{"id":"https://openalex.org/W3166178281","doi":"https://doi.org/10.1177/10943420221136848","title":"Mixed precision LU factorization on GPU tensor cores: reducing data movement and memory footprint","display_name":"Mixed precision LU factorization on GPU tensor cores: reducing data movement and memory footprint","publication_year":2023,"publication_date":"2023-01-03","ids":{"openalex":"https://openalex.org/W3166178281","doi":"https://doi.org/10.1177/10943420221136848","mag":"3166178281"},"language":"en","primary_location":{"id":"doi:10.1177/10943420221136848","is_oa":true,"landing_page_url":"https://doi.org/10.1177/10943420221136848","pdf_url":"https://journals.sagepub.com/doi/pdf/10.1177/10943420221136848?download=true","source":{"id":"https://openalex.org/S60606485","display_name":"The International Journal of High Performance Computing Applications","issn_l":"0890-2720","issn":["0890-2720","1094-3420","1741-2846"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320017","host_organization_name":"SAGE Publishing","host_organization_lineage":["https://openalex.org/P4310320017"],"host_organization_lineage_names":["SAGE Publishing"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"The International Journal of High Performance Computing Applications","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"bronze","oa_url":"https://journals.sagepub.com/doi/pdf/10.1177/10943420221136848?download=true","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5014227566","display_name":"Florent Lopez","orcid":"https://orcid.org/0000-0002-8479-5749"},"institutions":[{"id":"https://openalex.org/I21160419","display_name":"Ansys (United States)","ror":"https://ror.org/05cf5b117","country_code":"US","type":"company","lineage":["https://openalex.org/I21160419"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Florent Lopez","raw_affiliation_strings":["Livermore Software Technology, an ANSYS Company, USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Livermore Software Technology, an ANSYS Company, USA","institution_ids":["https://openalex.org/I21160419"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5065253192","display_name":"Th\u00e9o Mary","orcid":"https://orcid.org/0000-0001-9949-4634"},"institutions":[{"id":"https://openalex.org/I1294671590","display_name":"Centre National de la Recherche Scientifique","ror":"https://ror.org/02feahw73","country_code":"FR","type":"government","lineage":["https://openalex.org/I1294671590"]},{"id":"https://openalex.org/I39804081","display_name":"Sorbonne Universit\u00e9","ror":"https://ror.org/02en5vm52","country_code":"FR","type":"education","lineage":["https://openalex.org/I39804081"]}],"countries":["FR"],"is_corresponding":true,"raw_author_name":"Theo Mary","raw_affiliation_strings":["Sorbonne Universit\u00e9, CNRS, France"],"raw_orcid":"https://orcid.org/0000-0001-9949-4634","affiliations":[{"raw_affiliation_string":"Sorbonne Universit\u00e9, CNRS, France","institution_ids":["https://openalex.org/I1294671590","https://openalex.org/I39804081"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5065253192"],"corresponding_institution_ids":["https://openalex.org/I1294671590","https://openalex.org/I39804081"],"apc_list":null,"apc_paid":null,"fwci":1.6053,"has_fulltext":true,"cited_by_count":8,"citation_normalized_percentile":{"value":0.80973603,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":89,"max":98},"biblio":{"volume":"37","issue":"2","first_page":"165","last_page":"179"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9991999864578247,"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":0.9991999864578247,"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/T10792","display_name":"Matrix Theory and Algorithms","score":0.9988999962806702,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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/T10715","display_name":"Distributed and Parallel Computing Systems","score":0.9851999878883362,"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/memory-footprint","display_name":"Memory footprint","score":0.8187018632888794},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7294623851776123},{"id":"https://openalex.org/keywords/factorization","display_name":"Factorization","score":0.6808313727378845},{"id":"https://openalex.org/keywords/tensor","display_name":"Tensor (intrinsic definition)","score":0.658861517906189},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.6089890003204346},{"id":"https://openalex.org/keywords/matrix-decomposition","display_name":"Matrix decomposition","score":0.5751245617866516},{"id":"https://openalex.org/keywords/single-precision-floating-point-format","display_name":"Single-precision floating-point format","score":0.5599664449691772},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.48711585998535156},{"id":"https://openalex.org/keywords/matrix","display_name":"Matrix (chemical