{"id":"https://openalex.org/W3113443077","doi":"https://doi.org/10.1109/hpec43674.2020.9286218","title":"A Novel Inference Algorithm for Large Sparse Neural Network using Task Graph Parallelism","display_name":"A Novel Inference Algorithm for Large Sparse Neural Network using Task Graph Parallelism","publication_year":2020,"publication_date":"2020-09-22","ids":{"openalex":"https://openalex.org/W3113443077","doi":"https://doi.org/10.1109/hpec43674.2020.9286218","mag":"3113443077"},"language":"en","primary_location":{"id":"doi:10.1109/hpec43674.2020.9286218","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpec43674.2020.9286218","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2020 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/A5048830943","display_name":"Dian-Lun Lin","orcid":"https://orcid.org/0000-0003-3075-7437"},"institutions":[{"id":"https://openalex.org/I223532165","display_name":"University of Utah","ror":"https://ror.org/03r0ha626","country_code":"US","type":"education","lineage":["https://openalex.org/I223532165"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Dian-Lun Lin","raw_affiliation_strings":["University of Utah"],"affiliations":[{"raw_affiliation_string":"University of Utah","institution_ids":["https://openalex.org/I223532165"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5088685794","display_name":"Tsung\u2010Wei Huang","orcid":"https://orcid.org/0000-0001-9768-3378"},"institutions":[{"id":"https://openalex.org/I223532165","display_name":"University of Utah","ror":"https://ror.org/03r0ha626","country_code":"US","type":"education","lineage":["https://openalex.org/I223532165"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Tsung-Wei Huang","raw_affiliation_strings":["University of Utah"],"affiliations":[{"raw_affiliation_string":"University of Utah","institution_ids":["https://openalex.org/I223532165"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5048830943"],"corresponding_institution_ids":["https://openalex.org/I223532165"],"apc_list":null,"apc_paid":null,"fwci":2.6512,"has_fulltext":false,"cited_by_count":39,"citation_normalized_percentile":{"value":0.91988434,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":89,"max":99},"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/T11273","display_name":"Advanced Graph Neural Networks","score":0.9983999729156494,"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"}},"topics":[{"id":"https://openalex.org/T11273","display_name":"Advanced Graph Neural Networks","score":0.9983999729156494,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.9976000189781189,"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/T11948","display_name":"Machine Learning in Materials Science","score":0.9961000084877014,"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":0.8203808069229126},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.6753076314926147},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.647142767906189},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.6058350205421448},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.5321340560913086},{"id":"https://openalex.org/keywords/data-parallelism","display_name":"Data parallelism","score":0.5285590887069702},{"id":"https://openalex.org/keywords/sparse-matrix","display_name":"Sparse matrix","score":0.4438111186027527},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.4117930829524994},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.41178083419799805},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.3572409749031067},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.34826385974884033},{"id":"https://openalex.org/keywords/parallelism","display_name":"Parallelism (grammar)","score":0.25808560848236084},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.23405462503433228}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8203808069229126},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.6753076314926147},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.647142767906189},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.6058350205421448},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.5321340560913086},{"id":"https://openalex.org/C61483411","wikidata":"https://www.wikidata.org/wiki/Q3124522","display_name":"Data parallelism","level":3,"score":0.5285590887069702},{"id":"https://openalex.org/C56372850","wikidata":"https://www.wikidata.org/wiki/Q1050404","display_name":"Sparse matrix","level":3,"score":0.4438111186027527},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.4117930829524994},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.41178083419799805},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3572409749031067},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.34826385974884033},{"id":"https://openalex.org/C2781172179","wikidata":"https://www.wikidata.org/wiki/Q853109","display_name":"Parallelism (grammar)","level":2,"score":0.25808560848236084},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.23405462503433228},{"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/C555944384","wikidata":"https://www.wikidata.org/wiki/Q249","display_name":"Wireless","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/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0},{"id":"https://openalex.org/C77088390","wikidata":"https://www.wikidata.org/wiki/Q8513","display_name":"Database","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/hpec43674.2020.9286218","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpec43674.2020.9286218","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2020 IEEE High Performance Extreme Computing Conference (HPEC)","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":37,"referenced_works":["https://openalex.org/W1598866093","https://openalex.org/W2119144962","https://openalex.org/W2124653173","https://openalex.org/W2279098554","https://openalex.org/W2495819555","https://openalex.org/W2625457103","https://openalex.org/W2786066748","https://openalex.org/W2802166112","https://openalex.org/W2859678021","https://openalex.org/W2896457183","https://openalex.org/W2899244816","https://openalex.org/W2963341956","https://openalex.org/W2963351145","https://openalex.org/W2963366311","https://openalex.org/W2963416938","https://openalex.org/W2963756346","https://openalex.org/W2963959650","https://openalex.org/W2964299589","https://openalex.org/W2964537638","https://openalex.org/W2969388332","https://openalex.org/W2973134322","https://openalex.org/W2989899068","https://openalex.org/W2990514424","https://openalex.org/W2991040477","https://openalex.org/W3100809954","https://openalex.org/W3100839241","https://openalex.org/W3101812505","https://openalex.org/W4294149566","https://openalex.org/W4302343710","https://openalex.org/W6635810480","https://openalex.org/W6724186685","https://openalex.org/W6742080785","https://openalex.org/W6747917562","https://openalex.org/W6751238189","https://openalex.org/W6755207826","https://openalex.org/W6756379755","https://openalex.org/W6756718674"],"related_works":["https://openalex.org/W4399354997","https://openalex.org/W3062287","https://openalex.org/W2380390332","https://openalex.org/W2742145873","https://openalex.org/W2023572661","https://openalex.org/W4245975140","https://openalex.org/W2532592438","https://openalex.org/W2052993554","https://openalex.org/W2046125858","https://openalex.org/W1941712353"],"abstract_inverted_index":{"The":[0,63],"ever-increasing":[1],"size":[2],"of":[3,37,106,113,128,144,166,199,215],"modern":[4],"deep":[5],"neural":[6],"network":[7],"(DNN)":[8],"architectures":[9],"has":[10,70,189],"put":[11],"increasing":[12],"strain":[13],"on":[14,140],"the":[15,35,104,141,164,167,185],"hardware":[16],"needed":[17],"to":[18,59,73,109,125,159,163,205],"implement":[19],"them.":[20],"Sparsified":[21],"DNNs":[22,45],"can":[23,39,174],"greatly":[24],"reduce":[25],"memory":[26],"costs":[27],"and":[28,61,102,115,123,133,149],"increase":[29],"throughput":[30],"over":[31],"standard":[32],"DNNs,":[33],"if":[34],"loss":[36],"accuracy":[38],"be":[40],"adequately":[41],"controlled.":[42],"However,":[43],"sparse":[44,79,93],"present":[46],"unique":[47],"computational":[48],"challenges.":[49],"Efficient":[50],"model":[51,114,131],"or":[52],"data":[53,116,129],"parallelism":[54],"algorithms":[55],"are":[56],"extremely":[57],"hard":[58],"design":[60],"implement.":[62],"recent":[64],"effort":[65],"MIT/IEEE/Amazon":[66],"HPEC":[67,145,169],"Graph":[68],"Challenge":[69,148],"drawn":[71],"attention":[72],"high-performance":[74],"inference":[75,89,100,177],"methods":[76],"for":[77,91],"large":[78,92],"DNNs.":[80],"In":[81],"this":[82],"paper,":[83],"we":[84],"introduce":[85],"SNIG,":[86],"an":[87],"efficient":[88,111],"engine":[90],"DNN":[94,147,171],"s.":[95],"SNIG":[96,139,173,202],"develops":[97],"highly":[98],"optimized":[99],"kernels":[101],"leverages":[103],"power":[105],"CUDA":[107],"Graphs":[108],"enable":[110],"decomposition":[112,119],"parallelisms.":[117],"Our":[118],"strategy":[120],"is":[121,203],"flexible":[122],"scalable":[124,154],"different":[126],"partitions":[127],"volumes,":[130],"sizes,":[132],"GPU":[134,158],"numbers.":[135],"We":[136],"have":[137],"evaluated":[138],"official":[142],"benchmarks":[143],"Sparse":[146,170],"demonstrated":[150],"its":[151],"promising":[152],"performance":[153],"from":[155],"a":[156,181,209,213],"single":[157,182],"multiple":[160],"GPUs.":[161,217],"Compared":[162],"champion":[165],"2019":[168],"Challenge,":[172],"finish":[175],"all":[176],"workloads":[178],"using":[179],"only":[180],"GPU.":[183],"At":[184],"largest":[186],"DNN,":[187],"which":[188],"more":[190],"than":[191,208],"4":[192,216],"billion":[193],"parameters":[194],"across":[195],"1920":[196],"layers":[197],"each":[198],"65536":[200],"neurons,":[201],"up":[204],"2.3x":[206],"faster":[207],"state-of-the-art":[210],"baseline":[211],"under":[212],"machine":[214]},"counts_by_year":[{"year":2025,"cited_by_count":11},{"year":2024,"cited_by_count":8},{"year":2023,"cited_by_count":7},{"year":2022,"cited_by_count":7},{"year":2021,"cited_by_count":5},{"year":2020,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}