{"id":"https://openalex.org/W4386902741","doi":"https://doi.org/10.1109/tcad.2023.3317789","title":"Efficient <i>N:M</i> Sparse DNN Training Using Algorithm, Architecture, and Dataflow Co-Design","display_name":"Efficient <i>N:M</i> Sparse DNN Training Using Algorithm, Architecture, and Dataflow Co-Design","publication_year":2023,"publication_date":"2023-09-20","ids":{"openalex":"https://openalex.org/W4386902741","doi":"https://doi.org/10.1109/tcad.2023.3317789"},"language":"en","primary_location":{"id":"doi:10.1109/tcad.2023.3317789","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcad.2023.3317789","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":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2309.13015","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5066888122","display_name":"Chao Fang","orcid":"https://orcid.org/0000-0003-3430-1189"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chao Fang","raw_affiliation_strings":["School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"raw_orcid":"https://orcid.org/0000-0003-3430-1189","affiliations":[{"raw_affiliation_string":"School of Electronic Science and Engineering, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5059737143","display_name":"Wei Sun","orcid":"https://orcid.org/0000-0003-2873-6418"},"institutions":[{"id":"https://openalex.org/I83019370","display_name":"Eindhoven University of Technology","ror":"https://ror.org/02c2kyt77","country_code":"NL","type":"education","lineage":["https://openalex.org/I83019370"]}],"countries":["NL"],"is_corresponding":false,"raw_author_name":"Wei Sun","raw_affiliation_strings":["Electronic System Group, Eindhoven University of Technology, Eindhoven, The Netherlands"],"raw_orcid":"https://orcid.org/0000-0003-2873-6418","affiliations":[{"raw_affiliation_string":"Electronic System Group, Eindhoven University of Technology, Eindhoven, The Netherlands","institution_ids":["https://openalex.org/I83019370"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5041431476","display_name":"Aojun Zhou","orcid":"https://orcid.org/0000-0002-4742-8624"},"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":"Aojun Zhou","raw_affiliation_strings":["CUHK-Sensetime Joint Lab, The Chinese University of Hong Kong, Hong Kong, China"],"raw_orcid":"https://orcid.org/0000-0002-4742-8624","affiliations":[{"raw_affiliation_string":"CUHK-Sensetime Joint Lab, The Chinese University of Hong Kong, Hong Kong, China","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100696999","display_name":"Zhongfeng Wang","orcid":"https://orcid.org/0000-0002-7227-4786"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhongfeng Wang","raw_affiliation_strings":["School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"raw_orcid":"https://orcid.org/0000-0002-7227-4786","affiliations":[{"raw_affiliation_string":"School of Electronic Science and Engineering, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":2.0346,"has_fulltext":true,"cited_by_count":21,"citation_normalized_percentile":{"value":0.87934374,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":94,"max":100},"biblio":{"volume":"43","issue":"2","first_page":"506","last_page":"519"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9998000264167786,"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":0.9998000264167786,"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/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9955999851226807,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12676","display_name":"Machine Learning and ELM","score":0.9945999979972839,"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/dataflow","display_name":"Dataflow","score":0.8845611810684204},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7309409976005554},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.6732892394065857},{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.6227784156799316},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.5764995813369751},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.5670279264450073},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.564445972442627},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.5461429953575134},{"id":"https://openalex.org/keywords/sparse-matrix","display_name":"Sparse