{"id":"https://openalex.org/W4408891426","doi":"https://doi.org/10.1145/3689031.3696099","title":"T-MAC: CPU Renaissance via Table Lookup for Low-Bit LLM Deployment on Edge","display_name":"T-MAC: CPU Renaissance via Table Lookup for Low-Bit LLM Deployment on Edge","publication_year":2025,"publication_date":"2025-03-26","ids":{"openalex":"https://openalex.org/W4408891426","doi":"https://doi.org/10.1145/3689031.3696099"},"language":"en","primary_location":{"id":"doi:10.1145/3689031.3696099","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3689031.3696099","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Twentieth European Conference on Computer Systems","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/A5071351577","display_name":"Jianyu Wei","orcid":"https://orcid.org/0009-0009-0830-044X"},"institutions":[{"id":"https://openalex.org/I4210124949","display_name":"Microsoft Research (India)","ror":"https://ror.org/02w7f3w92","country_code":"IN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210124949"]}],"countries":["IN"],"is_corresponding":true,"raw_author_name":"Jianyu Wei","raw_affiliation_strings":["USTC / Microsoft Research"],"affiliations":[{"raw_affiliation_string":"USTC / Microsoft Research","institution_ids":["https://openalex.org/I4210124949"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5078290360","display_name":"Shijie Cao","orcid":"https://orcid.org/0009-0000-2001-3763"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Shijie Cao","raw_affiliation_strings":["Microsoft Research"],"affiliations":[{"raw_affiliation_string":"Microsoft Research","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101534161","display_name":"Ting Cao","orcid":"https://orcid.org/0000-0002-9107-013X"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Ting Cao","raw_affiliation_strings":["Microsoft Research"],"affiliations":[{"raw_affiliation_string":"Microsoft Research","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102874660","display_name":"Lingxiao Ma","orcid":"https://orcid.org/0009-0009-9524-5476"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Lingxiao Ma","raw_affiliation_strings":["Microsoft Research"],"affiliations":[{"raw_affiliation_string":"Microsoft Research","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5005666540","display_name":"Lei Wang","orcid":"https://orcid.org/0009-0006-2313-5348"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Lei Wang","raw_affiliation_strings":["UCAS / Microsoft Research"],"affiliations":[{"raw_affiliation_string":"UCAS / Microsoft Research","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103041851","display_name":"Yanyong Zhang","orcid":"https://orcid.org/0000-0001-9046-798X"},"institutions":[{"id":"https://openalex.org/I91090688","display_name":"University of Science and Technology Chittagong","ror":"https://ror.org/00w9tx359","country_code":"BD","type":"education","lineage":["https://openalex.org/I91090688"]}],"countries":["BD"],"is_corresponding":false,"raw_author_name":"Yanyong Zhang","raw_affiliation_strings":["USTC"],"affiliations":[{"raw_affiliation_string":"USTC","institution_ids":["https://openalex.org/I91090688"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100438310","display_name":"Mao Yang","orcid":"https://orcid.org/0009-0009-6455-3898"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Mao Yang","raw_affiliation_strings":["Microsoft Research"],"affiliations":[{"raw_affiliation_string":"Microsoft Research","institution_ids":["https://openalex.org/I4210164937"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5071351577"],"corresponding_institution_ids":["https://openalex.org/I4210124949"],"apc_list":null,"apc_paid":null,"fwci":15.5247,"has_fulltext":false,"cited_by_count":11,"citation_normalized_percentile":{"value":0.98944801,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"278","last_page":"292"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11181","display_name":"Advanced Data Storage Technologies","score":0.9983000159263611,"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":0.9983000159263611,"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":0.9958000183105469,"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/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9909999966621399,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic 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/lookup-table","display_name":"Lookup table","score":0.7865539789199829},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7376919984817505},{"id":"https://openalex.org/keywords/table","display_name":"Table (database)","score":0.7080560922622681},{"id":"https://openalex.org/keywords/software-deployment","display_name":"Software deployment","score":0.6946938037872314},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.6107405424118042},{"id":"https://openalex.org/keywords/bit","display_name":"Bit (key)","score":0.5466764569282532},{"id":"https://openalex.org/keywords/the-renaissance","display_name":"The Renaissance","score":0.4353828728199005},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.3959254026412964},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.3875635862350464},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.379843533039093},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.23837387561798096},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.17838028073310852},{"id":"https://openalex.org/keywords/telecommunications","display_name":"Telecommunications","score":0.11814472079277039},{"id":"https://openalex.org/keywords/art","display_name":"Art","score":0.08843153715133667},{"id":"https://openalex.org/keywords/database","display_name":"Database","score":0.07936739921569824}],"concepts":[{"id":"https://openalex.org/C134835016","wikidata":"https://www.wikidata.org/wiki/Q690265","display_name":"Lookup table","level":2,"score":0.7865539789199829},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7376919984817505},{"id":"https://openalex.org/C45235069","wikidata":"https://www.wikidata.org/wiki/Q278425","display_name":"Table (database)","level":2,"score":0.7080560922622681},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.6946938037872314},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.6107405424118042},{"id":"https://openalex.org/C117011727","wikidata":"https://www.wikidata.org/wiki/Q1278488","display_name":"Bit (key)","level":2,"score":0.5466764569282532},{"id":"https://openalex.org/C52069626","wikidata":"https://www.wikidata.org/wiki/Q4692","display_name":"The Renaissance","level":2,"score":0.4353828728199005},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3959254026412964},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.3875635862350464},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.379843533039093},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.23837387561798096},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.17838028073310852},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.11814472079277039},{"id":"https://openalex.org/C142362112","wikidata":"https://www.wikidata.org/wiki/Q735","display_name":"Art","level":0,"score":0.08843153715133667},{"id":"https://openalex.org/C77088390","wikidata":"https://www.wikidata.org/wiki/Q8513","display_name":"Database","level":1,"score":0.07936739921569824},{"id":"https://openalex.org/C52119013","wikidata":"https://www.wikidata.org/wiki/Q50637","display_name":"Art history","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3689031.3696099","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3689031.3696099","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Twentieth European Conference on Computer Systems","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":19,"referenced_works":["https://openalex.org/W2804032941","https://openalex.org/W2963015836","https://openalex.org/W2963259903","https://openalex.org/W3046444208","https://openalex.org/W4246166885","https://openalex.org/W4283313765","https://openalex.org/W4298422451","https://openalex.org/W4309591680","https://openalex.org/W4312056202","https://openalex.org/W4367060479","https://openalex.org/W4385326807","https://openalex.org/W4387212331","https://openalex.org/W4387321091","https://openalex.org/W4391942385","https://openalex.org/W4395474395","https://openalex.org/W4402667003","https://openalex.org/W6727099177","https://openalex.org/W6752057402","https://openalex.org/W6797565540"],"related_works":["https://openalex.org/W2770234245","https://openalex.org/W96612179","https://openalex.org/W4229499248","https://openalex.org/W2566006169","https://openalex.org/W1567818861","https://openalex.org/W2987774938","https://openalex.org/W4327546585","https://openalex.org/W4256492088","https://openalex.org/W632915154","https://openalex.org/W2055733372"],"abstract_inverted_index":{"The":[0],"deployment":[1],"of":[2,26,39],"Large":[3],"Language":[4],"Models":[5],"(LLMs)":[6],"on":[7,28],"edge":[8],"devices":[9],"is":[10,19],"increasingly":[11],"important":[12],"to":[13,57,70],"enhance":[14],"on-device":[15],"intelligence.":[16],"Weight":[17],"quantization":[18],"crucial":[20],"for":[21,54,60],"reducing":[22],"the":[23],"memory":[24],"footprint":[25],"LLMs":[27,32],"devices.":[29],"However,":[30],"low-bit":[31],"necessitate":[33],"mixed":[34],"precision":[35,41,45,62],"matrix":[36],"multiplication":[37],"(mpGEMM)":[38],"low":[40],"weights":[42,59],"and":[43],"high":[44,61],"activations":[46],"during":[47],"inference.":[48],"Existing":[49],"systems,":[50],"lacking":[51],"native":[52],"support":[53],"mpGEMM,":[55],"resort":[56],"dequantize":[58],"computation.":[63],"Such":[64],"an":[65],"indirect":[66],"way":[67],"can":[68],"lead":[69],"a":[71],"significant":[72],"inference":[73],"overhead.":[74]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":9}],"updated_date":"2026-04-11T08:14:18.477133","created_date":"2025-10-10T00:00:00"}
