{"id":"https://openalex.org/W7092205149","doi":"https://doi.org/10.1145/3725843.3756105","title":"TransFusion: End-to-End Transformer Acceleration via Graph Fusion and Pipelining","display_name":"TransFusion: End-to-End Transformer Acceleration via Graph Fusion and Pipelining","publication_year":2025,"publication_date":"2025-10-17","ids":{"openalex":"https://openalex.org/W7092205149","doi":"https://doi.org/10.1145/3725843.3756105"},"language":null,"primary_location":{"id":"doi:10.1145/3725843.3756105","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3725843.3756105","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3725843.3756105","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 58th IEEE/ACM International Symposium on Microarchitecture","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3725843.3756105","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Linxuan Zhang","orcid":"https://orcid.org/0009-0000-6760-5504"},"institutions":[{"id":"https://openalex.org/I154425047","display_name":"University of Alberta","ror":"https://ror.org/0160cpw27","country_code":"CA","type":"education","lineage":["https://openalex.org/I154425047"]}],"countries":["CA"],"is_corresponding":true,"raw_author_name":"Linxuan Zhang","raw_affiliation_strings":["The Department of Electrical and Computer Engineering, University of Alberta, Edmonton, Alberta, Canada"],"affiliations":[{"raw_affiliation_string":"The Department of Electrical and Computer Engineering, University of Alberta, Edmonton, Alberta, Canada","institution_ids":["https://openalex.org/I154425047"]}]},{"author_position":"middle","author":{"id":null,"display_name":"J. Nelson Amaral","orcid":"https://orcid.org/0000-0002-9943-1809"},"institutions":[{"id":"https://openalex.org/I154425047","display_name":"University of Alberta","ror":"https://ror.org/0160cpw27","country_code":"CA","type":"education","lineage":["https://openalex.org/I154425047"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"J. Nelson Amaral","raw_affiliation_strings":["Department of Computing Science, University of Alberta, Edmonton, Alberta, Canada"],"affiliations":[{"raw_affiliation_string":"Department of Computing Science, University of Alberta, Edmonton, Alberta, Canada","institution_ids":["https://openalex.org/I154425047"]}]},{"author_position":"last","author":{"id":null,"display_name":"Di Niu","orcid":"https://orcid.org/0000-0002-5250-7327"},"institutions":[{"id":"https://openalex.org/I154425047","display_name":"University of Alberta","ror":"https://ror.org/0160cpw27","country_code":"CA","type":"education","lineage":["https://openalex.org/I154425047"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Di Niu","raw_affiliation_strings":["The Department of Electrical and Computer Engineering, University of Alberta, Edmonton, Alberta, Canada"],"affiliations":[{"raw_affiliation_string":"The Department of Electrical and Computer Engineering, University of Alberta, Edmonton, Alberta, Canada","institution_ids":["https://openalex.org/I154425047"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I154425047"],"apc_list":null,"apc_paid":null,"fwci":2.6995,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.94036879,"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":"1491","last_page":"1504"},"is_retracted":false,"is_paratext":false,"is_xpac":true,"primary_topic":{"id":"https://openalex.org/T11273","display_name":"Advanced Graph Neural Networks","score":0.7131999731063843,"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"}},"topics":[{"id":"https://openalex.org/T11273","display_name":"Advanced Graph Neural Networks","score":0.7131999731063843,"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/T12292","display_name":"Graph Theory and Algorithms","score":0.03779999911785126,"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/T10887","display_name":"Bioinformatics and Genomic Networks","score":0.015399999916553497,"subfield":{"id":"https://openalex.org/subfields/1312","display_name":"Molecular Biology"},"field":{"id":"https://openalex.org/fields/13","display_name":"Biochemistry, Genetics and Molecular Biology"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.44369998574256897},{"id":"https://openalex.org/keywords/graph","display_name":"Graph","score":0.4334000051021576},{"id":"https://openalex.org/keywords/fusion","display_name":"Fusion","score":0.41589999198913574},{"id":"https://openalex.org/keywords/acceleration","display_name":"Acceleration","score":0.3856000006198883},{"id":"https://openalex.org/keywords/minification","display_name":"Minification","score":0.29269999265670776}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6164000034332275},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.477400004863739},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.44369998574256897},{"id":"https://openalex.org/C132525143","wikidata":"https://www.wikidata.org/wiki/Q141488","display_name":"Graph","level":2,"score":0.4334000051021576},{"id":"https://openalex.org/C158525013","wikidata":"https://www.wikidata.org/wiki/Q2593739","display_name":"Fusion","level":2,"score":0.41589999198913574},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.3856000006198883},{"id":"https://openalex.org/C147764199","wikidata":"https://www.wikidata.org/wiki/Q6865248","display_name":"Minification","level":2,"score":0.29269999265670776},{"id":"https://openalex.org/C33954974","wikidata":"https://www.wikidata.org/wiki/Q486494","display_name":"Sensor