{"id":"https://openalex.org/W7167032729","doi":"https://doi.org/10.1145/3797905.3807846","title":"CXL-CCL: Inter-Node Collective GPU-Communication Using a CXL Shared Memory Pool","display_name":"CXL-CCL: Inter-Node Collective GPU-Communication Using a CXL Shared Memory Pool","publication_year":2026,"publication_date":"2026-07-02","ids":{"openalex":"https://openalex.org/W7167032729","doi":"https://doi.org/10.1145/3797905.3807846"},"language":null,"primary_location":{"id":"doi:10.1145/3797905.3807846","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3797905.3807846","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/10.1145/3797905.3807846","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5095771051","display_name":"Dong Xu","orcid":"https://orcid.org/0000-0002-4093-628X"},"institutions":[{"id":"https://openalex.org/I156087764","display_name":"University of California, Merced","ror":"https://ror.org/00d9ah105","country_code":"US","type":"education","lineage":["https://openalex.org/I156087764"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Dong Xu","raw_affiliation_strings":["University of California, Merced, Merced, California, USA"],"raw_orcid":"https://orcid.org/0000-0002-4093-628X","affiliations":[{"raw_affiliation_string":"University of California, Merced, Merced, California, USA","institution_ids":["https://openalex.org/I156087764"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139869824","display_name":"Han Meng","orcid":"https://orcid.org/0009-0007-2532-4721"},"institutions":[{"id":"https://openalex.org/I156087764","display_name":"University of California, Merced","ror":"https://ror.org/00d9ah105","country_code":"US","type":"education","lineage":["https://openalex.org/I156087764"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Han Meng","raw_affiliation_strings":["University of California, Merced, Merced, California, USA"],"raw_orcid":"https://orcid.org/0009-0007-2532-4721","affiliations":[{"raw_affiliation_string":"University of California, Merced, Merced, California, USA","institution_ids":["https://openalex.org/I156087764"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5114734440","display_name":"Xinyu Chen","orcid":"https://orcid.org/0009-0004-8774-1617"},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xinyu Chen","raw_affiliation_strings":["Zhejiang University, Hangzhou, Zhejiang, China"],"raw_orcid":"https://orcid.org/0009-0004-8774-1617","affiliations":[{"raw_affiliation_string":"Zhejiang University, Hangzhou, Zhejiang, China","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5127470655","display_name":"Dengcheng Zhu","orcid":null},"institutions":[{"id":"https://openalex.org/I4210091247","display_name":"University for Peace","ror":"https://ror.org/00cxrw595","country_code":"CR","type":"education","lineage":["https://openalex.org/I1286959531","https://openalex.org/I4210091247"]}],"countries":["CR"],"is_corresponding":false,"raw_author_name":"Dengcheng Zhu","raw_affiliation_strings":["Bytedance, San Jose, California, USA"],"raw_orcid":"https://orcid.org/0009-0001-5809-4578","affiliations":[{"raw_affiliation_string":"Bytedance, San Jose, California, USA","institution_ids":["https://openalex.org/I4210091247"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5127013602","display_name":"Wei Tang","orcid":null},"institutions":[{"id":"https://openalex.org/I4210091247","display_name":"University for Peace","ror":"https://ror.org/00cxrw595","country_code":"CR","type":"education","lineage":["https://openalex.org/I1286959531","https://openalex.org/I4210091247"]}],"countries":["CR"],"is_corresponding":false,"raw_author_name":"Wei Tang","raw_affiliation_strings":["Bytedance, San Jose, California, USA"],"raw_orcid":"https://orcid.org/0009-0002-5441-5503","affiliations":[{"raw_affiliation_string":"Bytedance, San Jose, California, USA","institution_ids":["https://openalex.org/I4210091247"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100394591","display_name":"Fei Liu","orcid":"https://orcid.org/0000-0001-9533-3446"},"institutions":[{"id":"https://openalex.org/I4210091247","display_name":"University for Peace","ror":"https://ror.org/00cxrw595","country_code":"CR","type":"education","lineage":["https://openalex.org/I1286959531","https://openalex.org/I4210091247"]}],"countries":["CR"],"is_corresponding":false,"raw_author_name":"Fei