{"id":"https://openalex.org/W7117577863","doi":"https://doi.org/10.1109/iccd65941.2025.00084","title":"Flame: A Multiplier-Free LLM Accelerator with Dynamic Block Floating Point","display_name":"Flame: A Multiplier-Free LLM Accelerator with Dynamic Block Floating Point","publication_year":2025,"publication_date":"2025-11-10","ids":{"openalex":"https://openalex.org/W7117577863","doi":"https://doi.org/10.1109/iccd65941.2025.00084"},"language":null,"primary_location":{"id":"doi:10.1109/iccd65941.2025.00084","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccd65941.2025.00084","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE 43rd International Conference on Computer Design (ICCD)","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/A5117107477","display_name":"Ao Lyu","orcid":null},"institutions":[{"id":"https://openalex.org/I4210090176","display_name":"Institute of Computing Technology","ror":"https://ror.org/0090r4d87","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210090176"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Ao Lyu","raw_affiliation_strings":["Institute of Computing Technology, Chinese Academy of Sciences,State Key Laboratory of Processors"],"affiliations":[{"raw_affiliation_string":"Institute of Computing Technology, Chinese Academy of Sciences,State Key Laboratory of Processors","institution_ids":["https://openalex.org/I4210090176"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5009432381","display_name":"Haishuang Fan","orcid":"https://orcid.org/0000-0002-4633-4044"},"institutions":[{"id":"https://openalex.org/I4210090176","display_name":"Institute of Computing Technology","ror":"https://ror.org/0090r4d87","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210090176"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Haishuang Fan","raw_affiliation_strings":["Institute of Computing Technology, Chinese Academy of Sciences,State Key Laboratory of Processors"],"affiliations":[{"raw_affiliation_string":"Institute of Computing Technology, Chinese Academy of Sciences,State Key Laboratory of Processors","institution_ids":["https://openalex.org/I4210090176"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5102284748","display_name":"Guihai Yan","orcid":null},"institutions":[{"id":"https://openalex.org/I4210090176","display_name":"Institute of Computing Technology","ror":"https://ror.org/0090r4d87","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210090176"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Guihai Yan","raw_affiliation_strings":["Institute of Computing Technology, Chinese Academy of Sciences,State Key Laboratory of Processors"],"affiliations":[{"raw_affiliation_string":"Institute of Computing Technology, Chinese Academy of Sciences,State Key Laboratory of Processors","institution_ids":["https://openalex.org/I4210090176"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5117107477"],"corresponding_institution_ids":["https://openalex.org/I4210090176"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.81662726,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"549","last_page":"557"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.13940000534057617,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.13940000534057617,"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/T10028","display_name":"Topic Modeling","score":0.10040000081062317,"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/T12131","display_name":"Wireless Signal Modulation Classification","score":0.09690000116825104,"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/speedup","display_name":"Speedup","score":0.7282999753952026},{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.570900022983551},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5648000240325928},{"id":"https://openalex.org/keywords/floating-point","display_name":"Floating point","score":0.5232999920845032},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.4334999918937683},{"id":"https://openalex.org/keywords/block","display_name":"Block (permutation group theory)","score":0.40149998664855957},{"id":"https://openalex.org/keywords/checksum","display_name":"Checksum","score":0.3917999863624573},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.36970001459121704}],"concepts":[{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.7282999753952026},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6996999979019165},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.570900022983551},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5648000240325928},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5421000123023987},{"id":"https://openalex.org/C84211073","wikidata":"https://www.wikidata.org/wiki/Q117879","display_name":"Floating point","level":2,"score":0.5232999920845032},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.4334999918937683},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.40149998664855957},{"id":"https://openalex.org/C162372511","wikidata":"https://www.wikidata.org/wiki/Q218341","display_name":"Checksum","level":2,"score":0.3917999863624573},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.38749998807907104},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.36970001459121704},{"id":"https://openalex.org/C3826847","wikidata":"https://www.wikidata.org/wiki/Q188768","display_name":"FLOPS","level":2,"score":0.3416999876499176},{"id":"https://openalex.org/C2777472644","wikidata":"https://www.wikidata.org/wiki/Q16968992","display_name":"Approximate