{"id":"https://openalex.org/W1991056822","doi":"https://doi.org/10.1145/2465529.2465540","title":"Accelerating GPGPU architecture simulation","display_name":"Accelerating GPGPU architecture simulation","publication_year":2013,"publication_date":"2013-06-17","ids":{"openalex":"https://openalex.org/W1991056822","doi":"https://doi.org/10.1145/2465529.2465540","mag":"1991056822"},"language":"en","primary_location":{"id":"doi:10.1145/2465529.2465540","is_oa":false,"landing_page_url":"https://doi.org/10.1145/2465529.2465540","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM SIGMETRICS/international conference on Measurement and modeling of computer systems","raw_type":"proceedings-article"},"type":"article","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/A5048614443","display_name":"Zhibin Yu","orcid":"https://orcid.org/0000-0001-8067-9612"},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"funder","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210145761","display_name":"Shenzhen Institutes of Advanced Technology","ror":"https://ror.org/04gh4er46","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210145761"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Zhibin Yu","raw_affiliation_strings":["Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China","Shenzhen Institute of Advanced Technology Chinese Academy of Sciences, Shenzhen, China"],"affiliations":[{"raw_affiliation_string":"Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China","institution_ids":["https://openalex.org/I4210145761","https://openalex.org/I19820366"]},{"raw_affiliation_string":"Shenzhen Institute of Advanced Technology Chinese Academy of Sciences, Shenzhen, China","institution_ids":["https://openalex.org/I4210145761","https://openalex.org/I19820366"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5033119975","display_name":"Lieven Eeckhout","orcid":"https://orcid.org/0000-0001-8792-4473"},"institutions":[{"id":"https://openalex.org/I32597200","display_name":"Ghent University","ror":"https://ror.org/00cv9y106","country_code":"BE","type":"education","lineage":["https://openalex.org/I32597200"]}],"countries":["BE"],"is_corresponding":false,"raw_author_name":"Lieven Eeckhout","raw_affiliation_strings":["ELIS Department, Ghent University, Ghent, Belgium"],"affiliations":[{"raw_affiliation_string":"ELIS Department, Ghent University, Ghent, Belgium","institution_ids":["https://openalex.org/I32597200"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5080413066","display_name":"Nilanjan Goswami","orcid":null},"institutions":[{"id":"https://openalex.org/I33213144","display_name":"University of Florida","ror":"https://ror.org/02y3ad647","country_code":"US","type":"education","lineage":["https://openalex.org/I33213144"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Nilanjan Goswami","raw_affiliation_strings":["Intelligent Design of Efficient Architectures Lab, University of Florida, Gainesville, USA","Intelligent Design of Efficient Architectures Lab, University of Florida, Gainesville, USA#TAB#"],"affiliations":[{"raw_affiliation_string":"Intelligent Design of Efficient Architectures Lab, University of Florida, Gainesville, USA","institution_ids":["https://openalex.org/I33213144"]},{"raw_affiliation_string":"Intelligent Design of Efficient Architectures Lab, University of Florida, Gainesville, USA#TAB#","institution_ids":["https://openalex.org/I33213144"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100455376","display_name":"Tao Li","orcid":"https://orcid.org/0000-0003-1697-8022"},"institutions":[{"id":"https://openalex.org/I33213144","display_name":"University of Florida","ror":"https://ror.org/02y3ad647","country_code":"US","type":"education","lineage":["https://openalex.org/I33213144"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Tao Li","raw_affiliation_strings":["Intelligent Design of Efficient Architectures Lab, University of Florida, Gainesville, USA","Intelligent Design of Efficient Architectures Lab, University of Florida, Gainesville, USA#TAB#"],"affiliations":[{"raw_affiliation_string":"Intelligent Design of Efficient Architectures Lab, University of Florida, Gainesville, USA","institution_ids":["https://openalex.org/I33213144"]},{"raw_affiliation_string":"Intelligent Design of Efficient Architectures Lab, University of Florida, Gainesville, USA#TAB#","institution_ids":["https://openalex.org/I33213144"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5068885069","display_name":"Lizy K. John","orcid":"https://orcid.org/0000-0002-8747-5214"},"institutions":[{"id":"https://openalex.org/I86519309","display_name":"The University of Texas at Austin","ror":"https://ror.org/00hj54h04","country_code":"US","type":"education","lineage":["https://openalex.org/I86519309"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Lizy John","raw_affiliation_strings":["Department of Electrical and Computer Engineering, University of Texas at Austin, Austin, USA"],"affiliations":[{"raw_affiliation_string":"Department of Electrical and Computer Engineering, University