Liu","raw_affiliation_strings":["Bytedance, San Jose, California, USA"],"raw_orcid":"https://orcid.org/0000-0001-9533-3446","affiliations":[{"raw_affiliation_string":"Bytedance, San Jose, California, USA","institution_ids":["https://openalex.org/I4210091247"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5040158607","display_name":"Liguang Xie","orcid":"https://orcid.org/0000-0002-3764-9888"},"institutions":[{"id":"https://openalex.org/I4210162608","display_name":"Arzeda (United States)","ror":"https://ror.org/04zw4rj10","country_code":"US","type":"company","lineage":["https://openalex.org/I4210162608"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Liguang Xie","raw_affiliation_strings":["Bytedance technology company Limited, Seattle, Washington, USA"],"raw_orcid":"https://orcid.org/0000-0002-3764-9888","affiliations":[{"raw_affiliation_string":"Bytedance technology company Limited, Seattle, Washington, USA","institution_ids":["https://openalex.org/I4210162608"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101737431","display_name":"Wu Xiang","orcid":"https://orcid.org/0009-0002-6733-0947"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu Xiang","raw_affiliation_strings":["ByteDance Inc., Beijing, China"],"raw_orcid":"https://orcid.org/0009-0002-6733-0947","affiliations":[{"raw_affiliation_string":"ByteDance Inc., Beijing, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101785595","display_name":"Rui Shi","orcid":"https://orcid.org/0009-0003-9122-4703"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rui Shi","raw_affiliation_strings":["Bytedance, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0003-9122-4703","affiliations":[{"raw_affiliation_string":"Bytedance, Beijing, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5127215952","display_name":"Yue Li","orcid":null},"institutions":[{"id":"https://openalex.org/I4210091247","display_name":"University for Peace","ror":"https://ror.org/00cxrw595","country_code":"CR","type":"education","lineage":["https://openalex.org/I1286959531","https://openalex.org/I4210091247"]}],"countries":["CR"],"is_corresponding":false,"raw_author_name":"Yue Li","raw_affiliation_strings":["Bytedance, San Jose, California, USA"],"raw_orcid":"https://orcid.org/0009-0005-1169-7389","affiliations":[{"raw_affiliation_string":"Bytedance, San Jose, California, USA","institution_ids":["https://openalex.org/I4210091247"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5127185765","display_name":"Henry Hu","orcid":null},"institutions":[{"id":"https://openalex.org/I4210091247","display_name":"University for Peace","ror":"https://ror.org/00cxrw595","country_code":"CR","type":"education","lineage":["https://openalex.org/I1286959531","https://openalex.org/I4210091247"]}],"countries":["CR"],"is_corresponding":false,"raw_author_name":"Henry Hu","raw_affiliation_strings":["Bytedance, San Jose, California, USA"],"raw_orcid":"https://orcid.org/0009-0009-1505-2907","affiliations":[{"raw_affiliation_string":"Bytedance, San Jose, California, USA","institution_ids":["https://openalex.org/I4210091247"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5091925126","display_name":"Hui Zhang","orcid":"https://orcid.org/0000-0002-2358-7946"},"institutions":[{"id":"https://openalex.org/I4210091247","display_name":"University for Peace","ror":"https://ror.org/00cxrw595","country_code":"CR","type":"education","lineage":["https://openalex.org/I1286959531","https://openalex.org/I4210091247"]}],"countries":["CR"],"is_corresponding":false,"raw_author_name":"Hui Zhang","raw_affiliation_strings":["ByteDance, San Jose, California, USA"],"raw_orcid":"https://orcid.org/0000-0002-2358-7946","affiliations":[{"raw_affiliation_string":"ByteDance, San Jose, California, USA","institution_ids":["https://openalex.org/I4210091247"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139873048","display_name":"Jianping Jiang","orcid":"https://orcid.org/0009-0009-1750-660X"},"institutions":[{"id":"https://openalex.org/I4210132887","display_name":"III-N Technology (United States)","ror":"https://ror.org/03m9crh17","country_code":"US","type":"company","lineage":["https://openalex.org/I4210132887"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jianping Jiang","raw_affiliation_strings":["Xconn-Tech, San Jose, California, USA"],"raw_orcid":"https://orcid.org/0009-0009-1750-660X","affiliations":[{"raw_affiliation_string":"Xconn-Tech, San