{"id":"https://openalex.org/W4415250928","doi":"https://doi.org/10.1109/hpec67600.2025.11196674","title":"Performance-Energy Characterization of ML Inference on Heterogeneous Edge AI Platforms","display_name":"Performance-Energy Characterization of ML Inference on Heterogeneous Edge AI Platforms","publication_year":2025,"publication_date":"2025-09-15","ids":{"openalex":"https://openalex.org/W4415250928","doi":"https://doi.org/10.1109/hpec67600.2025.11196674"},"language":"en","primary_location":{"id":"doi:10.1109/hpec67600.2025.11196674","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpec67600.2025.11196674","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE High Performance Extreme Computing Conference (HPEC)","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":null,"display_name":"Palash Kohli","orcid":null},"institutions":[{"id":"https://openalex.org/I74796645","display_name":"Birla Institute of Technology and Science, Pilani","ror":"https://ror.org/001p3jz28","country_code":"IN","type":"education","lineage":["https://openalex.org/I74796645"]}],"countries":["IN"],"is_corresponding":true,"raw_author_name":"Palash Kohli","raw_affiliation_strings":["BITS Pilani,Pilani,India"],"affiliations":[{"raw_affiliation_string":"BITS Pilani,Pilani,India","institution_ids":["https://openalex.org/I74796645"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Rakshith Jayanth","orcid":null},"institutions":[{"id":"https://openalex.org/I1174212","display_name":"University of Southern California","ror":"https://ror.org/03taz7m60","country_code":"US","type":"education","lineage":["https://openalex.org/I1174212"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Rakshith Jayanth","raw_affiliation_strings":["University of Southern California,Los Angeles,USA"],"affiliations":[{"raw_affiliation_string":"University of Southern California,Los Angeles,USA","institution_ids":["https://openalex.org/I1174212"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103040030","display_name":"Neelesh Gupta","orcid":null},"institutions":[{"id":"https://openalex.org/I1174212","display_name":"University of Southern California","ror":"https://ror.org/03taz7m60","country_code":"US","type":"education","lineage":["https://openalex.org/I1174212"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Neelesh Gupta","raw_affiliation_strings":["University of Southern California,Los Angeles,USA"],"affiliations":[{"raw_affiliation_string":"University of Southern California,Los Angeles,USA","institution_ids":["https://openalex.org/I1174212"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5037580197","display_name":"Haoyang Fan","orcid":null},"institutions":[{"id":"https://openalex.org/I1174212","display_name":"University of Southern California","ror":"https://ror.org/03taz7m60","country_code":"US","type":"education","lineage":["https://openalex.org/I1174212"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Haoyang Fan","raw_affiliation_strings":["University of Southern California,Los Angeles,USA"],"affiliations":[{"raw_affiliation_string":"University of Southern California,Los Angeles,USA","institution_ids":["https://openalex.org/I1174212"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5033166029","display_name":"Viktor K. Prasanna","orcid":"https://orcid.org/0000-0002-1609-8589"},"institutions":[{"id":"https://openalex.org/I1174212","display_name":"University of Southern California","ror":"https://ror.org/03taz7m60","country_code":"US","type":"education","lineage":["https://openalex.org/I1174212"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Viktor Prasanna","raw_affiliation_strings":["University of Southern California,Los Angeles,USA"],"affiliations":[{"raw_affiliation_string":"University of Southern California,Los Angeles,USA","institution_ids":["https://openalex.org/I1174212"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":5,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I74796645"],"apc_list":null,"apc_paid":null,"fwci":2.326,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":{"value":0.90455957,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":91,"max":98},"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/T10036","display_name":"Advanced Neural Network Applications","score":0.715399980545044,"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"}},"topics":[{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.715399980545044,"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/T10320","display_name":"Neural Networks and Applications","score":0.6295999884605408,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"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/T10273","display_name":"IoT and Edge/Fog Computing","score":0.6074000000953674,"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/speedup","display_name":"Speedup","score":0.9277999997138977},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.8019999861717224},{"id":"https://openalex.org/keywords/multi-core-processor","display_name":"Multi-core processor","score":0.6380000114440918},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.5741999745368958},{"id":"https://openalex.org/keywords/symmetric-multiprocessor-system","display_name":"Symmetric multiprocessor system","score":0.5699999928474426},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.517799973487854},{"id":"https://openalex.org/keywords/edge-computing","display_name":"Edge computing","score":0.47279998660087585}],"concepts":[{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.9277999997138977},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.8019999861717224},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.795799970626831},{"id":"https://openalex.org/C78766204","wikidata":"https://www.wikidata.org/wiki/Q555032","display_name":"Multi-core