fusion","level":2,"score":0.27239999175071716},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.26269999146461487},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.24400000274181366}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3725843.3756105","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3725843.3756105","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3725843.3756105","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 58th IEEE/ACM International Symposium on Microarchitecture","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3725843.3756105","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3725843.3756105","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3725843.3756105","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 58th IEEE/ACM International Symposium on Microarchitecture","raw_type":"proceedings-article"},"sustainable_development_goals":[{"display_name":"Affordable and clean energy","score":0.8382651209831238,"id":"https://metadata.un.org/sdg/7"}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7092205149.pdf","grobid_xml":"https://content.openalex.org/works/W7092205149.grobid-xml"},"referenced_works_count":30,"referenced_works":["https://openalex.org/W2940862705","https://openalex.org/W2963341956","https://openalex.org/W2970454332","https://openalex.org/W2998732502","https://openalex.org/W3007007518","https://openalex.org/W3100985894","https://openalex.org/W3121523901","https://openalex.org/W3127736057","https://openalex.org/W3154851733","https://openalex.org/W3159727696","https://openalex.org/W3162542754","https://openalex.org/W3174544005","https://openalex.org/W3189877953","https://openalex.org/W3206453033","https://openalex.org/W3207411217","https://openalex.org/W4214686755","https://openalex.org/W4224215744","https://openalex.org/W4283811336","https://openalex.org/W4285601701","https://openalex.org/W4304014014","https://openalex.org/W4311327542","https://openalex.org/W4318541578","https://openalex.org/W4366341968","https://openalex.org/W4386071870","https://openalex.org/W4389491877","https://openalex.org/W4389500477","https://openalex.org/W4390873361","https://openalex.org/W4392265934","https://openalex.org/W4404640117","https://openalex.org/W4404955120"],"related_works":[],"abstract_inverted_index":{"Transformer":[0,18,50],"acceleration":[1],"has":[2],"increasingly":[3],"emphasized":[4],"local":[5],"fusion":[6],"within":[7],"isolated":[8],"submodules,":[9],"such":[10,29],"as":[11,17,59],"multi-head":[12],"attention":[13],"(MHA)":[14],"and":[15,26,57,69,86,120,126,140,157],"softmax.However,":[16],"models":[19],"continue":[20],"to":[21,33,132],"scale":[22],"in":[23],"both":[24,124],"depth":[25],"context":[27],"length,":[28],"fragmented":[30],"optimizations":[31],"fail":[32],"address":[34],"end-to-end":[35,49],"inefficiencies":[36],"across":[37,90,123],"the":[38,80,145],"full":[39],"encoder/decoder":[40],"stack.This":[41],"paper":[42],"presents":[43],"TransFusion,":[44],"a":[45,74,107],"comprehensive":[46],"framework":[47],"for":[48],"layers,":[51],"including":[52],"QKV":[53],"projections,":[54],"MHA,":[55],"LayerNorm,":[56],"FFN,":[58],"structured":[60],"Einsum":[61],"Cascades,":[62],"enabling":[63],"precise":[64],"modelling":[65],"of":[66,135],"data":[67,153],"dependencies":[68],"execution":[70,99],"order.TransFusion":[71],"introduces":[72],"DPipe,":[73],"unified":[75],"graph-based":[76],"scheduler":[77],"that":[78,116],"partitions":[79],"Einsum-centric":[81],"directed":[82],"acyclic":[83],"graph":[84],"(DAG)":[85],"applies":[87],"latency-aware":[88],"pipelining":[89],"hardware":[91],"hierarchies":[92],"using":[93],"dynamic":[94],"programming":[95],"(DP).To":[96],"enable":[97],"scalable":[98],"under":[100],"strict":[101],"memory":[102],"budgets,":[103],"TransFusion":[104,129],"integrates":[105],"TileSeek,":[106],"Monte":[108],"Carlo":[109],"Tree":[110],"Search":[111],"(MCTS)-based":[112],"tiling":[113],"search":[114],"algorithm":[115],"balances":[117],"buffer":[118],"reuse":[119],"system":[121],"constraints.Evaluated":[122],"cloud":[125,139],"edge":[127,143],"architecture,":[128],"achieves":[130],"up":[131],"an":[133],"average":[134],"1.6":[136],"speedup":[137],"on":[138,142],"2.2":[141],"over":[144],"prior":[146],"state-of-the-art,":[147],"FuseMax,":[148],"by":[149],"jointly":[150],"optimizing":[151],"inter-layer":[152],"reuse,":[154],"intra-layer":[155],"pipelining,":[156],"operator":[158],"scheduling.":[159]},"counts_by_year":[{"year":2025,"cited_by_count":1}],"updated_date":"2026-03-12T06:13:28.667946","created_date":"2025-10-17T00:00:00"}