analysis)","score":0.4808836281299591},{"id":"https://openalex.org/keywords/state","display_name":"State (computer science)","score":0.45075365900993347},{"id":"https://openalex.org/keywords/double-precision-floating-point-format","display_name":"Double-precision floating-point format","score":0.4468100965023041},{"id":"https://openalex.org/keywords/footprint","display_name":"Footprint","score":0.4413301646709442},{"id":"https://openalex.org/keywords/single-core","display_name":"Single-core","score":0.4251607060432434},{"id":"https://openalex.org/keywords/computational-science","display_name":"Computational science","score":0.3373323082923889},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.1975257694721222},{"id":"https://openalex.org/keywords/floating-point","display_name":"Floating point","score":0.17103669047355652}],"concepts":[{"id":"https://openalex.org/C74912251","wikidata":"https://www.wikidata.org/wiki/Q6815727","display_name":"Memory footprint","level":2,"score":0.8187018632888794},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7294623851776123},{"id":"https://openalex.org/C187834632","wikidata":"https://www.wikidata.org/wiki/Q188804","display_name":"Factorization","level":2,"score":0.6808313727378845},{"id":"https://openalex.org/C155281189","wikidata":"https://www.wikidata.org/wiki/Q3518150","display_name":"Tensor (intrinsic definition)","level":2,"score":0.658861517906189},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.6089890003204346},{"id":"https://openalex.org/C42355184","wikidata":"https://www.wikidata.org/wiki/Q1361088","display_name":"Matrix decomposition","level":3,"score":0.5751245617866516},{"id":"https://openalex.org/C133095886","wikidata":"https://www.wikidata.org/wiki/Q1307173","display_name":"Single-precision floating-point format","level":3,"score":0.5599664449691772},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.48711585998535156},{"id":"https://openalex.org/C106487976","wikidata":"https://www.wikidata.org/wiki/Q685816","display_name":"Matrix (chemical analysis)","level":2,"score":0.4808836281299591},{"id":"https://openalex.org/C48103436","wikidata":"https://www.wikidata.org/wiki/Q599031","display_name":"State (computer science)","level":2,"score":0.45075365900993347},{"id":"https://openalex.org/C35912277","wikidata":"https://www.wikidata.org/wiki/Q1243369","display_name":"Double-precision floating-point format","level":3,"score":0.4468100965023041},{"id":"https://openalex.org/C132943942","wikidata":"https://www.wikidata.org/wiki/Q2562511","display_name":"Footprint","level":2,"score":0.4413301646709442},{"id":"https://openalex.org/C2780365336","wikidata":"https://www.wikidata.org/wiki/Q25047934","display_name":"Single-core","level":2,"score":0.4251607060432434},{"id":"https://openalex.org/C459310","wikidata":"https://www.wikidata.org/wiki/Q117801","display_name":"Computational science","level":1,"score":0.3373323082923889},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.1975257694721222},{"id":"https://openalex.org/C84211073","wikidata":"https://www.wikidata.org/wiki/Q117879","display_name":"Floating point","level":2,"score":0.17103669047355652},{"id":"https://openalex.org/C158693339","wikidata":"https://www.wikidata.org/wiki/Q190524","display_name":"Eigenvalues and eigenvectors","level":2,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C202444582","wikidata":"https://www.wikidata.org/wiki/Q837863","display_name":"Pure mathematics","level":1,"score":0.0},{"id":"https://openalex.org/C151730666","wikidata":"https://www.wikidata.org/wiki/Q7205","display_name":"Paleontology","level":1,"score":0.0},{"id":"https://openalex.org/C159985019","wikidata":"https://www.wikidata.org/wiki/Q181790","display_name":"Composite material","level":1,"score":0.0},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0},{"id":"https://openalex.org/C62520636","wikidata":"https://www.wikidata.org/wiki/Q944","display_name":"Quantum mechanics","level":1,"score":0.0},{"id":"https://openalex.org/C192562407","wikidata":"https://www.wikidata.org/wiki/Q228736","display_name":"Materials science","level":0,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1177/10943420221136848","is_oa":true,"landing_page_url":"https://doi.org/10.1177/10943420221136848","pdf_url":"https://journals.sagepub.com/doi/pdf/10.1177/10943420221136848?download=true","source":{"id":"https://openalex.org/S60606485","display_name":"The International Journal of High Performance Computing Applications","issn_l":"0890-2720","issn":["0890-2720","1094-3420","1741-2846"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320017","host_organization_name":"SAGE Publishing","host_organization_lineage":["https://openalex.org/P4310320017"],"host_organization_lineage_names":["SAGE Publishing"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"The International Journal of High Performance Computing