{"id":"https://openalex.org/W7125845704","doi":"https://doi.org/10.1145/3779212.3790165","title":"FuseFlow: A Fusion-Centric Compilation Framework for Sparse Deep Learning on Streaming Dataflow","display_name":"FuseFlow: A Fusion-Centric Compilation Framework for Sparse Deep Learning on Streaming Dataflow","publication_year":2026,"publication_date":"2026-03-10","ids":{"openalex":"https://openalex.org/W7125845704","doi":"https://doi.org/10.1145/3779212.3790165"},"language":null,"primary_location":{"id":"doi:10.1145/3779212.3790165","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3779212.3790165","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 31st ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2","raw_type":"proceedings-article"},"type":"article","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3779212.3790165","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5032530949","display_name":"Rubens Lacouture","orcid":"https://orcid.org/0009-0008-2268-0074"},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Rubens Lacouture","raw_affiliation_strings":["Stanford University, Stanford, California, USA"],"raw_orcid":"https://orcid.org/0009-0008-2268-0074","affiliations":[{"raw_affiliation_string":"Stanford University, Stanford, California, USA","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128869924","display_name":"Nathan Zhang","orcid":null},"institutions":[{"id":"https://openalex.org/I4210099960","display_name":"Systems Control (United States)","ror":"https://ror.org/014tted12","country_code":"US","type":"company","lineage":["https://openalex.org/I4210099960"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Nathan Zhang","raw_affiliation_strings":["SambaNova Systems, Inc., Palo Alto, California, USA"],"raw_orcid":"https://orcid.org/0000-0002-9668-902X","affiliations":[{"raw_affiliation_string":"SambaNova Systems, Inc., Palo Alto, California, USA","institution_ids":["https://openalex.org/I4210099960"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054704326","display_name":"Ritvik Sharma","orcid":"https://orcid.org/0000-0002-5809-7031"},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ritvik Sharma","raw_affiliation_strings":["Stanford University, Stanford, California, USA"],"raw_orcid":"https://orcid.org/0000-0002-5809-7031","affiliations":[{"raw_affiliation_string":"Stanford University, Stanford, California, USA","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5053721069","display_name":"Marco Siracusa","orcid":"https://orcid.org/0000-0003-2782-837X"},"institutions":[{"id":"https://openalex.org/I2799803557","display_name":"Barcelona Supercomputing Center","ror":"https://ror.org/05sd8tv96","country_code":"ES","type":"facility","lineage":["https://openalex.org/I2799803557","https://openalex.org/I9617848"]},{"id":"https://openalex.org/I9617848","display_name":"Universitat Polit\u00e8cnica de Catalunya","ror":"https://ror.org/03mb6wj31","country_code":"ES","type":"education","lineage":["https://openalex.org/I9617848"]}],"countries":["ES"],"is_corresponding":false,"raw_author_name":"Marco Siracusa","raw_affiliation_strings":["Barcelona Supercomputing Center, Barcelona, Spain"],"raw_orcid":"https://orcid.org/0000-0003-2782-837X","affiliations":[{"raw_affiliation_string":"Barcelona Supercomputing Center, Barcelona, Spain","institution_ids":["https://openalex.org/I9617848","https://openalex.org/I2799803557"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128692397","display_name":"Fredrik Berg Kjolstad","orcid":null},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Fredrik Kjolstad","raw_affiliation_strings":["Stanford University, Stanford, California, USA"],"raw_orcid":"https://orcid.org/0000-0002-2267-903X","affiliations":[{"raw_affiliation_string":"Stanford University, Stanford, California, USA","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5023857198","display_name":"Kunle Olukotun","orcid":"https://orcid.org/0000-0002-8779-0636"},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kunle Olukotun","raw_affiliation_strings":["Stanford University, Stanford, California, USA"],"raw_orcid":"https://orcid.org/0000-0002-8779-0636","affiliations":[{"raw_affiliation_string":"Stanford University, Stanford, California, USA","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5039681071","display_name":"Olivia Hsu","orcid":"https://orcid.org/0000-0002-4195-8106"},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Olivia Hsu","raw_affiliation_strings":["Stanford University, Stanford, California, USA and Carnegie Mellon University, Pittsburgh, Pennsylvania, USA"],"raw_orcid":"https://orcid.org/0000-0002-4195-8106","affiliations":[{"raw_affiliation_string":"Stanford University, Stanford, California, USA and Carnegie Mellon University, Pittsburgh, Pennsylvania, USA","institution_ids":["https://openalex.org/I74973139"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":7,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":37.7,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.98391643,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":95,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"798","last_page":"820"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10904","display_name":"Embedded Systems Design Techniques","score":0.6682999730110168,"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/T10904","display_name":"Embedded Systems Design Techniques","score":0.6682999730110168,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.05790000036358833,"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/T11986","display_name":"Scientific Computing and Data Management","score":0.025599999353289604,"subfield":{"id":"https://openalex.org/subfields/1802","display_name":"Information Systems and Management"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/dataflow","display_name":"Dataflow","score":0.9887999892234802},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.66839998960495},{"id":"https://openalex.org/keywords/compiler","display_name":"Compiler","score":0.6438000202178955},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.5964000225067139},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.5939000248908997},{"id":"https://openalex.org/keywords/dataflow-architecture","display_name":"Dataflow