406699895858765},{"id":"https://openalex.org/C56372850","wikidata":"https://www.wikidata.org/wiki/Q1050404","display_name":"Sparse matrix","level":3,"score":0.6881136298179626},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.6792598962783813},{"id":"https://openalex.org/C188045654","wikidata":"https://www.wikidata.org/wiki/Q17148339","display_name":"Memory bandwidth","level":2,"score":0.578842282295227},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5199128985404968},{"id":"https://openalex.org/C85817219","wikidata":"https://www.wikidata.org/wiki/Q884772","display_name":"Block matrix","level":3,"score":0.49528399109840393},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.4937673509120941},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4814262092113495},{"id":"https://openalex.org/C74912251","wikidata":"https://www.wikidata.org/wiki/Q6815727","display_name":"Memory footprint","level":2,"score":0.45716750621795654},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.44693076610565186},{"id":"https://openalex.org/C83283714","wikidata":"https://www.wikidata.org/wiki/Q121117","display_name":"Supercomputer","level":2,"score":0.43758291006088257},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.35048627853393555},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.34939008951187134},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2169581651687622},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"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/C62520636","wikidata":"https://www.wikidata.org/wiki/Q944","display_name":"Quantum mechanics","level":1,"score":0.0},{"id":"https://openalex.org/C158693339","wikidata":"https://www.wikidata.org/wiki/Q190524","display_name":"Eigenvalues and eigenvectors","level":2,"score":0.0},{"id":"https://openalex.org/C163716315","wikidata":"https://www.wikidata.org/wiki/Q901177","display_name":"Gaussian","level":2,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/hpec55821.2022.9926290","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpec55821.2022.9926290","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2022 IEEE High Performance Extreme Computing Conference (HPEC)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.41999998688697815,"id":"https://metadata.un.org/sdg/7","display_name":"Affordable and clean energy"}],"awards":[{"id":"https://openalex.org/G2039107507","display_name":null,"funder_award_id":"61972444,61825202,62072195,61832006","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":28,"referenced_works":["https://openalex.org/W2112796928","https://openalex.org/W2484446135","https://openalex.org/W2560674852","https://openalex.org/W2582955817","https://openalex.org/W2898106867","https://openalex.org/W2910087390","https://openalex.org/W2915589364","https://openalex.org/W2954219808","https://openalex.org/W2964537638","https://openalex.org/W2973134322","https://openalex.org/W3035063907","https://openalex.org/W3090912412","https://openalex.org/W3100809954","https://openalex.org/W3100839241","https://openalex.org/W3113443077","https://openalex.org/W3115410382","https://openalex.org/W3117140649","https://openalex.org/W3137147200","https://openalex.org/W3157183890","https://openalex.org/W3205727737","https://openalex.org/W4200090124","https://openalex.org/W4220810253","https://openalex.org/W6691187937","https://openalex.org/W6717556742","https://openalex.org/W6734593296","https://openalex.org/W6748728731","https://openalex.org/W6759263581","https://openalex.org/W6791940793"],"related_works":["https://openalex.org/W2052993554","https://openalex.org/W2046125858","https://openalex.org/W2001342884","https://openalex.org/W2125326641","https://openalex.org/W4386302689","https://