Jose, California, USA","institution_ids":["https://openalex.org/I4210132887"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100407387","display_name":"Dong Li","orcid":"https://orcid.org/0000-0001-9336-0694"},"institutions":[{"id":"https://openalex.org/I156087764","display_name":"University of California, Merced","ror":"https://ror.org/00d9ah105","country_code":"US","type":"education","lineage":["https://openalex.org/I156087764"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Dong Li","raw_affiliation_strings":["University of California, Merced, Merced, California, USA"],"raw_orcid":"https://orcid.org/0000-0001-9336-0694","affiliations":[{"raw_affiliation_string":"University of California, Merced, Merced, California, USA","institution_ids":["https://openalex.org/I156087764"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":5,"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":"766","last_page":"779"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.34310001134872437,"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":0.34310001134872437,"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/T10829","display_name":"Interconnection Networks and Systems","score":0.11599999666213989,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.09529999643564224,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/infiniband","display_name":"InfiniBand","score":0.7621999979019165},{"id":"https://openalex.org/keywords/shared-memory","display_name":"Shared memory","score":0.6960999965667725},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.6406999826431274},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.6176000237464905},{"id":"https://openalex.org/keywords/interconnection","display_name":"Interconnection","score":0.47679999470710754},{"id":"https://openalex.org/keywords/memory-management","display_name":"Memory management","score":0.4708999991416931},{"id":"https://openalex.org/keywords/distributed-shared-memory","display_name":"Distributed shared memory","score":0.44850000739097595},{"id":"https://openalex.org/keywords/distributed-memory","display_name":"Distributed memory","score":0.41530001163482666},{"id":"https://openalex.org/keywords/shared-resource","display_name":"Shared resource","score":0.4090000092983246}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8363999724388123},{"id":"https://openalex.org/C2781030343","wikidata":"https://www.wikidata.org/wiki/Q922437","display_name":"InfiniBand","level":2,"score":0.7621999979019165},{"id":"https://openalex.org/C133875982","wikidata":"https://www.wikidata.org/wiki/Q764810","display_name":"Shared memory","level":2,"score":0.6960999965667725},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.6406999826431274},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.6176000237464905},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.5493000149726868},{"id":"https://openalex.org/C123745756","wikidata":"https://www.wikidata.org/wiki/Q1665949","display_name":"Interconnection","level":2,"score":0.47679999470710754},{"id":"https://openalex.org/C176649486","wikidata":"https://www.wikidata.org/wiki/Q2308807","display_name":"Memory management","level":3,"score":0.4708999991416931},{"id":"https://openalex.org/C39528615","wikidata":"https://www.wikidata.org/wiki/Q1229610","display_name":"Distributed shared memory","level":5,"score":0.44850000739097595},{"id":"https://openalex.org/C91481028","wikidata":"https://www.wikidata.org/wiki/Q1054686","display_name":"Distributed memory","level":3,"score":0.41530001163482666},{"id":"https://openalex.org/C51332947","wikidata":"https://www.wikidata.org/wiki/Q1172305","display_name":"Shared resource","level":2,"score":0.4090000092983246},{"id":"https://openalex.org/C67448173","wikidata":"https://www.wikidata.org/wiki/Q254217","display_name":"Collective memory","level":2,"score":0.3853999972343445},{"id":"https://openalex.org/C130795937","wikidata":"https://www.wikidata.org/wiki/Q2561570","display_name":"Remote direct memory access","level":2,"score":0.36250001192092896},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.3594000041484833},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.3582000136375427},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.3215999901294708},{"id":"https://openalex.org/C41036726","wikidata":"https://www.wikidata.org/wiki/Q844824","display_name":"Physical