{"id":"https://openalex.org/W4416996074","doi":"https://doi.org/10.1145/3779120","title":"Kernel Float: Unlocking Mixed-Precision GPU Programming","display_name":"Kernel Float: Unlocking Mixed-Precision GPU Programming","publication_year":2025,"publication_date":"2025-12-04","ids":{"openalex":"https://openalex.org/W4416996074","doi":"https://doi.org/10.1145/3779120"},"language":"en","primary_location":{"id":"doi:10.1145/3779120","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3779120","pdf_url":null,"source":{"id":"https://openalex.org/S163326656","display_name":"ACM Transactions on Mathematical Software","issn_l":"0098-3500","issn":["0098-3500","1557-7295"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ACM Transactions on Mathematical Software","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://scholarlypublications.universiteitleiden.nl/access/item%3A4286691/view","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5073239665","display_name":"Stijn Heldens","orcid":"https://orcid.org/0000-0001-8792-6305"},"institutions":[{"id":"https://openalex.org/I4210095242","display_name":"Netherlands eScience Center","ror":"https://ror.org/00rbjv475","country_code":"NL","type":"funder","lineage":["https://openalex.org/I2800991832","https://openalex.org/I4210090210","https://openalex.org/I4210095242"]}],"countries":["NL"],"is_corresponding":false,"raw_author_name":"Stijn Heldens","raw_affiliation_strings":["Netherlands eScience Center, Amsterdam, The Netherlands","Netherlands eScience Center, The Netherlands"],"raw_orcid":"https://orcid.org/0000-0001-8792-6305","affiliations":[{"raw_affiliation_string":"Netherlands eScience Center, Amsterdam, The Netherlands","institution_ids":["https://openalex.org/I4210095242"]},{"raw_affiliation_string":"Netherlands eScience Center, The Netherlands","institution_ids":["https://openalex.org/I4210095242"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5035572633","display_name":"Ben van Werkhoven","orcid":"https://orcid.org/0000-0002-7508-3272"},"institutions":[{"id":"https://openalex.org/I121797337","display_name":"Leiden University","ror":"https://ror.org/027bh9e22","country_code":"NL","type":"education","lineage":["https://openalex.org/I121797337"]}],"countries":["NL"],"is_corresponding":false,"raw_author_name":"Ben van Werkhoven","raw_affiliation_strings":["Leiden University, Leiden, The Netherlands","Leiden University, The Netherlands"],"raw_orcid":"https://orcid.org/0000-0002-7508-3272","affiliations":[{"raw_affiliation_string":"Leiden University, Leiden, The Netherlands","institution_ids":["https://openalex.org/I121797337"]},{"raw_affiliation_string":"Leiden University, The Netherlands","institution_ids":["https://openalex.org/I121797337"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.7437,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.79827919,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":94,"max":97},"biblio":{"volume":"52","issue":"1","first_page":"1","last_page":"36"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11697","display_name":"Numerical Methods and Algorithms","score":0.5842000246047974,"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"}},"topics":[{"id":"https://openalex.org/T11697","display_name":"Numerical Methods and Algorithms","score":0.5842000246047974,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.36070001125335693,"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/T11693","display_name":"Cryptography and Residue Arithmetic","score":0.006599999964237213,"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/kernel","display_name":"Kernel (algebra)","score":0.6518999934196472},{"id":"https://openalex.org/keywords/block","display_name":"Block (permutation group theory)","score":0.4986000061035156},{"id":"https://openalex.org/keywords/double-precision-floating-point-format","display_name":"Double-precision floating-point format","score":0.4187000095844269},{"id":"https://openalex.org/keywords/interface","display_name":"Interface (matter)","score":0.41519999504089355},{"id":"https://openalex.org/keywords/memory-bandwidth","display_name":"Memory bandwidth","score":0.41510000824928284},{"id":"https://openalex.org/keywords/trigonometric-functions","display_name":"Trigonometric functions","score":0.3953000009059906},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.3847000002861023},{"id":"https://openalex.org/keywords/source-code","display_name":"Source code","score":0.3540000021457672}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8474000096321106},{"id":"https://openalex.org/C74193536","wikidata":"https://www.wikidata.org/wiki/Q574844","display_name":"Kernel (algebra)","level":2,"score":0.6518999934196472},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.631600022315979},{"id":"https://openalex.org/C459310","wikidata":"https://www.wikidata.org/wiki/Q117801","display_name":"Computational science","level":1,"score":0.5357999801635742},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.4986000061035156},{"id":"https://openalex.org/C35912277","wikidata":"https://www.wikidata.org/wiki/Q1243369","display_name":"Double-precision floating-point format","level":3,"score":0.4187000095844269},{"id":"https://openalex.org/C113843644","wikidata":"https://www.wikidata.org/wiki/Q901882","display_name":"Interface (matter)","level":4,"score":0.41519999504089355},{"id":"https://openalex.org/C188045654","wikidata":"https://www.wikidata.org/wiki/Q17148339","display_name":"Memory bandwidth","level":2,"score":0.41510000824928284},{"id":"https://openalex.org/C178009071","wikidata":"https://www.wikidata.org/wiki/Q93344","display_name":"Trigonometric functions","level":2,"score":0.3953000009059906},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.3847000002861023},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.36340001225471497},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.3540000021457672},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.35040000081062317},{"id":"https://openalex.org/C83283714","wikidata":"https://www.wikidata.org/wiki/Q121117","display_name":"Supercomputer","level":2,"score":0.3450999855995178},{"id":"https://openalex.org/C2776257435","wikidata":"https://www.wikidata.org/wiki/Q1576430","display_name":"Bandwidth (computing)","level":2,"score":0.3206999897956848},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.31040000915527344},{"id":"https://openalex.org/C12267149","wikidata":"https://www.wikidata.org/wiki/Q282453","display_name":"Support vector machine","level":2,"score":0.3091000020503998},{"id":"https://openalex.org/C76752949","wikidata":"https://www.wikidata.org/wiki/Q7607499","display_name":"Stencil","level":2,"score":0.3052999973297119},{"id":"https://openalex.org/C50630238","wikidata":"https://www.wikidata.org/wiki/Q971505","display_name":"General-purpose computing on graphics processing units","level":3,"score":0.29760000109672546},{"id":"https://openalex.org/C122280245","wikidata":"https://www.wikidata.org/wiki/Q620622","display_name":"Kernel method","level":3,"score":0.2865000069141388},{"id":"https://openalex.org/C139571649","wikidata":"https://www.wikidata.org/wiki/Q1156793","display_name":"Program optimization","level":3,"score":0.27549999952316284},{"id":"https://openalex.org/C161824985","wikidata":"https://www.wikidata.org/wiki/Q919509","display_name":"Vector processor","level":2,"score":0.2734000086784363},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.26910001039505005},{"id":"https://openalex.org/C34165917","wikidata":"https://www.wikidata.org/wiki/Q188267","display_name":"Programming paradigm","level":2,"score":0.2680000066757202}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3779120","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3779120","pdf_url":null,"source":{"id":"https://openalex.org/S163326656","display_name":"ACM Transactions on Mathematical Software","issn_l":"0098-3500","issn":["0098-3500","1557-7295"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ACM Transactions on Mathematical Software","raw_type":"journal-article"},{"id":"pmh:oai:scholarlypublications.universiteitleiden.nl:item_4286690","is_oa":true,"landing_page_url":"https://hdl.handle.net/1887/4286690","pdf_url":"https://scholarlypublications.universiteitleiden.nl/access/item%3A4286691/view","source":{"id":"https://openalex.org/S4306400850","display_name":"Leiden Repository (Leiden University)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I121797337","host_organization_name":"Leiden University","host_organization_lineage":["https://openalex.org/I121797337"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"ACM Transactions on Mathematical Software","raw_type":"Article / Letter to editor"}],"best_oa_location":{"id":"pmh:oai:scholarlypublications.universiteitleiden.nl:item_4286690","is_oa":true,"landing_page_url":"https://hdl.handle.net/1887/4286690","pdf_url":"https://scholarlypublications.universiteitleiden.nl/access/item%3A4286691/view","source":{"id":"https://openalex.org/S4306400850","display_name":"Leiden Repository (Leiden University)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I121797337","host_organization_name":"Leiden University","host_organization_lineage":["https://openalex.org/I121797337"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"ACM Transactions on Mathematical Software","raw_type":"Article / Letter to editor"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1348636090","display_name":"CORTEX: the Center