{"id":"https://openalex.org/W4383749415","doi":"https://doi.org/10.14778/3598581.3598604","title":"SDPipe: A Semi-Decentralized Framework for Heterogeneity-Aware Pipeline-parallel Training","display_name":"SDPipe: A Semi-Decentralized Framework for Heterogeneity-Aware Pipeline-parallel Training","publication_year":2023,"publication_date":"2023-05-01","ids":{"openalex":"https://openalex.org/W4383749415","doi":"https://doi.org/10.14778/3598581.3598604"},"language":"en","primary_location":{"id":"doi:10.14778/3598581.3598604","is_oa":false,"landing_page_url":"https://doi.org/10.14778/3598581.3598604","pdf_url":null,"source":{"id":"https://openalex.org/S4210226185","display_name":"Proceedings of the VLDB Endowment","issn_l":"2150-8097","issn":["2150-8097"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the VLDB Endowment","raw_type":"journal-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/A5015552951","display_name":"Xupeng Miao","orcid":"https://orcid.org/0000-0002-9371-8358"},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Xupeng Miao","raw_affiliation_strings":["Carnegie Mellon University"],"affiliations":[{"raw_affiliation_string":"Carnegie Mellon University","institution_ids":["https://openalex.org/I74973139"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5114171391","display_name":"Yining Shi","orcid":null},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yining Shi","raw_affiliation_strings":["Peking University"],"affiliations":[{"raw_affiliation_string":"Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102862758","display_name":"Zhi Yang","orcid":"https://orcid.org/0000-0002-8219-4499"},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhi Yang","raw_affiliation_strings":["Peking University"],"affiliations":[{"raw_affiliation_string":"Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5062357883","display_name":"Bin Cui","orcid":"https://orcid.org/0000-0003-1681-4677"},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Bin Cui","raw_affiliation_strings":["Peking University"],"affiliations":[{"raw_affiliation_string":"Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5062437461","display_name":"Zhihao Jia","orcid":"https://orcid.org/0000-0002-1270-5185"},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhihao Jia","raw_affiliation_strings":["Carnegie Mellon University"],"affiliations":[{"raw_affiliation_string":"Carnegie Mellon University","institution_ids":["https://openalex.org/I74973139"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5015552951"],"corresponding_institution_ids":["https://openalex.org/I74973139"],"apc_list":null,"apc_paid":null,"fwci":5.0179,"has_fulltext":false,"cited_by_count":29,"citation_normalized_percentile":{"value":0.96251583,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":90,"max":100},"biblio":{"volume":"16","issue":"9","first_page":"2354","last_page":"2363"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9991000294685364,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9991000294685364,"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/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9983000159263611,"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/T11273","display_name":"Advanced Graph Neural Networks","score":0.9954000115394592,"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":0.8648315668106079},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.7571209073066711},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.7476744651794434},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.7035796642303467},{"id":"https://openalex.org/keywords/synchronization","display_name":"Synchronization (alternating current)","score":0.5785753726959229},{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.5680748820304871},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.5358093976974487},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.5316318273544312},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.4952806532382965},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.4765094816684723},{"id":"https://openalex.org/keywords/data-synchronization","display_name":"Data synchronization","score":0.44276002049446106},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.37185898423194885},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.2315787374973297},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.19881975650787354},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.176163911819458},{"id":"https://openalex.org/keywords/database","display_name":"Database","score":0.13176050782203674},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.10021200776100159}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8648315668106079},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.7571209073066711},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.7476744651794434},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.7035796642303467},{"id":"https://openalex.org/C2778562939","wikidata":"https://www.wikidata.org/wiki/Q1298791","display_name":"Synchronization (alternating