{"id":"https://openalex.org/W3209029431","doi":"https://doi.org/10.1007/s10586-021-03370-9","title":"Accelerating distributed deep neural network training with pipelined MPI allreduce","display_name":"Accelerating distributed deep neural network training with pipelined MPI allreduce","publication_year":2021,"publication_date":"2021-08-07","ids":{"openalex":"https://openalex.org/W3209029431","doi":"https://doi.org/10.1007/s10586-021-03370-9","mag":"3209029431"},"language":"en","primary_location":{"id":"doi:10.1007/s10586-021-03370-9","is_oa":true,"landing_page_url":"https://doi.org/10.1007/s10586-021-03370-9","pdf_url":"https://link.springer.com/content/pdf/10.1007/s10586-021-03370-9.pdf","source":{"id":"https://openalex.org/S106148199","display_name":"Cluster Computing","issn_l":"1386-7857","issn":["1386-7857","1573-7543"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319900","host_organization_name":"Springer Science+Business Media","host_organization_lineage":["https://openalex.org/P4310319900","https://openalex.org/P4310319965"],"host_organization_lineage_names":["Springer Science+Business Media","Springer Nature"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Cluster Computing","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://link.springer.com/content/pdf/10.1007/s10586-021-03370-9.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5033947083","display_name":"Adri\u00e1n Castell\u00f3","orcid":"https://orcid.org/0000-0002-8576-8451"},"institutions":[{"id":"https://openalex.org/I60053951","display_name":"Universitat Polit\u00e8cnica de Val\u00e8ncia","ror":"https://ror.org/01460j859","country_code":"ES","type":"education","lineage":["https://openalex.org/I60053951"]}],"countries":["ES"],"is_corresponding":true,"raw_author_name":"Adri\u00e1n Castell\u00f3","raw_affiliation_strings":["Universitat Polit\u00e8cnica de Val\u00e8ncia, Valencia, Spain"],"raw_orcid":"https://orcid.org/0000-0002-8576-8451","affiliations":[{"raw_affiliation_string":"Universitat Polit\u00e8cnica de Val\u00e8ncia, Valencia, Spain","institution_ids":["https://openalex.org/I60053951"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5012806004","display_name":"Enrique S. Quintana\u2013Ort\u0301\u0131","orcid":"https://orcid.org/0000-0002-5454-165X"},"institutions":[{"id":"https://openalex.org/I60053951","display_name":"Universitat Polit\u00e8cnica de Val\u00e8ncia","ror":"https://ror.org/01460j859","country_code":"ES","type":"education","lineage":["https://openalex.org/I60053951"]}],"countries":["ES"],"is_corresponding":false,"raw_author_name":"Enrique S. Quintana-Ort\u00ed","raw_affiliation_strings":["Universitat Polit\u00e8cnica de Val\u00e8ncia, Valencia, Spain"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Universitat Polit\u00e8cnica de Val\u00e8ncia, Valencia, Spain","institution_ids":["https://openalex.org/I60053951"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5040384183","display_name":"J. Duato","orcid":"https://orcid.org/0000-0002-7785-0607"},"institutions":[{"id":"https://openalex.org/I60053951","display_name":"Universitat Polit\u00e8cnica de Val\u00e8ncia","ror":"https://ror.org/01460j859","country_code":"ES","type":"education","lineage":["https://openalex.org/I60053951"]}],"countries":["ES"],"is_corresponding":false,"raw_author_name":"Jos\u00e9 Duato","raw_affiliation_strings":["Universitat Polit\u00e8cnica de Val\u00e8ncia, Valencia, Spain"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Universitat Polit\u00e8cnica de Val\u00e8ncia, Valencia, Spain","institution_ids":["https://openalex.org/I60053951"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":["https://openalex.org/A5033947083"],"corresponding_institution_ids":["https://openalex.org/I60053951"],"apc_list":{"value":2190,"currency":"EUR","value_usd":2790},"apc_paid":{"value":2190,"currency":"EUR","value_usd":2790},"fwci":1.2284,"has_fulltext":true,"cited_by_count":15,"citation_normalized_percentile":{"value":0.8339591,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":98},"biblio":{"volume":"24","issue":"4","first_page":"3797","last_page":"3813"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9997000098228455,"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":0.9997000098228455,"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/T12808","display_name":"Ferroelectric and Negative Capacitance Devices","score":0.9990000128746033,"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/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9987999796867371,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.9321880340576172},{"id":"https://openalex.org/keywords/blocking","display_name":"Blocking (statistics)","score":0.7126862406730652},{"id":"https://openalex.org/keywords/workload","display_name":"Workload","score":0.6482952237129211},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.6024577021598816},{"id":"https://openalex.org/keywords/synchronization","display_name":"Synchronization (alternating current)","score":0.5790510177612305},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.5647012591362},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.5289459228515625},{"id":"https://openalex.org/keywords/thread","display_name":"Thread (computing)","score":0.5192632675170898},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.42208027839660645},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.2713152766227722},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.20065462589263916},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.13355320692062378},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.08550766110420227}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.9321880340576172},{"id":"https://openalex.org/C144745244","wikidata":"https://www.wikidata.org/wiki/Q4927286","display_name":"Blocking (statistics)","level":2,"score":0.7126862406730652},{"id":"https://openalex.org/C2778476105","wikidata":"https://www.wikidata.org/wiki/Q628539","display_name":"Workload","level":2,"score":0.6482952237129211},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.6024577021598816},{"id":"https://openalex.org/C2778562939","wikidata":"https://www.wikidata.org/wiki/Q1298791","display_name":"Synchronization (alternating current)","level":3,"score":0.5790510177612305},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.5647012591362},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.5289459228515625},{"id":"https://openalex.org/C138101251","wikidata":"https://www.wikidata.org/wiki/Q213092","display_name":"Thread (computing)","level":2,"score":0.5192632675170898},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.42208027839660645},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2713152766227722},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.20065462589263916},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.13355320692062378},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.08550766110420227},{"id":"https://openalex.org/C127162648","wikidata":"https://www.wikidata.org/wiki/Q16858953","display_name":"Channel (broadcasting)","level":2,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1007/s10586-021-03370-9","is_oa":true,"landing_page_url":"https://doi.org/10.1007/s10586-021-03370-9","pdf_url":"https://link.springer.com/content/pdf/10.1007/s10586-021-03370-9.pdf","source":{"id":"https://openalex.org/S106148199","display_name":"Cluster Computing","issn_l":"1386-7857","issn":["1386-7857","1573-7543"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319900","host_organization_name":"Springer Science+Business Media","host_organization_lineage":["https://openalex.org/P4310319900","https://openalex.org/P4310319965"],"host_organization_lineage_names":["Springer Science+Business Media","Springer Nature"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Cluster Computing","raw_type":"journal-article"},{"id":"pmh:oai:riunet.upv.es:10251/186225","is_oa":true,"landing_page_url":"http://hdl.handle.net/10251/186225","pdf_url":null,"source":{"id":"https://openalex.org/S4306400639","display_name":"RiuNet (Universitat Polit\u00e8cnica de Val\u00e8ncia)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I60053951","host_organization_name":"Universitat Polit\u00e8cnica de Val\u00e8ncia","host_organization_lineage":["https://openalex.org/I60053951"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":"info:eu-repo/semantics/article"}],"best_oa_location":{"id":"doi:10.1007/s10586-021-03370-9","is_oa":true,"landing_page_url":"https://doi.org/10.1007/s10586-021-03370-9","pdf_url":"https://link.springer.com/content/pdf/10.1007/s10586-021-03370-9.pdf","source":{"id":"https://openalex.org/S106148199","display_name":"Cluster Computing","issn_l":"1386-7857","issn":["1386-7857","1573-7543"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319900","host_organization_name":"Springer Science+Business Media","host_organization_lineage":["https://openalex.org/P4310319900","https://openalex.org/P4310319965"],"host_organization_lineage_names":["Springer Science+Business Media","Springer Nature"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Cluster Computing","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1041213487","display_name":null,"funder_award_id":"FJC2019-039222-I","funder_id":"https://openalex.org/F4320315062","funder_display_name":"Ministerio