for Optimal, Real-Time Machine Studies of the Explosive Universe","funder_award_id":"NWA.1160.18.316","funder_id":"https://openalex.org/F4320321800","funder_display_name":"Nederlandse Organisatie voor Wetenschappelijk Onderzoek"},{"id":"https://openalex.org/G3810360413","display_name":null,"funder_award_id":"NLESC.OEC.2022.001","funder_id":"https://openalex.org/F4320314239","funder_display_name":"Netherlands eScience Center"},{"id":"https://openalex.org/G5555681916","display_name":"Direct numerical simulation of turbulent heat transfer with non-ideal gases and surface roughness","funder_award_id":"2022.001","funder_id":"https://openalex.org/F4320321800","funder_display_name":"Nederlandse Organisatie voor Wetenschappelijk Onderzoek"}],"funders":[{"id":"https://openalex.org/F4320314239","display_name":"Netherlands eScience Center","ror":"https://ror.org/00rbjv475"},{"id":"https://openalex.org/F4320321800","display_name":"Nederlandse Organisatie voor Wetenschappelijk Onderzoek","ror":"https://ror.org/04jsz6e67"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416996074.pdf","grobid_xml":"https://content.openalex.org/works/W4416996074.grobid-xml"},"referenced_works_count":28,"referenced_works":["https://openalex.org/W1969213662","https://openalex.org/W2080592089","https://openalex.org/W2118495140","https://openalex.org/W2137269967","https://openalex.org/W2265166184","https://openalex.org/W2357843841","https://openalex.org/W2554131156","https://openalex.org/W2555109231","https://openalex.org/W2766338242","https://openalex.org/W2771613338","https://openalex.org/W2793822144","https://openalex.org/W2887327791","https://openalex.org/W2902657696","https://openalex.org/W2937715180","https://openalex.org/W2948348275","https://openalex.org/W2952851876","https://openalex.org/W2981849318","https://openalex.org/W2995792110","https://openalex.org/W3000483857","https://openalex.org/W3010946599","https://openalex.org/W3037020375","https://openalex.org/W3038558128","https://openalex.org/W3126212707","https://openalex.org/W3209744653","https://openalex.org/W4308090436","https://openalex.org/W4322008031","https://openalex.org/W4385585415","https://openalex.org/W6912012926"],"related_works":[],"abstract_inverted_index":{"Modern":[0],"GPUs":[1,122],"feature":[2],"specialized":[3],"hardware":[4,90],"for":[5,76,83],"low-precision":[6,41,84],"floating-point":[7],"arithmetic":[8],"to":[9,128,140,143],"accelerate":[10],"compute-intensive":[11],"workloads":[12],"that":[13,57,87,155],"do":[14],"not":[15,159],"require":[16],"high":[17],"numerical":[18],"accuracy,":[19],"such":[20],"as":[21],"those":[22],"from":[23,111],"artificial":[24],"intelligence.":[25],"However,":[26],"despite":[27],"the":[28,59,94],"significant":[29],"gains":[30],"in":[31],"computational":[32],"throughput,":[33],"memory":[34],"bandwidth":[35],"utilization,":[36],"and":[37,80,119,146,176],"energy":[38],"efficiency,":[39],"integrating":[40],"formats":[42],"into":[43,107],"scientific":[44],"applications":[45],"remains":[46],"difficult.":[47],"We":[48],"introduce":[49],"Kernel":[50,66,105],"Float":[51,67,106],",":[52],"a":[53,69,73],"header-only":[54],"C++":[55],"library":[56],"simplifies":[58],"development":[60],"of":[61,96,126],"portable":[62],"mixed-precision":[63,97,156],"GPU":[64,109],"kernels.":[65],"provides":[68],"generic":[70],"vector":[71,177],"type,":[72],"unified":[74],"interface":[75],"common":[77],"mathematical":[78],"operations,":[79],"fast":[81],"approximations":[82],"transcendental":[85],"functions":[86],"lack":[88],"native":[89],"support.":[91],"To":[92],"demonstrate":[93],"potential":[95],"computing":[98],"unlocked":[99],"by":[100,138],"our":[101],"library,":[102],"we":[103],"integrated":[104],"nine":[108],"kernels":[110,145],"various":[112],"domains.":[113],"Our":[114,151],"evaluation":[115],"on":[116,161,168],"Nvidia":[117],"A100":[118],"AMD":[120],"MI250X":[121],"shows":[123],"performance":[124,157],"improvements":[125],"up":[127,139],"\\(12\\times\\)":[129],"over":[130],"double":[131],"precision,":[132],"while":[133],"reducing":[134],"source":[135],"code":[136],"length":[137],"50%":[141],"compared":[142],"handwritten":[144],"having":[147],"negligible":[148],"runtime":[149],"overhead.":[150],"results":[152],"further":[153],"show":[154],"depends":[158],"only":[160],"choosing":[162],"appropriate":[163],"data":[164],"types":[165],"but":[166],"also":[167],"tuning":[169],"traditional":[170],"optimization":[171],"parameters":[172],"(e.g.,":[173],"block":[174],"size":[175],"width)":[178],"and,":[179],"when":[180],"relevant,":[181],"even":[182],"domain-specific":[183],"parameters.":[184]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-08-01T09:00:35.917206","created_date":"2025-12-04T00:00:00"}