{"id":"https://openalex.org/W7167012197","doi":"https://doi.org/10.1145/3797905.3807833","title":"SHIRO: Near-Optimal Communication Strategies for Distributed Sparse Matrix Multiplication","display_name":"SHIRO: Near-Optimal Communication Strategies for Distributed Sparse Matrix Multiplication","publication_year":2026,"publication_date":"2026-07-02","ids":{"openalex":"https://openalex.org/W7167012197","doi":"https://doi.org/10.1145/3797905.3807833"},"language":null,"primary_location":{"id":"doi:10.1145/3797905.3807833","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3797905.3807833","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 40th 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://doi.org/10.1145/3797905.3807833","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5062023820","display_name":"Zhuang Chen","orcid":"https://orcid.org/0009-0006-4156-9879"},"institutions":[{"id":"https://openalex.org/I4210129730","display_name":"RIKEN Center for Computational Science","ror":"https://ror.org/03r519674","country_code":"JP","type":"facility","lineage":["https://openalex.org/I4210110652","https://openalex.org/I4210129730"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Chen Zhuang","raw_affiliation_strings":["Institute of Science Tokyo, Tokyo, Japan and RIKEN Center for Computational Science, Kobe, Japan"],"raw_orcid":"https://orcid.org/0009-0006-4156-9879","affiliations":[{"raw_affiliation_string":"Institute of Science Tokyo, Tokyo, Japan and RIKEN Center for Computational Science, Kobe, Japan","institution_ids":["https://openalex.org/I4210129730"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5078250463","display_name":"Lingqi Zhang","orcid":"https://orcid.org/0000-0002-2452-1551"},"institutions":[{"id":"https://openalex.org/I4210129730","display_name":"RIKEN Center for Computational Science","ror":"https://ror.org/03r519674","country_code":"JP","type":"facility","lineage":["https://openalex.org/I4210110652","https://openalex.org/I4210129730"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Lingqi Zhang","raw_affiliation_strings":["RIKEN Center for Computational Science, Kobe, Japan"],"raw_orcid":"https://orcid.org/0000-0002-2452-1551","affiliations":[{"raw_affiliation_string":"RIKEN Center for Computational Science, Kobe, Japan","institution_ids":["https://openalex.org/I4210129730"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057424239","display_name":"Benjamin Brock","orcid":"https://orcid.org/0000-0003-1488-1622"},"institutions":[{"id":"https://openalex.org/I1343180700","display_name":"Intel (United States)","ror":"https://ror.org/01ek73717","country_code":"US","type":"company","lineage":["https://openalex.org/I1343180700"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Benjamin Brock","raw_affiliation_strings":["Intel Corporation, San Francisco, USA"],"raw_orcid":"https://orcid.org/0000-0003-1488-1622","affiliations":[{"raw_affiliation_string":"Intel Corporation, San Francisco, USA","institution_ids":["https://openalex.org/I1343180700"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051506217","display_name":"Du Wu","orcid":"https://orcid.org/0000-0002-4002-0837"},"institutions":[{"id":"https://openalex.org/I4210129730","display_name":"RIKEN Center for Computational Science","ror":"https://ror.org/03r519674","country_code":"JP","type":"facility","lineage":["https://openalex.org/I4210110652","https://openalex.org/I4210129730"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Du Wu","raw_affiliation_strings":["Institute of Science Tokyo, Tokyo, Japan and RIKEN Center for Computational Science, Kobe, Japan"],"raw_orcid":"https://orcid.org/0000-0002-4002-0837","affiliations":[{"raw_affiliation_string":"Institute of Science Tokyo, Tokyo, Japan and RIKEN Center for Computational Science, Kobe, Japan","institution_ids":["https://openalex.org/I4210129730"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100338479","display_name":"Peng Chen","orcid":"https://orcid.org/0000-0003-1244-3151"},"institutions":[{"id":"https://openalex.org/I4210129730","display_name":"RIKEN Center for Computational Science","ror":"https://ror.org/03r519674","country_code":"JP","type":"facility","lineage":["https://openalex.org/I4210110652","https://openalex.org/I4210129730"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Peng Chen","raw_affiliation_strings":["RIKEN Center for Computational Science, Kobe, Japan"],"raw_orcid":"https://orcid.org/0000-0003-1244-3151","affiliations":[{"raw_affiliation_string":"RIKEN Center for Computational Science, Kobe, Japan","institution_ids":["https://openalex.org/I4210129730"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5011254074","display_name":"Toshio Endo","orcid":"https://orcid.org/0000-0001-7297-6211"},"institutions":[{"id":"https://openalex.org/I4400009020","display_name":"Institute of Science