{"id":"https://openalex.org/W3047537431","doi":"https://doi.org/10.1109/infocom41043.2020.9155446","title":"Preemptive All-reduce Scheduling for Expediting Distributed DNN Training","display_name":"Preemptive All-reduce Scheduling for Expediting Distributed DNN Training","publication_year":2020,"publication_date":"2020-07-01","ids":{"openalex":"https://openalex.org/W3047537431","doi":"https://doi.org/10.1109/infocom41043.2020.9155446","mag":"3047537431"},"language":"en","primary_location":{"id":"doi:10.1109/infocom41043.2020.9155446","is_oa":false,"landing_page_url":"https://doi.org/10.1109/infocom41043.2020.9155446","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE INFOCOM 2020 - IEEE Conference on Computer Communications","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/A5054679083","display_name":"Yixin Bao","orcid":"https://orcid.org/0000-0002-6921-2154"},"institutions":[{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Yixin Bao","raw_affiliation_strings":["Department of Computer Science, The University of Hong Kong"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science, The University of Hong Kong","institution_ids":["https://openalex.org/I889458895"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5013856812","display_name":"Yanghua Peng","orcid":"https://orcid.org/0000-0003-3989-4358"},"institutions":[{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Yanghua Peng","raw_affiliation_strings":["Department of Computer Science, The University of Hong Kong"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science, The University of Hong Kong","institution_ids":["https://openalex.org/I889458895"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5014390034","display_name":"Yangrui Chen","orcid":"https://orcid.org/0000-0003-4182-3444"},"institutions":[{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Yangrui Chen","raw_affiliation_strings":["Department of Computer Science, The University of Hong Kong"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science, The University of Hong Kong","institution_ids":["https://openalex.org/I889458895"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5012597518","display_name":"Chuan Wu","orcid":"https://orcid.org/0000-0002-3144-4398"},"institutions":[{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Chuan Wu","raw_affiliation_strings":["Department of Computer Science, The University of Hong Kong"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science, The University of Hong Kong","institution_ids":["https://openalex.org/I889458895"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I889458895"],"apc_list":null,"apc_paid":null,"fwci":4.2094,"has_fulltext":false,"cited_by_count":63,"citation_normalized_percentile":{"value":0.95371581,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":90,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"626","last_page":"635"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9983000159263611,"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":0.9983000159263611,"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":0.996399998664856,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9951000213623047,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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":0.8483890891075134},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.6838800311088562},{"id":"https://openalex.org/keywords/directed-acyclic-graph","display_name":"Directed acyclic graph","score":0.6736780405044556},{"id":"https://openalex.org/keywords/testbed","display_name":"Testbed","score":0.6384111642837524},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.566405713558197},{"id":"https://openalex.org/keywords/fixed-priority-pre-emptive-scheduling","display_name":"Fixed-priority pre-emptive scheduling","score":0.4736391603946686},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.47087371349334717},{"id":"https://openalex.org/keywords/oracle","display_name":"Oracle","score":0.4673220217227936},{"id":"https://openalex.org/keywords/two-level-scheduling","display_name":"Two-level scheduling","score":0.4138285219669342},{"id":"https://openalex.org/keywords/dynamic-priority-scheduling","display_name":"Dynamic priority