{"id":"https://openalex.org/W7133534962","doi":"https://doi.org/10.1109/hpca68181.2026.11408569","title":"The Cost of Dynamic Reasoning: Demystifying AI Agents and Test-Time Scaling from an AI Infrastructure Perspective","display_name":"The Cost of Dynamic Reasoning: Demystifying AI Agents and Test-Time Scaling from an AI Infrastructure Perspective","publication_year":2026,"publication_date":"2026-01-31","ids":{"openalex":"https://openalex.org/W7133534962","doi":"https://doi.org/10.1109/hpca68181.2026.11408569"},"language":null,"primary_location":{"id":"doi:10.1109/hpca68181.2026.11408569","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpca68181.2026.11408569","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 IEEE International Symposium on High Performance Computer Architecture (HPCA)","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":"https://openalex.org/A5041027404","display_name":"Junsu Kim","orcid":"https://orcid.org/0000-0001-9583-7978"},"institutions":[{"id":"https://openalex.org/I157485424","display_name":"Korea Advanced Institute of Science and Technology","ror":"https://ror.org/05apxxy63","country_code":"KR","type":"education","lineage":["https://openalex.org/I157485424"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jiin Kim","raw_affiliation_strings":["KAIST,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"KAIST,South Korea","institution_ids":["https://openalex.org/I157485424"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128108727","display_name":"Byeongjun Shin","orcid":null},"institutions":[{"id":"https://openalex.org/I157485424","display_name":"Korea Advanced Institute of Science and Technology","ror":"https://ror.org/05apxxy63","country_code":"KR","type":"education","lineage":["https://openalex.org/I157485424"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Byeongjun Shin","raw_affiliation_strings":["KAIST,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"KAIST,South Korea","institution_ids":["https://openalex.org/I157485424"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5055063660","display_name":"Jinha Chung","orcid":"https://orcid.org/0000-0002-0212-714X"},"institutions":[{"id":"https://openalex.org/I157485424","display_name":"Korea Advanced Institute of Science and Technology","ror":"https://ror.org/05apxxy63","country_code":"KR","type":"education","lineage":["https://openalex.org/I157485424"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jinha Chung","raw_affiliation_strings":["KAIST,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"KAIST,South Korea","institution_ids":["https://openalex.org/I157485424"]}]},{"author_position":"last","author":{"id":null,"display_name":"Minsoo Rhu","orcid":null},"institutions":[{"id":"https://openalex.org/I157485424","display_name":"Korea Advanced Institute of Science and Technology","ror":"https://ror.org/05apxxy63","country_code":"KR","type":"education","lineage":["https://openalex.org/I157485424"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Minsoo Rhu","raw_affiliation_strings":["KAIST,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"KAIST,South Korea","institution_ids":["https://openalex.org/I157485424"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":18.5949,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.98395183,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":95,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"16"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10906","display_name":"AI-based Problem Solving and Planning","score":0.12330000102519989,"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/T10906","display_name":"AI-based Problem Solving and Planning","score":0.12330000102519989,"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/T10456","display_name":"Multi-Agent Systems and Negotiation","score":0.10610000044107437,"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/T11010","display_name":"Logic, Reasoning, and Knowledge","score":0.09799999743700027,"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/perspective","display_name":"Perspective (graphical)","score":0.6256999969482422},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.3310000002384186},{"id":"https://openalex.org/keywords/scale","display_name":"Scale (ratio)","score":0.3249000012874603},{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.3075999915599823},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.27399998903274536}],"concepts":[{"id":"https://openalex.org/C12713177","wikidata":"https://www.wikidata.org/wiki/Q1900281","display_name":"Perspective (graphical)","level":2,"score":0.6256999969482422},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.554099977016449},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.35530000925064087},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3310000002384186},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.3249000012874603},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.3075999915599823},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.3070000112056732},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.27399998903274536},{"id":"https://openalex.org/C58328972","wikidata":"https://www.wikidata.org/wiki/Q184609","display_name":"Expert