{"id":"https://openalex.org/W2766338242","doi":"https://doi.org/10.1109/hpec.2017.8091072","title":"Exploiting half precision arithmetic in Nvidia GPUs","display_name":"Exploiting half precision arithmetic in Nvidia GPUs","publication_year":2017,"publication_date":"2017-09-01","ids":{"openalex":"https://openalex.org/W2766338242","doi":"https://doi.org/10.1109/hpec.2017.8091072","mag":"2766338242"},"language":"en","primary_location":{"id":"doi:10.1109/hpec.2017.8091072","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpec.2017.8091072","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2017 IEEE High Performance Extreme Computing Conference (HPEC)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5065637540","display_name":"Nhut-Minh Ho","orcid":"https://orcid.org/0000-0002-3864-8027"},"institutions":[{"id":"https://openalex.org/I165932596","display_name":"National University of Singapore","ror":"https://ror.org/01tgyzw49","country_code":"SG","type":"education","lineage":["https://openalex.org/I165932596"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Nhut-Minh Ho","raw_affiliation_strings":["Department of Computer Science, National University of Singapore, Singapore"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science, National University of Singapore, Singapore","institution_ids":["https://openalex.org/I165932596"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5023989495","display_name":"Weng\u2010Fai Wong","orcid":"https://orcid.org/0000-0002-4281-2053"},"institutions":[{"id":"https://openalex.org/I165932596","display_name":"National University of Singapore","ror":"https://ror.org/01tgyzw49","country_code":"SG","type":"education","lineage":["https://openalex.org/I165932596"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Weng-Fai Wong","raw_affiliation_strings":["Department of Computer Science, National University of Singapore, Singapore"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science, National University of Singapore, Singapore","institution_ids":["https://openalex.org/I165932596"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I165932596"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":69,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"7"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9998999834060669,"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":0.9998999834060669,"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/T11697","display_name":"Numerical Methods and Algorithms","score":0.9995999932289124,"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/T10363","display_name":"Low-power high-performance VLSI design","score":0.9948999881744385,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/pascal","display_name":"Pascal (unit)","score":0.8650376796722412},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8399572968482971},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.82329261302948},{"id":"https://openalex.org/keywords/single-precision-floating-point-format","display_name":"Single-precision floating-point format","score":0.713782787322998},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.7011418342590332},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.6497542262077332},{"id":"https://openalex.org/keywords/exploit","display_name":"Exploit","score":0.6022665500640869},{"id":"https://openalex.org/keywords/arbitrary-precision-arithmetic","display_name":"Arbitrary-precision arithmetic","score":0.5693962574005127},{"id":"https://openalex.org/keywords/double-precision-floating-point-format","display_name":"Double-precision floating-point format","score":0.5634720325469971},{"id":"https://openalex.org/keywords/floating-point","display_name":"Floating point","score":0.49997878074645996},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.4245457351207733},{"id":"https://openalex.org/keywords/general-purpose-computing-on-graphics-processing-units","display_name":"General-purpose computing on graphics processing units","score":0.42393261194229126},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.23853665590286255},{"id":"https://openalex.org/keywords/graphics","display_name":"Graphics","score":0.19023549556732178},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.11052221059799194},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.10138726234436035}],"concepts":[{"id":"https://openalex.org/C75608658","wikidata":"https://www.wikidata.org/wiki/Q44395","display_name":"Pascal (unit)","level":2,"score":0.8650376796722412},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8399572968482971},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.82329261302948},{"id":"https://openalex.org/C133095886","wikidata":"https://www.wikidata.org/wiki/Q1307173","display_name":"Single-precision floating-point format","level":3,"score":0.713782787322998},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.7011418342590332},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.6497542262077332},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.6022665500640869},{"id":"https://openalex.org/C83581934","wikidata":"https://www.wikidata.org/wiki/Q527381","display_name":"Arbitrary-precision