{"id":"https://openalex.org/W7203728331","doi":"https://doi.org/10.1109/access.2026.3725140","title":"DVFS-Aware Energy-Efficient Implementation of DNNs on Heterogeneous Hybrid Parallel Computers","display_name":"DVFS-Aware Energy-Efficient Implementation of DNNs on Heterogeneous Hybrid Parallel Computers","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7203728331","doi":"https://doi.org/10.1109/access.2026.3725140"},"language":"en","primary_location":{"id":"doi:10.1109/access.2026.3725140","doi":"https://doi.org/10.1109/access.2026.3725140","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2026.3725140","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"is_core":true,"listed_in":["anvur-scientific","cwts-core","doaj","jufo-1","ki-jl-1","norway-1","oarl","poland-100","russia-white-list-1","vabb-shw"],"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":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Access","raw_type":"journal-article"},"parent_works":[],"lineage":["https://openalex.org/W7203728331"],"child_works_count":0,"type":"article","is_traditional":true,"indexed_in":["crossref","doaj"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1109/access.2026.3725140","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5006106085","display_name":"Hamidreza Khaleghzadeh","orcid":"https://orcid.org/0000-0003-4070-7468","observed_orcids":["https://orcid.org/0000-0003-4070-7468"]},"institutions":[{"id":"https://openalex.org/I63072094","display_name":"University of Portsmouth","ror":"https://ror.org/03ykbk197","country_code":"GB","type":"education","lineage":["https://openalex.org/I63072094"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Hamidreza Khaleghzadeh","raw_affiliation_strings":["University of Portsmouth"],"raw_orcid":"https://orcid.org/0000-0003-4070-7468","affiliations":[{"raw_affiliation_string":"University of Portsmouth","institution_ids":["https://openalex.org/I63072094"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102916943","display_name":"Atefeh Khazaei","orcid":"https://orcid.org/0000-0001-7380-6126","observed_orcids":["https://orcid.org/0000-0001-7380-6126"]},"institutions":[{"id":"https://openalex.org/I63072094","display_name":"University of Portsmouth","ror":"https://ror.org/03ykbk197","country_code":"GB","type":"education","lineage":["https://openalex.org/I63072094"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Atefeh Khazaei","raw_affiliation_strings":["University of Portsmouth"],"raw_orcid":"https://orcid.org/0000-0001-7380-6126","affiliations":[{"raw_affiliation_string":"University of Portsmouth","institution_ids":["https://openalex.org/I63072094"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5084068586","display_name":"Alexey Lastovetsky","orcid":"https://orcid.org/0000-0001-9460-3897","observed_orcids":["https://orcid.org/0000-0001-9460-3897"]},"institutions":[{"id":"https://openalex.org/I100930933","display_name":"University College Dublin","ror":"https://ror.org/05m7pjf47","country_code":"IE","type":"education","lineage":["https://openalex.org/I100930933"]}],"countries":["IE"],"is_corresponding":false,"raw_author_name":"Alexey Lastovetsky","raw_affiliation_strings":["University College Dublin"],"raw_orcid":"https://orcid.org/0000-0001-9460-3897","affiliations":[{"raw_affiliation_string":"University College Dublin","institution_ids":["https://openalex.org/I100930933"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":{"value":2075,"currency":"USD","value_usd":2075},"apc_paid":{"value":2075,"currency":"USD","value_usd":2075},"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.0,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"14","issue":null,"first_page":"132028","last_page":"132044"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9965999722480774,"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":0.9965999722480774,"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/T14098","display_name":"Energy Efficiency in Computing","score":0.0007999999797903001,"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/T10101","display_name":"Cloud Computing and Resource