{"id":"https://openalex.org/W4413411892","doi":"https://doi.org/10.1145/3721145.3725779","title":"BitWeaver: Read-Time Truncation in Memory","display_name":"BitWeaver: Read-Time Truncation in Memory","publication_year":2025,"publication_date":"2025-06-08","ids":{"openalex":"https://openalex.org/W4413411892","doi":"https://doi.org/10.1145/3721145.3725779"},"language":"en","primary_location":{"id":"doi:10.1145/3721145.3725779","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3721145.3725779","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3721145.3725779","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 39th ACM International Conference on Supercomputing","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3721145.3725779","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5119381438","display_name":"Garrett Gagnon","orcid":"https://orcid.org/0009-0009-0754-9518"},"institutions":[{"id":"https://openalex.org/I165799507","display_name":"Rensselaer Polytechnic Institute","ror":"https://ror.org/01rtyzb94","country_code":"US","type":"education","lineage":["https://openalex.org/I165799507"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Garrett Gagnon","raw_affiliation_strings":["Samsung Semiconductor US, San Jose, California, USA and Rensselaer Polytechnic Institute, Troy, New York, USA"],"raw_orcid":"https://orcid.org/0009-0009-0754-9518","affiliations":[{"raw_affiliation_string":"Samsung Semiconductor US, San Jose, California, USA and Rensselaer Polytechnic Institute, Troy, New York, USA","institution_ids":["https://openalex.org/I165799507"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5009999184","display_name":"Srikanth Malla","orcid":"https://orcid.org/0000-0002-7431-3064"},"institutions":[{"id":"https://openalex.org/I100625452","display_name":"ON Semiconductor (United States)","ror":"https://ror.org/03nw6pt28","country_code":"US","type":"company","lineage":["https://openalex.org/I100625452"]},{"id":"https://openalex.org/I2250650973","display_name":"Samsung (South Korea)","ror":"https://ror.org/04w3jy968","country_code":"KR","type":"company","lineage":["https://openalex.org/I2250650973"]}],"countries":["KR","US"],"is_corresponding":false,"raw_author_name":"Srikanth Malla","raw_affiliation_strings":["Samsung Semiconductor US, San Jose, California, USA"],"raw_orcid":"https://orcid.org/0000-0002-7431-3064","affiliations":[{"raw_affiliation_string":"Samsung Semiconductor US, San Jose, California, USA","institution_ids":["https://openalex.org/I100625452","https://openalex.org/I2250650973"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5060280539","display_name":"Yangwook Kang","orcid":"https://orcid.org/0009-0007-9536-1894"},"institutions":[{"id":"https://openalex.org/I100625452","display_name":"ON Semiconductor (United States)","ror":"https://ror.org/03nw6pt28","country_code":"US","type":"company","lineage":["https://openalex.org/I100625452"]},{"id":"https://openalex.org/I2250650973","display_name":"Samsung (South Korea)","ror":"https://ror.org/04w3jy968","country_code":"KR","type":"company","lineage":["https://openalex.org/I2250650973"]}],"countries":["KR","US"],"is_corresponding":false,"raw_author_name":"Yangwook Kang","raw_affiliation_strings":["Samsung Semiconductor US, San Jose, California, USA"],"raw_orcid":"https://orcid.org/0009-0007-9536-1894","affiliations":[{"raw_affiliation_string":"Samsung Semiconductor US, San Jose, California, USA","institution_ids":["https://openalex.org/I100625452","https://openalex.org/I2250650973"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100324294","display_name":"Liu Liu","orcid":"https://orcid.org/0000-0003-0792-8146"},"institutions":[{"id":"https://openalex.org/I165799507","display_name":"Rensselaer Polytechnic Institute","ror":"https://ror.org/01rtyzb94","country_code":"US","type":"education","lineage":["https://openalex.org/I165799507"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Liu Liu","raw_affiliation_strings":["Rensselaer Polytechnic Institute, Troy, New York, USA"],"raw_orcid":"https://orcid.org/0000-0003-0792-8146","affiliations":[{"raw_affiliation_string":"Rensselaer Polytechnic Institute, Troy, New York, USA","institution_ids":["https://openalex.org/I165799507"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.20087423,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"13","last_page":"25"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9998999834060669,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9998999834060669,"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":0.9998000264167786,"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/T10101","display_name":"Cloud Computing and Resource Management","score":0.9951000213623047,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/computer-science","display_name":"Computer science","score":0.6932097673416138},{"id":"https://openalex.org/keywords/truncation","display_name":"Truncation (statistics)","score":0.6514977812767029},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.333680659532547},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.04877921938896179}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6932097673416138},{"id":"https://openalex.org/C106195933","wikidata":"https://www.wikidata.org/wiki/Q7847935","display_name":"Truncation (statistics)","level":2,"score":0.6514977812767029},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.333680659532547},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.04877921938896179}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3721145.3725779","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3721145.3725779","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3721145.3725779","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 39th ACM International Conference on Supercomputing","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3721145.3725779","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3721145.3725779","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3721145.3725779","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 39th ACM International Conference on Supercomputing","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G3642890131","display_name":"CAREER: Elastic Algorithm-Architecture Co-Design for Energy-Efficient and Scalable Artificial Intelligence via Dynamic Connectivity","funder_award_id":"2442271","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320332195","display_name":"Samsung","ror":"https://ror.org/04w3jy968"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4413411892.pdf","grobid_xml":"https://content.openalex.org/works/W4413411892.grobid-xml"},"referenced_works_count":22,"referenced_works":["https://openalex.org/W2002555321","https://openalex.org/W2081140905","https://openalex.org/W2120000030","https://openalex.org/W2497599918","https://openalex.org/W2606411279","https://openalex.org/W2622148586","https://openalex.org/W2888482885","https://openalex.org/W2890894339","https://openalex.org/W2923014074","https://openalex.org/W2946609015","https://openalex.org/W2951939640","https://openalex.org/W2996908057","https://openalex.org/W2998617917","https://openalex.org/W3159727696","https://openalex.org/W3208215217","https://openalex.org/W4214686755","https://openalex.org/W4285798357","https://openalex.org/W4292261931","https://openalex.org/W4324292875","https://openalex.org/W4380874652","https://openalex.org/W4402130049","https://openalex.org/W4404133752"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W1974935850","https://openalex.org/W2074350688","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109"],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2],"(LLMs)":[3],"have":[4],"demonstrated":[5],"remarkable":[6],"capabilities":[7],"in":[8,27,114,159],"generating":[9],"contextually":[10],"relevant":[11],"responses":[12],"to":[13,38,56,79,120,155,167,170],"prompts,":[14],"but":[15],"their":[16],"inference":[17],"performance":[18,98],"is":[19,33],"often":[20],"constrained":[21,68],"by":[22,69,178],"a":[23,86,117,133,140,156],"severe":[24],"memory":[25,70,78,115,123,151,160],"bottleneck":[26],"the":[28,45,48,58,74,81,180],"self-attention":[29],"stage.This":[30],"bottleneck,":[31],"which":[32,72],"inherently":[34],"memory-bound,":[35],"has":[36,111],"led":[37],"extensive":[39],"research":[40],"into":[41,84],"strategies":[42,89,177],"for":[43,143],"reducing":[44],"size":[46],"of":[47,165,182],"Key-Value":[49],"(KV)":[50],"cache.Many":[51],"existing":[52],"approaches":[53],"employ":[54],"quantization":[55],"lower":[57],"data":[59,63,75,82,110,148],"precision":[60,105,149],"and":[61,125,137,162],"reduce":[62,122],"volume.However,":[64],"these":[65],"methods":[66],"are":[67],"technology,":[71],"requires":[73],"read":[76],"from":[77],"match":[80],"written":[83],"it.As":[85],"result,":[87],"such":[88],"must":[90],"apply":[91],"reductions":[92],"at":[93,107],"write-time,":[94],"limiting":[95],"either":[96],"model":[97,127],"or":[99],"achievable":[100],"speedup.We":[101],"identify":[102],"that":[103],"enabling":[104],"scaling":[106],"read-time":[108,134],"-after":[109],"been":[112],"stored":[113],"-offers":[116],"unique":[118],"opportunity":[119],"simultaneously":[121],"traffic":[124],"retain":[126],"accuracy.To":[128],"this":[129],"end,":[130],"we":[131],"develop":[132],"precision-scaling":[135],"mechanism":[136],"introduce":[138],"BitWeaver,":[139],"hardware-enabled":[141],"solution":[142],"in-memory":[144],"truncation.BitWeaver":[145],"dynamically":[146],"reduces":[147],"during":[150],"reads,":[152],"achieving":[153],"up":[154,166],"3\u00d7":[157],"increase":[158],"throughput":[161],"execution":[163],"speedups":[164],"80%":[168],"compared":[169],"baseline.Additionally,":[171],"BitWeaver":[172],"enhances":[173],"sparse":[174],"KV":[175],"cache":[176],"improving":[179],"efficiency":[181],"state-of-the-art":[183],"sparsity":[184],"techniques.":[185]},"counts_by_year":[],"updated_date":"2026-07-31T08:31:51.225901","created_date":"2025-10-10T00:00:00"}