{"id":"https://openalex.org/W7161716988","doi":"https://doi.org/10.48550/arxiv.2605.17821","title":"TierCheck: Tiered Checkpointing for Fault Tolerance in Large Language Model Training","display_name":"TierCheck: Tiered Checkpointing for Fault Tolerance in Large Language Model Training","publication_year":2026,"publication_date":"2026-05-18","ids":{"openalex":"https://openalex.org/W7161716988","doi":"https://doi.org/10.48550/arxiv.2605.17821"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.17821","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.17821","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.17821","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5064254673","display_name":"Shujie Han","orcid":"https://orcid.org/0000-0001-5311-5782"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Han, Shujie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136480607","display_name":"Feng Jiang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiang, Feng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5012129385","display_name":"Patrick P. C. Lee","orcid":"https://orcid.org/0000-0002-4501-4364"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lee, Patrick P. C.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136469295","display_name":"Xiao Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Xiao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136474655","display_name":"Zhijie Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Zhijie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136466939","display_name":"Nannan Zhao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Nannan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101687035","display_name":"Xiaonan Zhao","orcid":"https://orcid.org/0009-0008-9481-4879"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Xiaonan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5091272448","display_name":"Lichen Pan","orcid":"https://orcid.org/0000-0001-7451-0140"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pan, Lichen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10772","display_name":"Distributed systems and fault tolerance","score":0.8119999766349792,"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/T10772","display_name":"Distributed systems and fault tolerance","score":0.8119999766349792,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.03739999979734421,"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"}},{"id":"https://openalex.org/T12127","display_name":"Software System Performance and Reliability","score":0.020600000396370888,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.7164999842643738},{"id":"https://openalex.org/keywords/fault-tolerance","display_name":"Fault tolerance","score":0.6596999764442444},{"id":"https://openalex.org/keywords/forcing","display_name":"Forcing (mathematics)","score":0.5978000164031982},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.4948999881744385},{"id":"https://openalex.org/keywords/consistency","display_name":"Consistency (knowledge bases)","score":0.439300000667572},{"id":"https://openalex.org/keywords/base","display_name":"Base (topology)","score":0.43700000643730164},{"id":"https://openalex.org/keywords/scheme","display_name":"Scheme (mathematics)","score":0.42320001125335693},{"id":"https://openalex.org/keywords/reliability","display_name":"Reliability (semiconductor)","score":0.39750000834465027},{"id":"https://openalex.org/keywords/software-fault-tolerance","display_name":"Software fault tolerance","score":0.36149999499320984}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.819599986076355},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.7164999842643738},{"id":"https://openalex.org/C63540848","wikidata":"https://www.wikidata.org/wiki/Q3140932","display_name":"Fault tolerance","level":2,"score":0.6596999764442444},{"id":"https://openalex.org/C197115733","wikidata":"https://www.wikidata.org/wiki/Q1003136","display_name":"Forcing (mathematics)","level":2,"score":0.5978000164031982},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.5925999879837036},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.4948999881744385},{"id":"https://openalex.org/C2776436953","wikidata":"https://www.wikidata.org/wiki/Q5163215","display_name":"Consistency (knowledge bases)","level":2,"score":0.439300000667572},{"id":"https://openalex.org/C42058472","wikidata":"https://www.wikidata.org/wiki/Q810214","display_name":"Base (topology)","level":2,"score":0.43700000643730164},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.42320001125335693},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.39750000834465027},{"id":"https://openalex.org/C50712370","wikidata":"https://www.wikidata.org/wiki/Q4269346","display_name":"Software