{"id":"https://openalex.org/W3202294838","doi":"https://doi.org/10.1145/3472456.3472496","title":"Optimizing Massively Parallel Winograd Convolution on ARM Processor","display_name":"Optimizing Massively Parallel Winograd Convolution on ARM Processor","publication_year":2021,"publication_date":"2021-08-09","ids":{"openalex":"https://openalex.org/W3202294838","doi":"https://doi.org/10.1145/3472456.3472496","mag":"3202294838"},"language":"en","primary_location":{"id":"doi:10.1145/3472456.3472496","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3472456.3472496","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"50th International Conference on Parallel Processing","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/A5100440903","display_name":"Dongsheng Li","orcid":"https://orcid.org/0000-0001-9743-2034"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Dongsheng Li","raw_affiliation_strings":["Sun Yat-sen University, China"],"affiliations":[{"raw_affiliation_string":"Sun Yat-sen University, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5041534890","display_name":"Dan Huang","orcid":"https://orcid.org/0000-0001-5582-1031"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dan Huang","raw_affiliation_strings":["Sun Yat-sen University, China"],"affiliations":[{"raw_affiliation_string":"Sun Yat-sen University, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101483479","display_name":"Zhiguang Chen","orcid":"https://orcid.org/0000-0002-9318-5715"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhiguang Chen","raw_affiliation_strings":["Sun Yat-sen University, China"],"affiliations":[{"raw_affiliation_string":"Sun Yat-sen University, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101633465","display_name":"Yutong Lu","orcid":"https://orcid.org/0000-0001-5315-3375"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yutong Lu","raw_affiliation_strings":["Sun Yat-sen University, China"],"affiliations":[{"raw_affiliation_string":"Sun Yat-sen University, China","institution_ids":["https://openalex.org/I157773358"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5100440903"],"corresponding_institution_ids":["https://openalex.org/I157773358"],"apc_list":null,"apc_paid":null,"fwci":0.7685,"has_fulltext":false,"cited_by_count":10,"citation_normalized_percentile":{"value":0.73844771,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":95,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"12"},"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/T10812","display_name":"Human Pose and Action Recognition","score":0.9914000034332275,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9900000095367432,"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/computer-science","display_name":"Computer science","score":0.8828284740447998},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.8037236928939819},{"id":"https://openalex.org/keywords/cache","display_name":"Cache","score":0.5927268862724304},{"id":"https://openalex.org/keywords/convolution","display_name":"Convolution (computer science)","score":0.5808824300765991},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.5304555296897888},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.4692164659500122},{"id":"https://openalex.org/keywords/multi-core-processor","display_name":"Multi-core processor","score":0.46245336532592773},{"id":"https://openalex.org/keywords/instruction-set","display_name":"Instruction set","score":0.41067180037498474},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.20478495955467224},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.14062249660491943},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.08401897549629211}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8828284740447998},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.8037236928939819},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.5927268862724304},{"id":"https://openalex.org/C45347329","wikidata":"https://www.wikidata.org/wiki/Q5166604","display_name":"Convolution (computer science)","level":3,"score":0.5808824300765991},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.5304555296897888},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.4692164659500122},{"id":"https://openalex.org/C78766204","wikidata":"https://www.wikidata.org/wiki/Q555032","display_name":"Multi-core processor","level":2,"score":0.46245336532592773},{"id":"https://openalex.org/C202491316","wikidata":"https://www.wikidata.org/wiki/Q272683","display_name":"Instruction set","level":2,"score":0.41067180037498474},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.20478495955467224},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.14062249660491943},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.08401897549629211}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3472456.3472496","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3472456.3472496","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"50th International Conference on Parallel Processing","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":15,"referenced_works":["https://openalex.org/W1487564550","https://openalex.org/W1832693441","https://openalex.org/W2020141429","https://openalex.org/W2068993993","https://openalex.org/W2114508814","https://openalex.org/W2172654076","https://openalex.org/W2194775991","https://openalex.org/W2269172429","https://openalex.org/W2304648132","https://openalex.org/W2582996697","https://openalex.org/W2748128350","https://openalex.org/W2786374423","https://openalex.org/W2899296278","https://openalex.org/W2963446712","https://openalex.org/W2963674387"],"related_works":["https://openalex.org/W2128523353","https://openalex.org/W2152099439","https://openalex.org/W1984163603","https://openalex.org/W2291648581","https://openalex.org/W3130422087","https://openalex.org/W2107644726","https://openalex.org/W2406856881","https://openalex.org/W3004195166","https://openalex.org/W2011868109","https://openalex.org/W2081416538"],"abstract_inverted_index":{"Convolution":[0],"Neural":[1],"Network":[2],"(CNN)":[3],"has":[4],"gained":[5],"a":[6,144,184,191],"great":[7],"success":[8],"in":[9,31,40],"deep":[10,41],"learning":[11,42],"applications":[12],"and":[13,70,85,114,129],"been":[14],"accelerated":[15],"by":[16,78],"dedicated":[17],"convolutional":[18,187],"algorithms.":[19],"Winograd-based":[20,97,140,176],"algorithm":[21,98,104,138],"can":[22,204],"greatly":[23],"reduce":[24],"the":[25,71,106,111,118,125,136,174,198],"number":[26],"of":[27,120,186],"arithmetic":[28,132],"operations":[29],"required":[30],"convolution.":[32],"However,":[33],"our":[34,162,202],"experiments":[35],"show":[36,157,200],"that":[37,158,201],"existing":[38,210],"implementations":[39,211],"libraries":[43],"cannot":[44],"achieve":[45,143,206],"expected":[46],"parallel":[47,72,137],"performance":[48,73,208],"on":[49,99,178,183,190,212],"ARM":[50,61,100,121,179,213],"manycore":[51,62,101,149],"CPUs":[52,63],"with":[53,154,194],"last-level":[54],"cache":[55,81],"(LLC).":[56],"Compared":[57],"to":[58,110,142,170],"multicore":[59],"processor,":[60,122],"have":[64],"more":[65,67,75,145],"cores,":[66],"NUMA":[68,83,196],"nodes":[69],"is":[74,115],"easily":[76],"restricted":[77],"memory":[79],"bandwidth,":[80],"contention,":[82],"configuration":[84],"etc.":[86],"In":[87],"this":[88],"paper,":[89],"we":[90],"propose":[91],"an":[92],"optimized":[93,116],"implementation":[94,147,163,203],"for":[95,117,139,148,159],"single-precision":[96],"CPUs.":[102,150],"Our":[103],"adjusts":[105],"data":[107],"layout":[108],"according":[109],"input":[112],"shape":[113],"characteristics":[119],"thus":[123],"reducing":[124],"matrix":[126],"transformation":[127],"overhead":[128],"achieving":[130],"high":[131],"intensity.":[133],"We":[134],"redesign":[135],"convolution":[141,177],"efficient":[146],"The":[151],"experimental":[152],"results":[153,199],"32":[155],"cores":[156],"modern":[160],"ConvNets,":[161],"achieves":[164],"speedups":[165],"ranging":[166],"from":[167],"3":[168],"\u00d7":[169,172],"5":[171],"over":[173],"state-of-the-art":[175],"processor.":[180,214],"Even":[181],"conducted":[182],"set":[185],"benchmarks":[188],"executing":[189],"128-core":[192],"system":[193],"4":[195],"nodes,":[197],"also":[205],"better":[207],"than":[209]},"counts_by_year":[{"year":2025,"cited_by_count":2},{"year":2024,"cited_by_count":3},{"year":2023,"cited_by_count":5}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}