{"id":"https://openalex.org/W7166661686","doi":"https://doi.org/10.1145/3774895.3819987","title":"Omni-MP: Practical PIM Matrix Mapping for Accelerating LLM Inference on a Real PIM Hardware","display_name":"Omni-MP: Practical PIM Matrix Mapping for Accelerating LLM Inference on a Real PIM Hardware","publication_year":2026,"publication_date":"2026-06-30","ids":{"openalex":"https://openalex.org/W7166661686","doi":"https://doi.org/10.1145/3774895.3819987"},"language":null,"primary_location":{"id":"doi:10.1145/3774895.3819987","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3774895.3819987","pdf_url":null,"source":null,"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 40th ACM International Conference on Supercomputing - Workshops","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3774895.3819987","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5028282859","display_name":"Daehyeon Baek","orcid":"https://orcid.org/0009-0008-0460-3809"},"institutions":[{"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"]},{"id":"https://openalex.org/I4387155180","display_name":"Samsung SDS (South Korea)","ror":"https://ror.org/0476bn305","country_code":null,"type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4387155180"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Daehyeon Baek","raw_affiliation_strings":["Samsung SDS, Seoul, Republic of Korea"],"raw_orcid":"https://orcid.org/0009-0008-0460-3809","affiliations":[{"raw_affiliation_string":"Samsung SDS, Seoul, Republic of Korea","institution_ids":["https://openalex.org/I2250650973","https://openalex.org/I4387155180"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102865647","display_name":"JunSeok Son","orcid":"https://orcid.org/0000-0001-8344-4536"},"institutions":[{"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"]},{"id":"https://openalex.org/I4387155180","display_name":"Samsung SDS (South Korea)","ror":"https://ror.org/0476bn305","country_code":null,"type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4387155180"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jimyoung Son","raw_affiliation_strings":["Samsung SDS, Seoul, Republic of Korea"],"raw_orcid":"https://orcid.org/0009-0007-0110-5866","affiliations":[{"raw_affiliation_string":"Samsung SDS, Seoul, Republic of Korea","institution_ids":["https://openalex.org/I2250650973","https://openalex.org/I4387155180"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103722801","display_name":"J O Choi","orcid":null},"institutions":[{"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"]},{"id":"https://openalex.org/I4387155180","display_name":"Samsung SDS (South Korea)","ror":"https://ror.org/0476bn305","country_code":null,"type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4387155180"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jieun Choi","raw_affiliation_strings":["Samsung SDS, Seoul, Republic of Korea"],"raw_orcid":"https://orcid.org/0009-0009-5652-9994","affiliations":[{"raw_affiliation_string":"Samsung SDS, Seoul, Republic of Korea","institution_ids":["https://openalex.org/I2250650973","https://openalex.org/I4387155180"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128364620","display_name":"Kihyo Moon","orcid":null},"institutions":[{"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"]},{"id":"https://openalex.org/I4387155180","display_name":"Samsung SDS (South Korea)","ror":"https://ror.org/0476bn305","country_code":null,"type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4387155180"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Kihyo Moon","raw_affiliation_strings":["Samsung SDS, Seoul, Republic of Korea"],"raw_orcid":"https://orcid.org/0009-0008-5145-8333","affiliations":[{"raw_affiliation_string":"Samsung SDS, Seoul, Republic of Korea","institution_ids":["https://openalex.org/I2250650973","https://openalex.org/I4387155180"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100986244","display_name":"Minsung Jang","orcid":null},"institutions":[{"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"]},{"id":"https://openalex.org/I4387155180","display_name":"Samsung SDS (South Korea)","ror":"https://ror.org/0476bn305","country_code":null,"type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4387155180"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Minsung Jang","raw_affiliation_strings":["Samsung SDS, Seoul, Republic of Korea"],"raw_orcid":"https://orcid.org/0009-0005-7805-6383","affiliations":[{"raw_affiliation_string":"Samsung SDS, Seoul, Republic of Korea","institution_ids":["https://openalex.org/I2250650973","https://openalex.org/I4387155180"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5041250916","display_name":"Hyojung Lee","orcid":"https://orcid.org/0000-0002-0471-6650"},"institutions":[{"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"]},{"id":"https://openalex.org/I4387155180","display_name":"Samsung SDS (South Korea)","ror":"https://ror.org/0476bn305","country_code":null,"type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4387155180"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Hyojung Lee","raw_affiliation_strings":["Samsung SDS, Seoul, Republic of Korea"],"raw_orcid":"https://orcid.org/0000-0001-5399-4334","affiliations":[{"raw_affiliation_string":"Samsung SDS, Seoul, Republic of Korea","institution_ids":["https://openalex.org/I2250650973","https://openalex.org/I4387155180"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.79051919,"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":"5"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.3903000056743622,"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":0.3903000056743622,"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/T12808","display_name":"Ferroelectric