fault tolerance","level":3,"score":0.36149999499320984},{"id":"https://openalex.org/C93226319","wikidata":"https://www.wikidata.org/wiki/Q193137","display_name":"Differential (mechanical device)","level":2,"score":0.36090001463890076},{"id":"https://openalex.org/C65813073","wikidata":"https://www.wikidata.org/wiki/Q1622420","display_name":"High availability","level":2,"score":0.3474000096321106},{"id":"https://openalex.org/C200601418","wikidata":"https://www.wikidata.org/wiki/Q2193887","display_name":"Reliability engineering","level":1,"score":0.323199987411499},{"id":"https://openalex.org/C167391956","wikidata":"https://www.wikidata.org/wiki/Q1401211","display_name":"Fault model","level":3,"score":0.3174999952316284},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.31709998846054077},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.3095000088214874},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3068999946117401},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.30410000681877136},{"id":"https://openalex.org/C84511453","wikidata":"https://www.wikidata.org/wiki/Q2914952","display_name":"Concurrency control","level":3,"score":0.2985000014305115},{"id":"https://openalex.org/C93361087","wikidata":"https://www.wikidata.org/wiki/Q4426698","display_name":"Data consistency","level":2,"score":0.29510000348091125},{"id":"https://openalex.org/C112987892","wikidata":"https://www.wikidata.org/wiki/Q5051574","display_name":"Catastrophic failure","level":2,"score":0.2786000072956085},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.27630001306533813},{"id":"https://openalex.org/C175551986","wikidata":"https://www.wikidata.org/wiki/Q47089","display_name":"Fault (geology)","level":2,"score":0.27219998836517334},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.2687999904155731},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2671999931335449},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.26249998807907104},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.25699999928474426},{"id":"https://openalex.org/C2780385302","wikidata":"https://www.wikidata.org/wiki/Q367158","display_name":"Protocol (science)","level":3,"score":0.2547000050544739},{"id":"https://openalex.org/C151319957","wikidata":"https://www.wikidata.org/wiki/Q752739","display_name":"Asynchronous communication","level":2,"score":0.25360000133514404},{"id":"https://openalex.org/C888380","wikidata":"https://www.wikidata.org/wiki/Q2427787","display_name":"Persistent data structure","level":2,"score":0.2515999972820282}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.17821","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.17821","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.17821","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.17821","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"Language":[1],"Model":[2],"(LLM)":[3],"training":[4,116],"is":[5],"frequently":[6],"interrupted":[7],"by":[8],"a":[9,32,43,57],"heterogeneous":[10],"spectrum":[11],"of":[12],"failures,":[13],"from":[14],"common":[15],"GPU":[16],"crashes":[17],"to":[18,80,107,122],"catastrophic":[19],"cluster-wide":[20],"outages.":[21],"Existing":[22],"checkpointing":[23,46,120],"systems":[24],"rely":[25],"on":[26,104],"monolithic,":[27],"single-tier":[28],"storage":[29,50],"backend,":[30],"forcing":[31],"trade-off":[33],"between":[34,134],"state-saving":[35],"overhead":[36],"and":[37,67,95,125,137],"recovery":[38],"speed.":[39],"We":[40],"propose":[41],"TierCheck,":[42],"cluster-aware":[44,98],"tiered":[45],"system":[47],"that":[48,60,112],"aligns":[49],"placement":[51],"with":[52],"failure":[53],"heterogeneity.":[54],"TierCheck":[55,113],"adopts":[56],"three-tier":[58],"design":[59],"maintains":[61],"lightweight":[62],"differential":[63],"checkpoints":[64,79],"in":[65],"local":[66],"peer":[68],"memory":[69],"for":[70],"fast":[71,97,138],"localized":[72],"recovery,":[73],"while":[74],"asynchronously":[75],"migrating":[76],"heavyweight":[77],"base":[78],"remote":[81],"persistent":[82],"storage.":[83],"It":[84],"also":[85],"ensures":[86],"strict":[87],"global":[88],"consistency":[89],"across":[90],"tiers":[91],"without":[92],"stalling":[93],"training,":[94],"achieves":[96,114],"checkpoint":[99],"restoration":[100],"during":[101],"recovery.":[102,139],"Evaluations":[103],"models":[105],"up":[106],"40":[108],"billion":[109],"parameters":[110],"show":[111],"low":[115],"overhead,":[117],"reduces":[118],"end-to-end":[119],"time":[121],"under":[123],"10s,":[124],"supports":[126],"high-frequency":[127],"checkpointing,":[128],"ultimately":[129],"striking":[130],"an":[131],"optimal":[132],"balance":[133],"low-overhead":[135],"persistence":[136]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-20T00:00:00"}
