{"id":"https://openalex.org/W4414054306","doi":"https://doi.org/10.1109/tnse.2025.3607331","title":"Parallel Gradient Computation and Synchronization: Enhancing the Efficiency of Distributed Training for LLMs","display_name":"Parallel Gradient Computation and Synchronization: Enhancing the Efficiency of Distributed Training for LLMs","publication_year":2025,"publication_date":"2025-09-08","ids":{"openalex":"https://openalex.org/W4414054306","doi":"https://doi.org/10.1109/tnse.2025.3607331"},"language":"en","primary_location":{"id":"doi:10.1109/tnse.2025.3607331","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tnse.2025.3607331","pdf_url":null,"source":{"id":"https://openalex.org/S2484352698","display_name":"IEEE Transactions on Network Science and Engineering","issn_l":"2327-4697","issn":["2327-4697","2334-329X"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Network Science and Engineering","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/A5077976838","display_name":"Hao Li","orcid":"https://orcid.org/0000-0002-6258-0878"},"institutions":[{"id":"https://openalex.org/I37461747","display_name":"Wuhan University","ror":"https://ror.org/033vjfk17","country_code":"CN","type":"education","lineage":["https://openalex.org/I37461747"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Hao Li","raw_affiliation_strings":["Electronic Information School, Wuhan University, Wuhan, China"],"affiliations":[{"raw_affiliation_string":"Electronic Information School, Wuhan University, Wuhan, China","institution_ids":["https://openalex.org/I37461747"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100616426","display_name":"Hao Jiang","orcid":"https://orcid.org/0000-0002-8533-1612"},"institutions":[{"id":"https://openalex.org/I37461747","display_name":"Wuhan University","ror":"https://ror.org/033vjfk17","country_code":"CN","type":"education","lineage":["https://openalex.org/I37461747"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hao Jiang","raw_affiliation_strings":["Electronic Information School, Wuhan University, Wuhan, China"],"affiliations":[{"raw_affiliation_string":"Electronic Information School, Wuhan University, Wuhan, China","institution_ids":["https://openalex.org/I37461747"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5086080556","display_name":"Jing Wu","orcid":"https://orcid.org/0000-0002-6955-2087"},"institutions":[{"id":"https://openalex.org/I37461747","display_name":"Wuhan University","ror":"https://ror.org/033vjfk17","country_code":"CN","type":"education","lineage":["https://openalex.org/I37461747"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jing Wu","raw_affiliation_strings":["Electronic Information School, Wuhan University, Wuhan, China"],"affiliations":[{"raw_affiliation_string":"Electronic Information School, Wuhan University, Wuhan, China","institution_ids":["https://openalex.org/I37461747"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Guiao Yang","orcid":"https://orcid.org/0009-0005-4919-5712"},"institutions":[{"id":"https://openalex.org/I37461747","display_name":"Wuhan University","ror":"https://ror.org/033vjfk17","country_code":"CN","type":"education","lineage":["https://openalex.org/I37461747"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Guiao Yang","raw_affiliation_strings":["Electronic Information School, Wuhan University, Wuhan, China"],"affiliations":[{"raw_affiliation_string":"Electronic Information School, Wuhan University, Wuhan, China","institution_ids":["https://openalex.org/I37461747"]}]},{"author_position":"last","author":{"id":null,"display_name":"Jian Zhang","orcid":"https://orcid.org/0000-0003-4394-0628"},"institutions":[{"id":"https://openalex.org/I37461747","display_name":"Wuhan University","ror":"https://ror.org/033vjfk17","country_code":"CN","type":"education","lineage":["https://openalex.org/I37461747"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jian Zhang","raw_affiliation_strings":["Electronic Information School, Wuhan University, Wuhan, China"],"affiliations":[{"raw_affiliation_string":"Electronic Information School, Wuhan University, Wuhan, China","institution_ids":["https://openalex.org/I37461747"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5077976838"],"corresponding_institution_ids":["https://openalex.org/I37461747"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.26673645,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"13","issue":null,"first_page":"1962","last_page":"1976"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10715","display_name":"Distributed and Parallel Computing Systems","score":0.9937000274658203,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T10715","display_name":"Distributed and Parallel Computing Systems","score":0.9937000274658203,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T11986","display_name":"Scientific Computing and Data Management","score":0.9465000033378601,"subfield":{"id":"https://openalex.org/subfields/1802","display_name":"Information Systems and Management"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9158999919891357,"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/synchronization","display_name":"Synchronization (alternating current)","score":0.8167999982833862},{"id":"https://openalex.org/keywords/data-synchronization","display_name":"Data synchronization","score":0.6238999962806702},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.5807999968528748},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.5486000180244446},{"id":"https://openalex.org/keywords/load-balancing","display_name":"Load balancing (electrical power)","score":0.4526999890804291},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.4300000071525574},{"id":"https://openalex.org/keywords/bandwidth","display_name":"Bandwidth (computing)","score":0.39570000767707825},{"id":"https://openalex.org/keywords/gradient-descent","display_name":"Gradient descent","score":0.3495999872684479}],"concepts":[{"id":"https://openalex.org/C2778562939","wikidata":"https://www.wikidata.org/wiki/Q1298791","display_name":"Synchronization (alternating current)","level":3,"score":0.8167999982833862},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7702999711036682},{"id":"https://openalex.org/C108734733","wikidata":"https://www.wikidata.org/wiki/Q1172333","display_name":"Data