{"id":"https://openalex.org/W3207427731","doi":"https://doi.org/10.1145/3545008.3545011","title":"EmbRace: Accelerating Sparse Communication for Distributed Training of Deep Neural Networks","display_name":"EmbRace: Accelerating Sparse Communication for Distributed Training of Deep Neural Networks","publication_year":2022,"publication_date":"2022-08-29","ids":{"openalex":"https://openalex.org/W3207427731","doi":"https://doi.org/10.1145/3545008.3545011","mag":"3207427731"},"language":"en","primary_location":{"id":"doi:10.1145/3545008.3545011","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3545008.3545011","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 51st International Conference on Parallel Processing","raw_type":"proceedings-article"},"type":"preprint","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2110.09132","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5101629962","display_name":"Shengwei Li","orcid":"https://orcid.org/0000-0001-9330-7113"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shengwei Li","raw_affiliation_strings":["National University of Defense Technology, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5055426388","display_name":"Zhiquan Lai","orcid":"https://orcid.org/0000-0002-3458-4732"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhiquan Lai","raw_affiliation_strings":["National University of Defense Technology, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100440920","display_name":"Dongsheng Li","orcid":"https://orcid.org/0000-0003-3103-8442"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dongsheng Li","raw_affiliation_strings":["National University of Defense Technology, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042229671","display_name":"Yiming Zhang","orcid":"https://orcid.org/0000-0002-9570-8962"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]},{"id":"https://openalex.org/I191208505","display_name":"Xiamen University","ror":"https://ror.org/00mcjh785","country_code":"CN","type":"education","lineage":["https://openalex.org/I191208505"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yiming Zhang","raw_affiliation_strings":["National University of Defense Technology, China and Xiamen University, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, China and Xiamen University, China","institution_ids":["https://openalex.org/I170215575","https://openalex.org/I191208505"]}]},{"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/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiangyu Ye","raw_affiliation_strings":["National University of Defense Technology, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5054823053","display_name":"Yabo Duan","orcid":"https://orcid.org/0009-0002-8370-1277"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yabo Duan","raw_affiliation_strings":["National University of Defense Technology, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, China","institution_ids":["https://openalex.org/I170215575"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":6,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.8324,"has_fulltext":false,"cited_by_count":6,"citation_normalized_percentile":{"value":0.76632978,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"11"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11273","display_name":"Advanced Graph Neural Networks","score":0.9998000264167786,"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":0.9998000264167786,"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":0.9995999932289124,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.9994000196456909,"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":0.8432974219322205},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.7817690968513489},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.6404691934585571},{"id":"https://openalex.org/keywords/embedding","display_name":"Embedding","score":0.6149164438247681},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.5692330598831177},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.5390694737434387},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.5347453951835632},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5343787670135498},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.5242000818252563},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.5045188665390015},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.4864882826805115},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.4536133110523224},{"id":"https://openalex.org/keywords/models-of-communication","display_name":"Models of communication","score":0.437465101480484},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.4088980555534363},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.38530418276786804},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.13318347930908203}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8432974219322205},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.7817690968513489},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.6404691934585571},{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.6149164438247681},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.5692330598831177},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.5390694737434387},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.5347453951835632},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5343787670135498},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.5242000818252563},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5045188665390015},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.4864882826805115},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.4536133110523224},{"id":"https://openalex.org/C158156997","wikidata":"https://www.wikidata.org/wiki/Q1416645","display_name":"Models of communication","level":2,"score":0.437465101480484},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4088980555534363},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.38530418276786804},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.13318347930908203},{"id":"https://openalex.org/C77088390","wikidata":"https://www.wikidata.org/wiki/Q8513","display_name":"Database","level":1,"score":0.0},{"id":"https://openalex.org/C144024400","wikidata":"https://www.wikidata.org/wiki/Q21201","display_name":"Sociology","level":0,"score":0.0},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.0},{"id":"https://openalex.org/C21547014","wikidata":"https://www.wikidata.org/wiki/Q1423657","display_name":"Operations