{"id":"https://openalex.org/W4413755221","doi":"https://doi.org/10.1109/isvlsi65124.2025.11130212","title":"Adaptive Multi-Precision Inference for Large-Scale AI Using IEEE P3109 FP8","display_name":"Adaptive Multi-Precision Inference for Large-Scale AI Using IEEE P3109 FP8","publication_year":2025,"publication_date":"2025-07-06","ids":{"openalex":"https://openalex.org/W4413755221","doi":"https://doi.org/10.1109/isvlsi65124.2025.11130212"},"language":"en","primary_location":{"id":"doi:10.1109/isvlsi65124.2025.11130212","is_oa":false,"landing_page_url":"https://doi.org/10.1109/isvlsi65124.2025.11130212","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","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/A5119133566","display_name":"Arnav Ramamoorthy","orcid":null},"institutions":[{"id":"https://openalex.org/I11912373","display_name":"Micron (United States)","ror":"https://ror.org/02fv52296","country_code":"US","type":"company","lineage":["https://openalex.org/I11912373"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Arnav Ramamoorthy","raw_affiliation_strings":["Micron Technology, Inc"],"affiliations":[{"raw_affiliation_string":"Micron Technology, Inc","institution_ids":["https://openalex.org/I11912373"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5081338858","display_name":"Kiran Gunnam","orcid":"https://orcid.org/0000-0001-7219-7234"},"institutions":[{"id":"https://openalex.org/I11912373","display_name":"Micron (United States)","ror":"https://ror.org/02fv52296","country_code":"US","type":"company","lineage":["https://openalex.org/I11912373"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kiran Kumar Gunnam","raw_affiliation_strings":["Micron Technology, Inc"],"affiliations":[{"raw_affiliation_string":"Micron Technology, Inc","institution_ids":["https://openalex.org/I11912373"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5119133566"],"corresponding_institution_ids":["https://openalex.org/I11912373"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.1299182,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"2"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10711","display_name":"Target Tracking and Data Fusion in Sensor Networks","score":0.9235000014305115,"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/T10711","display_name":"Target Tracking and Data Fusion in Sensor Networks","score":0.9235000014305115,"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/T10876","display_name":"Fault Detection and Control Systems","score":0.9132000207901001,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6872051954269409},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6270825266838074},{"id":"https://openalex.org/keywords/scale","display_name":"Scale (ratio)","score":0.5247911810874939},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.400879442691803},{"id":"https://openalex.org/keywords/physics","display_name":"Physics","score":0.060173988342285156}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6872051954269409},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6270825266838074},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.5247911810874939},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.400879442691803},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.060173988342285156},{"id":"https://openalex.org/C62520636","wikidata":"https://www.wikidata.org/wiki/Q944","display_name":"Quantum mechanics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/isvlsi65124.2025.11130212","is_oa":false,"landing_page_url":"https://doi.org/10.1109/isvlsi65124.2025.11130212","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"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":{"Accelerating":[0],"inference":[1,64],"for":[2,33,77,96],"large-scale":[3],"AI":[4],"models":[5,99],"remains":[6],"a":[7,50,111],"critical":[8],"challenge":[9],"in":[10,41],"both":[11],"research":[12],"and":[13,21,29,36],"industry.":[14],"Traditional":[15],"optimizations-static":[16],"quantization,":[17],"pruning,":[18],"knowledge":[19],"distillation,":[20],"parallelization-offer":[22],"substantial":[23],"improvements,":[24],"but":[25],"often":[26],"treat":[27],"precision":[28],"compression":[30],"as":[31],"fixed":[32],"all":[34,37],"layers":[35],"inputs.":[38],"Recent":[39],"advancements":[40],"floating-point":[42],"formats,":[43],"notably":[44],"IEEE":[45],"P3109":[46],"FP8":[47,70],"[4],":[48],"present":[49],"compelling":[51],"alternative":[52],"to":[53,101,123],"conventional":[54],"FP16":[55,125],"or":[56],"BF16.":[57],"This":[58],"paper":[59],"introduces":[60],"an":[61,89,128],"adaptive":[62],"multi-precision":[63],"strategy":[65],"that":[66],"leverages":[67],"the":[68,116],"new":[69],"format":[71],"with":[72,93],"dynamically":[73],"tuned":[74],"scaling":[75],"factors":[76],"each":[78],"tensor":[79],"based":[80],"on":[81,110],"real-time":[82],"range":[83],"analysis.":[84],"We":[85],"show":[86],"how":[87],"integrating":[88],"online":[90],"policy":[91],"controller":[92],"partial":[94],"caching":[95],"large":[97],"transformer":[98],"leads":[100],"significant":[102],"speedups":[103],"while":[104],"preserving":[105],"near-original":[106],"accuracy.":[107],"Experimental":[108],"results":[109],"scaled-down":[112],"GPT-style":[113],"model":[114],"confirm":[115],"promise":[117],"of":[118],"FP8-based":[119],"dynamic":[120],"quantization":[121],"compared":[122],"uniform":[124],"approaches,":[126],"demonstrating":[127],"average":[129],"$1.9":[130],"\\times$":[131],"throughput":[132],"increase":[133],"at":[134],"under":[135],"1%":[136],"accuracy":[137],"drop.":[138]},"counts_by_year":[],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}