architecture","score":0.4984000027179718},{"id":"https://openalex.org/keywords/granularity","display_name":"Granularity","score":0.47920000553131104},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.43290001153945923}],"concepts":[{"id":"https://openalex.org/C96324660","wikidata":"https://www.wikidata.org/wiki/Q205446","display_name":"Dataflow","level":2,"score":0.9887999892234802},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.854200005531311},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.66839998960495},{"id":"https://openalex.org/C169590947","wikidata":"https://www.wikidata.org/wiki/Q47506","display_name":"Compiler","level":2,"score":0.6438000202178955},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5964000225067139},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.5939000248908997},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.5907999873161316},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5485000014305115},{"id":"https://openalex.org/C176727019","wikidata":"https://www.wikidata.org/wiki/Q1172415","display_name":"Dataflow architecture","level":3,"score":0.4984000027179718},{"id":"https://openalex.org/C177774035","wikidata":"https://www.wikidata.org/wiki/Q1246948","display_name":"Granularity","level":2,"score":0.47920000553131104},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.43290001153945923},{"id":"https://openalex.org/C173801870","wikidata":"https://www.wikidata.org/wiki/Q201413","display_name":"Heuristic","level":2,"score":0.42750000953674316},{"id":"https://openalex.org/C56372850","wikidata":"https://www.wikidata.org/wiki/Q1050404","display_name":"Sparse matrix","level":3,"score":0.40779998898506165},{"id":"https://openalex.org/C88468194","wikidata":"https://www.wikidata.org/wiki/Q1172416","display_name":"Data-flow analysis","level":3,"score":0.3767000138759613},{"id":"https://openalex.org/C83283714","wikidata":"https://www.wikidata.org/wiki/Q121117","display_name":"Supercomputer","level":2,"score":0.3165000081062317},{"id":"https://openalex.org/C184596265","wikidata":"https://www.wikidata.org/wiki/Q2651576","display_name":"Model of computation","level":3,"score":0.28630000352859497},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2791000008583069},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.275299996137619},{"id":"https://openalex.org/C158525013","wikidata":"https://www.wikidata.org/wiki/Q2593739","display_name":"Fusion","level":2,"score":0.272599995136261},{"id":"https://openalex.org/C124066611","wikidata":"https://www.wikidata.org/wiki/Q28684319","display_name":"Sparse approximation","level":2,"score":0.25940001010894775},{"id":"https://openalex.org/C489000","wikidata":"https://www.wikidata.org/wiki/Q747385","display_name":"Data flow diagram","level":2,"score":0.2565000057220459},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.2565000057220459},{"id":"https://openalex.org/C78766204","wikidata":"https://www.wikidata.org/wiki/Q555032","display_name":"Multi-core processor","level":2,"score":0.2563999891281128}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3779212.3790165","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3779212.3790165","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 31st ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2511.04768","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2511.04768","pdf_url":"https://arxiv.org/pdf/2511.04768","source":{"id":"https://openalex.org/S4393918464","display_name":"ArXiv.org","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"}],"best_oa_location":{"id":"doi:10.1145/3779212.3790165","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3779212.3790165","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 31st ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2","raw_type":"proceedings-article"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/9","score":0.46534496545791626,"display_name":"Industry, innovation and infrastructure"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"As":[0],"deep":[1,6],"learning":[2,7,39,107],"models":[3,9,40],"scale,":[4],"sparse":[5,37,46,65,130],"(DL)":[8],"that":[10,35,112],"exploit":[11],"sparsity":[12,83],"in":[13,42,121],"weights,":[14],"activations,":[15],"or":[16],"inputs":[17],"and":[18,82,145],"specialized":[19],"dataflow":[20,47,51,80,89],"hardware":[21],"have":[22],"emerged":[23],"as":[24],"powerful":[25],"solutions":[26],"to":[27,44,59,69,143],"address":[28],"efficiency.":[29],"We":[30,97],"propose":[31],"FuseFlow,":[32,150],"a":[33,87,141,156],"compiler":[34,58],"converts":[36],"machine":[38,106],"written":[41],"PyTorch":[43],"fused":[45],"graphs":[48],"for":[49,91,100,129,163],"reconfigurable":[50],"architectures":[52],"(RDAs).":[53],"FuseFlow":[54,74,99,138],"is":[55,125],"the":[56,135],"first":[57],"support":[60],"general":[61],"cross-expression":[62,116],"fusion":[63,70,95,114,117],"of":[64,94],"operations.":[66],"In":[67],"addition":[68],"across":[71,103,118],"kernels":[72],"(expressions),":[73],"also":[75,139],"supports":[76],"optimizations":[77],"like":[78],"parallelization,":[79],"ordering,":[81],"blocking.":[84],"It":[85],"targets":[86],"cycle-accurate":[88],"simulator":[90],"microarchitectural":[92],"analysis":[93],"strategies.":[96],"use":[98],"design-space":[101],"exploration":[102],"four":[104],"real-world":[105],"applications":[108],"with":[109,165],"sparsity,":[110],"showing":[111],"full":[113],"(entire":[115],"all":[119],"computation":[120],"an":[122,160],"end-to-end":[123],"model)":[124],"not":[126],"always":[127],"optimal":[128],"models\u2014fusion":[131],"granularity":[132],"depends":[133],"on":[134],"model":[136],"itself.":[137],"provides":[140],"heuristic":[142],"identify":[144],"prune":[146],"suboptimal":[147],"configurations.":[148],"Using":[149],"we":[151],"achieve":[152],"performance":[153],"improvements,":[154],"including":[155],"~2.7x":[157],"speedup":[158],"over":[159],"unfused":[161],"baseline":[162],"GPT-3":[164],"BigBird":[166],"block-sparse":[167],"attention.":[168]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-01-28T00:00:00"}