of Texas at Austin, Austin, USA","institution_ids":["https://openalex.org/I86519309"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5022262922","display_name":"Hai Jin","orcid":"https://orcid.org/0000-0002-3934-7605"},"institutions":[{"id":"https://openalex.org/I4210144487","display_name":"Cloud Computing Center","ror":"https://ror.org/04aa0zm65","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210144487"]},{"id":"https://openalex.org/I47720641","display_name":"Huazhong University of Science and Technology","ror":"https://ror.org/00p991c53","country_code":"CN","type":"education","lineage":["https://openalex.org/I47720641"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hai Jin","raw_affiliation_strings":["Service Computing Technologies and System Lab/Cluster and Grid Computing Lab, HUST, Wuhan, China","Service Computing Technologies and System Lab/Cluster and Grid Computing Lab, HUST, Wuhan, China#TAB#"],"affiliations":[{"raw_affiliation_string":"Service Computing Technologies and System Lab/Cluster and Grid Computing Lab, HUST, Wuhan, China","institution_ids":["https://openalex.org/I4210144487"]},{"raw_affiliation_string":"Service Computing Technologies and System Lab/Cluster and Grid Computing Lab, HUST, Wuhan, China#TAB#","institution_ids":["https://openalex.org/I47720641"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5012773300","display_name":"Chengzhong Xu","orcid":"https://orcid.org/0000-0001-9480-0356"},"institutions":[{"id":"https://openalex.org/I185443292","display_name":"Wayne State University","ror":"https://ror.org/01070mq45","country_code":"US","type":"education","lineage":["https://openalex.org/I185443292"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Chengzhong Xu","raw_affiliation_strings":["Department of Electrical and Computer Engineering, Wayne State University, Detroit, USA"],"affiliations":[{"raw_affiliation_string":"Department of Electrical and Computer Engineering, Wayne State University, Detroit, USA","institution_ids":["https://openalex.org/I185443292"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5048614443"],"corresponding_institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210145761"],"apc_list":null,"apc_paid":null,"fwci":0.9495,"has_fulltext":false,"cited_by_count":13,"citation_normalized_percentile":{"value":0.74780362,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":89,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"331","last_page":"332"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":1.0,"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":1.0,"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/T10829","display_name":"Interconnection Networks and Systems","score":0.9976999759674072,"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"}},{"id":"https://openalex.org/T11181","display_name":"Advanced Data Storage Technologies","score":0.9959999918937683,"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/general-purpose-computing-on-graphics-processing-units","display_name":"General-purpose computing on graphics processing units","score":0.9217627048492432},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8481044173240662},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.7843213081359863},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.6822052001953125},{"id":"https://openalex.org/keywords/kernel","display_name":"Kernel (algebra)","score":0.571405291557312},{"id":"https://openalex.org/keywords/graphics-processing-unit","display_name":"Graphics processing unit","score":0.4947773814201355},{"id":"https://openalex.org/keywords/thread","display_name":"Thread (computing)","score":0.4566035568714142},{"id":"https://openalex.org/keywords/graphics","display_name":"Graphics","score":0.4544851779937744},{"id":"https://openalex.org/keywords/microarchitecture","display_name":"Microarchitecture","score":0.43161484599113464},{"id":"https://openalex.org/keywords/computer-graphics","display_name":"Computer graphics (images)","score":0.1709691286087036},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.10589626431465149}],"concepts":[{"id":"https://openalex.org/C50630238","wikidata":"https://www.wikidata.org/wiki/Q971505","display_name":"General-purpose computing on graphics processing units","level":3,"score":0.9217627048492432},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8481044173240662},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.7843213081359863},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.6822052001953125},{"id":"https://openalex.org/C74193536","wikidata":"https://www.wikidata.org/wiki/Q574844","display_name":"Kernel (algebra)","level":2,"score":0.571405291557312},{"id":"https://openalex.org/C2779851693","wikidata":"https://www.wikidata.org/wiki/Q183484","display_name":"Graphics processing unit","level":2,"score":0.4947773814201355},{"id":"https://openalex.org/C138101251","wikidata":"https://www.wikidata.org/wiki/Q213092","display_name":"Thread (computing)","level":2,"score":0.4566035568714142},{"id":"https://openalex.org/C21442007","wikidata":"https://www.wikidata.org/wiki/Q1027879","display_name":"Graphics","level":2,"score":0.4544851779937744},{"id":"https://openalex.org/C107598950","wikidata":"https://www.wikidata.org/wiki/Q259864","display_name":"Microarchitecture","level":2,"score":0.43161484599113464},{"id":"https://openalex.org/C121684516","wikidata":"https://www.wikidata.org/wiki/Q7600677","display_name":"Computer graphics (images)","level":1,"score":0.1709691286087036},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.10589626431465149},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.0},{"id":"https://openalex.org/C114614502","wikidata":"https://www.wikidata.org/wiki/Q76592","display_name":"Combinatorics","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/2465529.2465540","is_oa":false,"landing_page_url":"https://doi.org/10.1145/2465529.2465540","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM SIGMETRICS/international conference on Measurement and modeling of computer systems","raw_type":"proceedings-article"},{"id":"pmh:oai:archive.ugent.be:4235676","is_oa":false,"landing_page_url":"http://hdl.handle.net/1854/LU-4235676","pdf_url":null,"source":{"id":"https://openalex.org/S4306400478","display_name":"Ghent University Academic Bibliography (Ghent University)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I32597200","host_organization_name":"Ghent University","host_organization_lineage":["https://openalex.org/I32597200"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"ISBN: 9781450319003","raw_type":"info:eu-repo/semantics/publishedVersion"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Industry, innovation and infrastructure","score":0.4399999976158142,"id":"https://metadata.un.org/sdg/9"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":4,"referenced_works":["https://openalex.org/W1979527452","https://openalex.org/W2123037616","https://openalex.org/W2153456949","https://openalex.org/W2158924248"],"related_works":["https://openalex.org/W1963859303","https://openalex.org/W2364044215","https://openalex.org/W2389600408","https://openalex.org/W240129890","https://openalex.org/W3048701459","https://openalex.org/W2149078538","https://openalex.org/W2370314112","https://openalex.org/W1912958759","https://openalex.org/W2792081825","https://openalex.org/W2893308117"],"abstract_inverted_index":{"Recently,":[0],"graphics":[1],"processing":[2],"units":[3],"(GPUs)":[4],"have":[5],"opened":[6],"up":[7,12,23,180],"new":[8],"opportunities":[9],"for":[10],"speeding":[11],"general-purpose":[13],"parallel":[14,49],"applications":[15],"due":[16,38],"to":[17,24,39,54,127,146,181],"their":[18],"massive":[19],"computational":[20],"power":[21],"and":[22,51,67,103,179],"hundreds":[25],"of":[26,28,43,95,112,132,150,158,175,189],"thousands":[27],"threads":[29],"enabled":[30],"by":[31,82,153,172],"programming":[32],"models":[33],"such":[34],"as":[35],"CUDA.":[36],"However,":[37],"the":[40,77,92,109,113,133,148],"serial":[41],"nature":[42],"existing":[44,97],"micro-architecture":[45],"simulators,":[46],"these":[47],"massively":[48],"architectures":[50,63],"workloads":[52],"need":[53],"be":[55],"simulated":[56,151],"sequentially.":[57],"As":[58],"a":[59,101,122,129,173],"result,":[60],"simulating":[61],"GPGPU":[62,78,87,98,123,169],"with":[64,183],"typical":[65],"benchmarks":[66],"input":[68],"data":[69],"sets":[70],"is":[71,145],"extremely":[72],"time-consuming.":[73],"This":[74],"paper":[75],"addresses":[76],"architecture":[79,170],"simulation":[80,139,171],"challenge":[81],"generating":[83],"miniature,":[84],"yet":[85],"representative":[86],"kernels.":[88],"We":[89,119],"first":[90],"summarize":[91],"static":[93],"characteristics":[94],"an":[96,184],"kernel":[99,124],"in":[100],"profile,":[102],"analyze":[104],"its":[105],"dynamic":[106],"behavior":[107],"using":[108],"novel":[110],"concept":[111],"divergence":[114],"flow":[115],"statistics":[116],"graph":[117],"(DFSG).":[118],"subsequently":[120],"use":[121],"synthesizing":[125],"framework":[126],"generate":[128],"miniature":[130],"proxy":[131],"original":[134],"kernel,":[135],"which":[136],"can":[137,167],"reduce":[138,147],"time":[140],"significantly.":[141],"The":[142],"key":[143],"idea":[144],"number":[149],"instructions":[152],"decreasing":[154],"per-thread":[155],"iteration":[156],"counts":[157],"loops.":[159],"Our":[160],"experimental":[161],"results":[162],"show":[163],"that":[164],"our":[165],"approach":[166],"accelerate":[168],"factor":[174],"88X":[176],"on":[177],"average":[178,185],"589X":[182],"IPC":[186],"relative":[187],"error":[188],"5.6%.":[190]},"counts_by_year":[{"year":2025,"cited_by_count":2},{"year":2023,"cited_by_count":2},{"year":2022,"cited_by_count":1},{"year":2021,"cited_by_count":1},{"year":2020,"cited_by_count":1},{"year":2019,"cited_by_count":1},{"year":2018,"cited_by_count":1},{"year":2017,"cited_by_count":1},{"year":2016,"cited_by_count":1},{"year":2015,"cited_by_count":1},{"year":2014,"cited_by_count":1}],"updated_date":"2026-04-04T16:13:02.066488","created_date":"2025-10-10T00:00:00"}