openalex.org/W2068996722","https://openalex.org/W1541167181","https://openalex.org/W128997104","https://openalex.org/W2595503355","https://openalex.org/W1928489786"],"abstract_inverted_index":{"As":[0],"the":[1,10,38,41,61,66,70,86,93,100,125,137,151,156,161,165,177,215,220,231,245,251,276,309,314],"model":[2,83,174,208],"scale":[3],"of":[4,15,44,63,88,96,139,169,313],"Deep":[5,21],"Neural":[6,22],"Networks":[7,23],"(DNNs)":[8],"increases,":[9],"memory":[11,162,167,178,192],"and":[12,112,135,179,182,191,219,227,278,287,305,316],"computational":[13],"cost":[14,207,213],"DNNs":[16],"become":[17],"overwhelmingly":[18],"large.":[19],"Sparse":[20,89],"(SpDNNs)":[24],"are":[25],"promising":[26],"to":[27,50,84,108,123,163,187,209,261,275,285,303],"cope":[28],"with":[29,129],"this":[30,76],"challenge":[31],"by":[32],"using":[33],"fewer":[34],"weights":[35],"while":[36],"preserving":[37],"accuracy.":[39],"However,":[40],"sparsity":[42],"nature":[43],"SpDNN":[45,64,71,97,246,252],"models":[46],"makes":[47],"it":[48],"difficult":[49],"run":[51],"efficiently":[52],"on":[53,194,250,268,290],"GPUs.":[54],"To":[55,196],"stimulate":[56],"technical":[57],"advances":[58],"for":[59,234,244],"improving":[60],"efficiency":[62,87],"inference,":[65],"MIT/IEEE/Amazon":[67],"GraphChallenge":[68],"proposes":[69],"Challenge":[72,253],"in":[73,224],"2019.":[74],"In":[75],"paper,":[77],"we":[78,118,149,204],"present":[79,205],"a":[80,199,206,269,291,298],"hybrid":[81,172],"compute":[82,173],"improve":[85],"Matrix":[90],"Multiplications":[91],"(SpMMs),":[92],"core":[94],"computation":[95],"inference.":[98,247],"First,":[99],"given":[101],"sparse":[102,116,126],"weight":[103],"matrix":[104,201],"will":[105,241],"be":[106,143,242],"divided":[107],"generate":[109],"many":[110],"(sparse":[111],"dense)":[113],"submatrices.":[114],"For":[115,146],"submatrices,":[117,148],"leverage":[119],"compile-time":[120],"data":[121,127,157],"embedding":[122],"compile":[124],"together":[128],"their":[130],"corresponding":[131],"computations":[132],"into":[133],"instructions":[134],"hence":[136],"number":[138],"random":[140],"accesses":[141,193],"can":[142],"reduced":[144],"significantly.":[145],"dense":[147],"follow":[150],"traditional":[152,216],"computing":[153,189,217,222,232],"mode":[154,218,223,233],"where":[155],"is":[158,202,237],"obtained":[159],"from":[160],"exploit":[164],"high":[166],"bandwidth":[168],"GPU.":[170,195,273],"This":[171],"effectively":[175],"balances":[176],"instruction":[180],"bottlenecks,":[181],"offers":[183,283],"more":[184],"scheduling":[185],"opportunities":[186],"overlap":[188],"operations":[190],"determine":[197],"whether":[198],"sub":[200],"sparse,":[203],"estimate":[210],"its":[211],"time":[212],"under":[214],"data-embedded":[221],"an":[225],"accurate":[226],"efficient":[228],"manner.":[229],"Once":[230],"all":[235],"submatrices":[236],"determined,":[238],"customized":[239],"codes":[240],"generated":[243],"Experimental":[248],"results":[249],"benchmarks":[254],"show":[255],"that":[256],"our":[257,281],"approach":[258,282],"achieves":[259],"up":[260,284,302],"197.86":[262],"tera-edges":[263],"per":[264],"second":[265],"inference":[266],"throughput":[267],"single":[270,292],"NVIDIA":[271],"A100":[272],"Compared":[274],"2021":[277,315],"2020":[279,317],"champions,":[280],"6.37x":[286],"89.94x":[288],"speedups":[289,307],"GPU,":[293],"respectively.":[294],"We":[295],"also":[296],"implement":[297],"16-GPU":[299,311],"version,":[300],"showing":[301],"9.49x":[304],"80.11x":[306],"over":[308],"former":[310],"baselines":[312],"champions.":[318]},"counts_by_year":[{"year":2024,"cited_by_count":2},{"year":2023,"cited_by_count":3}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}