{"id":"https://openalex.org/W4239685157","doi":"https://doi.org/10.1109/mlhpc.2016.005","title":"Performance-Portable Autotuning of OpenCL Kernels for Convolutional Layers of Deep Neural Networks","display_name":"Performance-Portable Autotuning of OpenCL Kernels for Convolutional Layers of Deep Neural Networks","alternate_titles":[],"publication_year":2016,"publication_date":"2016-11-01","ids":{"openalex":"https://openalex.org/W4239685157","doi":"https://doi.org/10.1109/mlhpc.2016.005","mag":"2562339416"},"language":"en","primary_location":{"id":"doi:10.1109/mlhpc.2016.005","doi":"https://doi.org/10.1109/mlhpc.2016.005","is_oa":false,"landing_page_url":"https://doi.org/10.1109/mlhpc.2016.005","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2016 2nd Workshop on Machine Learning in HPC Environments (MLHPC)","raw_type":"proceedings-article"},"parent_works":[],"lineage":["https://openalex.org/W4239685157"],"child_works_count":0,"type":"conference-paper","is_traditional":true,"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/A5062373552","display_name":"Yaohung Mike Tsai","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I75027704","display_name":"University of Tennessee at Knoxville","ror":"https://ror.org/020f3ap87","country_code":"US","type":"education","lineage":["https://openalex.org/I75027704"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Yaohung M. Tsai","raw_affiliation_strings":["University of Tennessee Knoxville"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Tennessee Knoxville","institution_ids":["https://openalex.org/I75027704"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073990539","display_name":"Piotr \u0141uszczek","orcid":"https://orcid.org/0000-0002-0089-6965","observed_orcids":["https://orcid.org/0000-0002-0089-6965"]},"institutions":[{"id":"https://openalex.org/I75027704","display_name":"University of Tennessee at Knoxville","ror":"https://ror.org/020f3ap87","country_code":"US","type":"education","lineage":["https://openalex.org/I75027704"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Piotr Luszczek","raw_affiliation_strings":["University of Tennessee Knoxville"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Tennessee Knoxville","institution_ids":["https://openalex.org/I75027704"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5070456582","display_name":"Jakub Kurzak","orcid":"https://orcid.org/0000-0002-9697-0145","observed_orcids":["https://orcid.org/0000-0002-9697-0145"]},"institutions":[{"id":"https://openalex.org/I75027704","display_name":"University of Tennessee at Knoxville","ror":"https://ror.org/020f3ap87","country_code":"US","type":"education","lineage":["https://openalex.org/I75027704"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Jakub Kurzak","raw_affiliation_strings":["University of Tennessee Knoxville"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Tennessee Knoxville","institution_ids":["https://openalex.org/I75027704"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5075517045","display_name":"Jack J. Dongarra","orcid":"https://orcid.org/0000-0003-3247-1782","observed_orcids":["https://orcid.org/0000-0003-3247-1782"]},"institutions":[{"id":"https://openalex.org/I1289243028","display_name":"Oak Ridge National Laboratory","ror":"https://ror.org/01qz5mb56","country_code":"US","type":"facility","lineage":["https://openalex.org/I1289243028","https://openalex.org/I1330989302","https://openalex.org/I39565521","https://openalex.org/I4210159294"]},{"id":"https://openalex.org/I75027704","display_name":"University of Tennessee at Knoxville","ror":"https://ror.org/020f3ap87","country_code":"US","type":"education","lineage":["https://openalex.org/I75027704"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Jack Dongarra","raw_affiliation_strings":["University of Tennessee Knoxville and Oak Ridge National Laboratory"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Tennessee Knoxville and Oak Ridge National Laboratory","institution_ids":["https://openalex.org/I1289243028","https://openalex.org/I75027704"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":1.4672,"has_fulltext":false,"cited_by_count":16,"reviewed_by_count":0,"citation_normalized_percentile":{"value":0.81383989,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":86,"max":98},"biblio":{"volume":"2016","issue":null,"first_page":"9","last_page":"18"},"is_retracted":false,"is_pulled":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9994999766349792,"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":0.9994999766349792,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.00019999999494757503,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10904","display_name":"Embedded