inference","level":3,"score":0.33340001106262207},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.3321000039577484},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.31529998779296875},{"id":"https://openalex.org/C2742236","wikidata":"https://www.wikidata.org/wiki/Q924713","display_name":"Efficient energy use","level":2,"score":0.30640000104904175},{"id":"https://openalex.org/C3770464","wikidata":"https://www.wikidata.org/wiki/Q775963","display_name":"Smoothing","level":2,"score":0.2953999936580658},{"id":"https://openalex.org/C35912277","wikidata":"https://www.wikidata.org/wiki/Q1243369","display_name":"Double-precision floating-point format","level":3,"score":0.29490000009536743},{"id":"https://openalex.org/C186370098","wikidata":"https://www.wikidata.org/wiki/Q442787","display_name":"Energy (signal processing)","level":2,"score":0.2863999903202057},{"id":"https://openalex.org/C459310","wikidata":"https://www.wikidata.org/wiki/Q117801","display_name":"Computational science","level":1,"score":0.27630001306533813},{"id":"https://openalex.org/C84462506","wikidata":"https://www.wikidata.org/wiki/Q173142","display_name":"Digital signal processing","level":2,"score":0.2727999985218048},{"id":"https://openalex.org/C13164978","wikidata":"https://www.wikidata.org/wiki/Q600158","display_name":"Hardware acceleration","level":3,"score":0.271699994802475},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.2563999891281128},{"id":"https://openalex.org/C155281189","wikidata":"https://www.wikidata.org/wiki/Q3518150","display_name":"Tensor (intrinsic definition)","level":2,"score":0.25360000133514404}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/iccd65941.2025.00084","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccd65941.2025.00084","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE 43rd International Conference on Computer Design (ICCD)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.9109507203102112,"id":"https://metadata.un.org/sdg/7","display_name":"Affordable and clean energy"}],"awards":[{"id":"https://openalex.org/G3969026231","display_name":null,"funder_award_id":"Y201923","funder_id":"https://openalex.org/F4320322847","funder_display_name":"Youth Innovation Promotion Association of the Chinese Academy of Sciences"},{"id":"https://openalex.org/G6031024966","display_name":null,"funder_award_id":"62002340,62090020","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320322847","display_name":"Youth Innovation Promotion Association of the Chinese Academy of Sciences","ror":"https://ror.org/031141b54"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":11,"referenced_works":["https://openalex.org/W2093647425","https://openalex.org/W2606722458","https://openalex.org/W2725159389","https://openalex.org/W2946355854","https://openalex.org/W2963367920","https://openalex.org/W3003537648","https://openalex.org/W3190062760","https://openalex.org/W3208633927","https://openalex.org/W4221149185","https://openalex.org/W4385245566","https://openalex.org/W4393407021"],"related_works":[],"abstract_inverted_index":{"Recently,":[0],"large":[1],"language":[2],"models":[3],"(LLMs)":[4],"have":[5],"achieved":[6],"remarkable":[7],"success":[8],"across":[9],"various":[10],"machine":[11],"learning":[12],"tasks.":[13],"However,":[14],"deploying":[15],"LLM":[16,44],"inference":[17,45,70],"still":[18],"remains":[19],"challenging":[20],"due":[21],"to":[22,60,65,105,122,136],"the":[23,88],"high":[24],"computing":[25],"overhead":[26],"and":[27,69,128],"substantial":[28],"memory":[29],"footprint.":[30],"In":[31],"this":[32],"paper,":[33],"we":[34,80,96],"propose":[35,81],"Flame,":[36],"a":[37,54,82,98],"software-hardware":[38],"co-design":[39],"accelerator":[40,110],"that":[41,86,118],"enables":[42],"efficient":[43,108],"with":[46],"dynamic":[47],"block":[48],"floating":[49],"point":[50],"(BFP).":[51],"Flame":[52,119],"employs":[53],"layer-wise":[55],"adaptive":[56],"precision":[57],"search":[58],"algorithm":[59,104],"optimize":[61],"BFP":[62,78],"mantissa":[63],"bitwidth":[64],"balance":[66],"model":[67,141],"accuracy":[68,74],"speedup.":[71],"To":[72],"mitigate":[73],"degradation":[75],"introduced":[76],"by":[77],"conversion,":[79],"channel":[83],"reorder":[84],"approach":[85],"adjusts":[87],"value":[89],"distribution":[90],"within":[91],"each":[92],"tensor":[93],"group.":[94],"Finally,":[95],"leverage":[97],"novel":[99],"hardware-friendly":[100],"linear":[101],"complexity":[102],"multiplication":[103],"implement":[106],"an":[107],"hardware":[109],"featuring":[111],"multiplierfree":[112],"processing":[113],"units.":[114],"Evaluation":[115],"results":[116],"show":[117],"achieves":[120],"up":[121],"<tex":[123,129],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[124,130],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">$6.32":[125],"\\times$</tex>":[126,132],"speedup":[127],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">$4.04":[131],"energy":[133],"efficiency":[134],"compared":[135],"GPU,":[137],"while":[138],"maintaining":[139],"superior":[140],"accuracy.":[142]},"counts_by_year":[],"updated_date":"2026-04-09T08:11:56.329763","created_date":"2025-12-30T00:00:00"}