{"id":"https://openalex.org/W4400682098","doi":"https://doi.org/10.1145/3672198.3673804","title":"OmNICCL: Zero-cost Sparse AllReduce with Direct Cache Access and SmartNICs","display_name":"OmNICCL: Zero-cost Sparse AllReduce with Direct Cache Access and SmartNICs","publication_year":2024,"publication_date":"2024-07-16","ids":{"openalex":"https://openalex.org/W4400682098","doi":"https://doi.org/10.1145/3672198.3673804"},"language":"en","primary_location":{"id":"doi:10.1145/3672198.3673804","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3672198.3673804","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3672198.3673804?download=true","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 2024 SIGCOMM Workshop on Networks for AI Computing","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3672198.3673804?download=true","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5104601957","display_name":"Tongzhou Gu","orcid":"https://orcid.org/0000-0002-1401-9170"},"institutions":[{"id":"https://openalex.org/I4210099236","display_name":"Kootenay Association for Science & Technology","ror":"https://ror.org/011pv9p44","country_code":"CA","type":"nonprofit","lineage":["https://openalex.org/I4210099236"]},{"id":"https://openalex.org/I71920554","display_name":"King Abdullah University of Science and Technology","ror":"https://ror.org/01q3tbs38","country_code":"SA","type":"education","lineage":["https://openalex.org/I71920554"]}],"countries":["CA","SA"],"is_corresponding":false,"raw_author_name":"Tongzhou Gu","raw_affiliation_strings":["KAUST"],"raw_orcid":"https://orcid.org/0000-0002-1401-9170","affiliations":[{"raw_affiliation_string":"KAUST","institution_ids":["https://openalex.org/I4210099236","https://openalex.org/I71920554"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5014208203","display_name":"Jiawei Fei","orcid":"https://orcid.org/0000-0001-9325-0516"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiawei Fei","raw_affiliation_strings":["NUDT"],"raw_orcid":"https://orcid.org/0000-0001-9325-0516","affiliations":[{"raw_affiliation_string":"NUDT","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5042255975","display_name":"Marco Canini","orcid":"https://orcid.org/0000-0002-5051-4283"},"institutions":[{"id":"https://openalex.org/I4210099236","display_name":"Kootenay Association for Science & Technology","ror":"https://ror.org/011pv9p44","country_code":"CA","type":"nonprofit","lineage":["https://openalex.org/I4210099236"]},{"id":"https://openalex.org/I71920554","display_name":"King Abdullah University of Science and Technology","ror":"https://ror.org/01q3tbs38","country_code":"SA","type":"education","lineage":["https://openalex.org/I71920554"]}],"countries":["CA","SA"],"is_corresponding":false,"raw_author_name":"Marco Canini","raw_affiliation_strings":["KAUST"],"raw_orcid":"https://orcid.org/0000-0002-5051-4283","affiliations":[{"raw_affiliation_string":"KAUST","institution_ids":["https://openalex.org/I4210099236","https://openalex.org/I71920554"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.3122,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.56728115,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":91,"max":95},"biblio":{"volume":null,"issue":null,"first_page":"75","last_page":"83"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11478","display_name":"Caching and Content Delivery","score":0.9993000030517578,"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"}},"topics":[{"id":"https://openalex.org/T11478","display_name":"Caching and Content Delivery","score":0.9993000030517578,"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/T11181","display_name":"Advanced Data Storage Technologies","score":0.9988999962806702,"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/T10772","display_name":"Distributed systems and fault tolerance","score":0.9976999759674072,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6759271621704102},{"id":"https://openalex.org/keywords/cache","display_name":"Cache","score":0.6263315677642822},{"id":"https://openalex.org/keywords/zero","display_name":"Zero (linguistics)","score":0.46313148736953735},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.4306317865848541}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6759271621704102},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.6263315677642822},{"id":"https://openalex.org/C2780813799","wikidata":"https://www.wikidata.org/wiki/Q3274237","display_name":"Zero (linguistics)","level":2,"score":0.46313148736953735},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.4306317865848541},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3672198.3673804","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3672198.3673804","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3672198.3673804?download=true","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 2024 SIGCOMM Workshop on Networks for AI Computing","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3672198.3673804","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3672198.3673804","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3672198.3673804?download=true","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 2024 SIGCOMM Workshop on Networks for AI Computing","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G2345396535","display_name":null,"funder_award_id":"CRG2020","funder_id":"https://openalex.org/F4320322320","funder_display_name":"King Abdullah University of Science and Technology"}],"funders":[{"id":"https://openalex.org/F4320322320","display_name":"King Abdullah University of Science and Technology","ror":"https://ror.org/01q3tbs38"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4400682098.pdf","grobid_xml":"https://content.openalex.org/works/W4400682098.grobid-xml"},"referenced_works_count":24,"referenced_works":["https://openalex.org/W1572016165","https://openalex.org/W1825216778","https://openalex.org/W2133654146","https://openalex.org/W2143295630","https://openalex.org/W2769986458","https://openalex.org/W2968450015","https://openalex.org/W2985025837","https://openalex.org/W2985108934","https://openalex.org/W3013190171","https://openalex.org/W3083442288","https://openalex.org/W3086105743","https://openalex.org/W3091097978","https://openalex.org/W3165940308","https://openalex.org/W3193034122","https://openalex.org/W3206778761","https://openalex.org/W4206204369","https://openalex.org/W4210645050","https://openalex.org/W4242589337","https://openalex.org/W4242634435","https://openalex.org/W4289401659","https://openalex.org/W4297097375","https://openalex.org/W4327911839","https://openalex.org/W4396814849","https://openalex.org/W4401352466"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"AllReduce":[0,48,91],"is":[1],"a":[2,34,85,94],"collective":[3],"communication":[4],"pattern":[5],"commonly":[6],"used":[7],"in":[8],"Distributed":[9],"Deep":[10],"Learning":[11],"(DDL)":[12],"and":[13,67,98],"High":[14],"Performance":[15],"Computing":[16],"(HPC).":[17],"Sparse":[18],"AllReduce,":[19],"which":[20],"compresses":[21],"the":[22,41,73],"data":[23,74],"transmitted,":[24],"achieves":[25],"significant":[26],"acceleration":[27],"on":[28,72],"specific":[29],"workloads.":[30],"However,":[31],"compression":[32,66],"introduces":[33],"non-negligible":[35],"performance":[36,100],"overhead.":[37],"Therefore,":[38],"we":[39],"propose":[40],"OmNICreduce":[42],"algorithm,":[43],"an":[44],"efficient":[45],"inter-node":[46],"sparse":[47,107],"method,":[49],"as":[50,52],"well":[51],"its":[53],"implementation,":[54],"OmNICCL.":[55],"It":[56],"utilizes":[57],"Direct":[58],"Cache":[59],"Access":[60],"(DCA)":[61],"to":[62,84],"achieve":[63],"zero-overhead":[64],"lossless":[65],"employs":[68],"SmartNICs":[69],"for":[70],"aggregation":[71],"plane.":[75],"We":[76],"demonstrate":[77],"that":[78],"our":[79],"method":[80],"can":[81],"provide":[82],"up":[83],"7.24\u00d7":[86],"speedup":[87],"over":[88,102],"conventional":[89],"dense":[90],"methods":[92],"under":[93],"100Gbps":[95],"RoCEv2":[96],"network":[97],"1.76-17.37\u00d7":[99],"improvement":[101],"state-of-the-art":[103],"implementations":[104],"when":[105],"performing":[106],"AllReduce.":[108]},"counts_by_year":[{"year":2025,"cited_by_count":1}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}