synchronization","level":3,"score":0.6238999962806702},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5807999968528748},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5486000180244446},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.5080999732017517},{"id":"https://openalex.org/C138959212","wikidata":"https://www.wikidata.org/wiki/Q1806783","display_name":"Load balancing (electrical power)","level":3,"score":0.4526999890804291},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.4300000071525574},{"id":"https://openalex.org/C2776257435","wikidata":"https://www.wikidata.org/wiki/Q1576430","display_name":"Bandwidth (computing)","level":2,"score":0.39570000767707825},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.3495999872684479},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.33889999985694885},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.321399986743927},{"id":"https://openalex.org/C70061542","wikidata":"https://www.wikidata.org/wiki/Q989016","display_name":"Distributed database","level":2,"score":0.3075000047683716},{"id":"https://openalex.org/C115680565","wikidata":"https://www.wikidata.org/wiki/Q5977448","display_name":"Gradient method","level":2,"score":0.3059999942779541},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.3041999936103821},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.30219998955726624},{"id":"https://openalex.org/C130120984","wikidata":"https://www.wikidata.org/wiki/Q2835898","display_name":"Distributed algorithm","level":2,"score":0.2935999929904938},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.2770000100135803},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2750000059604645},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2736000120639801},{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.26429998874664307},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.26100000739097595},{"id":"https://openalex.org/C183322885","wikidata":"https://www.wikidata.org/wiki/Q17007702","display_name":"Context model","level":3,"score":0.25929999351501465}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tnse.2025.3607331","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tnse.2025.3607331","pdf_url":null,"source":{"id":"https://openalex.org/S2484352698","display_name":"IEEE Transactions on Network Science and Engineering","issn_l":"2327-4697","issn":["2327-4697","2334-329X"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Network Science and Engineering","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G2758118447","display_name":null,"funder_award_id":"2022YFB2901202","funder_id":"https://openalex.org/F4320335777","funder_display_name":"National Key Research and Development Program of China"}],"funders":[{"id":"https://openalex.org/F4320335777","display_name":"National Key Research and Development Program of China","ror":null}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":41,"referenced_works":["https://openalex.org/W2299112634","https://openalex.org/W2970918514","https://openalex.org/W2981114289","https://openalex.org/W2982668898","https://openalex.org/W2991040477","https://openalex.org/W3035748372","https://openalex.org/W3043699074","https://openalex.org/W3092260380","https://openalex.org/W3154431492","https://openalex.org/W3159351344","https://openalex.org/W3176617251","https://openalex.org/W4214711850","https://openalex.org/W4294433898","https://openalex.org/W4294762998","https://openalex.org/W4312968147","https://openalex.org/W4321392172","https://openalex.org/W4386707689","https://openalex.org/W4388756711","https://openalex.org/W4390692489","https://openalex.org/W4391855109","https://openalex.org/W4393186514","https://openalex.org/W4394564255","https://openalex.org/W4394945032","https://openalex.org/W4396814955","https://openalex.org/W4396817294","https://openalex.org/W4400056121","https://openalex.org/W4400680326","https://openalex.org/W4402401807","https://openalex.org/W4402742293","https://openalex.org/W4403442463","https://openalex.org/W4404783473","https://openalex.org/W4406459136","https://openalex.org/W4407213345","https://openalex.org/W4408355016","https://openalex.org/W4409076673","https://openalex.org/W4409761834","https://openalex.org/W4409882952","https://openalex.org/W4409991262","https://openalex.org/W4410229517","https://openalex.org/W4411485189","https://openalex.org/W4412536562"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W230091440","https://openalex.org/W2390279801","https://openalex.org/W2233261550","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2059217922"],"abstract_inverted_index":{"As":[0],"the":[1,9,32,44,50,106,123,131],"size":[2],"of":[3,11,46,125,133],"large":[4],"language":[5],"models":[6],"(LLMs)":[7],"increases,":[8],"limitations":[10],"a":[12,36,62,96],"single":[13],"data":[14,30,55,120],"center,":[15],"such":[16],"as":[17],"constrained":[18],"computational":[19],"resources":[20],"and":[21,72,115,127,141],"storage":[22],"capacity,":[23],"have":[24],"made":[25],"distributed":[26,63,134],"training":[27,51,64,135],"across":[28,53,118],"multiple":[29,54,119],"centers":[31],"preferred":[33],"solution.":[34],"However,":[35],"primary":[37],"challenge":[38],"in":[39,92,129],"this":[40,58],"context":[41],"is":[42,80,102,136],"reducing":[43],"impact":[45],"gradient":[47,70,90,97,108],"synchronization":[48,73,91,98,109],"on":[49],"efficiency":[52,132],"centers.":[56,121],"In":[57,94],"work,":[59],"we":[60],"propose":[61],"scheme":[65],"for":[66],"LLMs,":[67],"named":[68,100],"parallel":[69],"computation":[71],"(PGCS).":[74],"Specifically,":[75],"while":[76],"one":[77],"expert":[78,87],"model":[79,88],"being":[81],"trained":[82],"to":[83,104],"compute":[84],"gradients,":[85],"another":[86],"performs":[89],"parallel.":[93],"addition,":[95],"algorithm":[99],"BLP":[101,128],"developed":[103],"find":[105],"optimal":[107],"strategy":[110],"under":[111],"arbitrary":[112],"network":[113],"connectivity":[114],"limited":[116],"bandwidth":[117],"Ultimately,":[122],"effectiveness":[124],"PGCS":[126],"enhancing":[130],"demonstrated":[137],"through":[138],"comprehensive":[139],"simulations":[140],"physical":[142],"experiments.":[143]},"counts_by_year":[],"updated_date":"2026-04-09T08:11:56.329763","created_date":"2025-10-10T00:00:00"}