{"id":"https://openalex.org/W7133306715","doi":"https://doi.org/10.1109/isscc49663.2026.11408953","title":"A 28nm Speculative-Decoding LLM Processor Achieving 105-to-685\u00b5s/Token Latency for Billion-Parameter Models","display_name":"A 28nm Speculative-Decoding LLM Processor Achieving 105-to-685\u00b5s/Token Latency for Billion-Parameter Models","publication_year":2026,"publication_date":"2026-02-15","ids":{"openalex":"https://openalex.org/W7133306715","doi":"https://doi.org/10.1109/isscc49663.2026.11408953"},"language":null,"primary_location":{"id":"doi:10.1109/isscc49663.2026.11408953","is_oa":false,"landing_page_url":"https://doi.org/10.1109/isscc49663.2026.11408953","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 IEEE International Solid-State Circuits Conference (ISSCC)","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/A5128008324","display_name":"Yang Wang","orcid":null},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Yang Wang","raw_affiliation_strings":["Tsinghua University,Beijing,China"],"affiliations":[{"raw_affiliation_string":"Tsinghua University,Beijing,China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100696526","display_name":"Huanyu Wang","orcid":"https://orcid.org/0000-0001-8962-0393"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Huanyu Wang","raw_affiliation_strings":["Tsinghua University,Beijing,China"],"affiliations":[{"raw_affiliation_string":"Tsinghua University,Beijing,China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124146592","display_name":"Jiaxin Yang","orcid":null},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiaxin Yang","raw_affiliation_strings":["Tsinghua University,Beijing,China"],"affiliations":[{"raw_affiliation_string":"Tsinghua University,Beijing,China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113293648","display_name":"Yutong Su","orcid":"https://orcid.org/0009-0008-6059-5778"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yutong Su","raw_affiliation_strings":["Tsinghua University,Beijing,China"],"affiliations":[{"raw_affiliation_string":"Tsinghua University,Beijing,China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5085105148","display_name":"Ruiqi Guo","orcid":"https://orcid.org/0000-0003-4729-7385"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ruiqi Guo","raw_affiliation_strings":["Tsinghua University,Beijing,China"],"affiliations":[{"raw_affiliation_string":"Tsinghua University,Beijing,China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5009692777","display_name":"Zhiheng Yue","orcid":null},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhiheng Yue","raw_affiliation_strings":["Tsinghua University,Beijing,China"],"affiliations":[{"raw_affiliation_string":"Tsinghua University,Beijing,China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5022808670","display_name":"Jiangyuan Gu","orcid":"https://orcid.org/0000-0003-1190-7524"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiangyuan Gu","raw_affiliation_strings":["Tsinghua University,Beijing,China"],"affiliations":[{"raw_affiliation_string":"Tsinghua University,Beijing,China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5127882839","display_name":"Shaojun Wei","orcid":null},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shaojun Wei","raw_affiliation_strings":["Tsinghua University,Beijing,China"],"affiliations":[{"raw_affiliation_string":"Tsinghua University,Beijing,China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5068155662","display_name":"Yuanming Hu","orcid":"https://orcid.org/0000-0002-1136-9909"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang Hu","raw_affiliation_strings":["Tsinghua University,Beijing,China"],"affiliations":[{"raw_affiliation_string":"Tsinghua University,Beijing,China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5078843314","display_name":"Shouyi Yin","orcid":"https://orcid.org/0000-0002-8438-8588"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shouyi Yin","raw_affiliation_strings":["Tsinghua University,Beijing,China"],"affiliations":[{"raw_affiliation_string":"Tsinghua University,Beijing,China","institution_ids":["https://openalex.org/I99065089"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":10,"corresponding_author_ids":["https://openalex.org/A5128008324"],"corresponding_institution_ids":["https://openalex.org/I99065089"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.8768021,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"546","last_page":"548"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12326","display_name":"Network Packet Processing and Optimization","score":0.2538999915122986,"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"}},"topics":[{"id":"https://openalex.org/T12326","display_name":"Network Packet Processing and Optimization","score":0.2538999915122986,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.15049999952316284,"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/T11181","display_name":"Advanced Data Storage Technologies","score":0.148499995470047,"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/decoding-methods","display_name":"Decoding methods","score":0.7910000085830688},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.7268999814987183},{"id":"https://openalex.org/keywords/exponent","display_name":"Exponent","score":0.46320000290870667},{"id":"https://openalex.org/keywords/face","display_name":"Face (sociological concept)","score":0.3580000102519989},{"id":"https://openalex.org/keywords/low-latency","display_name":"Low latency (capital markets)","score":0.29820001125335693}],"concepts":[{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.7910000085830688},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7594000101089478},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.7268999814987183},{"id":"https://openalex.org/C2780388253","wikidata":"https://www.wikidata.org/wiki/Q5421508","display_name":"Exponent","level":2,"score":0.46320000290870667},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.4449000060558319},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.40230000019073486},{"id":"https://openalex.org/C2779304628","wikidata":"https://www.wikidata.org/wiki/Q3503480","display_name":"Face (sociological concept)","level":2,"score":0.3580000102519989},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.3098999857902527},{"id":"https://openalex.org/C46637626","wikidata":"https://www.wikidata.org/wiki/Q6693015","display_name":"Low latency (capital markets)","level":2,"score":0.29820001125335693},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.2897000014781952},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.2822999954223633},{"id":"https://openalex.org/C107598950","wikidata":"https://www.wikidata.org/wiki/Q259864","display_name":"Microarchitecture","level":2,"score":0.25870001316070557},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.2574999928474426}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/isscc49663.2026.11408953","is_oa":false,"landing_page_url":"https://doi.org/10.1109/isscc49663.2026.11408953","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 IEEE International Solid-State Circuits Conference (ISSCC)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":6,"referenced_works":["https://openalex.org/W4360605545","https://openalex.org/W4395112660","https://openalex.org/W4408105642","https://openalex.org/W4408183246","https://openalex.org/W4411120473","https://openalex.org/W4412945506"],"related_works":[],"abstract_inverted_index":{"LLMs":[0],"face":[1],"decoding":[2,16],"bottlenecks.":[3],"Speculative":[4],"Decoding":[5],"(SD)":[6],"reduces":[7],"latency":[8],"via":[9],"a":[10,18],"small":[11],"draft":[12,53,59],"model":[13,21],"for":[14],"serial":[15],"and":[17,37,58],"large":[19],"target":[20,55],"to":[22],"verify":[23],"in":[24],"parallel.":[25],"Despite":[26],"this":[27],"advantage,":[28],"it":[29],"still":[30],"suffers":[31],"from":[32],"exponent":[33,50],"redundancy,":[34],"inefficient":[35],"weights/KV,":[36],"hardware":[38],"underutilization.":[39],"We":[40],"propose":[41],"an":[42,49],"SD":[43],"LLM":[44],"processor":[45],"with":[46],"three":[47],"innovations:":[48],"dual-reuse":[51],"MAC,":[52],"back-propagation":[54],"mixed-precision":[56],"dataflow,":[57],"early":[60],"start":[61],"parallel":[62],"compute,":[63],"achieving":[64],"<tex":[65],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[66],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">$10.29":[67],"\\times$</tex>":[68],"speed":[69],"increase.":[70]},"counts_by_year":[],"updated_date":"2026-03-05T07:30:30.508283","created_date":"2026-03-04T00:00:00"}