de Ciencia, Innovaci\u00f3n y Universidades"},{"id":"https://openalex.org/G5181993914","display_name":null,"funder_award_id":"TIN-2017-82972-R","funder_id":"https://openalex.org/F4320315062","funder_display_name":"Ministerio de Ciencia, Innovaci\u00f3n y Universidades"}],"funders":[{"id":"https://openalex.org/F4320315062","display_name":"Ministerio de Ciencia, Innovaci\u00f3n y Universidades","ror":null},{"id":"https://openalex.org/F4320321047","display_name":"Partnership for Advanced Computing in Europe AISBL","ror":"https://ror.org/038j4hz26"},{"id":"https://openalex.org/F4320321697","display_name":"LOEWE Zentrum AdRIA","ror":"https://ror.org/05n911h24"},{"id":"https://openalex.org/F4320336419","display_name":"Ag\u00e8ncia Valenciana de la Innovaci\u00f3","ror":null}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W3209029431.pdf","grobid_xml":"https://content.openalex.org/works/W3209029431.grobid-xml"},"referenced_works_count":36,"referenced_works":["https://openalex.org/W31923072","https://openalex.org/W1589918049","https://openalex.org/W1686810756","https://openalex.org/W2056476515","https://openalex.org/W2108598243","https://openalex.org/W2131613942","https://openalex.org/W2152783013","https://openalex.org/W2163605009","https://openalex.org/W2170796499","https://openalex.org/W2194775991","https://openalex.org/W2443267493","https://openalex.org/W2562749854","https://openalex.org/W2604319603","https://openalex.org/W2740001873","https://openalex.org/W2785957273","https://openalex.org/W2787998955","https://openalex.org/W2889569921","https://openalex.org/W2898319404","https://openalex.org/W2908125138","https://openalex.org/W2916947863","https://openalex.org/W2942929621","https://openalex.org/W2953384591","https://openalex.org/W2955052486","https://openalex.org/W2969976451","https://openalex.org/W2970971581","https://openalex.org/W2972087877","https://openalex.org/W3038827379","https://openalex.org/W3041089513","https://openalex.org/W3092494433","https://openalex.org/W3103417206","https://openalex.org/W3111877393","https://openalex.org/W3118608800","https://openalex.org/W3159351747","https://openalex.org/W6600203831","https://openalex.org/W6610640258","https://openalex.org/W6713134421"],"related_works":["https://openalex.org/W986318368","https://openalex.org/W2000785801","https://openalex.org/W2384410913","https://openalex.org/W2352878646","https://openalex.org/W2004734601","https://openalex.org/W2130149817","https://openalex.org/W2990194547","https://openalex.org/W1480123525","https://openalex.org/W2023832055","https://openalex.org/W2082485924"],"abstract_inverted_index":{"Abstract":[0],"TensorFlow":[1],"(TF)":[2],"is":[3],"usually":[4],"combined":[5,40],"with":[6,41,48,91],"the":[7,62,66,74,79,97,101,109,115],"Horovod":[8],"(HVD)":[9],"workload":[10],"distribution":[11],"package":[12],"to":[13,18,35,45,59,72,114,124],"obtain":[14],"a":[15,31,42,55,88],"parallel":[16],"tool":[17],"train":[19],"deep":[20],"neural":[21],"network":[22],"on":[23],"clusters":[24],"of":[25,64,111,145],"computers.":[26],"HVD":[27],"in":[28,69],"turn":[29],"utilizes":[30],"blocking":[32,89,98],"Allreduce":[33,75,90],"primitive":[34],"share":[36],"information":[37],"among":[38],"processes,":[39],"communication":[43,47,116],"thread":[44],"overlap":[46],"computation.":[49],"In":[50],"this":[51],"work,":[52],"we":[53,107,131],"perform":[54],"thorough":[56],"experimental":[57],"analysis":[58],"expose":[60],"(1)":[61],"importance":[63],"selecting":[65],"best":[67],"algorithm":[68],"MPI":[70],"libraries":[71],"realize":[73],"operation;":[76],"and":[77,151],"(2)":[78],"performance":[80,138],"acceleration":[81],"that":[82,119,133,141],"can":[83,135],"be":[84],"attained":[85],"when":[86],"replacing":[87],"its":[92],"non-blocking":[93],"counterpart":[94],"(while":[95],"maintaining":[96],"behaviour":[99],"via":[100,127],"appropriate":[102],"synchronization":[103],"mechanism).":[104],"Furthermore,":[105],"(3)":[106],"explore":[108],"benefits":[110],"applying":[112],"pipelining":[113,134],"exchange,":[117],"demonstrating":[118],"these":[120],"improvements":[121],"carry":[122],"over":[123],"distributed":[125],"training":[126],"TF+HVD.":[128],"Finally,":[129],"(4)":[130],"show":[132],"also":[136],"boost":[137],"for":[139],"applications":[140],"make":[142],"heavy":[143],"use":[144],"other":[146],"collectives,":[147],"such":[148],"as":[149],"Broadcast":[150],"Reduce-Scatter.":[152]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":5},{"year":2024,"cited_by_count":1},{"year":2023,"cited_by_count":4},{"year":2022,"cited_by_count":4}],"updated_date":"2026-05-21T06:26:12.895304","created_date":"2025-10-10T00:00:00"}