current)","level":3,"score":0.5785753726959229},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.5680748820304871},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.5358093976974487},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.5316318273544312},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.4952806532382965},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.4765094816684723},{"id":"https://openalex.org/C108734733","wikidata":"https://www.wikidata.org/wiki/Q1172333","display_name":"Data synchronization","level":3,"score":0.44276002049446106},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.37185898423194885},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2315787374973297},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.19881975650787354},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.176163911819458},{"id":"https://openalex.org/C77088390","wikidata":"https://www.wikidata.org/wiki/Q8513","display_name":"Database","level":1,"score":0.13176050782203674},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.10021200776100159},{"id":"https://openalex.org/C127162648","wikidata":"https://www.wikidata.org/wiki/Q16858953","display_name":"Channel (broadcasting)","level":2,"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/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.14778/3598581.3598604","is_oa":false,"landing_page_url":"https://doi.org/10.14778/3598581.3598604","pdf_url":null,"source":{"id":"https://openalex.org/S4210226185","display_name":"Proceedings of the VLDB Endowment","issn_l":"2150-8097","issn":["2150-8097"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the VLDB Endowment","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":30,"referenced_works":["https://openalex.org/W1616857247","https://openalex.org/W1788418780","https://openalex.org/W2029463628","https://openalex.org/W2044212084","https://openalex.org/W2057332538","https://openalex.org/W2066332749","https://openalex.org/W2108598243","https://openalex.org/W2140655807","https://openalex.org/W2340222647","https://openalex.org/W2523435939","https://openalex.org/W2612026221","https://openalex.org/W2765170276","https://openalex.org/W2962863496","https://openalex.org/W2963433607","https://openalex.org/W2969388332","https://openalex.org/W2972087877","https://openalex.org/W3086105743","https://openalex.org/W3159219445","https://openalex.org/W3177263144","https://openalex.org/W4200068255","https://openalex.org/W4214549129","https://openalex.org/W4221079727","https://openalex.org/W4226328099","https://openalex.org/W4290991121","https://openalex.org/W4310282800","https://openalex.org/W4311209912","https://openalex.org/W4313479568","https://openalex.org/W4362566357","https://openalex.org/W4364382874","https://openalex.org/W6757193083"],"related_works":["https://openalex.org/W2361361118","https://openalex.org/W2087937280","https://openalex.org/W2364921833","https://openalex.org/W2546696010","https://openalex.org/W1992741870","https://openalex.org/W2170649215","https://openalex.org/W2131630752","https://openalex.org/W4251849966","https://openalex.org/W3009421818","https://openalex.org/W2131840028"],"abstract_inverted_index":{"The":[0],"increasing":[1],"size":[2],"of":[3,32,74,104,134,143,171],"both":[4,105,108],"deep":[5],"learning":[6],"models":[7],"and":[8,27,112,157,175],"training":[9,18],"data":[10,28],"necessitates":[11],"the":[12,49,56,65,71,75,102,124,131,141,168],"ability":[13],"to":[14,55,100],"scale":[15],"out":[16],"model":[17,25,51,126],"through":[19],"pipeline-parallel":[20,116],"training,":[21],"which":[22,128,146],"combines":[23],"pipelined":[24],"parallelism":[26],"parallelism.":[29],"However,":[30],"most":[31],"them":[33],"assume":[34],"an":[35],"ideal":[36],"homogeneous":[37],"dedicated":[38],"cluster.":[39],"As":[40],"for":[41,87,130,154],"real":[42],"cloud":[43],"clusters,":[44],"these":[45],"approaches":[46],"suffer":[47],"from":[48],"intensive":[50],"synchronization":[52,89,135],"overheads":[53],"due":[54],"dynamic":[57],"environment":[58],"heterogeneity.":[59,161],"Such":[60],"a":[61,68,96,151,165],"huge":[62],"challenge":[63],"leaves":[64],"design":[66],"in":[67,115],"dilemma:":[69],"either":[70],"performance":[72,82,156,174],"bottleneck":[73],"central":[76],"parameter":[77],"server":[78],"(PS)":[79],"or":[80],"severe":[81],"degradation":[83],"caused":[84],"by":[85],"stragglers":[86],"decentralized":[88],"(like":[90],"All-Reduce).":[91],"This":[92],"approach":[93],"presents":[94],"SDPipe,":[95],"new":[97],"semi-decentralized":[98],"framework":[99],"get":[101],"best":[103],"worlds,":[106],"achieving":[107],"high":[109,120],"heterogeneity":[110],"tolerance":[111],"convergence":[113,158],"efficiency":[114],"training.":[117],"To":[118],"provide":[119],"performance,":[121],"we":[122,139],"decentralize":[123],"communication":[125],"synchronization,":[127],"accounts":[129],"largest":[132],"proportion":[133],"overhead.":[136],"In":[137],"contrast,":[138],"centralize":[140],"process":[142],"group":[144],"scheduling,":[145],"is":[147],"lightweight":[148],"but":[149],"needs":[150],"global":[152],"view":[153],"better":[155],"speed":[159],"against":[160],"We":[162],"show":[163],"via":[164],"prototype":[166],"implementation":[167],"significant":[169],"advantage":[170],"SDPipe":[172],"on":[173],"scalability,":[176],"facing":[177],"different":[178],"environments.":[179]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":23},{"year":2024,"cited_by_count":3},{"year":2023,"cited_by_count":1}],"updated_date":"2026-04-09T08:11:56.329763","created_date":"2025-10-10T00:00:00"}