matrix","score":0.5195060968399048},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.4499545693397522},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.3658800721168518},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.276086688041687},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.2352384328842163}],"concepts":[{"id":"https://openalex.org/C96324660","wikidata":"https://www.wikidata.org/wiki/Q205446","display_name":"Dataflow","level":2,"score":0.8845611810684204},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7309409976005554},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.6732892394065857},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.6227784156799316},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5764995813369751},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.5670279264450073},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.564445972442627},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.5461429953575134},{"id":"https://openalex.org/C56372850","wikidata":"https://www.wikidata.org/wiki/Q1050404","display_name":"Sparse matrix","level":3,"score":0.5195060968399048},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.4499545693397522},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3658800721168518},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.276086688041687},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.2352384328842163},{"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/C6557445","wikidata":"https://www.wikidata.org/wiki/Q173113","display_name":"Agronomy","level":1,"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/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1109/tcad.2023.3317789","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcad.2023.3317789","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"},{"id":"pmh:oai:arXiv.org:2309.13015","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2309.13015","pdf_url":"https://arxiv.org/pdf/2309.13015","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"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":"text"},{"id":"pmh:oai:pure.tue.nl:openaire/a0a0e678-c270-4c6f-b608-68d3f66ff83f","is_oa":true,"landing_page_url":"https://research.tue.nl/en/publications/a0a0e678-c270-4c6f-b608-68d3f66ff83f","pdf_url":"https://research.tue.nl/files/353401303/Efficient_NM_Sparse_DNN_Training_Using_Algorithm_Architecture_and_Dataflow_Co-Design.pdf","source":{"id":"https://openalex.org/S4406922641","display_name":"TU/e Research Portal","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Fang, C, Sun, W, Zhou, A & Wang, Z 2024, 'Efficient N:M Sparse DNN Training Using Algorithm, Architecture, and Dataflow Co-Design', IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, vol. 43, no. 2, 10256041, pp. 506-519. https://doi.org/10.1109/TCAD.2023.3317789","raw_type":"info:eu-repo/semantics/article"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2309.13015","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2309.13015","pdf_url":"https://arxiv.org/pdf/2309.13015","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"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":"text"},"sustainable_development_goals":[{"display_name":"Affordable and clean energy","id":"https://metadata.un.org/sdg/7","score":0.8999999761581421}],"awards":[{"id":"https://openalex.org/G2250357272","display_name":null,"funder_award_id":"62174084","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G2635439592","display_name":null,"funder_award_id":"2022YFB4400604","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G4187776313","display_name":null,"funder_award_id":"2022YFB4400604","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/W4386902741.pdf","grobid_xml":"https://content.openalex.org/works/W4386902741.grobid-xml"},"referenced_works_count":62,"referenced_works":["https://openalex.org/W2194775991","https://openalex.org/W2490345425","https://openalex.org/W2606722458","https://openalex.org/W2787094505","https://openalex.org/W2940399336","https://openalex.org/W2946529874","https://openalex.org/W2972054167","https://openalex.org/W2988246957","https://openalex.org/W2988559961","https://openalex.org/W2990001850","https://openalex.org/W2996874060","https://openalex.org/W2998218113","https://openalex.org/W3004303566","https://openalex.org/W3006098440","https://openalex.org/W3085447111","https://openalex.org/W3090288569","https://openalex.org/W3092156174","https://openalex.org/W3092357178","https://openalex.org/W3098417701","https://openalex.org/W3104849992","https://openalex.org/W3107472389","https://openalex.org/W3112412839","https://openalex.org/W3130875094","https://openalex.org/W3173340834","https://openalex.org/W3173370496","https://openalex.org/W3184191382","https://openalex.org/W3189487902","https://openalex.org/W3194419730","https://openalex.org/W3203827521","https://openalex.org/W3208633927","https://openalex.org/W3211878177","https://openalex.org/W4210823275","https://openalex.org/W4213446428","https://openalex.org/W4225905866","https://openalex.org/W4239669814","https://openalex.org/W4280635517","https://openalex.org/W4280640895","