and Negative Capacitance Devices","score":0.21559999883174896,"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"}},{"id":"https://openalex.org/T10502","display_name":"Advanced Memory and Neural Computing","score":0.16269999742507935,"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/inference","display_name":"Inference","score":0.6636000275611877},{"id":"https://openalex.org/keywords/dram","display_name":"Dram","score":0.6575000286102295},{"id":"https://openalex.org/keywords/software","display_name":"Software","score":0.5285999774932861},{"id":"https://openalex.org/keywords/memory-bandwidth","display_name":"Memory bandwidth","score":0.41920000314712524},{"id":"https://openalex.org/keywords/offset","display_name":"Offset (computer science)","score":0.39989998936653137},{"id":"https://openalex.org/keywords/bandwidth","display_name":"Bandwidth (computing)","score":0.3571999967098236}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8026999831199646},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6636000275611877},{"id":"https://openalex.org/C7366592","wikidata":"https://www.wikidata.org/wiki/Q1255620","display_name":"Dram","level":2,"score":0.6575000286102295},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.5285999774932861},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.45350000262260437},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.4323999881744385},{"id":"https://openalex.org/C188045654","wikidata":"https://www.wikidata.org/wiki/Q17148339","display_name":"Memory bandwidth","level":2,"score":0.41920000314712524},{"id":"https://openalex.org/C175291020","wikidata":"https://www.wikidata.org/wiki/Q1156822","display_name":"Offset (computer science)","level":2,"score":0.39989998936653137},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.37619999051094055},{"id":"https://openalex.org/C2776257435","wikidata":"https://www.wikidata.org/wiki/Q1576430","display_name":"Bandwidth (computing)","level":2,"score":0.3571999967098236},{"id":"https://openalex.org/C17349429","wikidata":"https://www.wikidata.org/wiki/Q1049914","display_name":"Matrix multiplication","level":3,"score":0.34689998626708984},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.29499998688697815},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.29100000858306885},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.2897000014781952},{"id":"https://openalex.org/C106487976","wikidata":"https://www.wikidata.org/wiki/Q685816","display_name":"Matrix (chemical analysis)","level":2,"score":0.28049999475479126},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.2770000100135803},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2612000107765198},{"id":"https://openalex.org/C58013763","wikidata":"https://www.wikidata.org/wiki/Q5754574","display_name":"High-level synthesis","level":3,"score":0.2612000107765198},{"id":"https://openalex.org/C46743427","wikidata":"https://www.wikidata.org/wiki/Q1341685","display_name":"Inference engine","level":3,"score":0.25380000472068787}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3774895.3819987","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3774895.3819987","pdf_url":null,"source":null,"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 40th ACM International Conference on Supercomputing - Workshops","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3774895.3819987","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3774895.3819987","pdf_url":null,"source":null,"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 40th ACM International Conference on Supercomputing - Workshops","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":4,"referenced_works":["https://openalex.org/W3189166979","https://openalex.org/W4297097348","https://openalex.org/W4401211704","https://openalex.org/W4401880111"],"related_works":[],"abstract_inverted_index":{"The":[0],"rapid":[1],"growth":[2],"of":[3,102],"large":[4],"language":[5],"models":[6],"(LLMs)":[7],"has":[8],"led":[9],"to":[10,26,58],"heavy":[11],"demands":[12],"for":[13,64,130],"data":[14,33],"and":[15,49,117],"computation.":[16],"Processing-in-Memory":[17],"(PIM)":[18],"technologies":[19],"integrate":[20],"processing":[21],"units":[22],"into":[23],"DRAM":[24],"chips":[25],"maximize":[27],"memory":[28],"bandwidth":[29],"by":[30,108],"eliminating":[31],"off-chip":[32],"movement,":[34],"thereby":[35],"accelerating":[36],"memory-intensive":[37],"operations":[38,55,70,105],"such":[39],"as":[40],"General":[41],"Matrix-Vector":[42],"Multiplication":[43],"(GEMV).":[44],"However,":[45],"existing":[46],"PIM":[47,82,96],"hardware":[48],"software":[50],"stacks":[51],"provide":[52],"non-optimized":[53],"GEMV":[54,69,104],"with":[56],"respect":[57],"matrix":[59,112],"dimensions,":[60],"which":[61],"poses":[62],"challenges":[63],"transformer-based":[65],"LLM":[66,127],"inference":[67,107,128],"where":[68],"occur":[71],"across":[72],"diverse":[73,110],"dimensions.":[74],"To":[75],"address":[76],"this":[77],"limitation,":[78],"we":[79],"propose":[80],"Omni-Matrix":[81],"mapping":[83],"(Omni-MP),":[84],"a":[85,92],"flexible":[86],"PIM-GEMV":[87],"offloading":[88],"mechanism":[89],"evaluated":[90],"on":[91,114,132],"pre-commercial":[93],"prototype":[94],"HBM2-based":[95],"accelerator.":[97],"Omni-MP":[98,123],"enhances":[99],"the":[100],"performance":[101,129],"all":[103],"during":[106],"optimizing":[109],"weight":[111],"transformations":[113],"both":[115],"fully-connected":[116],"attention":[118],"layers.":[119],"With":[120],"these":[121],"optimizations,":[122],"significantly":[124],"improves":[125],"end-to-end":[126],"40.2%":[131],"average,":[133],"over":[134],"GPU-only":[135],"inference.":[136]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-07-01T00:00:00"}
