{"id":"https://openalex.org/W4406858787","doi":"https://doi.org/10.1109/ipccc59868.2024.10850044","title":"Characterizing and Understanding the Performance of Small Language Models on Edge Devices","display_name":"Characterizing and Understanding the Performance of Small Language Models on Edge Devices","publication_year":2024,"publication_date":"2024-11-22","ids":{"openalex":"https://openalex.org/W4406858787","doi":"https://doi.org/10.1109/ipccc59868.2024.10850044"},"language":"en","primary_location":{"id":"doi:10.1109/ipccc59868.2024.10850044","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ipccc59868.2024.10850044","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 IEEE International Performance, Computing, and Communications Conference (IPCCC)","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/A5019806120","display_name":"Md Romyull Islam","orcid":"https://orcid.org/0000-0003-0550-2248"},"institutions":[{"id":"https://openalex.org/I172980758","display_name":"Kennesaw State University","ror":"https://ror.org/00jeqjx33","country_code":"US","type":"education","lineage":["https://openalex.org/I172980758"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Md Romyull Islam","raw_affiliation_strings":["Kennesaw State University,Department of Computer Science,GA,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Kennesaw State University,Department of Computer Science,GA,USA","institution_ids":["https://openalex.org/I172980758"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069953290","display_name":"Nobel Dhar","orcid":"https://orcid.org/0009-0000-5467-7082"},"institutions":[{"id":"https://openalex.org/I172980758","display_name":"Kennesaw State University","ror":"https://ror.org/00jeqjx33","country_code":"US","type":"education","lineage":["https://openalex.org/I172980758"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Nobel Dhar","raw_affiliation_strings":["Kennesaw State University,Department of Computer Science,GA,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Kennesaw State University,Department of Computer Science,GA,USA","institution_ids":["https://openalex.org/I172980758"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087218093","display_name":"Bobin Deng","orcid":"https://orcid.org/0000-0001-8361-9025"},"institutions":[{"id":"https://openalex.org/I172980758","display_name":"Kennesaw State University","ror":"https://ror.org/00jeqjx33","country_code":"US","type":"education","lineage":["https://openalex.org/I172980758"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Bobin Deng","raw_affiliation_strings":["Kennesaw State University,Department of Computer Science,GA,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Kennesaw State University,Department of Computer Science,GA,USA","institution_ids":["https://openalex.org/I172980758"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5066987043","display_name":"Tu N. Nguyen","orcid":"https://orcid.org/0000-0001-7184-4102"},"institutions":[{"id":"https://openalex.org/I172980758","display_name":"Kennesaw State University","ror":"https://ror.org/00jeqjx33","country_code":"US","type":"education","lineage":["https://openalex.org/I172980758"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Tu N. Nguyen","raw_affiliation_strings":["Kennesaw State University,Department of Computer Science,GA,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Kennesaw State University,Department of Computer Science,GA,USA","institution_ids":["https://openalex.org/I172980758"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087326137","display_name":"Selena He","orcid":"https://orcid.org/0000-0002-2332-9816"},"institutions":[{"id":"https://openalex.org/I172980758","display_name":"Kennesaw State University","ror":"https://ror.org/00jeqjx33","country_code":"US","type":"education","lineage":["https://openalex.org/I172980758"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Selena He","raw_affiliation_strings":["Kennesaw State University,Department of Computer Science,GA,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Kennesaw State University,Department of Computer Science,GA,USA","institution_ids":["https://openalex.org/I172980758"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5054721271","display_name":"Kun Suo","orcid":"https://orcid.org/0000-0001-8562-0492"},"institutions":[{"id":"https://openalex.org/I172980758","display_name":"Kennesaw State University","ror":"https://ror.org/00jeqjx33","country_code":"US","type":"education","lineage":["https://openalex.org/I172980758"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kun Suo","raw_affiliation_strings":["Kennesaw State University,Department of Computer Science,GA,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Kennesaw State University,Department of Computer