Systems Design Techniques","score":9.999999747378752e-05,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/auto-tuning","display_name":"auto-tuning","score":0.9160000085830688},{"id":"https://openalex.org/keywords/opencl","display_name":"OpenCL","score":0.7419999837875366},{"id":"https://openalex.org/keywords/convolutional-layers","display_name":"convolutional layers","score":0.8169999718666077},{"id":"https://openalex.org/keywords/deep-neural-network","display_name":"deep neural network","score":0.9760000109672546},{"id":"https://openalex.org/keywords/performance-portability","display_name":"performance portability","score":0.9300000071525574},{"id":"https://openalex.org/keywords/gpu-acceleration","display_name":"GPU acceleration","score":0.45899999141693115},{"id":"https://openalex.org/keywords/data-layout-optimization","display_name":"data layout optimization","score":0.8029999732971191},{"id":"https://openalex.org/keywords/low-level-optimization","display_name":"low-level optimization","score":0.6779999732971191},{"id":"https://openalex.org/keywords/cudnn","display_name":"cuDNN","score":0.9559999704360962},{"id":"https://openalex.org/keywords/high-bandwidth-memory","display_name":"high-bandwidth memory","score":0.6150000095367432},{"id":"https://openalex.org/keywords/deep-learning-training","display_name":"deep learning training","score":0.5220000147819519}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8672881126403809},{"id":"https://openalex.org/C63000827","wikidata":"https://www.wikidata.org/wiki/Q3080428","display_name":"Software portability","level":2,"score":0.7693042159080505},{"id":"https://openalex.org/C2777338717","wikidata":"https://www.wikidata.org/wiki/Q1762621","display_name":"Vendor","level":2,"score":0.6925657987594604},{"id":"https://openalex.org/C177212765","wikidata":"https://www.wikidata.org/wiki/Q627335","display_name":"Workflow","level":2,"score":0.6019489169120789},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.5504032373428345},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.48458898067474365},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.47059714794158936},{"id":"https://openalex.org/C188045654","wikidata":"https://www.wikidata.org/wiki/Q17148339","display_name":"Memory bandwidth","level":2,"score":0.46426984667778015},{"id":"https://openalex.org/C179518139","wikidata":"https://www.wikidata.org/wiki/Q5140297","display_name":"Coding (social sciences)","level":2,"score":0.44694775342941284},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.42273736000061035},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.38382241129875183},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.3432368040084839},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.2997629642486572},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2172389030456543},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.0},{"id":"https://openalex.org/C162853370","wikidata":"https://www.wikidata.org/wiki/Q39809","display_name":"Marketing","level":1,"score":0.0},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.0},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.0},{"id":"https://openalex.org/C77088390","wikidata":"https://www.wikidata.org/wiki/Q8513","display_name":"Database","level":1,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1109/mlhpc.2016.005","doi":"https://doi.org/10.1109/mlhpc.2016.005","is_oa":false,"landing_page_url":"https://doi.org/10.1109/mlhpc.2016.005","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2016 2nd Workshop on Machine Learning in HPC Environments (MLHPC)","raw_type":"proceedings-article"},{"id":"mag:2562339416","doi":null,"is_oa":false,"landing_page_url":"https://dblp.uni-trier.de/db/conf/sc/mlhpc2016.html#TsaiLKD16","pdf_url":null,"source":{"id":"https://openalex.org/S4306418683","display_name":"IEEE International Conference on High Performance Computing, Data, and Analytics","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"conference"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":"IEEE International Conference on High Performance Computing, Data, and