{"id":"https://openalex.org/W7125271546","doi":"https://doi.org/10.1109/cloudcom67567.2025.11331396","title":"Vectorization of Narrow Matrix Multiplication for Ascend AI Inference Acceleration","display_name":"Vectorization of Narrow Matrix Multiplication for Ascend AI Inference Acceleration","publication_year":2025,"publication_date":"2025-11-14","ids":{"openalex":"https://openalex.org/W7125271546","doi":"https://doi.org/10.1109/cloudcom67567.2025.11331396"},"language":null,"primary_location":{"id":"doi:10.1109/cloudcom67567.2025.11331396","is_oa":false,"landing_page_url":"https://doi.org/10.1109/cloudcom67567.2025.11331396","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 lEEE International Conference on Cloud Computing Technology and Science (CloudCom)","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/A5123483621","display_name":"Shurygin Anton","orcid":null},"institutions":[{"id":"https://openalex.org/I153845743","display_name":"Moscow Institute of Physics and Technology","ror":"https://ror.org/00v0z9322","country_code":"RU","type":"education","lineage":["https://openalex.org/I153845743"]}],"countries":["RU"],"is_corresponding":true,"raw_author_name":"Shurygin Anton","raw_affiliation_strings":["Moscow Institute of Physics and Technology,Moscow,Russia"],"affiliations":[{"raw_affiliation_string":"Moscow Institute of Physics and Technology,Moscow,Russia","institution_ids":["https://openalex.org/I153845743"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5123468573","display_name":"Frolov Aleksandr","orcid":null},"institutions":[{"id":"https://openalex.org/I153845743","display_name":"Moscow Institute of Physics and Technology","ror":"https://ror.org/00v0z9322","country_code":"RU","type":"education","lineage":["https://openalex.org/I153845743"]}],"countries":["RU"],"is_corresponding":false,"raw_author_name":"Frolov Aleksandr","raw_affiliation_strings":["Moscow Institute of Physics and Technology,Moscow,Russia"],"affiliations":[{"raw_affiliation_string":"Moscow Institute of Physics and Technology,Moscow,Russia","institution_ids":["https://openalex.org/I153845743"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5123483621"],"corresponding_institution_ids":["https://openalex.org/I153845743"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.76291374,"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":"7"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.6636000275611877,"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":0.6636000275611877,"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":0.04690000042319298,"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":0.040800001472234726,"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/vectorization","display_name":"Vectorization (mathematics)","score":0.7091000080108643},{"id":"https://openalex.org/keywords/multiplication","display_name":"Multiplication (music)","score":0.6394000053405762},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.5153999924659729},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.5123000144958496},{"id":"https://openalex.org/keywords/matrix-multiplication","display_name":"Matrix multiplication","score":0.4950999915599823},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.44859999418258667},{"id":"https://openalex.org/keywords/matrix","display_name":"Matrix (chemical analysis)","score":0.426800012588501},{"id":"https://openalex.org/keywords/operator","display_name":"Operator (biology)","score":0.4104999899864197}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7339000105857849},{"id":"https://openalex.org/C41681595","wikidata":"https://www.wikidata.org/wiki/Q7917855","display_name":"Vectorization (mathematics)","level":2,"score":0.7091000080108643},{"id":"https://openalex.org/C2780595030","wikidata":"https://www.wikidata.org/wiki/Q3860309","display_name":"Multiplication (music)","level":2,"score":0.6394000053405762},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.5153999924659729},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.5123000144958496},{"id":"https://openalex.org/C17349429","wikidata":"https://www.wikidata.org/wiki/Q1049914","display_name":"Matrix