scheduling","score":0.39240846037864685},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.36408597230911255},{"id":"https://openalex.org/keywords/schedule","display_name":"Schedule","score":0.3401090204715729},{"id":"https://openalex.org/keywords/rate-monotonic-scheduling","display_name":"Rate-monotonic scheduling","score":0.24733585119247437},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.20672741532325745},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.1652655303478241},{"id":"https://openalex.org/keywords/mathematical-optimization","display_name":"Mathematical optimization","score":0.09272897243499756}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8483890891075134},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.6838800311088562},{"id":"https://openalex.org/C74197172","wikidata":"https://www.wikidata.org/wiki/Q1195339","display_name":"Directed acyclic graph","level":2,"score":0.6736780405044556},{"id":"https://openalex.org/C31395832","wikidata":"https://www.wikidata.org/wiki/Q1318674","display_name":"Testbed","level":2,"score":0.6384111642837524},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.566405713558197},{"id":"https://openalex.org/C122141398","wikidata":"https://www.wikidata.org/wiki/Q5456330","display_name":"Fixed-priority pre-emptive scheduling","level":5,"score":0.4736391603946686},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.47087371349334717},{"id":"https://openalex.org/C55166926","wikidata":"https://www.wikidata.org/wiki/Q2892946","display_name":"Oracle","level":2,"score":0.4673220217227936},{"id":"https://openalex.org/C119948110","wikidata":"https://www.wikidata.org/wiki/Q7858726","display_name":"Two-level scheduling","level":4,"score":0.4138285219669342},{"id":"https://openalex.org/C107568181","wikidata":"https://www.wikidata.org/wiki/Q5319000","display_name":"Dynamic priority scheduling","level":3,"score":0.39240846037864685},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.36408597230911255},{"id":"https://openalex.org/C68387754","wikidata":"https://www.wikidata.org/wiki/Q7271585","display_name":"Schedule","level":2,"score":0.3401090204715729},{"id":"https://openalex.org/C127456818","wikidata":"https://www.wikidata.org/wiki/Q238879","display_name":"Rate-monotonic scheduling","level":4,"score":0.24733585119247437},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.20672741532325745},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.1652655303478241},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.09272897243499756},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.0},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0},{"id":"https://openalex.org/C115903868","wikidata":"https://www.wikidata.org/wiki/Q80993","display_name":"Software engineering","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/infocom41043.2020.9155446","is_oa":false,"landing_page_url":"https://doi.org/10.1109/infocom41043.2020.9155446","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE INFOCOM 2020 - IEEE Conference on Computer Communications","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":68,"referenced_works":["https://openalex.org/W104184427","https://openalex.org/W114517082","https://openalex.org/W1446909886","https://openalex.org/W1525444182","https://openalex.org/W1557310162","https://openalex.org/W1686810756","https://openalex.org/W1984788566","https://openalex.org/W2007126746","https://openalex.org/W2057332538","https://openalex.org/W2063511957","https://openalex.org/W2130942839","https://openalex.org/W2132737349","https://openalex.org/W2168231600","https://openalex.org/W2186615578","https://openalex.org/W2194775991","https://openalex.org/W2402144811","https://openalex.org/W2606891064","https://openalex.org/W2617411258","https://openalex.org/W2617766261","https://openalex.org/W2742145794","https://openalex.org/W2763421725","https://openalex.org/W2769644379","https://openalex.org/W2787998955","https://openalex.org/W2789197829","https://openalex.org/W2798515322","https://openalex.org/W2899771611","https://openalex.org/W2904556356","https://openalex.org/W2919897868","https://openalex.org/W2920397365","https://openalex.org/W2947198853","https://openalex.org/W2953384591","https://openalex.org/W2962684017","https://openalex.org/W2962758