Science,GA,USA","institution_ids":["https://openalex.org/I172980758"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I172980758"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":12,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"10"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.7358999848365784,"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"}},"topics":[{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.7358999848365784,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6867755651473999},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.5781318545341492},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.2224082350730896}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6867755651473999},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.5781318545341492},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2224082350730896}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/ipccc59868.2024.10850044","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ipccc59868.2024.10850044","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 IEEE International Performance, Computing, and Communications Conference (IPCCC)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Quality Education","score":0.6200000047683716,"id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":14,"referenced_works":["https://openalex.org/W2998309782","https://openalex.org/W3210365128","https://openalex.org/W4312868944","https://openalex.org/W4318777262","https://openalex.org/W4390659326","https://openalex.org/W4392943591","https://openalex.org/W4394725454","https://openalex.org/W4395704222","https://openalex.org/W6727099177","https://openalex.org/W6755843862","https://openalex.org/W6768086466","https://openalex.org/W6793442341","https://openalex.org/W6847478871","https://openalex.org/W6867310301"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"In":[0],"recent":[1],"years,":[2],"significant":[3],"advancements":[4],"in":[5,69,105,188],"computing":[6,37,236],"power,":[7],"data":[8],"richness,":[9],"algorithmic":[10],"development,":[11],"and":[12,23,47,55,65,77,102,132,159,171,190,201,214,220,230],"the":[13,20,53,70,72,98,109,166,195,218,223,232],"growing":[14],"demand":[15],"for":[16],"applications":[17],"have":[18,50,91],"catalyzed":[19],"rapid":[21],"emergence":[22],"proliferation":[24],"of":[25,57,74,108,128,198,222,234],"large":[26],"language":[27,59],"models":[28,60,200,215],"(LLMs)":[29],"across":[30],"various":[31],"scenarios.":[32],"Concurrently,":[33],"factors":[34],"such":[35,152],"as":[36,153],"resource":[38,79,103,191],"limitations,":[39],"cost":[40],"considerations,":[41],"real-time":[42],"application":[43],"requirements,":[44],"task-specific":[45],"customization,":[46],"privacy":[48],"concerns":[49],"also":[51],"driven":[52],"development":[54],"deployment":[56],"small":[58],"(SLMs).":[61],"Unlike":[62],"extensively":[63],"researched":[64],"widely":[66,140],"deployed":[67],"LLMs":[68],"cloud,":[71],"performance":[73,189,228],"SLM":[75],"workloads":[76,178],"their":[78,202],"impact":[80,175],"on":[81,120,148,204,217],"edge":[82,121,130,150,176,205,224,235],"environments":[83],"remain":[84],"poorly":[85],"understood.":[86],"More":[87],"detailed":[88],"studies":[89],"will":[90],"to":[92,96,226],"be":[93],"carried":[94],"out":[95],"understand":[97],"advantages,":[99],"constraints,":[100],"performances,":[101],"consumption":[104],"different":[106,169],"settings":[107],"edge.This":[110],"paper":[111],"addresses":[112],"this":[113],"gap":[114],"by":[115],"comprehensively":[116],"analyzing":[117],"representative":[118],"SLMs":[119,172],"platforms.":[122,206],"Initially,":[123],"we":[124,136],"provide":[125],"a":[126],"summary":[127],"contemporary":[129],"hardware":[131,170,213],"popular":[133,149],"SLMs.":[134],"Subsequently,":[135],"quantitatively":[137],"evaluate":[138],"several":[139],"used":[141],"SLMs,":[142],"including":[143],"TinyLlama,":[144],"Phi-3,":[145],"Llama-3,":[146],"etc.,":[147],"platforms":[151],"Raspberry":[154],"Pi,":[155],"Nvidia":[156],"Jetson":[157],"Orin,":[158],"Mac":[160],"mini.":[161],"Our":[162,183],"findings":[163],"reveal":[164],"that":[165,186],"interaction":[167],"between":[168],"can":[173],"significantly":[174],"AI":[177],"while":[179],"introducing":[180],"non-negligible":[181],"overhead.":[182],"experiments":[184],"demonstrate":[185],"variations":[187],"usage":[192],"might":[193],"constrain":[194],"workload":[196],"capabilities":[197],"specific":[199],"feasibility":[203],"Therefore,":[207],"users":[208],"must":[209],"judiciously":[210],"match":[211],"appropriate":[212],"based":[216],"requirements":[219],"characteristics":[221],"environment":[225],"avoid":[227],"bottlenecks":[229],"optimize":[231],"utility":[233],"capabilities.":[237]},"counts_by_year":[{"year":2026,"cited_by_count":7},{"year":2025,"cited_by_count":5}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}