arithmetic","level":2,"score":0.5693962574005127},{"id":"https://openalex.org/C35912277","wikidata":"https://www.wikidata.org/wiki/Q1243369","display_name":"Double-precision floating-point format","level":3,"score":0.5634720325469971},{"id":"https://openalex.org/C84211073","wikidata":"https://www.wikidata.org/wiki/Q117879","display_name":"Floating point","level":2,"score":0.49997878074645996},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.4245457351207733},{"id":"https://openalex.org/C50630238","wikidata":"https://www.wikidata.org/wiki/Q971505","display_name":"General-purpose computing on graphics processing units","level":3,"score":0.42393261194229126},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.23853665590286255},{"id":"https://openalex.org/C21442007","wikidata":"https://www.wikidata.org/wiki/Q1027879","display_name":"Graphics","level":2,"score":0.19023549556732178},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.11052221059799194},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.10138726234436035},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.0},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/hpec.2017.8091072","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpec.2017.8091072","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2017 IEEE High Performance Extreme Computing Conference (HPEC)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Affordable and clean energy","score":0.8600000143051147,"id":"https://metadata.un.org/sdg/7"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":21,"referenced_works":["https://openalex.org/W1525058640","https://openalex.org/W1782174992","https://openalex.org/W2028499920","https://openalex.org/W2040281526","https://openalex.org/W2077143534","https://openalex.org/W2080592089","https://openalex.org/W2090268225","https://openalex.org/W2099404643","https://openalex.org/W2104508714","https://openalex.org/W2110195531","https://openalex.org/W2112760171","https://openalex.org/W2118031182","https://openalex.org/W2147423491","https://openalex.org/W2147654959","https://openalex.org/W2157273783","https://openalex.org/W2167182394","https://openalex.org/W2410071687","https://openalex.org/W3010528205","https://openalex.org/W3138798301","https://openalex.org/W4230890341","https://openalex.org/W4245302940"],"related_works":["https://openalex.org/W2116803521","https://openalex.org/W2773283032","https://openalex.org/W3150370983","https://openalex.org/W3150959508","https://openalex.org/W2239119680","https://openalex.org/W1564887326","https://openalex.org/W1571090276","https://openalex.org/W2185760795","https://openalex.org/W2930605373","https://openalex.org/W3215589575"],"abstract_inverted_index":{"With":[0],"the":[1,26,66,123,130,145],"growing":[2],"importance":[3],"of":[4,46],"deep":[5],"learning":[6],"and":[7,99,114,142],"energy-saving":[8],"approximate":[9],"computing,":[10],"half":[11,52,93,111],"precision":[12,48,53,70,94,108,112],"floating":[13],"point":[14],"arithmetic":[15,109],"(FP16)":[16],"is":[17],"fast":[18],"gaining":[19],"popularity.":[20],"Nvidia's":[21],"recent":[22],"Pascal":[23,146],"architecture":[24],"was":[25,133],"first":[27],"GPU":[28],"that":[29,42,144],"offered":[30],"FP16":[31],"support.":[32],"However,":[33],"when":[34],"actual":[35],"products":[36],"were":[37],"shipped,":[38],"programmers":[39],"soon":[40],"realized":[41],"a":[43,126],"na\u00efve":[44],"replacement":[45],"single":[47,107],"(FP32)":[49],"code":[50,88],"with":[51],"led":[54],"to":[55,64,82,89,110],"disappointing":[56],"performance":[57],"results,":[58],"even":[59],"if":[60],"they":[61],"are":[62],"willing":[63],"tolerate":[65],"increase":[67],"in":[68,119],"error":[69],"reduction":[71],"brings.":[72],"In":[73,122],"this":[74],"paper,":[75],"we":[76,101],"developed":[77],"an":[78],"automated":[79],"conversion":[80],"framework":[81],"help":[83],"users":[84],"migrate":[85],"their":[86],"CUDA":[87],"better":[90],"exploit":[91],"Pascal's":[92],"capability.":[95],"Using":[96],"our":[97],"tools":[98],"techniques,":[100],"successfully":[102],"convert":[103],"many":[104,120],"benchmarks":[105],"from":[106],"equivalent,":[113],"achieved":[115],"significant":[116],"speedup":[117,128],"improvement":[118],"cases.":[121],"best":[124],"case,":[125],"3\u00d7":[127],"over":[129],"FP32":[131],"version":[132],"achieved.":[134],"We":[135],"shall":[136],"also":[137],"discuss":[138],"some":[139],"new":[140],"issues":[141],"opportunities":[143],"GPUs":[147],"brought.":[148]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":4},{"year":2024,"cited_by_count":9},{"year":2023,"cited_by_count":8},{"year":2022,"cited_by_count":8},{"year":2021,"cited_by_count":12},{"year":2020,"cited_by_count":7},{"year":2019,"cited_by_count":11},{"year":2018,"cited_by_count":8},{"year":2012,"cited_by_count":1}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}