system","level":2,"score":0.2646999955177307},{"id":"https://openalex.org/C539667460","wikidata":"https://www.wikidata.org/wiki/Q2414942","display_name":"Management science","level":1,"score":0.257099986076355}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/hpca68181.2026.11408569","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpca68181.2026.11408569","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 IEEE International Symposium on High Performance Computer Architecture (HPCA)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Industry, innovation and infrastructure","score":0.6299622654914856,"id":"https://metadata.un.org/sdg/9"}],"awards":[{"id":"https://openalex.org/G1623312020","display_name":null,"funder_award_id":"RS-202400438851","funder_id":"https://openalex.org/F4320328359","funder_display_name":"Ministry of Science and ICT, South Korea"},{"id":"https://openalex.org/G8076470811","display_name":null,"funder_award_id":"SRFC-IT2402-03","funder_id":"https://openalex.org/F4320332195","funder_display_name":"Samsung"}],"funders":[{"id":"https://openalex.org/F4320328359","display_name":"Ministry of Science and ICT, South Korea","ror":"https://ror.org/01wpjm123"},{"id":"https://openalex.org/F4320332195","display_name":"Samsung","ror":"https://ror.org/04w3jy968"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":24,"referenced_works":["https://openalex.org/W1714211023","https://openalex.org/W2407386500","https://openalex.org/W2889787757","https://openalex.org/W3175111539","https://openalex.org/W4366564134","https://openalex.org/W4387321091","https://openalex.org/W4387494819","https://openalex.org/W4389518760","https://openalex.org/W4389520103","https://openalex.org/W4393160302","https://openalex.org/W4394945037","https://openalex.org/W4401176373","https://openalex.org/W4401211704","https://openalex.org/W4404400709","https://openalex.org/W4408848702","https://openalex.org/W4408893361","https://openalex.org/W4408901326","https://openalex.org/W4411092043","https://openalex.org/W4412944812","https://openalex.org/W4415797413","https://openalex.org/W4415800514","https://openalex.org/W7103754027","https://openalex.org/W7133209886","https://openalex.org/W7133222210"],"related_works":[],"abstract_inverted_index":{"Large-language-model":[0],"(LLM)-based":[1],"AI":[2,59,86,140],"agents":[3,107],"have":[4],"recently":[5],"showcased":[6],"impressive":[7],"versatility":[8],"by":[9,139],"employing":[10],"dynamic":[11],"reasoning,":[12,98,160],"an":[13],"adaptive,":[14],"multi-step":[15],"process":[16],"that":[17,105],"coordinates":[18],"with":[19,110,163],"external":[20],"tools.":[21],"This":[22,50],"shift":[23,154],"from":[24,115],"static,":[25],"single-turn":[26],"inference":[27],"to":[28],"agentic,":[29],"multi-turn":[30],"workflows":[31],"broadens":[32],"task":[33],"generalization":[34],"and":[35,48,69,78,96,122],"behavioral":[36],"flexibility,":[37],"but":[38],"it":[39],"also":[40],"introduces":[41],"serious":[42],"concerns":[43],"about":[44],"system-level":[45,56],"cost,":[46],"efficiency,":[47],"sustainability.":[49],"paper":[51],"presents":[52],"the":[53,134],"first":[54],"comprehensive":[55],"analysis":[57],"of":[58,129],"agents,":[60,131],"quantifying":[61],"their":[62],"resource":[63],"usage,":[64],"latency":[65,120],"behavior,":[66],"energy":[67],"consumption,":[68],"datacenter-wide":[70],"power":[71],"consumption":[72],"demands":[73,137],"across":[74],"diverse":[75],"agent":[76,87,141,156],"designs":[77],"test-time":[79],"scaling":[80],"strategies.":[81],"We":[82],"further":[83],"characterize":[84],"how":[85],"design":[88,157],"choices,":[89],"such":[90],"as":[91],"few-shot":[92],"prompting,":[93],"reflection":[94],"depth,":[95],"parallel":[97],"impact":[99],"accuracy-cost":[100],"tradeoffs.":[101],"Our":[102],"findings":[103],"reveal":[104],"while":[106],"improve":[108],"accuracy":[109],"increased":[111],"compute,":[112],"they":[113],"suffer":[114],"rapidly":[116],"diminishing":[117],"returns,":[118],"widening":[119],"variance,":[121],"unsustainable":[123],"infrastructure":[124],"costs.":[125],"Through":[126],"detailed":[127],"evaluation":[128],"representative":[130],"we":[132],"highlight":[133],"profound":[135],"computational":[136],"introduced":[138],"workflows,":[142],"uncovering":[143],"a":[144,152],"looming":[145],"sustainability":[146],"crisis.":[147],"These":[148],"results":[149],"call":[150],"for":[151],"paradigm":[153],"in":[155],"toward":[158],"compute-efficient":[159],"balancing":[161],"performance":[162],"deployability":[164],"under":[165],"real-world":[166],"constraints.":[167]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-06-16T09:24:06.705377","created_date":"2026-03-05T00:00:00"}