{"id":"https://openalex.org/W4411486293","doi":"https://doi.org/10.1145/3695053.3731060","title":"GPUs All Grown-Up: Fully Device-Driven SpMV Using GPU Work Graphs","display_name":"GPUs All Grown-Up: Fully Device-Driven SpMV Using GPU Work Graphs","publication_year":2025,"publication_date":"2025-06-20","ids":{"openalex":"https://openalex.org/W4411486293","doi":"https://doi.org/10.1145/3695053.3731060"},"language":"en","primary_location":{"id":"doi:10.1145/3695053.3731060","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3695053.3731060","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3695053.3731060","source":null,"license":"cc-by-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-sa","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 52nd Annual International Symposium on Computer Architecture","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3695053.3731060","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5118454220","display_name":"Fabian Wildgrube","orcid":"https://orcid.org/0009-0006-6874-1747"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fabian Wildgrube","raw_affiliation_strings":["AMD, Munich, Germany"],"raw_orcid":"https://orcid.org/0009-0006-6874-1747","affiliations":[{"raw_affiliation_string":"AMD, Munich, Germany","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5029473578","display_name":"Pete Ehrett","orcid":null},"institutions":[{"id":"https://openalex.org/I4210137977","display_name":"Advanced Micro Devices (United States)","ror":"https://ror.org/04kd6c783","country_code":"US","type":"company","lineage":["https://openalex.org/I4210137977"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Pete Ehrett","raw_affiliation_strings":["AMD, Austin, TX, USA"],"raw_orcid":"https://orcid.org/0000-0002-4587-3716","affiliations":[{"raw_affiliation_string":"AMD, Austin, TX, USA","institution_ids":["https://openalex.org/I4210137977"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Paul Trojahn","orcid":"https://orcid.org/0009-0007-9457-0424"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Paul Trojahn","raw_affiliation_strings":["AMD, Munich, Germany"],"raw_orcid":"https://orcid.org/0009-0007-9457-0424","affiliations":[{"raw_affiliation_string":"AMD, Munich, Germany","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024954324","display_name":"Richard Membarth","orcid":"https://orcid.org/0000-0002-9979-7579"},"institutions":[{"id":"https://openalex.org/I4210106192","display_name":"Technische Hochschule Ingolstadt","ror":"https://ror.org/02bxzcy64","country_code":"DE","type":"education","lineage":["https://openalex.org/I4210106192"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Richard Membarth","raw_affiliation_strings":["THI, Ingolstadt, Germany and DFKI, Saarbr\u00fccken, Germany"],"raw_orcid":"https://orcid.org/0000-0002-9979-7579","affiliations":[{"raw_affiliation_string":"THI, Ingolstadt, Germany and DFKI, Saarbr\u00fccken, Germany","institution_ids":["https://openalex.org/I4210106192"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5077056823","display_name":"Bradford M. Beckmann","orcid":"https://orcid.org/0000-0002-5444-6521"},"institutions":[{"id":"https://openalex.org/I174135032","display_name":"Bellevue College","ror":"https://ror.org/05gr4yv49","country_code":"US","type":"education","lineage":["https://openalex.org/I174135032"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Bradford Beckmann","raw_affiliation_strings":["AMD, Bellevue, WA, USA"],"raw_orcid":"https://orcid.org/0000-0002-5444-6521","affiliations":[{"raw_affiliation_string":"AMD, Bellevue, WA, USA","institution_ids":["https://openalex.org/I174135032"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5106380196","display_name":"Dominik Baumeister","orcid":"https://orcid.org/0009-0006-4346-8334"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dominik Baumeister","raw_affiliation_strings":["AMD, Munich, Germany"],"raw_orcid":"https://orcid.org/0009-0006-4346-8334","affiliations":[{"raw_affiliation_string":"AMD, Munich, Germany","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5081551649","display_name":"Matth\u00e4us G. Chajdas","orcid":"https://orcid.org/0000-0003-4689-2932"},"institutions":[{"id":"https://openalex.org/I4210094487","display_name":"Intel (Germany)","ror":"https://ror.org/00m2x0g47","country_code":"DE","type":"company","lineage":["https://openalex.org/I1343180700","https://openalex.org/I4210094487"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Matth\u00e4us Chajdas","raw_affiliation_strings":["AMD (Now at Intel), Munich, Germany"],"raw_orcid":"https://orcid.org/0000-0003-4689-2932","affiliations":[{"raw_affiliation_string":"AMD (Now at Intel), Munich, Germany","institution_ids":["https://openalex.org/I4210094487"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":3.0301,"has_fulltext":true,"cited_by_count":2,"citation_normalized_percentile":{"value":0.89209062,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":97,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"1777","last_page":"1791"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9926999807357788,"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":0.9926999807357788,"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/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9825999736785889,"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"}},{"id":"https://openalex.org/T10101","display_name":"Cloud Computing and Resource Management","score":0.9793000221252441,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/computer-science","display_name":"Computer