{"id":"https://openalex.org/W7128043230","doi":"https://doi.org/10.1145/3748173.3779187","title":"Enabling Efficient SpMM for Sparse Attention on GEMM-Optimized Hardware with Block Aggregation","display_name":"Enabling Efficient SpMM for Sparse Attention on GEMM-Optimized Hardware with Block Aggregation","publication_year":2026,"publication_date":"2026-02-05","ids":{"openalex":"https://openalex.org/W7128043230","doi":"https://doi.org/10.1145/3748173.3779187"},"language":null,"primary_location":{"id":"doi:10.1145/3748173.3779187","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3748173.3779187","pdf_url":null,"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 2026 ACM/SIGDA International Symposium on Field Programmable Gate Arrays","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3748173.3779187","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5125135226","display_name":"Tianchu Ji","orcid":null},"institutions":[{"id":"https://openalex.org/I59553526","display_name":"Stony Brook University","ror":"https://ror.org/05qghxh33","country_code":"US","type":"education","lineage":["https://openalex.org/I59553526"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Tianchu Ji","raw_affiliation_strings":["Stony Brook University, Stony Brook, NY, USA"],"affiliations":[{"raw_affiliation_string":"Stony Brook University, Stony Brook, NY, USA","institution_ids":["https://openalex.org/I59553526"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125247598","display_name":"Niranjan Balasubramanian","orcid":null},"institutions":[{"id":"https://openalex.org/I59553526","display_name":"Stony Brook University","ror":"https://ror.org/05qghxh33","country_code":"US","type":"education","lineage":["https://openalex.org/I59553526"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Niranjan Balasubramanian","raw_affiliation_strings":["Stony Brook University, Stony Brook, NY, USA"],"affiliations":[{"raw_affiliation_string":"Stony Brook University, Stony Brook, NY, USA","institution_ids":["https://openalex.org/I59553526"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5010051312","display_name":"Michael Ferdman","orcid":"https://orcid.org/0000-0001-5808-1040"},"institutions":[{"id":"https://openalex.org/I59553526","display_name":"Stony Brook University","ror":"https://ror.org/05qghxh33","country_code":"US","type":"education","lineage":["https://openalex.org/I59553526"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Michael Ferdman","raw_affiliation_strings":["Stony Brook University, Stony Brook, NY, USA"],"affiliations":[{"raw_affiliation_string":"Stony Brook University, Stony Brook, NY, USA","institution_ids":["https://openalex.org/I59553526"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5125244607","display_name":"Peter Milder","orcid":null},"institutions":[{"id":"https://openalex.org/I59553526","display_name":"Stony Brook University","ror":"https://ror.org/05qghxh33","country_code":"US","type":"education","lineage":["https://openalex.org/I59553526"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Peter Milder","raw_affiliation_strings":["Stony Brook University, Stony Brook, NY, USA"],"affiliations":[{"raw_affiliation_string":"Stony Brook University, Stony Brook, NY, USA","institution_ids":["https://openalex.org/I59553526"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5125135226"],"corresponding_institution_ids":["https://openalex.org/I59553526"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.64310646,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"67","last_page":"78"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.8341000080108643,"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":0.8341000080108643,"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/T10904","display_name":"Embedded Systems Design Techniques","score":0.053199999034404755,"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/T14347","display_name":"Big Data and Digital Economy","score":0.0430000014603138,"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/block","display_name":"Block (permutation group theory)","score":0.6230000257492065},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.5539000034332275},{"id":"https://openalex.org/keywords/sparse-matrix","display_name":"Sparse matrix","score":0.5080000162124634},{"id":"https://openalex.org/keywords/acceleration","display_name":"Acceleration","score":0.45260000228881836},{"id":"https://openalex.org/keywords/data-access","display_name":"Data access","score":0.448199987411499},{"id":"https://openalex.org/keywords/hardware-acceleration","display_name":"Hardware acceleration","score":0.41839998960494995},{"id":"https://openalex.org/keywords/dynamic-data","display_name":"Dynamic data","score":0.34619998931884766},{"id":"https://openalex.org/keywords/direct-memory-access","display_name":"Direct memory access","score":0.29840001463890076}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8079000115394592},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.6230000257492065},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5539000034332275},{"id":"https://openalex.org/C56372850","wikidata":"https://www.wikidata.org/wiki/Q1050404","display_name":"Sparse matrix","level":3,"score":0.5080000162124634},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.45260000228881836},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.44859999418258667},{"id":"https://openalex.org/C47487241","wikidata":"https://www.wikidata.org/wiki/Q5227230","display_name":"Data access","level":2,"score":0.448199987411499},{"id":"https://openalex.org/C13164978","wikidata":"https://www.wikidata.org/wiki/Q600158","display_name":"Hardware