Applications","raw_type":"journal-article"},{"id":"pmh:oai:HAL:hal-02937325v2","is_oa":true,"landing_page_url":"https://hal.science/hal-02937325","pdf_url":"https://hal.science/hal-02937325/document","source":{"id":"https://openalex.org/S4306402512","display_name":"HAL (Le Centre pour la Communication Scientifique Directe)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1294671590","host_organization_name":"Centre National de la Recherche Scientifique","host_organization_lineage":["https://openalex.org/I1294671590"],"host_organization_lineage_names":[],"type":"repository"},"license":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"International Journal of High Performance Computing Applications, 2023, 37 (2), pp.165-179. ⟨10.1177/10943420221136848⟩","raw_type":"info:eu-repo/semantics/article"}],"best_oa_location":{"id":"doi:10.1177/10943420221136848","is_oa":true,"landing_page_url":"https://doi.org/10.1177/10943420221136848","pdf_url":"https://journals.sagepub.com/doi/pdf/10.1177/10943420221136848?download=true","source":{"id":"https://openalex.org/S60606485","display_name":"The International Journal of High Performance Computing Applications","issn_l":"0890-2720","issn":["0890-2720","1094-3420","1741-2846"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320017","host_organization_name":"SAGE Publishing","host_organization_lineage":["https://openalex.org/P4310320017"],"host_organization_lineage_names":["SAGE Publishing"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"The International Journal of High Performance Computing Applications","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W3166178281.pdf","grobid_xml":"https://content.openalex.org/works/W3166178281.grobid-xml"},"referenced_works_count":18,"referenced_works":["https://openalex.org/W2020804487","https://openalex.org/W2026237859","https://openalex.org/W2035080386","https://openalex.org/W2067580655","https://openalex.org/W2599973081","https://openalex.org/W2752879223","https://openalex.org/W2895305554","https://openalex.org/W2903589674","https://openalex.org/W2910374893","https://openalex.org/W3000958723","https://openalex.org/W3009762362","https://openalex.org/W3022909171","https://openalex.org/W3023393331","https://openalex.org/W3037775421","https://openalex.org/W3045823895","https://openalex.org/W3084211317","https://openalex.org/W3108141724","https://openalex.org/W3162768929"],"related_works":["https://openalex.org/W3150370983","https://openalex.org/W2239119680","https://openalex.org/W2963207152","https://openalex.org/W1638830944","https://openalex.org/W1564887326","https://openalex.org/W2185760795","https://openalex.org/W2044409366","https://openalex.org/W1980667043","https://openalex.org/W3166178281","https://openalex.org/W2534590166"],"abstract_inverted_index":{"Modern":[0],"GPUs":[1],"equipped":[2],"with":[3,140,154],"mixed":[4,24,122],"precision":[5,26,59,97,115,123],"tensor":[6,76,128],"core":[7,77],"units":[8],"present":[9],"great":[10],"potential":[11],"to":[12,34,41,62,65,91,171],"accelerate":[13],"dense":[14],"linear":[15],"algebra":[16],"operations":[17],"such":[18],"as":[19],"LU":[20,27],"factorization.":[21],"However,":[22],"state-of-the-art":[23],"half/single":[25],"factorization":[28,86],"algorithms":[29],"all":[30,71],"require":[31],"the":[32,52,57,93,131,141,144,146,157],"matrix":[33,94],"be":[35],"stored":[36],"in":[37,95,130],"single":[38,61,114],"precision,":[39],"leading":[40],"expensive":[42],"data":[43,158],"movement":[44,159],"and":[45,120,160,163,173,178],"storage":[46,58],"costs.":[47],"This":[48],"is":[49,89,107,149],"explained":[50],"by":[51],"fact":[53],"that":[54,88,138],"simply":[55],"switching":[56],"from":[60,74],"half":[63,96,156],"leads":[64],"significant":[66,101],"loss":[67,102],"of":[68,103,117,143,150],"accuracy,":[69],"forfeiting":[70],"accuracy":[72,152],"benefits":[73],"using":[75],"technology.":[78],"In":[79],"this":[80],"article,":[81],"we":[82],"propose":[83],"a":[84,110,121],"new":[85],"algorithm":[87,126],"able":[90],"store":[92],"without":[98],"incurring":[99],"any":[100],"accuracy.":[104],"Our":[105,134],"approach":[106,148],"based":[108],"on":[109,176],"left-looking":[111],"scheme":[112],"employing":[113],"buffers":[116],"controlled":[118],"size":[119],"doubly":[124],"partitioned":[125],"exploiting":[127],"cores":[129],"panel":[132],"factorizations.":[133],"numerical":[135],"results":[136],"show":[137],"compared":[139],"state":[142],"art,":[145],"proposed":[147],"similar":[151],"but":[153],"only":[155],"memory":[161],"footprint,":[162],"hence":[164],"potentially":[165],"much":[166],"faster:":[167],"it":[168],"achieves":[169],"up":[170],"2\u00d7":[172],"3.5\u00d7":[174],"speedups":[175],"V100":[177],"A100":[179],"GPUs,":[180],"respectively.":[181]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2024,"cited_by_count":5},{"year":2023,"cited_by_count":1},{"year":2022,"cited_by_count":1}],"updated_date":"2026-08-04T08:18:43.703281","created_date":"2025-10-10T00:00:00"}