826","https://openalex.org/W2962835968","https://openalex.org/W2963112338","https://openalex.org/W2963403868","https://openalex.org/W2963786636","https://openalex.org/W2964004663","https://openalex.org/W2972874238","https://openalex.org/W2975712713","https://openalex.org/W3037338728","https://openalex.org/W3037442399","https://openalex.org/W3037875189","https://openalex.org/W3101036738","https://openalex.org/W3124352525","https://openalex.org/W4210431029","https://openalex.org/W4245217724","https://openalex.org/W4288357791","https://openalex.org/W4301239768","https://openalex.org/W4385245566","https://openalex.org/W6604254268","https://openalex.org/W6628367914","https://openalex.org/W6679393576","https://openalex.org/W6679436768","https://openalex.org/W6684859321","https://openalex.org/W6686509673","https://openalex.org/W6713134421","https://openalex.org/W6738144653","https://openalex.org/W6738460352","https://openalex.org/W6739693220","https://openalex.org/W6739901393","https://openalex.org/W6745245109","https://openalex.org/W6746200960","https://openalex.org/W6748645090","https://openalex.org/W6754818105","https://openalex.org/W6756040250","https://openalex.org/W6762211661","https://openalex.org/W6767869616"],"related_works":["https://openalex.org/W2545511463","https://openalex.org/W1545991362","https://openalex.org/W2167574351","https://openalex.org/W3184267879","https://openalex.org/W2978148977","https://openalex.org/W1935185185","https://openalex.org/W2106332846","https://openalex.org/W2257422103","https://openalex.org/W4300092030","https://openalex.org/W1552447518"],"abstract_inverted_index":{"Data-parallel":[0],"training":[1,8,147,186],"is":[2,161],"widely":[3],"used":[4],"for":[5,91,163,173],"scaling":[6],"DNN":[7,27,115],"over":[9],"large":[10],"datasets,":[11],"using":[12],"the":[13,38,77,81,112,139,146,150],"parameter":[14],"server":[15],"or":[16],"all-reduce":[17,108],"architecture.":[18],"Communication":[19],"scheduling":[20,37,50,93,153],"has":[21,56],"been":[22,57],"promising":[23],"to":[24,31,193],"accelerate":[25],"distributed":[26],"training,":[28,116],"which":[29,168],"aims":[30],"overlap":[32],"communication":[33,41,49,101,121,142,152,171],"with":[34,122,187,197],"computation":[35,54,123],"by":[36],"order":[39],"of":[40,47,84,114,120],"operations.":[42],"We":[43,97],"identify":[44,138],"two":[45,131],"limitations":[46],"previous":[48],"work.":[51],"First,":[52],"layer-wise":[53],"graph":[55,73],"a":[58,69,100,135,157,165,179],"common":[59],"assumption,":[60],"while":[61],"modern":[62],"machine":[63],"learning":[64],"frameworks":[65],"(e.g.,":[66],"TensorFlow)":[67],"use":[68],"sophisticated":[70],"directed":[71],"acyclic":[72],"(DAG)":[74],"representation":[75],"as":[76,154],"execution":[78],"model.":[79],"Second,":[80],"default":[82],"sizes":[83],"tensors":[85,109,172],"are":[86],"often":[87],"less":[88],"than":[89],"optimal":[90,151],"transmission":[92],"and":[94,124],"bandwidth":[95,126,175],"utilization.":[96,127,176],"propose":[98],"PACE,":[99],"scheduler":[102,129],"that":[103,144,183],"preemptively":[104],"schedules":[105],"(potentially":[106],"fused)":[107],"based":[110],"on":[111],"DAG":[113],"guaranteeing":[117],"maximal":[118],"overlapping":[119],"high":[125],"The":[128],"contains":[130],"integrated":[132],"modules:":[133],"given":[134],"DAG,":[136,167],"we":[137],"best":[140],"tensor-preemptive":[141],"schedule":[143],"minimizes":[145],"time;":[148],"exploiting":[149],"an":[155],"oracle,":[156],"dynamic":[158],"programming":[159],"approach":[160],"developed":[162],"generating":[164],"good":[166],"merges":[169],"small":[170],"efficient":[174],"Experiments":[177],"in":[178],"GPU":[180],"testbed":[181],"show":[182],"PACE":[184],"accelerates":[185],"representative":[188],"system":[189],"configurations,":[190],"achieving":[191],"up":[192],"36%":[194],"speed-up":[195],"compared":[196],"state-of-the-art":[198],"solutions.":[199]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":9},{"year":2024,"cited_by_count":8},{"year":2023,"cited_by_count":12},{"year":2022,"cited_by_count":11},{"year":2021,"cited_by_count":13},{"year":2020,"cited_by_count":7},{"year":2019,"cited_by_count":1}],"updated_date":"2026-06-26T08:34:08.712188","created_date":"2025-10-10T00:00:00"}