{"id":"https://openalex.org/W4413382536","doi":"https://doi.org/10.1145/3759244","title":"TorchBraid: High-Performance Layer-Parallel Training of Deep Neural Networks with MPI and GPU Acceleration","display_name":"TorchBraid: High-Performance Layer-Parallel Training of Deep Neural Networks with MPI and GPU Acceleration","publication_year":2025,"publication_date":"2025-08-14","ids":{"openalex":"https://openalex.org/W4413382536","doi":"https://doi.org/10.1145/3759244"},"language":"en","primary_location":{"id":"doi:10.1145/3759244","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3759244","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3759244","source":{"id":"https://openalex.org/S163326656","display_name":"ACM Transactions on Mathematical Software","issn_l":"0098-3500","issn":["0098-3500","1557-7295"],"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":"ACM Transactions on Mathematical Software","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"bronze","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3759244","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5052883660","display_name":"Eric C. Cyr","orcid":"https://orcid.org/0000-0003-3833-9598"},"institutions":[{"id":"https://openalex.org/I4210104735","display_name":"Sandia National Laboratories","ror":"https://ror.org/01apwpt12","country_code":"US","type":"facility","lineage":["https://openalex.org/I1330989302","https://openalex.org/I198811213","https://openalex.org/I4210104735"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Eric C. Cyr","raw_affiliation_strings":["Sandia National Laboratories, Albuquerque, New Mexico, USA","Sandia National Laboratories, USA"],"raw_orcid":"https://orcid.org/0000-0003-3833-9598","affiliations":[{"raw_affiliation_string":"Sandia National Laboratories, Albuquerque, New Mexico, USA","institution_ids":["https://openalex.org/I4210104735"]},{"raw_affiliation_string":"Sandia National Laboratories, USA","institution_ids":["https://openalex.org/I4210104735"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5022271457","display_name":"Jens Hahne","orcid":"https://orcid.org/0000-0002-7711-7894"},"institutions":[{"id":"https://openalex.org/I167360494","display_name":"University of Wuppertal","ror":"https://ror.org/00613ak93","country_code":"DE","type":"education","lineage":["https://openalex.org/I167360494"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Jens Hahne","raw_affiliation_strings":["Bergische Universit\u00e4t Wuppertal, Wuppertal, Germany","Bergische Universit\u00e4t Wuppertal, Germany"],"raw_orcid":"https://orcid.org/0000-0002-7711-7894","affiliations":[{"raw_affiliation_string":"Bergische Universit\u00e4t Wuppertal, Wuppertal, Germany","institution_ids":["https://openalex.org/I167360494"]},{"raw_affiliation_string":"Bergische Universit\u00e4t Wuppertal, Germany","institution_ids":["https://openalex.org/I167360494"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5107167253","display_name":"Nicholas S. Moore","orcid":"https://orcid.org/0000-0001-5224-9113"},"institutions":[{"id":"https://openalex.org/I188674872","display_name":"West Texas A&M University","ror":"https://ror.org/04gnp7x40","country_code":"US","type":"education","lineage":["https://openalex.org/I188674872"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Nicholas S. Moore","raw_affiliation_strings":["West Texas A&M University, Canyon, Texas, USA","West Texas A&M University, USA"],"raw_orcid":"https://orcid.org/0000-0001-5224-9113","affiliations":[{"raw_affiliation_string":"West Texas A&M University, Canyon, Texas, USA","institution_ids":["https://openalex.org/I188674872"]},{"raw_affiliation_string":"West Texas A&M University, USA","institution_ids":["https://openalex.org/I188674872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5038235793","display_name":"Jacob B. Schroder","orcid":"https://orcid.org/0000-0002-1076-9206"},"institutions":[{"id":"https://openalex.org/I169521973","display_name":"University of New Mexico","ror":"https://ror.org/05fs6jp91","country_code":"US","type":"education","lineage":["https://openalex.org/I169521973"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jacob B. Schroder","raw_affiliation_strings":["Department of Mathematics and Statistics, University of New Mexico, Albuquerque, New Mexico, USA","Dept. of Mathematics and Statistics, University of New Mexico, USA"],"raw_orcid":"https://orcid.org/0000-0002-1076-9206","affiliations":[{"raw_affiliation_string":"Department of Mathematics and Statistics, University of New Mexico, Albuquerque, New Mexico, USA","institution_ids":["https://openalex.org/I169521973"]},{"raw_affiliation_string":"Dept. of Mathematics and Statistics, University of New Mexico, USA","institution_ids":["https://openalex.org/I169521973"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5013739420","display_name":"Ben S. Southworth","orcid":"https://orcid.org/0000-0002-0283-4928"},"institutions":[{"id":"https://openalex.org/I1343871089","display_name":"Los Alamos National Laboratory","ror":"https://ror.org/01e41cf67","country_code":"US","type":"facility","lineage":["https://openalex.org/I1330989302","https://openalex.org/I1343871089","https://openalex.org/I198811213","https://openalex.org/I4210120050"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ben S. Southworth","raw_affiliation_strings":["Los