https://openalex.org/W4285130446","https://openalex.org/W4285247940","https://openalex.org/W4285506346","https://openalex.org/W4285590283","https://openalex.org/W4287208846","https://openalex.org/W4287391717","https://openalex.org/W4291653336","https://openalex.org/W4293195632","https://openalex.org/W4293219697","https://openalex.org/W4308083526","https://openalex.org/W4312929503","https://openalex.org/W4321466385","https://openalex.org/W4321608192","https://openalex.org/W4322501379","https://openalex.org/W4379115862","https://openalex.org/W6637373629","https://openalex.org/W6770699648","https://openalex.org/W6778883912","https://openalex.org/W6781359717","https://openalex.org/W6784333009","https://openalex.org/W6788811087","https://openalex.org/W6789946416","https://openalex.org/W6794457956","https://openalex.org/W6803607599","https://openalex.org/W6811466168"],"related_works":["https://openalex.org/W2293118914","https://openalex.org/W2998381397","https://openalex.org/W4236419692","https://openalex.org/W2058965144","https://openalex.org/W2164382479","https://openalex.org/W3167919718","https://openalex.org/W4251718783","https://openalex.org/W2171015181","https://openalex.org/W4239447582","https://openalex.org/W2187181201"],"abstract_inverted_index":{"Sparse":[0],"training":[1,59,86,199,237,267,272],"is":[2,67,109,150,201],"one":[3],"of":[4,13,30,45,70,116,124,178,192,197,260],"the":[5,10,52,99,113,131,139,156,161,167,189,195,230,234,251,271,277],"promising":[6],"techniques":[7],"to":[8,39,54,111,152,187],"reduce":[9,130],"computational":[11,132,190],"cost":[12,133],"deep":[14],"neural":[15],"networks":[16],"(DNNs)":[17],"while":[18,134],"retaining":[19],"high":[20,48],"accuracy.":[21,137],"In":[22],"particular,":[23],"N:M":[24,56,74,89,114,163,179],"fine-grained":[25],"structured":[26],"sparsity,":[27],"where":[28],"only${N}$out":[29],"consecutive${M}$elements":[31],"can":[32,128],"be":[33],"nonzero,":[34],"has":[35,60],"attracted":[36],"attention":[37],"due":[38],"its":[40],"hardware-friendly":[41],"pattern":[42],"and":[43,65,95,121,160,182,217,219,225],"capability":[44],"achieving":[46],"a":[47,68,84,102,142,204,256],"sparse":[49,57,75,90,143,164,180,236,241],"ratio.":[50],"However,":[51],"potential":[53],"accelerate":[55],"DNN":[58,125,146,211],"not":[61],"been":[62],"fully":[63],"exploited,":[64],"there":[66],"lack":[69],"efficient":[71],"hardware":[72],"supporting":[73],"training.":[76],"To":[77],"tackle":[78],"these":[79],"challenges,":[80],"this":[81],"article":[82],"presents":[83],"computation-efficient":[85,162],"scheme":[87,200,268],"for":[88,145],"DNNs":[91],"using":[92,209],"algorithm,":[93],"architecture,":[94],"dataflow":[96,168],"co-design.":[97],"At":[98,138,166],"algorithm":[100],"level,":[101,141,169],"bidirectional":[103],"weight":[104],"pruning":[105],"method,":[106],"dubbed":[107],"BDWP,":[108],"proposed":[110,186,266],"leverage":[112],"sparsity":[115],"weights":[117],"during":[118],"both":[119,155],"forward":[120],"backward":[122],"passes":[123],"training,":[126,147,253],"which":[127],"significantly":[129,269],"maintaining":[135],"model":[136],"architecture":[140],"accelerator":[144,232],"namely,":[148],"SAT,":[149],"developed":[151],"neatly":[153],"support":[154],"regular":[157],"dense":[158,252],"operations":[159],"operations.":[165],"multiple":[170],"optimization":[171],"methods":[172],"ranging":[173],"from":[174],"interleave":[175],"mapping,":[176],"pregeneration":[177],"weights,":[181],"offline":[183],"scheduling,":[184],"are":[185],"boost":[188],"efficiency":[191,279],"SAT.":[193],"Finally,":[194],"effectiveness":[196],"our":[198,265],"evaluated":[202],"on":[203,262],"Xilinx":[205],"VCU1525":[206],"FPGA":[207],"card":[208],"various":[210],"models":[212],"(ResNet9,":[213],"ViT,":[214],"VGG19,":[215],"ResNet18,":[216],"ResNet50)":[218],"datasets":[220],"(CIFAR-10,":[221],"CIFAR-100,":[222],"Tiny":[223],"ImageNet,":[224],"ImageNet).":[226],"Experimental":[227],"results":[228],"show":[229],"SAT":[231],"with":[233,250],"BDWP":[235],"method":[238],"under":[239],"2:8":[240],"ratio":[242],"achieves":[243],"an":[244],"average":[245],"speedup":[246],"of$1.75\\times":[247],"$over":[248,282],"that":[249],"accompanied":[254],"by":[255],"negligible":[257],"accuracy":[258],"loss":[259],"0.56%":[261],"average.":[263],"Furthermore,":[264],"improves":[270],"throughput":[273],"by$2.97\\times":[274],"$\u2013$25.22\\times":[275],"$and":[276],"energy":[278],"by$1.36\\times":[280],"$\u2013$3.58\\times":[281],"prior":[283],"FPGA-based":[284],"accelerators.":[285]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":17},{"year":2024,"cited_by_count":3}],"updated_date":"2026-08-01T09:00:35.917206","created_date":"2023-09-21T00:00:00"}