Tokyo","ror":"https://ror.org/05dqf9946","country_code":null,"type":"education","lineage":["https://openalex.org/I4400009020"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Toshio Endo","raw_affiliation_strings":["Institute of Science Tokyo, Tokyo, Japan"],"raw_orcid":"https://orcid.org/0000-0001-7297-6211","affiliations":[{"raw_affiliation_string":"Institute of Science Tokyo, Tokyo, Japan","institution_ids":["https://openalex.org/I4400009020"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100634486","display_name":"Satoshi Matsuoka","orcid":"https://orcid.org/0000-0003-1910-8532"},"institutions":[{"id":"https://openalex.org/I4210129730","display_name":"RIKEN Center for Computational Science","ror":"https://ror.org/03r519674","country_code":"JP","type":"facility","lineage":["https://openalex.org/I4210110652","https://openalex.org/I4210129730"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Satoshi Matsuoka","raw_affiliation_strings":["RIKEN Center for Computational Science, Kobe, Japan and Institute of Science Tokyo, Tokyo, Japan"],"raw_orcid":"https://orcid.org/0000-0003-1910-8532","affiliations":[{"raw_affiliation_string":"RIKEN Center for Computational Science, Kobe, Japan and Institute of Science Tokyo, Tokyo, Japan","institution_ids":["https://openalex.org/I4210129730"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5002208999","display_name":"Mohamed Wahib","orcid":"https://orcid.org/0000-0002-7165-2095"},"institutions":[{"id":"https://openalex.org/I4210129730","display_name":"RIKEN Center for Computational Science","ror":"https://ror.org/03r519674","country_code":"JP","type":"facility","lineage":["https://openalex.org/I4210110652","https://openalex.org/I4210129730"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Mohamed Wahib","raw_affiliation_strings":["RIKEN Center for Computational Science, Kobe, Japan"],"raw_orcid":"https://orcid.org/0000-0002-7165-2095","affiliations":[{"raw_affiliation_string":"RIKEN Center for Computational Science, Kobe, Japan","institution_ids":["https://openalex.org/I4210129730"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"422","last_page":"435"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.4618000090122223,"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":0.4618000090122223,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.17839999496936798,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"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":0.14229999482631683,"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/scalability","display_name":"Scalability","score":0.6851000189781189},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.6593999862670898},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.5311999917030334},{"id":"https://openalex.org/keywords/sparse-matrix","display_name":"Sparse matrix","score":0.5131999850273132},{"id":"https://openalex.org/keywords/matrix-multiplication","display_name":"Matrix multiplication","score":0.49070000648498535},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.44369998574256897},{"id":"https://openalex.org/keywords/distributed-database","display_name":"Distributed database","score":0.39739999175071716},{"id":"https://openalex.org/keywords/inefficiency","display_name":"Inefficiency","score":0.3955000042915344}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8087000250816345},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.6851000189781189},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.6593999862670898},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.5311999917030334},{"id":"https://openalex.org/C56372850","wikidata":"https://www.wikidata.org/wiki/Q1050404","display_name":"Sparse matrix","level":3,"score":0.5131999850273132},{"id":"https://openalex.org/C17349429","wikidata":"https://www.wikidata.org/wiki/Q1049914","display_name":"Matrix multiplication","level":3,"score":0.49070000648498535},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.4544000029563904},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.44369998574256897},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.40400001406669617},{"id":"https://openalex.org/C70061542","wikidata":"https://www.wikidata.org/wiki/Q989016","display_name":"Distributed database","level":2,"score":0.39739999175071716},{"id":"https://openalex.org/C2778869765","wikidata":"https://www.wikidata.org/wiki/Q6028363","display_name":"Inefficiency","level":2,"score":0.3955000042915344},{"id":"https://openalex.org/C106516650","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm design","level":2,"score":0.3804999887943268},{"id":"https://openalex.org/C2780595030","wikidata":"https://www.wikidata.org/wiki/Q3860309","display_name":"Multiplication (music)","level":2,"score":0.3626999855041504},{"id":"https://openalex.org/C192126672","wikidata":"https://www.wikidata.org/wiki/Q1068715","display_name":"Telecommunications network","level":2,"score":0.357699990272522},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.3456999957561493},{"id":"https://openalex.org/C101765175","wikidata":"https://www.wikidata.org/wiki/Q577764","display_name":"Communications