processor","level":2,"score":0.6380000114440918},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.5741999745368958},{"id":"https://openalex.org/C172430144","wikidata":"https://www.wikidata.org/wiki/Q17111997","display_name":"Symmetric multiprocessor system","level":2,"score":0.5699999928474426},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.517799973487854},{"id":"https://openalex.org/C2778456923","wikidata":"https://www.wikidata.org/wiki/Q5337692","display_name":"Edge computing","level":3,"score":0.47279998660087585},{"id":"https://openalex.org/C26713055","wikidata":"https://www.wikidata.org/wiki/Q245962","display_name":"Implementation","level":2,"score":0.4702000021934509},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.46380001306533813},{"id":"https://openalex.org/C50630238","wikidata":"https://www.wikidata.org/wiki/Q971505","display_name":"General-purpose computing on graphics processing units","level":3,"score":0.4036000072956085},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3799000084400177},{"id":"https://openalex.org/C2780165032","wikidata":"https://www.wikidata.org/wiki/Q16869822","display_name":"Energy consumption","level":2,"score":0.373199999332428},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3303000032901764},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.3125},{"id":"https://openalex.org/C83283714","wikidata":"https://www.wikidata.org/wiki/Q121117","display_name":"Supercomputer","level":2,"score":0.3107999861240387},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.29739999771118164},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.2919999957084656},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.2806999981403351},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.2727000117301941},{"id":"https://openalex.org/C112313634","wikidata":"https://www.wikidata.org/wiki/Q7886648","display_name":"Complement (music)","level":5,"score":0.27250000834465027},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.25999999046325684}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/hpec67600.2025.11196674","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpec67600.2025.11196674","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE High Performance Extreme Computing Conference (HPEC)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":25,"referenced_works":["https://openalex.org/W2964477674","https://openalex.org/W3126236051","https://openalex.org/W3130875629","https://openalex.org/W3216299549","https://openalex.org/W3217045543","https://openalex.org/W4211165432","https://openalex.org/W4297097386","https://openalex.org/W4367147545","https://openalex.org/W4400489024","https://openalex.org/W4403905010","https://openalex.org/W4403961418","https://openalex.org/W4406152263","https://openalex.org/W4406412004","https://openalex.org/W4407156715","https://openalex.org/W4408226143","https://openalex.org/W4409310868","https://openalex.org/W4409325865","https://openalex.org/W4409766004","https://openalex.org/W4410197190","https://openalex.org/W4410294862","https://openalex.org/W4411119478","https://openalex.org/W4411848355","https://openalex.org/W4412945272","https://openalex.org/W4413144996","https://openalex.org/W4414290345"],"related_works":[],"abstract_inverted_index":{"Edge":[0],"computing":[1],"enables":[2],"lower":[3],"latency":[4],"and":[5,26,65,83,105,122,183,211],"secure":[6],"real-time":[7],"processing":[8],"for":[9,42,117,136],"machine":[10],"learning":[11],"(ML)":[12],"workloads.":[13,48,73],"State-of-the-art":[14],"edge":[15],"platforms":[16],"are":[17],"heterogeneous":[18,45,56,107,216],"systems":[19,94],"with":[20,187],"compute":[21,40],"cores":[22,41],"of":[23,38,47,59,80,92,115,120,170,180,191,199],"diverse":[24],"performance":[25,69,77],"energy":[27,181],"capabilities.":[28],"This":[29],"diversity":[30],"motivates":[31],"the":[32,36,89,100,113,127,142,150,154,161,196],"need":[33],"to":[34,87,131,146,175,205],"identify":[35],"strengths":[37],"individual":[39],"more":[43],"informed":[44],"implementations":[46],"In":[49],"this":[50],"study,":[51],"we":[52],"benchmark":[53],"a":[54,60,106,157],"state-of-the-art":[55],"platform":[57],"consisting":[58],"multicore":[61],"CPU,":[62,101],"integrated":[63],"GPU":[64,102,155],"NPU,":[66],"on":[67,70,99,160,215],"its":[68],"various":[71,96],"ML":[72],"We":[74,124,164,194],"measure":[75],"system":[76],"in":[78,178,208],"terms":[79,179],"latency,":[81,186],"throughput,":[82],"Energy-Delay":[84],"Product":[85],"(EDP)":[86],"analyze":[88],"performance-energy":[90],"tradeoffs":[91],"such":[93],"across":[95],"configurations":[97],"-independently":[98],"or":[103],"NPU":[104,128,143],"NPU-GPU":[108,168],"configuration.":[109],"Our":[110],"results":[111,201],"indicate":[112],"superiority":[114],"NPUs":[116],"energy-efficient":[118],"inference":[119,172,177,207],"LLMs":[121,210],"CNNs.":[123],"find":[125],"that":[126,166],"offers":[129,156],"up":[130,145],"5.2\u00d7":[132],"speedup":[133,148,159],"over":[134],"CPU":[135],"CNN":[137],"inference.":[138],"For":[139],"LLM":[140,171],"inference,":[141],"achieves":[144],"4\u00d7":[147],"during":[149],"prefill":[151],"stage,":[152],"while":[153],"22\u00d7":[158],"decode":[162],"stage.":[163],"observe":[165],"hybrid":[167],"implementation":[169],"is":[173],"superior":[174],"CPU-only":[176],"consumption":[182],"end-to-end":[184],"(E2E)":[185],"an":[188],"EDP":[189],"decrease":[190],"about":[192],"93%.":[193],"highlight":[195],"broader":[197],"applicability":[198],"our":[200],"by":[202],"discussing":[203],"extensions":[204],"long-context":[206],"causal":[209],"emerging":[212],"state-space":[213],"models":[214],"platforms.":[217]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":1}],"updated_date":"2026-04-21T08:09:41.155169","created_date":"2025-10-16T00:00:00"}