multiplication","level":3,"score":0.4950999915599823},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.47999998927116394},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.44859999418258667},{"id":"https://openalex.org/C106487976","wikidata":"https://www.wikidata.org/wiki/Q685816","display_name":"Matrix (chemical analysis)","level":2,"score":0.426800012588501},{"id":"https://openalex.org/C17020691","wikidata":"https://www.wikidata.org/wiki/Q139677","display_name":"Operator (biology)","level":5,"score":0.4104999899864197},{"id":"https://openalex.org/C16320812","wikidata":"https://www.wikidata.org/wiki/Q1812200","display_name":"Idle","level":2,"score":0.38359999656677246},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.3659999966621399},{"id":"https://openalex.org/C122637931","wikidata":"https://www.wikidata.org/wiki/Q118084","display_name":"Unit (ring theory)","level":2,"score":0.361299991607666},{"id":"https://openalex.org/C134765980","wikidata":"https://www.wikidata.org/wiki/Q879126","display_name":"Bitwise operation","level":2,"score":0.3441999852657318},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.3199000060558319},{"id":"https://openalex.org/C53051483","wikidata":"https://www.wikidata.org/wiki/Q861555","display_name":"Cube (algebra)","level":2,"score":0.3149000108242035},{"id":"https://openalex.org/C60798267","wikidata":"https://www.wikidata.org/wiki/Q1226939","display_name":"Division (mathematics)","level":2,"score":0.30889999866485596},{"id":"https://openalex.org/C94375191","wikidata":"https://www.wikidata.org/wiki/Q11205","display_name":"Arithmetic","level":1,"score":0.3050000071525574},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.30090001225471497},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.2985000014305115},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.288100004196167},{"id":"https://openalex.org/C84316537","wikidata":"https://www.wikidata.org/wiki/Q36255","display_name":"Unit vector","level":2,"score":0.26829999685287476}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/cloudcom67567.2025.11331396","is_oa":false,"landing_page_url":"https://doi.org/10.1109/cloudcom67567.2025.11331396","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 lEEE International Conference on Cloud Computing Technology and Science (CloudCom)","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":5,"referenced_works":["https://openalex.org/W2953212265","https://openalex.org/W3157657667","https://openalex.org/W3164200338","https://openalex.org/W4246677555","https://openalex.org/W4385245566"],"related_works":[],"abstract_inverted_index":{"This":[0],"research":[1],"proposes":[2],"and":[3,77,90,128],"evaluates":[4],"a":[5,19,96,103,111],"novel":[6],"approach":[7],"to":[8,61,69,79],"optimizing":[9],"matrix":[10],"multiplication":[11,24],"(MatMul)":[12],"on":[13,117],"Huawei":[14],"Ascend":[15],"NPUs,":[16],"motivated":[17],"by":[18],"key":[20],"insight:":[21],"during":[22],"matrix-vector":[23],"(narrow":[25],"MatMul),":[26],"the":[27,35,44,52,66,70,81,115,124,129],"Cube":[28,67],"Unit":[29,37,68],"(AIC)":[30],"is":[31],"often":[32],"underutilized,":[33],"while":[34],"Vector":[36,71],"(AIV)":[38],"remains":[39],"idle":[40],"for":[41,102,121],"most":[42],"of":[43,59,83,100,126],"operator":[45],"runtime.":[46],"In":[47],"this":[48],"paper,":[49],"we":[50],"introduce":[51],"MatMul":[53],"algorithm,":[54],"which":[55],"uses":[56],"vector":[57],"instructions":[58],"AscendC":[60],"effectively":[62],"offload":[63],"computations":[64],"from":[65],"Unit.":[72],"The":[73],"algorithm":[74],"was":[75],"tested":[76],"applied":[78],"accelerating":[80],"inference":[82],"MLA":[84],"DeepSeek-V3operator.":[85],"By":[86],"successfully":[87],"overlapping":[88],"AIV":[89],"AIC":[91],"computations,":[92],"our":[93],"optimization":[94,119],"showed":[95],"mean":[97],"performance":[98],"gain":[99],"20%":[101],"single":[104],"token":[105],"processing":[106],"scenario.":[107],"Our":[108],"work":[109],"addresses":[110],"significant":[112],"gap":[113],"in":[114],"literature":[116],"practical":[118],"techniques":[120],"AscendC,":[122],"despite":[123],"availability":[125],"documentation":[127],"active":[130],"CANN":[131],"community.":[132]},"counts_by_year":[],"updated_date":"2026-01-22T23:33:04.759266","created_date":"2026-01-22T00:00:00"}