{"id":"https://openalex.org/W4384833556","doi":"https://doi.org/10.1145/3597031.3597057","title":"cuSCNN : an Efficient CUDA Implementation of Sparse CNNs","display_name":"cuSCNN : an Efficient CUDA Implementation of Sparse CNNs","publication_year":2023,"publication_date":"2023-06-14","ids":{"openalex":"https://openalex.org/W4384833556","doi":"https://doi.org/10.1145/3597031.3597057"},"language":"en","primary_location":{"id":"doi:10.1145/3597031.3597057","is_oa":false,"landing_page_url":"http://dx.doi.org/10.1145/3597031.3597057","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 13th International Symposium on Highly Efficient Accelerators and Reconfigurable Technologies","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/A5063878178","display_name":"Mohamed A. Elgammal","orcid":"https://orcid.org/0000-0001-8555-7331"},"institutions":[{"id":"https://openalex.org/I185261750","display_name":"University of Toronto","ror":"https://ror.org/03dbr7087","country_code":"CA","type":"education","lineage":["https://openalex.org/I185261750"]}],"countries":["CA"],"is_corresponding":true,"raw_author_name":"Mohamed A. Elgammal","raw_affiliation_strings":["University of Toronto, Canada"],"raw_orcid":"https://orcid.org/0000-0001-8555-7331","affiliations":[{"raw_affiliation_string":"University of Toronto, Canada","institution_ids":["https://openalex.org/I185261750"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5000665425","display_name":"Omar Mohamed Awad","orcid":"https://orcid.org/0000-0003-0676-5867"},"institutions":[{"id":"https://openalex.org/I185261750","display_name":"University of Toronto","ror":"https://ror.org/03dbr7087","country_code":"CA","type":"education","lineage":["https://openalex.org/I185261750"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Omar Mohamed Awad","raw_affiliation_strings":["University of Toronto, Canada"],"raw_orcid":"https://orcid.org/0000-0003-0676-5867","affiliations":[{"raw_affiliation_string":"University of Toronto, Canada","institution_ids":["https://openalex.org/I185261750"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5078152312","display_name":"Isak Edo Vivancos","orcid":"https://orcid.org/0009-0004-1189-2259"},"institutions":[{"id":"https://openalex.org/I185261750","display_name":"University of Toronto","ror":"https://ror.org/03dbr7087","country_code":"CA","type":"education","lineage":["https://openalex.org/I185261750"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Isak Edo Vivancos","raw_affiliation_strings":["University of Toronto, Canada"],"raw_orcid":"https://orcid.org/0009-0004-1189-2259","affiliations":[{"raw_affiliation_string":"University of Toronto, Canada","institution_ids":["https://openalex.org/I185261750"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5072544251","display_name":"Andreas Moshovos","orcid":"https://orcid.org/0000-0001-7768-367X"},"institutions":[{"id":"https://openalex.org/I185261750","display_name":"University of Toronto","ror":"https://ror.org/03dbr7087","country_code":"CA","type":"education","lineage":["https://openalex.org/I185261750"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Andreas Moshovos","raw_affiliation_strings":["University of Toronto, Canada"],"raw_orcid":"https://orcid.org/0000-0001-7768-367X","affiliations":[{"raw_affiliation_string":"University of Toronto, Canada","institution_ids":["https://openalex.org/I185261750"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5030184404","display_name":"Vaughn Betz","orcid":"https://orcid.org/0000-0003-0528-6493"},"institutions":[{"id":"https://openalex.org/I185261750","display_name":"University of Toronto","ror":"https://ror.org/03dbr7087","country_code":"CA","type":"education","lineage":["https://openalex.org/I185261750"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Vaughn Betz","raw_affiliation_strings":["University of Toronto, Canada"],"raw_orcid":"https://orcid.org/0000-0003-0528-6493","affiliations":[{"raw_affiliation_string":"University of Toronto, Canada","institution_ids":["https://openalex.org/I185261750"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5063878178"],"corresponding_institution_ids":["https://openalex.org/I185261750"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.07168959,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"4","issue":null,"first_page":"107","last_page":"113"},"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9973999857902527,"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/T12702","display_name":"Brain Tumor Detection and Classification","score":0.9968000054359436,"subfield":{"id":"https://openalex.org/subfields/2808","display_name":"Neurology"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8878542184829712},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.8625372052192688},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.7243649959564209},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.7091230154037476},{"id":"https://openalex.org/keywords/sparse-matrix","display_name":"Sparse matrix","score":0.5770352482795715},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.5588861703872681},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5584868788719177},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.54836505651474},{"id":"https://openalex.org/keywords/memory-bandwidth","display_name":"Memory bandwidth","score":0.5467532873153687},{"id":"https://openalex.org/keywords/kernel","display_name":"Kernel (algebra)","score":0.46074405312538147},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.44498491287231445},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.41037458181381226},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.35427576303482056},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.2489221692085266},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.1809716820716858}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8878542184829712},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.8625372052192688},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.7243649959564209},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.7091230154037476},{"id":"https://openalex.org/C56372850","wikidata":"https://www.wikidata.org/wiki/Q1050404","display_name":"Sparse matrix","level":3,"score":0.5770352482795715},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.5588861703872681},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5584868788719177},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.54836505651474},{"id":"https://openalex.org/C188045654","wikidata":"https://www.wikidata.org/wiki/Q17148339","display_name":"Memory bandwidth","level":2,"score":0.5467532873153687},{"id":"https://openalex.org/C74193536","wikidata":"https://www.wikidata.org/wiki/Q574844","display_name":"Kernel (algebra)","level":2,"score":0.46074405312538147},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.44498491287231445},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.41037458181381226},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.35427576303482056},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2489221692085266},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.1809716820716858},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"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/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0},{"id":"https://openalex.org/C555944384","wikidata":"https://www.wikidata.org/wiki/Q249","display_name":"Wireless","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},{"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/C114614502","wikidata":"https://www.wikidata.org/wiki/Q76592","display_name":"Combinatorics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3597031.3597057","is_oa":false,"landing_page_url":"http://dx.doi.org/10.1145/3597031.3597057","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 13th International Symposium on Highly Efficient Accelerators and Reconfigurable Technologies","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":17,"referenced_works":["https://openalex.org/W2121927366","https://openalex.org/W2155893237","https://openalex.org/W2516141709","https://openalex.org/W2589026418","https://openalex.org/W2625457103","https://openalex.org/W2899244816","https://openalex.org/W2940612399","https://openalex.org/W2971711696","https://openalex.org/W3019166713","https://openalex.org/W3091174280","https://openalex.org/W3162988700","https://openalex.org/W3169893408","https://openalex.org/W3216623900","https://openalex.org/W4236868170","https://openalex.org/W4240168186","https://openalex.org/W4319862704","https://openalex.org/W6713134421"],"related_works":["https://openalex.org/W1973046741","https://openalex.org/W2983282793","https://openalex.org/W4376647684","https://openalex.org/W2052993554","https://openalex.org/W2046125858","https://openalex.org/W1541167181","https://openalex.org/W4389888442","https://openalex.org/W2149529325","https://openalex.org/W3103168911","https://openalex.org/W4308090217"],"abstract_inverted_index":{"Deep":[0],"Neural":[1],"Network":[2],"models":[3,61],"are":[4],"becoming":[5],"much":[6],"larger":[7,161],"which":[8],"greatly":[9],"increases":[10],"their":[11],"computation":[12],"and":[13,25,62,104,110,127,171],"memory":[14,39,108],"requirements.":[15],"Sparsity":[16],"offers":[17],"great":[18],"opportunities":[19],"to":[20,88,118,122,132,147],"reduce":[21,105],"unnecessary":[22],"data":[23],"transfers":[24],"computations.":[26],"However,":[27],"exploiting":[28],"sparsity":[29,58],"in":[30,38],"CNN":[31,52],"inference":[32,53],"presents":[33],"challenges":[34],"such":[35],"as":[36],"irregularities":[37],"access":[40],"patterns.":[41],"To":[42],"overcome":[43],"this":[44],"challenge,":[45],"we":[46,173],"propose":[47],"cuSCNN,":[48],"an":[49,90,123],"efficient":[50,91,124],"sparse":[51],"engine":[54],"that":[55,100],"leverages":[56],"the":[57,77,81,106,140,168],"of":[59,116,142],"both":[60],"activations":[63],"using":[64],"optimized":[65],"sparse-sparse":[66],"matrix":[67],"convolution":[68],"kernels":[69],"with":[70,151],"compressed":[71],"operands.":[72],"cuSCNN":[73,112],"is":[74],"motivated":[75],"by":[76,80],"concepts":[78],"introduced":[79],"SCNN":[82],"hardware":[83],"accelerator[21]":[84],"but":[85],"modified":[86],"appropriately":[87],"achieve":[89],"software":[92],"implementation":[93,126,136],"for":[94],"GPUs.":[95],"We":[96],"develop":[97],"GPU":[98,156],"optimizations":[99],"boost":[101],"execution":[102],"performance":[103],"required":[107],"size":[109,165],"bandwidth.":[111],"achieves":[113,167],"a":[114,133],"speedup":[115,130],"up":[117],"171":[119],"\u00d7":[120,129],"compared":[121,131],"CPU":[125,135],"30":[128],"multi-threaded":[134],"without":[137],"batching,":[138],"enabling":[139],"use":[141],"inexpensive":[143],"low-end":[144],"memory-constrained":[145],"GPUs":[146],"implement":[148],"large":[149],"networks":[150],"near":[152],"real-time":[153],"latency.":[154],"Although":[155],"throughput":[157],"can":[158],"benefit":[159],"from":[160],"batch":[162,164],"sizes,":[163],"1":[166],"lowest":[169],"latency":[170],"hence":[172],"focus":[174],"on":[175],"it.":[176]},"counts_by_year":[],"updated_date":"2026-03-27T05:58:40.876381","created_date":"2025-10-10T00:00:00"}