Management","score":0.0007999999797903001,"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/dynamic-voltage-and-frequency-scaling","display_name":"dynamic voltage and frequency scaling","score":0.890999972820282},{"id":"https://openalex.org/keywords/deep-neural-network-training","display_name":"deep neural network training","score":0.8849999904632568},{"id":"https://openalex.org/keywords/energy-efficiency","display_name":"energy efficiency","score":0.9100000262260437},{"id":"https://openalex.org/keywords/bi-objective-optimization","display_name":"bi-objective optimization","score":0.8399999737739563},{"id":"https://openalex.org/keywords/pareto-front","display_name":"Pareto front","score":0.6430000066757202},{"id":"https://openalex.org/keywords/model-based-optimization","display_name":"model-based optimization","score":0.6600000262260437}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8374999761581421},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.44209998846054077},{"id":"https://openalex.org/C106515295","wikidata":"https://www.wikidata.org/wiki/Q26806595","display_name":"Parallel processing","level":2,"score":0.34049999713897705},{"id":"https://openalex.org/C18789546","wikidata":"https://www.wikidata.org/wiki/Q1341206","display_name":"Hybrid computer","level":2,"score":0.2759000062942505},{"id":"https://openalex.org/C106516650","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm design","level":2,"score":0.2750999927520752},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.2581000030040741},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.25769999623298645},{"id":"https://openalex.org/C104267543","wikidata":"https://www.wikidata.org/wiki/Q208163","display_name":"Signal processing","level":3,"score":0.2547999918460846},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.25220000743865967},{"id":"https://openalex.org/C2984822820","wikidata":"https://www.wikidata.org/wiki/Q1123036","display_name":"Processor scheduling","level":3,"score":0.2515999972820282}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/access.2026.3725140","doi":"https://doi.org/10.1109/access.2026.3725140","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2026.3725140","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"is_core":true,"listed_in":["anvur-scientific","cwts-core","doaj","jufo-1","ki-jl-1","norway-1","oarl","poland-100","russia-white-list-1","vabb-shw"],"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":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Access","raw_type":"journal-article"},{"id":"pmh:oai:doaj.org/article:ceccb49c7d2d42d7870bb5b9b78091ba","doi":null,"is_oa":false,"landing_page_url":"https://doaj.org/article/ceccb49c7d2d42d7870bb5b9b78091ba","pdf_url":null,"source":{"id":"https://openalex.org/S4306401280","display_name":"DOAJ (DOAJ: Directory of Open Access Journals)","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":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"IEEE Access, Vol 14, Pp 132028-132044 (2026)","raw_type":"article"}],"best_oa_location":{"id":"doi:10.1109/access.2026.3725140","doi":"https://doi.org/10.1109/access.2026.3725140","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2026.3725140","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"is_core":true,"listed_in":["anvur-scientific","cwts-core","doaj","jufo-1","ki-jl-1","norway-1","oarl","poland-100","russia-white-list-1","vabb-shw"],"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":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Access","raw_type":"journal-article"},"sustainable_development_goals":[{"display_name":"Affordable and clean