{"id":"https://openalex.org/W4414898674","doi":"https://doi.org/10.1109/icdcs63083.2025.00085","title":"RapidScribe: Bandwidth-aware Parallel Checkpoint for Distributed Neural-Network Training","display_name":"RapidScribe: Bandwidth-aware Parallel Checkpoint for Distributed Neural-Network Training","publication_year":2025,"publication_date":"2025-07-21","ids":{"openalex":"https://openalex.org/W4414898674","doi":"https://doi.org/10.1109/icdcs63083.2025.00085"},"language":"en","primary_location":{"id":"doi:10.1109/icdcs63083.2025.00085","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icdcs63083.2025.00085","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE 45th International Conference on Distributed Computing Systems (ICDCS)","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/A5101564554","display_name":"Wenxuan Liu","orcid":"https://orcid.org/0000-0002-1528-4224"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Wenxuan Liu","raw_affiliation_strings":["Shanghai Jiao Tong University,Institute of Image Communication and Network Engineering"],"affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University,Institute of Image Communication and Network Engineering","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5011017270","display_name":"Qihang Chen","orcid":"https://orcid.org/0000-0001-9034-9895"},"institutions":[{"id":"https://openalex.org/I63966007","display_name":"Massachusetts Institute of Technology","ror":"https://ror.org/042nb2s44","country_code":"US","type":"education","lineage":["https://openalex.org/I63966007"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Qihang Chen","raw_affiliation_strings":["Massachusetts Institute of Technology"],"affiliations":[{"raw_affiliation_string":"Massachusetts Institute of Technology","institution_ids":["https://openalex.org/I63966007"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5066106377","display_name":"Shuotao Xu","orcid":"https://orcid.org/0009-0000-8593-5443"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Shuotao Xu","raw_affiliation_strings":["Microsoft Research"],"affiliations":[{"raw_affiliation_string":"Microsoft Research","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101850449","display_name":"Hua Yang","orcid":"https://orcid.org/0000-0002-0417-234X"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hua Yang","raw_affiliation_strings":["Shanghai Jiao Tong University,Institute of Image Communication and Network Engineering"],"affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University,Institute of Image Communication and Network Engineering","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101421201","display_name":"Yuqing Yang","orcid":"https://orcid.org/0000-0003-3518-5212"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Yuqing Yang","raw_affiliation_strings":["Microsoft Research"],"affiliations":[{"raw_affiliation_string":"Microsoft Research","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101478278","display_name":"Peng Cheng","orcid":"https://orcid.org/0000-0003-4014-4757"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Peng Cheng","raw_affiliation_strings":["Microsoft Research"],"affiliations":[{"raw_affiliation_string":"Microsoft Research","institution_ids":["https://openalex.org/I4210164937"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5101564554"],"corresponding_institution_ids":["https://openalex.org/I183067930"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.34260146,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"824","last_page":"834"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12702","display_name":"Brain Tumor Detection and Classification","score":0.9444000124931335,"subfield":{"id":"https://openalex.org/subfields/2808","display_name":"Neurology"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},"topics":[{"id":"https://openalex.org/T12702","display_name":"Brain Tumor Detection and Classification","score":0.9444000124931335,"subfield":{"id":"https://openalex.org/subfields/2808","display_name":"Neurology"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},{"id":"https://openalex.org/T10320","display_name":"Neural Networks and Applications","score":0.9158999919891357,"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/fault-tolerance","display_name":"Fault tolerance","score":0.5533000230789185},{"id":"https://openalex.org/keywords/redundancy","display_name":"Redundancy (engineering)","score":0.5218999981880188},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.4551999866962433},{"id":"https://openalex.org/keywords/bandwidth","display_name":"Bandwidth (computing)","score":0.41940000653266907},{"id":"https://openalex.org/keywords/copying","display_name":"Copying","score":0.38929998874664307},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.38530001044273376}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8458999991416931},{"id":"https://openalex.org/C63540848","wikidata":"https://www.wikidata.org/wiki/Q3140932","display_name":"Fault tolerance","level":2,"score":0.5533000230789185},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.5218999981880188},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.4551999866962433},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.45249998569488525},{"id":"https://openalex.org/C2776257435","wikidata":"https://www.wikidata.org/wiki/Q1576430","display_name":"Bandwidth (computing)","level":2,"score":0.41940000653266907},{"id":"https://openalex.org/C2779151265","wikidata":"https://www.wikidata.org/wiki/Q1156791","display_name":"Copying","level":2,"score":0.38929998874664307},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.38530001044273376},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.36550000309944153},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.35850000381469727},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.33980000019073486},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3294000029563904},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.3215999901294708},{"id":"https://openalex.org/C167391956","wikidata":"https://www.wikidata.org/wiki/Q1401211","display_name":"Fault model","level":3,"score":0.30390000343322754},{"id":"https://openalex.org/C2779585090","wikidata":"https://www.wikidata.org/wiki/Q3457762","display_name":"Resilience (materials science)","level":2,"score":0.302700012922287},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.27959999442100525},{"id":"https://openalex.org/C83283714","wikidata":"https://www.wikidata.org/wiki/Q121117","display_name":"Supercomputer","level":2,"score":0.26030001044273376},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.25609999895095825}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icdcs63083.2025.00085","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icdcs63083.2025.00085","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE 45th International Conference on Distributed Computing Systems (ICDCS)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320321885","display_name":"Science and Technology Commission of Shanghai Municipality","ror":"https://ror.org/03kt66j61"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":21,"referenced_works":["https://openalex.org/W1987884672","https://openalex.org/W2025024269","https://openalex.org/W2032933946","https://openalex.org/W2045271686","https://openalex.org/W2064675550","https://openalex.org/W2130004097","https://openalex.org/W2131053137","https://openalex.org/W2194775991","https://openalex.org/W2798289565","https://openalex.org/W2963446712","https://openalex.org/W2991040477","https://openalex.org/W3021124033","https://openalex.org/W3129831491","https://openalex.org/W3204998121","https://openalex.org/W4248578633","https://openalex.org/W4381612905","https://openalex.org/W4386730038","https://openalex.org/W4387302750","https://openalex.org/W4388187930","https://openalex.org/W4389519413","https://openalex.org/W4401018280"],"related_works":[],"abstract_inverted_index":{"As":[0],"machine":[1],"learning":[2],"models":[3],"grow":[4],"in":[5,33,96],"complexity":[6],"and":[7,36,119],"size,":[8],"necessitating":[9],"the":[10,16,51,102,107],"use":[11,44],"of":[12,18,104],"extensive":[13],"GPU":[14,45,59,85,93,185,194],"clusters,":[15],"challenge":[17,80],"managing":[19],"frequent":[20,67,162],"system":[21],"failures":[22,27],"becomes":[23],"increasingly":[24],"critical.":[25],"These":[26],"can":[28],"lead":[29],"to":[30,88,94,127,134,193],"substantial":[31],"losses":[32],"computational":[34],"time":[35],"resources.":[37],"Large-scale":[38],"model":[39,120],"training":[40,76,121,132,136,171,179],"often":[41],"does":[42],"not":[43],"redundancy":[46],"for":[47,53,58,150],"fault":[48,195],"tolerance,":[49],"underscoring":[50],"necessity":[52],"an":[54],"efficient":[55],"checkpoint":[56,112,158,163],"mechanism":[57],"failure":[60,63],"recovery.":[61],"Enhanced":[62],"resilience":[64],"demands":[65],"more":[66],"checkpoints;":[68],"however":[69],"suspend-and-resume":[70],"based":[71],"checkpoints":[72],"could":[73],"severely":[74],"lower":[75],"throughput.RapidScribe":[77],"addresses":[78],"this":[79],"head-on":[81],"by":[82,115,142],"leveraging":[83],"unused":[84],"bus":[86],"bandwidth":[87,140],"facilitate":[89],"tensor":[90],"transfers":[91],"from":[92],"host":[95],"concurrent":[97],"with":[98,168],"ongoing":[99],"training.":[100],"At":[101],"heart":[103],"RapidScribe":[105],"is":[106],"Steno":[108,123,143],"protocol,":[109],"which":[110],"lowers":[111],"traffic":[113],"aggressively":[114],"copying":[116],"both":[117],"gradients":[118],"states.":[122],"generates":[124],"copy":[125],"schedules":[126],"be":[128],"fully":[129],"interleaved":[130],"within":[131],"processes":[133],"minimize":[135],"interruptions.":[137],"The":[138],"low":[139],"usage":[141],"makes":[144],"it":[145],"even":[146],"a":[147,182,189],"viable":[148],"solution":[149,192],"disk-based":[151],"checkpoints.":[152],"Experiments":[153],"show":[154],"that":[155,161],"RapidScribe\u2019s":[156],"parallel":[157],"techniques":[159],"ensure":[160],"operations":[164],"are":[165],"deeply":[166],"synchronized":[167],"neural":[169],"network":[170],"cycles,":[172],"preserving":[173],"near-baseline":[174],"throughput":[175],"across":[176],"various":[177],"distributed":[178],"configurations":[180],"on":[181],"AMD":[183],"MI250":[184],"cluster,":[186],"thus":[187],"providing":[188],"robust":[190],"checkpoint-based":[191],"tolerance.":[196]},"counts_by_year":[],"updated_date":"2026-04-09T08:11:56.329763","created_date":"2025-10-10T00:00:00"}