Alamos National Laboratory, Los Alamos, New Mexico, USA","Los Alamos National Laboratory, USA"],"raw_orcid":"https://orcid.org/0000-0002-0283-4928","affiliations":[{"raw_affiliation_string":"Los Alamos National Laboratory, Los Alamos, New Mexico, USA","institution_ids":["https://openalex.org/I1343871089"]},{"raw_affiliation_string":"Los Alamos National Laboratory, USA","institution_ids":["https://openalex.org/I1343871089"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5020044747","display_name":"David A. Vargas","orcid":"https://orcid.org/0000-0002-7786-5753"},"institutions":[{"id":"https://openalex.org/I4210104735","display_name":"Sandia National Laboratories","ror":"https://ror.org/01apwpt12","country_code":"US","type":"facility","lineage":["https://openalex.org/I1330989302","https://openalex.org/I198811213","https://openalex.org/I4210104735"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"David A. Vargas","raw_affiliation_strings":["Sandia National Laboratories, Albuquerque, New Mexico, USA","Sandia National Laboratories, USA"],"raw_orcid":"https://orcid.org/0000-0002-7786-5753","affiliations":[{"raw_affiliation_string":"Sandia National Laboratories, Albuquerque, New Mexico, USA","institution_ids":["https://openalex.org/I4210104735"]},{"raw_affiliation_string":"Sandia National Laboratories, USA","institution_ids":["https://openalex.org/I4210104735"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":5,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.18002154,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"51","issue":"3","first_page":"1","last_page":"30"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9991000294685364,"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":0.9991000294685364,"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/T10320","display_name":"Neural Networks and Applications","score":0.9955999851226807,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9897000193595886,"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":0.7861225008964539},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.6961176991462708},{"id":"https://openalex.org/keywords/acceleration","display_name":"Acceleration","score":0.6365389227867126},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.5133008360862732},{"id":"https://openalex.org/keywords/layer","display_name":"Layer (electronics)","score":0.512003481388092},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.47082382440567017},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.45215320587158203},{"id":"https://openalex.org/keywords/computational-science","display_name":"Computational science","score":0.4381944239139557},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.2317616045475006},{"id":"https://openalex.org/keywords/physics","display_name":"Physics","score":0.07244101166725159},{"id":"https://openalex.org/keywords/materials-science","display_name":"Materials science","score":0.052289485931396484}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7861225008964539},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.6961176991462708},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.6365389227867126},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.5133008360862732},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.512003481388092},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.47082382440567017},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.45215320587158203},{"id":"https://openalex.org/C459310","wikidata":"https://www.wikidata.org/wiki/Q117801","display_name":"Computational science","level":1,"score":0.4381944239139557},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2317616045475006},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.07244101166725159},{"id":"https://openalex.org/C192562407","wikidata":"https://www.wikidata.org/wiki/Q228736","display_name":"Materials science","level":0,"score":0.052289485931396484},{"id":"https://openalex.org/C74650414","wikidata":"https://www.wikidata.org/wiki/Q11397","display_name":"Classical mechanics","level":1,"score":0.0},{"id":"https://openalex.org/C153294291","wikidata":"https://www.wikidata.org/wiki/Q25261","display_name":"Meteorology","level":1,"score":0.0},{"id":"https://openalex.org/C159985019","wikidata":"https://www.wikidata.org/wiki/Q181790","display_name":"Composite material","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3759244","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3759244","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3759244","source":{"id":"https://openalex.org/S163326656","display_name":"ACM Transactions on Mathematical Software","issn_l":"0098-3500","issn":["0098-3500","1557-7295"],"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":"ACM Transactions on Mathematical Software","raw_type":"journal-article"}],"best_oa_location":{"id":"doi:10.1145/3759244","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3759244","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3759244","source":{"id":"https://openalex.org/S163326656","display_name":"ACM Transactions on Mathematical Software","issn_l":"0098-3500","issn":["0098-3500","1557-7295"],"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":"ACM Transactions on Mathematical Software","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G5726497713","display_name":null,"funder_award_id":"INST 