{"id":"https://openalex.org/W4416203809","doi":"https://doi.org/10.1145/3712285.3759802","title":"BurstEngine: An efficient distributed framework for training transformers On extremely Long sequences of over 1M tokens","display_name":"BurstEngine: An efficient distributed framework for training transformers On extremely Long sequences of over 1M tokens","publication_year":2025,"publication_date":"2025-11-12","ids":{"openalex":"https://openalex.org/W4416203809","doi":"https://doi.org/10.1145/3712285.3759802"},"language":null,"primary_location":{"id":"doi:10.1145/3712285.3759802","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3712285.3759802","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3712285.3759802","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5120351183","display_name":"Ao Sun","orcid":"https://orcid.org/0009-0002-3631-1233"},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Ao Sun","raw_affiliation_strings":["Beijing University of Posts and Telecommunications, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications, Beijing, China","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5091743314","display_name":"Weilin Zhao","orcid":"https://orcid.org/0000-0001-8016-1952"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Weilin Zhao","raw_affiliation_strings":["Tsinghua University, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100668503","display_name":"Xu Han","orcid":"https://orcid.org/0000-0002-4726-7621"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xu Han","raw_affiliation_strings":["Tsinghua University, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5060417049","display_name":"Cheng Yang","orcid":"https://orcid.org/0000-0001-7821-0030"},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Cheng Yang","raw_affiliation_strings":["Beijing University of Posts and Telecommunications, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications, Beijing, China","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100320723","display_name":"Zhiyuan Liu","orcid":"https://orcid.org/0000-0002-7709-2543"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhiyuan Liu","raw_affiliation_strings":["Tsinghua University, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100705849","display_name":"Chuan Shi","orcid":"https://orcid.org/0000-0002-3734-0266"},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chuan Shi","raw_affiliation_strings":["Beijing University of Posts and Telecommunications, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications, Beijing, China","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5046448314","display_name":"Maosong Sun","orcid":"https://orcid.org/0000-0002-6011-6115"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Maosong Sun","raw_affiliation_strings":["Tsinghua University, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5120351183"],"corresponding_institution_ids":["https://openalex.org/I139759216"],"apc_list":null,"apc_paid":null,"fwci":2.5505,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.91639222,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":91,"max":95},"biblio":{"volume":null,"issue":null,"first_page":"1429","last_page":"1445"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.39660000801086426,"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"}},"topics":[{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.39660000801086426,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.14910000562667847,"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/T12808","display_name":"Ferroelectric and Negative Capacitance Devices","score":0.05490000173449516,"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/workload","display_name":"Workload","score":0.5306000113487244},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.5023999810218811},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.4666999876499176},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.40389999747276306},{"id":"https://openalex.org/keywords/sequence","display_name":"Sequence (biology)","score":0.3855000138282776},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.384799987077713},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.36160001158714294},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.35499998927116394}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8321999907493591},{"id":"https://openalex.org/C2778476105","wikidata":"https://www.wikidata.org/wiki/Q628539","display_name":"Workload","level":2,"score":0.5306000113487244},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.5023999810218811},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.49459999799728394},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.4666999876499176},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.4383000135421753},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.40389999747276306},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.3855000138282776},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.384799987077713},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.37059998512268066},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.36160001158714294},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.35499998927116394},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.337799996137619},{"id":"https://openalex.org/C854659","wikidata":"https://www.wikidata.org/wiki/Q1859284","display_name":"Message passing","level":2,"score":0.33169999718666077},{"id":"https://openalex.org/C183322885","wikidata":"https://www.wikidata.org/wiki/Q17007702","display_name":"Context model","level":3,"score":0.3034000098705292},{"id":"https://openalex.org/C3826847","wikidata":"https://www.wikidata.org/wiki/Q188768","display_name":"FLOPS","level":2,"score":0.2980000078678131},{"id":"https://openalex.org/C2776257435","wikidata":"https://www.wikidata.org/wiki/Q1576430","display_name":"Bandwidth (computing)","level":2,"score":0.29409998655319214},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2930000126361847},{"id":"https://openalex.org/C55282118","wikidata":"https://www.wikidata.org/wiki/Q252683","display_name":"Snapshot (computer storage)","level":2,"score":0.2912999987602234},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.25999999046325684},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.2538999915122986}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3712285.3759802","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3712285.3759802","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3712285.3759802","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3712285.3759802","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":10,"referenced_works":["https://openalex.org/W2045271686","https://openalex.org/W3015286511","https://openalex.org/W3129831491","https://openalex.org/W3198659451","https://openalex.org/W3204998121","https://openalex.org/W4206706211","https://openalex.org/W4213019189","https://openalex.org/W4306955484","https://openalex.org/W4386768656","https://openalex.org/W4402713111"],"related_works":[],"abstract_inverted_index":{"Existing":[0],"methods":[1],"for":[2,106],"training":[3,132],"LLMs":[4,48,133],"on":[5,49,134],"long-sequence":[6,50],"data,":[7],"such":[8],"as":[9,20],"Tensor":[10],"Parallelism":[11],"and":[12,23,75,86],"Context":[13],"Parallelism,":[14],"exhibit":[15],"low":[16],"Model":[17],"FLOPs":[18],"Utilization":[19],"sequence":[21,30],"lengths":[22,31],"number":[24],"of":[25,109,138],"GPUs":[26],"increase,":[27],"especially":[28],"when":[29,131],"exceed":[32],"1M":[33,140],"tokens.":[34,141],"To":[35],"address":[36],"these":[37,114],"challenges,":[38],"we":[39],"propose":[40],"BurstEngine,":[41],"an":[42,55],"efficient":[43],"framework":[44],"designed":[45],"to":[46,70,96],"train":[47],"data.":[51],"BurstEngine":[52,81,101,116],"introduces":[53,82,102],"BurstAttention,":[54],"optimized":[56],"distributed":[57],"attention":[58,110],"with":[59,92,122],"lower":[60,124],"communication":[61,69],"cost":[62],"than":[63,127],"RingAttention.":[64],"BurstAttention":[65],"leverages":[66],"topology-aware":[67],"ring":[68],"fully":[71],"utilize":[72],"network":[73],"bandwidth":[74],"incorporates":[76],"fine-grained":[77],"communication-computation":[78],"overlap.":[79],"Furthermore,":[80],"sequence-level":[83],"selective":[84],"checkpointing":[85],"fuses":[87],"the":[88,93,128],"language":[89],"modeling":[90],"head":[91],"loss":[94],"function":[95],"reduce":[97],"memory":[98,125],"cost.":[99],"Additionally,":[100],"workload":[103],"balance":[104],"optimization":[105],"various":[107],"types":[108],"masking.":[111],"By":[112],"integrating":[113],"optimizations,":[115],"achieves":[117],"a":[118],"1.2":[119],"\u00d7":[120],"speedup":[121],"much":[123],"overhead":[126],"state-of-the-art":[129],"baselines":[130],"extremely":[135],"long":[136],"sequences":[137],"over":[139]},"counts_by_year":[{"year":2025,"cited_by_count":1}],"updated_date":"2026-03-07T16:01:11.037858","created_date":"2025-11-12T00:00:00"}