science","score":0.8239051699638367},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.6793573498725891},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.4889334738254547}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8239051699638367},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.6793573498725891},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.4889334738254547}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3695053.3731060","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3695053.3731060","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3695053.3731060","source":null,"license":"cc-by-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-sa","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 52nd Annual International Symposium on Computer Architecture","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3695053.3731060","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3695053.3731060","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3695053.3731060","source":null,"license":"cc-by-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-sa","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 52nd Annual International Symposium on Computer Architecture","raw_type":"proceedings-article"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/7","score":0.4300000071525574,"display_name":"Affordable and clean energy"}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4411486293.pdf","grobid_xml":"https://content.openalex.org/works/W4411486293.grobid-xml"},"referenced_works_count":29,"referenced_works":["https://openalex.org/W1508286210","https://openalex.org/W1974621770","https://openalex.org/W1990832096","https://openalex.org/W1991709203","https://openalex.org/W1995041765","https://openalex.org/W1996320382","https://openalex.org/W2009654791","https://openalex.org/W2063384380","https://openalex.org/W2087507944","https://openalex.org/W2088866486","https://openalex.org/W2098602979","https://openalex.org/W2153190325","https://openalex.org/W2263374743","https://openalex.org/W2294932020","https://openalex.org/W2435668395","https://openalex.org/W2461193710","https://openalex.org/W2581020339","https://openalex.org/W2617403939","https://openalex.org/W2810120553","https://openalex.org/W2898442528","https://openalex.org/W2903351422","https://openalex.org/W2964357930","https://openalex.org/W2984946966","https://openalex.org/W3098650335","https://openalex.org/W3190445506","https://openalex.org/W4233930397","https://openalex.org/W4394998995","https://openalex.org/W4403334264","https://openalex.org/W4405755154"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W4399354997","https://openalex.org/W3062287","https://openalex.org/W2380390332","https://openalex.org/W2005148983","https://openalex.org/W2012954338","https://openalex.org/W2096672917","https://openalex.org/W2392023973"],"abstract_inverted_index":{"Sparse":[0],"matrix-vector":[1],"multiplication":[2],"(SpMV)":[3],"is":[4,211],"a":[5,133,153,212],"key":[6],"operation":[7],"across":[8,170],"high-performance":[9],"computing,":[10],"graph":[11],"analytics,":[12],"and":[13,30,64,100,117,128,179],"many":[14],"more":[15,167],"applications.In":[16],"these":[17],"applications,":[18],"the":[19,65,70,74,139,175,182],"matrix":[20,218,229],"characteristics,":[21],"notably":[22],"non-zero":[23],"elements":[24],"per":[25],"row,":[26],"can":[27],"vary":[28],"widely":[29],"impact":[31],"which":[32],"algorithm":[33,51,186],"performs":[34],"best.Thus,":[35],"Graphics":[36],"Processing":[37],"Unit":[38],"(GPU)":[39],"SpMV":[40,62,146,155,192],"algorithms":[41],"often":[42],"rely":[43],"on":[44,69],"costly":[45],"preprocessing":[46,63,105],"to":[47,52,54,158,189,196,221],"determine":[48],"what":[49],"per-row":[50,67],"select":[53],"achieve":[55],"high":[56],"performance.In":[57],"this":[58],"work":[59],"we":[60],"combine":[61],"subsequent":[66],"processing":[68,113],"GPU":[71,78],"by":[72,156,203],"leveraging":[73],"novel":[75,143],"\"Work":[76],"Graphs\"":[77],"programming":[79],"model-initially":[80],"designed":[81],"for":[82,90,152,187,207],"graphics":[83],"applications-for":[84],"dynamic":[85],"on-device":[86],"self-scheduling.Work":[87],"Graphs":[88,145],"allow":[89],"fine-grain":[91],"dataflow":[92],"execution":[93],"of":[94,111,122,135,141,217],"individual":[95,112],"workgroups":[96,110],"using":[97],"emerging":[98],"hardware":[99],"firmware":[101],"support.As":[102],"soon":[103],"as":[104],"has":[106],"generated":[107],"sufficient":[108],"work,":[109],"kernels":[114],"are":[115],"self-scheduled":[116],"executed,":[118],"interleaved":[119],"with":[120,228],"those":[121],"other":[123],"kernels.This":[124],"improves":[125],"cache":[126],"locality":[127],"eliminates":[129],"host":[130],"interaction":[131],"altogether.Across":[132],"suite":[134],"59":[136],"sparse":[137],"matrices,":[138],"best":[140],"various":[142,171],"Work":[144],"implementations":[147],"outperforms":[148],"state-of-the-art":[149],"rocSPARSE":[150,176,184,197,222],"\"LRB\"":[151],"single":[154],"up":[157,188],"7.19":[159],"(mean:":[160],"3.35,":[161],"SD:":[162],"1.89).Furthermore,":[163],"it":[164,199],"achieves":[165],"much":[166],"stable":[168],"performance":[169],"sparsity":[172],"patterns":[173],"than":[174],"CSR-General":[177],"algorithm,":[178],"even":[180],"beats":[181],"advanced":[183],"CSR-Adaptive":[185],"92":[190],"consecutive":[191],"calculations.In":[193],"addition,":[194],"compared":[195,220],"LRB,":[198],"reduces":[200],"code":[201],"complexity":[202],"75%.Its":[204],"memory":[205],"footprint":[206],"supporting":[208],"data":[209,224],"structures":[210,225],"fixed":[213],"25":[214],"MiB":[215],"independent":[216],"size,":[219],"LRB's":[223],"that":[226],"scale":[227]},"counts_by_year":[{"year":2026,"cited_by_count":2}],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}