acceleration","level":3,"score":0.41839998960494995},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.36570000648498535},{"id":"https://openalex.org/C197298091","wikidata":"https://www.wikidata.org/wiki/Q5318963","display_name":"Dynamic data","level":2,"score":0.34619998931884766},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.33980000019073486},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.32089999318122864},{"id":"https://openalex.org/C37724790","wikidata":"https://www.wikidata.org/wiki/Q210813","display_name":"Direct memory access","level":3,"score":0.29840001463890076},{"id":"https://openalex.org/C106516650","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm design","level":2,"score":0.2906000018119812},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.28769999742507935},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.28139999508857727},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.28060001134872437},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2770000100135803},{"id":"https://openalex.org/C138827492","wikidata":"https://www.wikidata.org/wiki/Q6661985","display_name":"Data processing","level":2,"score":0.2759999930858612},{"id":"https://openalex.org/C194739806","wikidata":"https://www.wikidata.org/wiki/Q66221","display_name":"Computer data storage","level":2,"score":0.26840001344680786},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.26750001311302185},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.26499998569488525},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.2590999901294708},{"id":"https://openalex.org/C118021083","wikidata":"https://www.wikidata.org/wiki/Q610398","display_name":"System on a chip","level":2,"score":0.25699999928474426},{"id":"https://openalex.org/C138958017","wikidata":"https://www.wikidata.org/wiki/Q190087","display_name":"Data type","level":2,"score":0.2533999979496002},{"id":"https://openalex.org/C2985331491","wikidata":"https://www.wikidata.org/wiki/Q5227298","display_name":"Data format","level":2,"score":0.25270000100135803}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3748173.3779187","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3748173.3779187","pdf_url":null,"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 2026 ACM/SIGDA International Symposium on Field Programmable Gate Arrays","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3748173.3779187","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3748173.3779187","pdf_url":null,"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 2026 ACM/SIGDA International Symposium on Field Programmable Gate Arrays","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":47,"referenced_works":["https://openalex.org/W2794952988","https://openalex.org/W2912695634","https://openalex.org/W2946567085","https://openalex.org/W2950656546","https://openalex.org/W2962780877","https://openalex.org/W2970120757","https://openalex.org/W2970777192","https://openalex.org/W3015468748","https://openalex.org/W3016542674","https://openalex.org/W3017024317","https://openalex.org/W3040024858","https://openalex.org/W3129734321","https://openalex.org/W3158831985","https://openalex.org/W3159727696","https://openalex.org/W3173222967","https://openalex.org/W3189877953","https://openalex.org/W3205192296","https://openalex.org/W3207622241","https://openalex.org/W4211147898","https://openalex.org/W4214686755","https://openalex.org/W4220984504","https://openalex.org/W4224010033","https://openalex.org/W4224267386","https://openalex.org/W4285236966","https://openalex.org/W4293023328","https://openalex.org/W4295791054","https://openalex.org/W4296473245","https://openalex.org/W4303443398","https://openalex.org/W4308083827","https://openalex.org/W4316252394","https://openalex.org/W4318540786","https://openalex.org/W4375928954","https://openalex.org/W4380881077","https://openalex.org/W4384918448","https://openalex.org/W4390356856","https://openalex.org/W4390723197","https://openalex.org/W4391455288","https://openalex.org/W4393576971","https://openalex.org/W4393578753","https://openalex.org/W4401212175","https://openalex.org/W4401895034","https://openalex.org/W4402671766","https://openalex.org/W4404133827","https://openalex.org/W4404954358","https://openalex.org/W4407953517","https://openalex.org/W4415308159","https://openalex.org/W4415797534"],"related_works":[],"abstract_inverted_index":{"Rapidly":[0],"growing":[1],"context":[2],"lengths":[3],"have":[4],"amplified":[5],"the":[6,10,19,56],"inherent":[7],"sparsity":[8],"in":[9],"attention":[11,27],"mechanism":[12],"of":[13,61],"popular":[14],"Large":[15],"Language":[16],"Models.":[17],"However,":[18,64],"dynamic":[20,74],"data":[21,34,51,58,75],"access":[22,59],"patterns":[23,60],"required":[24],"by":[25,48],"sparse":[26,62],"are":[28,67],"challenging":[29],"to":[30,37,53,70],"realize":[31],"using":[32],"static":[33],"paths,":[35],"leading":[36],"execution":[38],"inefficiency.":[39],"Existing":[40],"SpMM":[41],"hardware":[42,80],"acceleration":[43],"techniques":[44],"address":[45],"these":[46,65],"inefficiencies":[47],"dynamically":[49],"configuring":[50],"paths":[52,76],"align":[54],"with":[55],"unstructured":[57],"attention.":[63],"approaches":[66],"not":[68],"applicable":[69],"GEMM-optimized":[71],"hardware,":[72],"where":[73],"would":[77],"introduce":[78],"unacceptable":[79],"complexity":[81],"and":[82],"frequency":[83],"degradation.":[84]},"counts_by_year":[],"updated_date":"2026-02-07T06:15:42.627816","created_date":"2026-02-07T00:00:00"}