management","level":1,"score":0.0},{"id":"https://openalex.org/C46312422","wikidata":"https://www.wikidata.org/wiki/Q11024","display_name":"Communication","level":1,"score":0.0},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3545008.3545011","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3545008.3545011","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 51st International Conference on Parallel Processing","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2110.09132","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2110.09132","pdf_url":"https://arxiv.org/pdf/2110.09132","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2110.09132","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2110.09132","pdf_url":"https://arxiv.org/pdf/2110.09132","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":62,"referenced_works":["https://openalex.org/W1522301498","https://openalex.org/W1554378292","https://openalex.org/W1825216778","https://openalex.org/W1984788566","https://openalex.org/W2057332538","https://openalex.org/W2060393849","https://openalex.org/W2083842231","https://openalex.org/W2146502635","https://openalex.org/W2168231600","https://openalex.org/W2257408573","https://openalex.org/W2259472270","https://openalex.org/W2427527485","https://openalex.org/W2512924740","https://openalex.org/W2523246573","https://openalex.org/W2525778437","https://openalex.org/W2526749411","https://openalex.org/W2611669587","https://openalex.org/W2626778328","https://openalex.org/W2734941459","https://openalex.org/W2769644379","https://openalex.org/W2774000609","https://openalex.org/W2787998955","https://openalex.org/W2884711234","https://openalex.org/W2885185669","https://openalex.org/W2896457183","https://openalex.org/W2906007643","https://openalex.org/W2915943718","https://openalex.org/W2920668770","https://openalex.org/W2944793600","https://openalex.org/W2951471551","https://openalex.org/W2951714314","https://openalex.org/W2952313066","https://openalex.org/W2962758826","https://openalex.org/W2963250244","https://openalex.org/W2963748441","https://openalex.org/W2963803379","https://openalex.org/W2964324519","https://openalex.org/W2970971581","https://openalex.org/W2973727699","https://openalex.org/W2975712713","https://openalex.org/W2982536084","https://openalex.org/W2994144272","https://openalex.org/W3037461905","https://openalex.org/W3037585619","https://openalex.org/W3037875189","https://openalex.org/W3047537431","https://openalex.org/W3084569308","https://openalex.org/W3091097978","https://openalex.org/W3096403968","https://openalex.org/W3124352525","https://openalex.org/W3156636320","https://openalex.org/W3204434815","https://openalex.org/W4245217724","https://openalex.org/W4253931064","https://openalex.org/W4288093768","https://openalex.org/W4288357791","https://openalex.org/W4295312788","https://openalex.org/W4301239768","https://openalex.org/W4307961379","https://openalex.org/W4385245566","https://openalex.org/W6691646088","https://openalex.org/W6725857009"],"related_works":["https://openalex.org/W2058965144","https://openalex.org/W2164382479","https://openalex.org/W2146343568","https://openalex.org/W98480971","https://openalex.org/W2150291671","https://openalex.org/W2013643406","https://openalex.org/W2027972911","https://openalex.org/W2157978810","https://openalex.org/W4391547476","https://openalex.org/W1966837078"],"abstract_inverted_index":{"Distributed":[0],"data-parallel":[1,101],"training":[2,85,194],"has":[3],"been":[4],"widely":[5],"adopted":[6],"for":[7,21,38,86],"deep":[8,15],"neural":[9],"network":[10],"(DNN)":[11],"models.":[12,88,177],"Although":[13],"current":[14],"learning":[16],"(DL)":[17],"frameworks":[18,33],"scale":[19],"well":[20],"dense":[22],"models":[23,40,46],"like":[24,41],"image":[25],"classification":[26],"models,":[27],"we":[28,73],"find":[29],"that":[30,47,181],"these":[31],"DL":[32],"have":[34,48,158],"relatively":[35],"low":[36],"scalability":[37],"sparse":[39,50,87,112,117,147],"natural":[42],"language":[43],"processing":[44],"(NLP)":[45],"highly":[49,111],"embedding":[51,151],"tables.":[52],"Most":[53],"existing":[54],"works":[55],"overlook":[56],"the":[57,107,135,140,146,191],"sparsity":[58],"of":[59,83,110,142,149,162],"model":[60,98,136],"parameters":[61],"thus":[62],"suffering":[63],"from":[64],"significant":[65],"but":[66],"unnecessary":[67],"communication":[68,78,108,118],"overhead.":[69],"In":[70],"this":[71],"paper,":[72],"propose":[74],"EmbRace,":[75],"an":[76],"efficient":[77],"framework":[79],"to":[80,105,185,190],"accelerate":[81],"communications":[82,148],"distributed":[84,193],"EmbRace":[89,125,163,182],"introduces":[90],"Sparsity-aware":[91],"Hybrid":[92],"Communication,":[93],"which":[94,133],"integrates":[95],"AlltoAll":[96],"and":[97,122,144,167,169],"parallelism":[99],"into":[100],"training,":[102],"so":[103],"as":[104],"reduce":[106],"overhead":[109],"parameters.":[113],"To":[114],"effectively":[115],"overlap":[116],"with":[119,153,173],"both":[120],"backward":[121],"forward":[123],"computation,":[124],"further":[126],"designs":[127],"a":[128,154,160],"2D":[129],"Communication":[130],"Scheduling":[131],"approach":[132],"optimizes":[134],"computation":[137],"procedure,":[138],"relaxes":[139],"dependency":[141],"embeddings,":[143],"schedules":[145],"each":[150],"row":[152],"priority":[155],"queue.":[156],"We":[157],"implemented":[159],"prototype":[161],"based":[164],"on":[165],"PyTorch":[166],"Horovod,":[168],"conducted":[170],"comprehensive":[171],"evaluations":[172],"four":[174],"representative":[175],"NLP":[176],"Experimental":[178],"results":[179],"show":[180],"achieves":[183],"up":[184],"2.41":[186],"\u00d7":[187],"speedup":[188],"compared":[189],"state-of-the-art":[192],"baselines.":[195]},"counts_by_year":[{"year":2025,"cited_by_count":2},{"year":2024,"cited_by_count":1},{"year":2023,"cited_by_count":3}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2021-10-25T00:00:00"}