{"id":"https://openalex.org/W7147699467","doi":"https://doi.org/10.1109/icaiic68212.2026.11454159","title":"Beyond System Calls: Uncovering Hidden I/O Behavior of Mmap-Based On-Device LLMs","display_name":"Beyond System Calls: Uncovering Hidden I/O Behavior of Mmap-Based On-Device LLMs","publication_year":2026,"publication_date":"2026-02-24","ids":{"openalex":"https://openalex.org/W7147699467","doi":"https://doi.org/10.1109/icaiic68212.2026.11454159"},"language":null,"primary_location":{"id":"doi:10.1109/icaiic68212.2026.11454159","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icaiic68212.2026.11454159","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","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/A5100734358","display_name":"Heejin Kim","orcid":"https://orcid.org/0000-0002-3486-3571"},"institutions":[{"id":"https://openalex.org/I138925566","display_name":"Ewha Womans University","ror":"https://ror.org/053fp5c05","country_code":"KR","type":"education","lineage":["https://openalex.org/I138925566"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Heejin Kim","raw_affiliation_strings":["Ewha Womans University,Department of Computer Engineering,Seoul,Republic of Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Ewha Womans University,Department of Computer Engineering,Seoul,Republic of Korea","institution_ids":["https://openalex.org/I138925566"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Jeongha Lee","orcid":null},"institutions":[{"id":"https://openalex.org/I138925566","display_name":"Ewha Womans University","ror":"https://ror.org/053fp5c05","country_code":"KR","type":"education","lineage":["https://openalex.org/I138925566"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jeongha Lee","raw_affiliation_strings":["Ewha Womans University,Department of Computer Engineering,Seoul,Republic of Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Ewha Womans University,Department of Computer Engineering,Seoul,Republic of Korea","institution_ids":["https://openalex.org/I138925566"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5132587137","display_name":"Hyokyung Bahn","orcid":null},"institutions":[{"id":"https://openalex.org/I138925566","display_name":"Ewha Womans University","ror":"https://ror.org/053fp5c05","country_code":"KR","type":"education","lineage":["https://openalex.org/I138925566"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Hyokyung Bahn","raw_affiliation_strings":["Ewha Womans University,Department of Computer Engineering,Seoul,Republic of Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Ewha Womans University,Department of Computer Engineering,Seoul,Republic of Korea","institution_ids":["https://openalex.org/I138925566"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I138925566"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"425","last_page":"430"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11181","display_name":"Advanced Data Storage Technologies","score":0.42500001192092896,"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"}},"topics":[{"id":"https://openalex.org/T11181","display_name":"Advanced Data Storage Technologies","score":0.42500001192092896,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.26179999113082886,"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/T12127","display_name":"Software System Performance and Reliability","score":0.06430000066757202,"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/software-deployment","display_name":"Software deployment","score":0.7135999798774719},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.6617000102996826},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5782999992370605},{"id":"https://openalex.org/keywords/tracing","display_name":"Tracing","score":0.44040000438690186},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.4207000136375427},{"id":"https://openalex.org/keywords/dram","display_name":"Dram","score":0.41449999809265137},{"id":"https://openalex.org/keywords/mobile-device","display_name":"Mobile device","score":0.4088999927043915}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.714900016784668},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.7135999798774719},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.6617000102996826},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5782999992370605},{"id":"https://openalex.org/C138673069","wikidata":"https://www.wikidata.org/wiki/Q322229","display_name":"Tracing","level":2,"score":0.44040000438690186},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.42739999294281006},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4207000136375427},{"id":"https://openalex.org/C7366592","wikidata":"https://www.wikidata.org/wiki/Q1255620","display_name":"Dram","level":2,"score":0.41449999809265137},{"id":"https://openalex.org/C186967261","wikidata":"https://www.wikidata.org/wiki/Q5082128","display_name":"Mobile