{"id":"https://openalex.org/W4410583578","doi":"https://doi.org/10.23919/date64628.2025.10993129","title":"FineQ: Software-Hardware Co-Design for Low-Bit Fine-Grained Mixed-Precision Quantization of LLMs","display_name":"FineQ: Software-Hardware Co-Design for Low-Bit Fine-Grained Mixed-Precision Quantization of LLMs","publication_year":2025,"publication_date":"2025-03-31","ids":{"openalex":"https://openalex.org/W4410583578","doi":"https://doi.org/10.23919/date64628.2025.10993129"},"language":"en","primary_location":{"id":"doi:10.23919/date64628.2025.10993129","is_oa":false,"landing_page_url":"https://doi.org/10.23919/date64628.2025.10993129","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 Design, Automation &amp;amp; Test in Europe Conference (DATE)","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/A5029636421","display_name":"Xilong Xie","orcid":"https://orcid.org/0009-0005-9988-2940"},"institutions":[{"id":"https://openalex.org/I82880672","display_name":"Beihang University","ror":"https://ror.org/00wk2mp56","country_code":"CN","type":"education","lineage":["https://openalex.org/I82880672"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Xilong Xie","raw_affiliation_strings":["School of Computer Science and Engineering, Beihang University,Beijing,China,100191"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Beihang University,Beijing,China,100191","institution_ids":["https://openalex.org/I82880672"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100456598","display_name":"Liang Wang","orcid":"https://orcid.org/0000-0003-3936-607X"},"institutions":[{"id":"https://openalex.org/I82880672","display_name":"Beihang University","ror":"https://ror.org/00wk2mp56","country_code":"CN","type":"education","lineage":["https://openalex.org/I82880672"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Liang Wang","raw_affiliation_strings":["School of Computer Science and Engineering, Beihang University,Beijing,China,100191"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Beihang University,Beijing,China,100191","institution_ids":["https://openalex.org/I82880672"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101586078","display_name":"Limin Xiao","orcid":"https://orcid.org/0000-0001-9438-9181"},"institutions":[{"id":"https://openalex.org/I82880672","display_name":"Beihang University","ror":"https://ror.org/00wk2mp56","country_code":"CN","type":"education","lineage":["https://openalex.org/I82880672"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Limin Xiao","raw_affiliation_strings":["School of Computer Science and Engineering, Beihang University,Beijing,China,100191"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Beihang University,Beijing,China,100191","institution_ids":["https://openalex.org/I82880672"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101781069","display_name":"Meng Han","orcid":"https://orcid.org/0000-0002-8085-9047"},"institutions":[{"id":"https://openalex.org/I82880672","display_name":"Beihang University","ror":"https://ror.org/00wk2mp56","country_code":"CN","type":"education","lineage":["https://openalex.org/I82880672"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Meng Han","raw_affiliation_strings":["School of Computer Science and Engineering, Beihang University,Beijing,China,100191"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Beihang University,Beijing,China,100191","institution_ids":["https://openalex.org/I82880672"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101913610","display_name":"Lin Sun","orcid":"https://orcid.org/0000-0002-5725-180X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lin Sun","raw_affiliation_strings":["Jiangsu Shuguang Optoelectric Co., Ltd"],"affiliations":[{"raw_affiliation_string":"Jiangsu Shuguang Optoelectric Co., Ltd","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042419001","display_name":"Shuai Zheng","orcid":"https://orcid.org/0000-0001-8560-8135"},"institutions":[{"id":"https://openalex.org/I82880672","display_name":"Beihang University","ror":"https://ror.org/00wk2mp56","country_code":"CN","type":"education","lineage":["https://openalex.org/I82880672"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shuai Zheng","raw_affiliation_strings":["School of Computer Science and Engineering, Beihang University,Beijing,China,100191"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Beihang University,Beijing,China,100191","institution_ids":["https://openalex.org/I82880672"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5102020758","display_name":"Xiangrong Xu","orcid":"https://orcid.org/0000-0001-7650-5169"},"institutions":[{"id":"https://openalex.org/I82880672","display_name":"Beihang University","ror":"https://ror.org/00wk2mp56","country_code":"CN","type":"education","lineage":["https://openalex.org/I82880672"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiangrong Xu","raw_affiliation_strings":["School of Computer Science and Engineering, Beihang University,Beijing,China,100191"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Beihang University,Beijing,China,100191","institution_ids":["https://openalex.org/I82880672"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5029636421"],"corresponding_institution_ids":["https://openalex.org/I82880672"],"apc_list":null,"apc_paid":null,"fwci":1.398,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":{"value":0.80633058,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":95,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"7"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11301","display_name":"Advanced Surface Polishing Techniques","score":0.9387999773025513,"subfield":{"id":"https://openalex.org/subfields/2204","display_name":"Biomedical Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11301","display_name":"Advanced Surface Polishing Techniques","score":0.9387999773025513,"subfield":{"id":"https://openalex.org/subfields/2204","display_name":"Biomedical Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T13114","display_name":"Image Processing Techniques and