Analytics","raw_type":null},{"id":"mag:2755072197","doi":null,"is_oa":false,"landing_page_url":"http://jglobal.jst.go.jp/public/20090422/201702264513076442","pdf_url":null,"source":{"id":"https://openalex.org/S4306512817","display_name":"IEEE Conference Proceedings","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":"IEEE Conference Proceedings","raw_type":null}],"best_oa_location":null,"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320309480","display_name":"Nvidia","ror":"https://ror.org/03jdj4y14"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":34,"referenced_works":["https://openalex.org/W60493759","https://openalex.org/W182691100","https://openalex.org/W587794757","https://openalex.org/W753012316","https://openalex.org/W1005811612","https://openalex.org/W1487583988","https://openalex.org/W1563686443","https://openalex.org/W1598866093","https://openalex.org/W1606347560","https://openalex.org/W1667652561","https://openalex.org/W1825672851","https://openalex.org/W1841592590","https://openalex.org/W1843951383","https://openalex.org/W1922123711","https://openalex.org/W1978642402","https://openalex.org/W2052863973","https://openalex.org/W2059745360","https://openalex.org/W2097117768","https://openalex.org/W2101159990","https://openalex.org/W2107911628","https://openalex.org/W2112796928","https://openalex.org/W2117539524","https://openalex.org/W2117731089","https://openalex.org/W2155893237","https://openalex.org/W2162390675","https://openalex.org/W2163605009","https://openalex.org/W2168231600","https://openalex.org/W2172654076","https://openalex.org/W2306526421","https://openalex.org/W2500595894","https://openalex.org/W2546241758","https://openalex.org/W2949245006","https://openalex.org/W2950179405","https://openalex.org/W2963340555"],"related_works":["https://openalex.org/W4224303483","https://openalex.org/W3091795552","https://openalex.org/W2549301733","https://openalex.org/W4280563761","https://openalex.org/W4415282691","https://openalex.org/W3167739300","https://openalex.org/W7062678914","https://openalex.org/W3004820010","https://openalex.org/W3008278996","https://openalex.org/W2911588913"],"abstract_inverted_index":{"We":[0],"present":[1],"a":[2,18,95,119,134,139,148,153],"portable":[3],"and":[4,11,34,50,66,106,164],"highly-optimized":[5],"Deep":[6],"Neural":[7],"Network":[8],"(DNN)":[9],"algorithm":[10],"its":[12],"implementation":[13,65],"techniques.":[14],"Our":[15,73],"approach":[16,163],"is":[17,44,69],"novel":[19],"combination":[20],"of":[21,85,115,141,161,167],"existing":[22],"HPC":[23],"techniques":[24],"that":[25,37],"methodically":[26],"applies":[27],"autotuning":[28],"as":[29,31],"well":[30],"data":[32],"layout":[33],"low-level":[35],"optimizations":[36],"achieve":[38],"performance":[39,116],"matching":[40],"and/or":[41],"exceeding":[42],"what":[43],"possible":[45],"with":[46,122],"either":[47],"reverse":[48],"engineering":[49],"manual":[51],"assembly":[52],"coding":[53],"or":[54],"proprietary":[55],"vendor":[56,155],"libraries.":[57],"The":[58],"former":[59],"was":[60],"done":[61],"inside":[62],"the":[63,67,80,89,113,123,142,159,165],"maxDNN":[64],"latter":[68],"represented":[70],"by":[71],"cuDNN.":[72],"work":[74],"may":[75],"be":[76],"directly":[77],"applied":[78],"to":[79,158],"most":[81],"time":[82],"consuming":[83],"part":[84],"DNN":[86],"workflow,":[87],"namely":[88],"training":[90],"process":[91],"which":[92],"often":[93],"needs":[94],"restart":[96],"when":[97],"it":[98],"stagnates":[99],"due":[100],"to,":[101],"for":[102],"example,":[103],"diminishing":[104],"gradients":[105],"getting":[107],"stuck":[108],"in":[109],"local":[110],"minima.":[111],"With":[112],"result":[114],"tests":[117],"on":[118,147],"consumer-grade":[120],"GPU":[121,150],"latest":[124],"High":[125],"Bandwidth":[126],"Memory":[127],"(HBM)":[128],"stack,":[129],"our":[130,162,168],"methodology":[131],"can":[132],"match":[133],"server":[135],"grade":[136],"hardware":[137],"at":[138],"fraction":[140],"price.":[143],"Another":[144],"tuning":[145],"sweep":[146],"new":[149],"architecture":[151],"from":[152],"different":[154],"also":[156],"attests":[157],"portability":[160],"quality":[166],"implementation.":[169]},"cited_by_api_url":"https://api.openalex.org/works?filter=cites:W4239685157","counts_by_year":[{"year":2024,"cited_by_count":1},{"year":2022,"cited_by_count":1},{"year":2021,"cited_by_count":1},{"year":2020,"cited_by_count":4},{"year":2019,"cited_by_count":1},{"year":2018,"cited_by_count":5},{"year":2017,"cited_by_count":3}],"updated_date":"2026-11-21T10:27:49.728292","created_date":"2022-05-12T00:00:00"}