device","level":2,"score":0.4088999927043915},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.31790000200271606},{"id":"https://openalex.org/C2778227907","wikidata":"https://www.wikidata.org/wiki/Q1893186","display_name":"Eviction","level":2,"score":0.3156000077724457},{"id":"https://openalex.org/C47487241","wikidata":"https://www.wikidata.org/wiki/Q5227230","display_name":"Data access","level":2,"score":0.2892000079154968},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.28369998931884766},{"id":"https://openalex.org/C193343404","wikidata":"https://www.wikidata.org/wiki/Q1928607","display_name":"Page fault","level":5,"score":0.28369998931884766},{"id":"https://openalex.org/C115051666","wikidata":"https://www.wikidata.org/wiki/Q6522493","display_name":"Ranging","level":2,"score":0.2793999910354614},{"id":"https://openalex.org/C2779662365","wikidata":"https://www.wikidata.org/wiki/Q5416694","display_name":"Event (particle physics)","level":2,"score":0.2750000059604645},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.26600000262260437}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icaiic68212.2026.11454159","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icaiic68212.2026.11454159","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G5283422203","display_name":null,"funder_award_id":"2024-00461678,2025-02214322","funder_id":"https://openalex.org/F4320322120","funder_display_name":"National Research Foundation of Korea"}],"funders":[{"id":"https://openalex.org/F4320322120","display_name":"National Research Foundation of Korea","ror":"https://ror.org/013aysd81"},{"id":"https://openalex.org/F4320328359","display_name":"Ministry of Science and ICT, South Korea","ror":"https://ror.org/01wpjm123"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":13,"referenced_works":["https://openalex.org/W4389401304","https://openalex.org/W4401379991","https://openalex.org/W4404955137","https://openalex.org/W4405717632","https://openalex.org/W4405794828","https://openalex.org/W4408146411","https://openalex.org/W4408147143","https://openalex.org/W4408529725","https://openalex.org/W4410738817","https://openalex.org/W4410794689","https://openalex.org/W4413014944","https://openalex.org/W4414736038","https://openalex.org/W4416309904"],"related_works":[],"abstract_inverted_index":{"The":[0,185],"growing":[1],"deployment":[2,193],"of":[3,15,104,140,155,168],"on-device":[4,79,134],"large":[5,205],"language":[6],"models":[7,206],"(LLMs)":[8],"on":[9,207],"smartphones":[10],"has":[11],"accelerated":[12],"the":[13,20,45,130,138,148,153,188],"adoption":[14],"memory-mapped":[16],"I/O":[17,156],"(mmap)":[18],"as":[19],"default":[21],"mechanism":[22],"for":[23,67,133,190],"loading":[24,36,65],"model":[25,69,145,169,192],"weights.":[26],"Unlike":[27],"traditional":[28],"system":[29],"call":[30],"based":[31],"access,":[32],"mmap":[33],"defers":[34],"data":[35],"to":[37,52,61,75,98,122,143,178,202],"page":[38,149,175],"faults,":[39],"making":[40],"storage":[41],"activity":[42],"invisible":[43],"at":[44,107],"system-call":[46],"level":[47],"and":[48,71,88,182,199],"requiring":[49],"kernel-level":[50],"tracing":[51],"understand":[53],"real":[54],"execution":[55,96],"behavior.":[56],"This":[57],"paper":[58],"uses":[59,72],"ftrace":[60],"collect":[62],"page-fault-driven":[63],"page-cache":[64],"events":[66],"mmap-based":[68],"access":[70],"this":[73],"approach":[74],"analyze":[76],"four":[77],"representative":[78],"LLM":[80,135],"workloads:":[81],"text":[82],"interaction,":[83],"vision-language":[84],"reasoning,":[85],"audio":[86],"transcription,":[87],"text-to-image":[89],"generation.":[90],"Our":[91],"measurements":[92],"show":[93],"that":[94,129],"mmapbased":[95],"continues":[97],"perform":[99],"large,":[100],"predominantly":[101],"sequential":[102],"scans":[103],"model-weight":[105],"files":[106],"each":[108],"configuration":[109],"or":[110],"inference":[111,160],"phase,":[112],"with":[113],"effective":[114],"working-set":[115],"sizes":[116],"ranging":[117],"from":[118],"approximately":[119],"500":[120],"MB":[121],"over":[123],"3":[124],"GB.":[125],"These":[126],"findings":[127],"indicate":[128],"primary":[131],"bottleneck":[132],"performance":[136,161],"is":[137],"amount":[139],"DRAM":[141],"available":[142],"maintain":[144],"pages":[146],"in":[147,172],"cache,":[150],"rather":[151],"than":[152],"choice":[154],"interface.":[157],"Ensuring":[158],"sustained":[159],"thus":[162],"requires":[163],"keeping":[164],"a":[165],"substantial":[166],"portion":[167],"weights":[170],"resident":[171],"memory.":[173],"Otherwise,":[174],"eviction":[176],"leads":[177],"repeated":[179],"major":[180],"faults":[181],"degraded":[183],"latency.":[184],"results":[186],"highlight":[187],"need":[189],"memory-aware":[191],"strategies,":[194],"OS-level":[195],"LLM-informed":[196],"caching":[197],"policies,":[198],"footprint-reduction":[200],"techniques":[201],"support":[203],"increasingly":[204],"resource-constrained":[208],"mobile":[209],"devices.":[210]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2026-04-02T00:00:00"}