Applications","score":0.9334999918937683,"subfield":{"id":"https://openalex.org/subfields/2214","display_name":"Media Technology"},"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":0.6426365971565247},{"id":"https://openalex.org/keywords/quantization","display_name":"Quantization (signal processing)","score":0.6041907072067261},{"id":"https://openalex.org/keywords/software","display_name":"Software","score":0.519711971282959},{"id":"https://openalex.org/keywords/software-implementation","display_name":"Software implementation","score":0.4728567600250244},{"id":"https://openalex.org/keywords/bit","display_name":"Bit (key)","score":0.470800518989563},{"id":"https://openalex.org/keywords/computer-architecture","display_name":"Computer architecture","score":0.3855648934841156},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.37364661693573},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.3403785526752472},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.1545408070087433},{"id":"https://openalex.org/keywords/computer-security","display_name":"Computer security","score":0.13689196109771729},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.06619682908058167}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6426365971565247},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.6041907072067261},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.519711971282959},{"id":"https://openalex.org/C2983609787","wikidata":"https://www.wikidata.org/wiki/Q10534782","display_name":"Software implementation","level":3,"score":0.4728567600250244},{"id":"https://openalex.org/C117011727","wikidata":"https://www.wikidata.org/wiki/Q1278488","display_name":"Bit (key)","level":2,"score":0.470800518989563},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.3855648934841156},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.37364661693573},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.3403785526752472},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.1545408070087433},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.13689196109771729},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.06619682908058167}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.23919/date64628.2025.10993129","is_oa":false,"landing_page_url":"https://doi.org/10.23919/date64628.2025.10993129","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 Design, Automation &amp;amp; Test in Europe Conference (DATE)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G5430715549","display_name":null,"funder_award_id":"62272026,62104014","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/F4320335787","display_name":"Fundamental Research Funds for the Central Universities","ror":null}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":10,"referenced_works":["https://openalex.org/W2583522611","https://openalex.org/W2980200167","https://openalex.org/W4242118294","https://openalex.org/W4281695175","https://openalex.org/W4292474994","https://openalex.org/W4366341968","https://openalex.org/W4387321091","https://openalex.org/W4392181806","https://openalex.org/W4393147284","https://openalex.org/W4401211807"],"related_works":["https://openalex.org/W4327546585","https://openalex.org/W2411923897","https://openalex.org/W4394546135","https://openalex.org/W4285347720","https://openalex.org/W4200259850","https://openalex.org/W2333831899","https://openalex.org/W2484894494","https://openalex.org/W2367385042","https://openalex.org/W2883113558","https://openalex.org/W2556125783"],"abstract_inverted_index":{"Large":[0],"language":[1,9],"models":[2],"(LLMs)":[3],"have":[4],"significantly":[5],"advanced":[6,36],"the":[7,25,97,104,163,168,171,181,192,202,205],"natural":[8],"processing":[10],"paradigm":[11],"but":[12],"impose":[13],"substantial":[14,68],"demands":[15],"on":[16],"memory":[17,31,69,119,150],"and":[18,102,118,137,144,200],"computational":[19],"resources.":[20],"Quantization":[21],"is":[22],"one":[23],"of":[24,33,92,106,204],"most":[26],"effective":[27],"ways":[28],"to":[29,46,67,134,147,180,196],"reduce":[30],"consumption":[32],"LLMs.":[34,93],"However,":[35],"single-precision":[37],"quantization":[38,51,91,164,184],"methods":[39,52],"experience":[40],"significant":[41],"accuracy":[42,117,178],"degradation":[43],"when":[44],"quantizing":[45],"ultra-low":[47],"bits.":[48],"Existing":[49],"mixed-precision":[50,90,183],"are":[53],"quantized":[54],"by":[55,208],"groups":[56],"with":[57],"coarse":[58],"granularity.":[59],"Employing":[60],"high":[61],"precision":[62,73],"for":[63,87,142],"group":[64],"data":[65,145],"leads":[66],"overhead,":[70],"whereas":[71],"low":[72],"severely":[74],"impacts":[75],"model":[76,116,177],"accuracy.":[77],"To":[78],"address":[79],"this":[80],"issue,":[81],"we":[82,122,153],"propose":[83,123],"FineQ,":[84],"software-hardware":[85],"co-design":[86],"low-bit":[88],"fine-grained":[89],"First,":[94],"FineQ":[95,174],"partitions":[96],"weights":[98],"into":[99],"finer-grained":[100],"clusters":[101,129],"considers":[103],"distribution":[105],"outliers":[107,136],"within":[108,128],"these":[109],"clusters,":[110],"thus":[111],"achieving":[112],"a":[113,187],"balance":[114],"between":[115],"overhead.":[120],"Then,":[121],"an":[124,139,155],"outlier":[125],"protection":[126],"mechanism":[127],"that":[130,160],"uses":[131],"3":[132],"bits":[133],"represent":[135],"introduce":[138,154],"encoding":[140],"scheme":[141],"index":[143],"concatenation":[146],"enable":[148],"aligned":[149],"access.":[151],"Finally,":[152],"accelerator":[156,193],"utilizing":[157],"temporal":[158],"coding":[159],"effectively":[161],"supports":[162],"algorithm":[165,185],"while":[166],"simplifying":[167],"multipliers":[169],"in":[170],"systolic":[172,206],"array.":[173],"achieves":[175,194],"higher":[176],"compared":[179],"SOTA":[182],"at":[186],"close":[188],"average":[189],"bit-width.":[190],"Meanwhile,":[191],"up":[195],"1.79x":[197],"energy":[198],"efficiency":[199],"reduces":[201],"area":[203],"array":[207],"61.2%.":[209]},"counts_by_year":[{"year":2025,"cited_by_count":2}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}