system","level":2,"score":0.3395000100135803},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.33149999380111694},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3188999891281128},{"id":"https://openalex.org/C130120984","wikidata":"https://www.wikidata.org/wiki/Q2835898","display_name":"Distributed algorithm","level":2,"score":0.3052000105381012},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.30090001225471497},{"id":"https://openalex.org/C557945733","wikidata":"https://www.wikidata.org/wiki/Q389772","display_name":"Data transmission","level":2,"score":0.28780001401901245},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.25619998574256897},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.25270000100135803},{"id":"https://openalex.org/C138959212","wikidata":"https://www.wikidata.org/wiki/Q1806783","display_name":"Load balancing (electrical power)","level":3,"score":0.2524999976158142},{"id":"https://openalex.org/C42355184","wikidata":"https://www.wikidata.org/wiki/Q1361088","display_name":"Matrix decomposition","level":3,"score":0.25049999356269836}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3797905.3807833","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3797905.3807833","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 40th ACM International Conference on Supercomputing","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3797905.3807833","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3797905.3807833","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 40th ACM International Conference on Supercomputing","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":29,"referenced_works":["https://openalex.org/W1566326249","https://openalex.org/W1986951886","https://openalex.org/W1990238127","https://openalex.org/W2003105834","https://openalex.org/W2035080386","https://openalex.org/W2040425799","https://openalex.org/W2469279958","https://openalex.org/W2483598939","https://openalex.org/W2529031466","https://openalex.org/W2605350416","https://openalex.org/W2759136286","https://openalex.org/W3034899595","https://openalex.org/W3130398772","https://openalex.org/W3169307980","https://openalex.org/W4213060883","https://openalex.org/W4285503871","https://openalex.org/W4316813731","https://openalex.org/W4327911434","https://openalex.org/W4381832157","https://openalex.org/W4391952904","https://openalex.org/W4391986945","https://openalex.org/W4394998401","https://openalex.org/W4395106472","https://openalex.org/W4399304401","https://openalex.org/W4401408827","https://openalex.org/W4405754808","https://openalex.org/W4405754817","https://openalex.org/W4413411877","https://openalex.org/W4413412033"],"related_works":[],"abstract_inverted_index":{"Distributed":[0],"Sparse":[1],"Matrix-Matrix":[2],"Multiplication":[3],"(SpMM)":[4],"is":[5],"a":[6,68,88,115],"fundamental":[7],"operation":[8],"in":[9,20,24,45,114,150],"high-performance":[10],"computing":[11],"and":[12,32,55,86,147,159],"deep":[13],"learning":[14],"applications.":[15],"The":[16],"major":[17],"performance":[18,31],"bottleneck":[19],"distributed":[21,46,117],"SpMM":[22,118,151],"lies":[23],"substantial":[25],"communication":[26,43,71,75,90],"overhead,":[27],"which":[28],"limits":[29],"both":[30],"scalability.":[33],"In":[34],"this":[35,163],"paper,":[36],"we":[37,65],"identify":[38],"two":[39],"key":[40],"sources":[41],"of":[42,58,82,140],"inefficiency":[44],"SpMM:":[47],"redundant":[48,103],"data":[49,104],"transfer":[50],"due":[51],"to":[52,133],"sparsity":[53,80],"unawareness,":[54],"suboptimal":[56],"utilization":[57],"hierarchical":[59,89],"network":[60,100],"topology.":[61],"To":[62],"address":[63],"these,":[64],"propose":[66],"(1)":[67],"fine-grained,":[69],"sparsity-aware":[70,95],"strategy":[72,91,96],"that":[73,92,127],"reduces":[74],"overhead":[76],"by":[77],"exploiting":[78],"the":[79,83,94],"pattern":[81],"sparse":[84],"matrix,":[85],"(2)":[87],"maps":[93],"onto":[97],"two-tier":[98],"GPU":[99],"architectures,":[101],"minimizing":[102],"movement":[105],"across":[106],"slower":[107],"inter-node":[108],"links.":[109],"We":[110],"implement":[111],"these":[112],"optimizations":[113],"comprehensive":[116],"framework,":[119],"SHIRO.":[120],"Extensive":[121],"evaluations":[122],"on":[123],"real-world":[124],"datasets":[125],"show":[126],"SHIRO":[128],"demonstrates":[129],"strong":[130],"scalability":[131],"up":[132],"128":[134],"GPUs,":[135],"achieving":[136],"geometric":[137],"mean":[138],"speedups":[139],"221.5":[141],"\u00d7,":[142,144,146],"56.0":[143],"23.4":[145],"8.8":[148],"\u00d7":[149],"over":[152],"four":[153],"state-of-the-art":[154],"baselines":[155],"(CAGNET,":[156],"SPA,":[157],"BCL,":[158],"CoLa,":[160],"respectively)":[161],"at":[162],"scale.":[164]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2026-07-03T00:00:00"}