218/78-1 FUGG","funder_id":"https://openalex.org/F4320320879","funder_display_name":"Deutsche Forschungsgemeinschaft"},{"id":"https://openalex.org/G6684765602","display_name":null,"funder_award_id":"LA-UR-24-20385","funder_id":"https://openalex.org/F4320338304","funder_display_name":"Los Alamos National Laboratory"}],"funders":[{"id":"https://openalex.org/F4320320879","display_name":"Deutsche Forschungsgemeinschaft","ror":"https://ror.org/018mejw64"},{"id":"https://openalex.org/F4320321114","display_name":"Bundesministerium f\u00fcr Bildung und Forschung","ror":"https://ror.org/04pz7b180"},{"id":"https://openalex.org/F4320338304","display_name":"Los Alamos National Laboratory","ror":"https://ror.org/01e41cf67"}],"has_content":{"grobid_xml":false,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4413382536.pdf"},"referenced_works_count":47,"referenced_works":["https://openalex.org/W1559687661","https://openalex.org/W1853291804","https://openalex.org/W1920022804","https://openalex.org/W1967134278","https://openalex.org/W1995343013","https://openalex.org/W2001032894","https://openalex.org/W2006700733","https://openalex.org/W2012630580","https://openalex.org/W2059919683","https://openalex.org/W2079440004","https://openalex.org/W2083206954","https://openalex.org/W2099001231","https://openalex.org/W2104266030","https://openalex.org/W2120891799","https://openalex.org/W2128084896","https://openalex.org/W2128556617","https://openalex.org/W2194775991","https://openalex.org/W2524221349","https://openalex.org/W2600297185","https://openalex.org/W2612252966","https://openalex.org/W2765729611","https://openalex.org/W2914454439","https://openalex.org/W2962727772","https://openalex.org/W2963359731","https://openalex.org/W2963524234","https://openalex.org/W2963889991","https://openalex.org/W2964224480","https://openalex.org/W2964350391","https://openalex.org/W2969388332","https://openalex.org/W2972087877","https://openalex.org/W3004424267","https://openalex.org/W3010710024","https://openalex.org/W3033311037","https://openalex.org/W3035965352","https://openalex.org/W3098011980","https://openalex.org/W3099878876","https://openalex.org/W3164436820","https://openalex.org/W3186990848","https://openalex.org/W3217179853","https://openalex.org/W4210830077","https://openalex.org/W4223967157","https://openalex.org/W4231154391","https://openalex.org/W4287076657","https://openalex.org/W4378505344","https://openalex.org/W4386709668","https://openalex.org/W6852417290","https://openalex.org/W6926573507"],"related_works":["https://openalex.org/W4399354997","https://openalex.org/W2005148983","https://openalex.org/W2012954338","https://openalex.org/W2096672917","https://openalex.org/W2392023973","https://openalex.org/W3189307731","https://openalex.org/W1428699136","https://openalex.org/W2949962288","https://openalex.org/W2364686214","https://openalex.org/W1998560227"],"abstract_inverted_index":{"TorchBraid":[0,50,113,131],"is":[1,148],"a":[2,121,173,192],"high-performance":[3],"implementation":[4],"of":[5,34,39,70,117,143,158],"layer-parallel":[6,138,144,170,177],"training":[7,20,116,185],"for":[8,76,96,150,169,186],"deep":[9,162],"neural":[10,54,110],"networks":[11],"(DNNs)":[12],"supporting":[13],"MPI-based":[14],"parallelism":[15],"and":[16,31,68,92,100,123],"GPU":[17,94],"acceleration.":[18],"Layer-parallel":[19],"has":[21],"been":[22],"developed":[23,168],"to":[24,74,129,183],"overcome":[25],"the":[26,43,52,58,66,78,97,130,151,155,159],"serialization":[27],"inherent":[28,81],"in":[29,42,72,82,120,136,181],"forward":[30],"backward":[32],"propagation":[33],"DNNs":[35],"that":[36],"limits":[37],"utilization":[38],"computational":[40,156],"resources":[41],"strong":[44,122],"scaling":[45,125],"limit.":[46],"To":[47],"achieve":[48],"this,":[49],"integrates":[51],"PyTorch":[53],"network":[55],"framework":[56],"with":[57,86,91,145,178],"state-of-the-art":[59],"XBraid":[60],"time-parallel":[61],"library.":[62],"This":[63],"article":[64],"presents":[65],"use":[67],"performance":[69],"TorchBraid,":[71],"addition":[73,128],"solutions":[75],"overcoming":[77],"algorithmic":[79],"challenges":[80],"combining":[83,176],"automatic":[84],"differentiation":[85],"layer-parallel.":[87],"Results":[88],"are":[89,140,167],"presented":[90,149],"without":[93],"acceleration":[95],"Tiny":[98],"ImageNet":[99],"MNIST":[101],"image":[102,188],"classification":[103,189],"data":[104],"sets,":[105],"as":[106,108],"well":[107],"recurrent":[109],"networks.":[111],"Overall,":[112],"enables":[114],"fast":[115],"DNNs,":[118],"both":[119],"weak":[124],"context.":[126],"In":[127],"software,":[132],"several":[133],"new":[134,174],"advances":[135],"applying":[137],"algorithms":[139,147],"detailed.":[141],"Integration":[142],"data-parallel":[146],"first":[152],"time,":[153],"showing":[154],"advantages":[157],"combination.":[160],"Standard":[161],"learning":[163],"techniques,":[164],"like":[165],"batch-normalization,":[166],"training.":[171],"Finally,":[172],"approach":[175],"spatial":[179],"coarsening":[180],"order":[182],"accelerate":[184],"3D":[187],"shows":[190],"roughly":[191],"10\u00d7":[193],"speedup":[194],"over":[195],"serial":[196],"execution.":[197]},"counts_by_year":[],"updated_date":"2026-05-21T06:26:12.895304","created_date":"2025-10-10T00:00:00"}