energy","id":"https://metadata.un.org/sdg/7","score":0.9853000044822693}],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":35,"referenced_works":["https://openalex.org/W795794625","https://openalex.org/W1483717008","https://openalex.org/W1489811591","https://openalex.org/W1972270941","https://openalex.org/W1973923004","https://openalex.org/W1974573565","https://openalex.org/W1987232633","https://openalex.org/W1988888548","https://openalex.org/W2007339694","https://openalex.org/W2018198861","https://openalex.org/W2058892023","https://openalex.org/W2101366119","https://openalex.org/W2117449401","https://openalex.org/W2118046912","https://openalex.org/W2130388542","https://openalex.org/W2139469116","https://openalex.org/W2140179600","https://openalex.org/W2148459868","https://openalex.org/W2405627360","https://openalex.org/W2518008961","https://openalex.org/W2519627041","https://openalex.org/W2522697730","https://openalex.org/W2544171875","https://openalex.org/W2552249601","https://openalex.org/W2733081033","https://openalex.org/W2748740248","https://openalex.org/W2751498991","https://openalex.org/W2948541941","https://openalex.org/W3090575287","https://openalex.org/W4294589493","https://openalex.org/W4327808413","https://openalex.org/W4387494819","https://openalex.org/W4391962614","https://openalex.org/W4399563636","https://openalex.org/W7140320753"],"related_works":["https://openalex.org/W7140320753","https://openalex.org/W7166283903","https://openalex.org/W6891884944","https://openalex.org/W3003893379","https://openalex.org/W7203470452","https://openalex.org/W4391451805","https://openalex.org/W3205418814","https://openalex.org/W3159436656","https://openalex.org/W4413109816","https://openalex.org/W4415443977"],"abstract_inverted_index":{"High-performance":[0],"heterogeneous":[1,79,167],"computing":[2,80],"has":[3],"emerged":[4],"as":[5,104,137],"a":[6,90,125,192],"key":[7],"resource":[8,28],"for":[9,52,63],"training":[10],"increasingly":[11],"Deep":[12],"Neural":[13],"Networks":[14],"(DNNs).":[15],"Achieving":[16],"optimal":[17],"performance":[18,95,199],"and":[19,33,60,74,96,111,114,133,173,184,200,211],"energy":[20,75,98,201,231],"efficiency":[21,232],"on":[22,43,77,165],"such":[23],"platforms,":[24,54],"however,":[25],"requires":[26],"careful":[27],"optimisation":[29,39,66,86,144],"at":[30],"both":[31,58,198],"application":[32],"platform":[34],"levels.":[35],"While":[36],"state-of-the-art":[37],"DNN":[38,68,101],"approaches":[40],"typically":[41],"focus":[42],"only":[44,191],"one":[45],"of":[46,67,100,106,109,157,197,221],"these":[47],"dimensions":[48],"or":[49],"are":[50,121],"designed":[51],"homogeneous":[53,212],"this":[55],"paper":[56],"considers":[57],"application-":[59,183],"platform-level":[61,185],"configurations":[62,186],"the":[64,84,94,107,142,155,222],"bi-objective":[65,85,143],"training,":[69],"to":[70,92,140],"minimise":[71],"execution":[72,149],"time":[73],"consumption":[76,151],"single-node":[78],"platforms.":[81],"After":[82],"formulating":[83],"problem,":[87],"we":[88],"propose":[89],"methodology":[91],"model":[93],"dynamic":[97],"profiles":[99],"applications,":[102],"expressed":[103],"functions":[105],"number":[108],"batches":[110],"Dynamic":[112],"Voltage":[113],"Frequency":[115],"Scaling":[116],"(DVFS)":[117],"states.":[118],"These":[119],"models":[120],"then":[122],"exploited":[123],"by":[124],"model-based":[126],"algorithm":[127,147],"that":[128,179,189,225],"employs":[129],"batch":[130],"distribution":[131,213],"(application-level)":[132],"DVFS":[134],"states":[135],"(platform-level)":[136],"decision":[138],"variables":[139],"solve":[141],"problem.":[145],"The":[146,160],"produces":[148],"time-energy":[150],"trade-off":[152],"solutions":[153,180],"in":[154,195,230,234],"form":[156],"Pareto":[158],"fronts.":[159],"proposed":[161],"approach":[162,206],"is":[163],"evaluated":[164],"two":[166],"hybrid":[168],"servers":[169],"comprising":[170],"multicore":[171],"CPUs":[172],"many-core":[174],"GPUs.":[175],"Experimental":[176],"results":[177,223],"demonstrate":[178],"incorporating":[181],"integrated":[182],"outperform":[187],"those":[188],"consider":[190],"single":[193],"dimension,":[194],"terms":[196],"efficiency.":[202],"Furthermore,":[203],"comparing":[204],"our":[205],"with":[207],"TensorFlow":[208],"data":[209],"parallelism":[210],"strategies":[214],"demonstrates":[215],"its":[216],"superiority.":[217],"A":[218],"comprehensive":[219],"analysis":[220],"reveals":[224],"DNNs":[226],"achieve":[227],"greater":[228],"improvements":[229],"than":[233],"performance.":[235]},"counts_by_year":[],"updated_date":"2026-11-14T11:04:58.165337","created_date":"2026-08-19T00:00:00"}