address","level":3,"score":0.3068000078201294},{"id":"https://openalex.org/C74426580","wikidata":"https://www.wikidata.org/wiki/Q719484","display_name":"Memory map","level":3,"score":0.2953000068664551},{"id":"https://openalex.org/C2778915421","wikidata":"https://www.wikidata.org/wiki/Q3643177","display_name":"Performance improvement","level":2,"score":0.28850001096725464},{"id":"https://openalex.org/C93446704","wikidata":"https://www.wikidata.org/wiki/Q449328","display_name":"Registered memory","level":3,"score":0.2867000102996826},{"id":"https://openalex.org/C118702147","wikidata":"https://www.wikidata.org/wiki/Q189396","display_name":"Dynamic random-access memory","level":3,"score":0.2863999903202057},{"id":"https://openalex.org/C51290061","wikidata":"https://www.wikidata.org/wiki/Q1936765","display_name":"Uniform memory access","level":4,"score":0.28209999203681946},{"id":"https://openalex.org/C854659","wikidata":"https://www.wikidata.org/wiki/Q1859284","display_name":"Message passing","level":2,"score":0.27379998564720154},{"id":"https://openalex.org/C12186640","wikidata":"https://www.wikidata.org/wiki/Q6815743","display_name":"Memory model","level":3,"score":0.27309998869895935},{"id":"https://openalex.org/C206345919","wikidata":"https://www.wikidata.org/wiki/Q20380951","display_name":"Resource (disambiguation)","level":2,"score":0.2676999866962433},{"id":"https://openalex.org/C57863822","wikidata":"https://www.wikidata.org/wiki/Q905488","display_name":"Flat memory model","level":4,"score":0.25780001282691956}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3797905.3807846","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3797905.3807846","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:10.1145/3797905.3807846","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3797905.3807846","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":38,"referenced_works":["https://openalex.org/W2419664341","https://openalex.org/W2604595776","https://openalex.org/W2611027062","https://openalex.org/W2615718892","https://openalex.org/W2756531049","https://openalex.org/W2798652486","https://openalex.org/W2891584720","https://openalex.org/W2903504144","https://openalex.org/W2934584470","https://openalex.org/W3046474425","https://openalex.org/W3109760540","https://openalex.org/W3131379896","https://openalex.org/W3157864729","https://openalex.org/W3169244019","https://openalex.org/W3170642333","https://openalex.org/W3204998121","https://openalex.org/W4280587521","https://openalex.org/W4282555269","https://openalex.org/W4310385035","https://openalex.org/W4318541517","https://openalex.org/W4318541522","https://openalex.org/W4320067983","https://openalex.org/W4321446282","https://openalex.org/W4327930439","https://openalex.org/W4387302703","https://openalex.org/W4387321556","https://openalex.org/W4388581411","https://openalex.org/W4391034987","https://openalex.org/W4394923336","https://openalex.org/W4400909502","https://openalex.org/W4401212200","https://openalex.org/W4405755164","https://openalex.org/W4407197341","https://openalex.org/W4408901606","https://openalex.org/W4408903192","https://openalex.org/W4409248382","https://openalex.org/W4411374195","https://openalex.org/W4414976983"],"related_works":[],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2],"(LLMs)":[3],"training":[4,165],"or":[5],"inference":[6],"across":[7,35,114],"multiple":[8,92],"nodes":[9,93],"introduces":[10],"significant":[11],"pressure":[12],"on":[13,63,91],"GPU":[14,59,122],"memory":[15,25,33,54,85,103],"and":[16,39,76,99,146],"interconnect":[17,178],"bandwidth.":[18],"The":[19],"Compute":[20],"Express":[21],"Link":[22],"(CXL)":[23],"shared":[24,53,84],"pool":[26,55,86],"offers":[27],"a":[28,46,95],"scalable":[29],"solution":[30],"by":[31,80,181],"enabling":[32],"sharing":[34],"nodes,":[36],"reducing":[37,177],"over-provisioning":[38],"improving":[40],"resource":[41],"utilization.":[42],"We":[43],"propose":[44],"CXL-CCL,":[45],"collective":[47,88,112],"communication":[48,77],"library,":[49],"leveraging":[50],"the":[51,70,82,153,174],"CXL":[52,83,97],"to":[56,152],"support":[57],"cross-node":[58],"operations":[60,113],"without":[61],"relying":[62],"traditional":[64],"RDMA-based":[65,155],"networking.":[66],"Our":[67,124],"design":[68],"addresses":[69],"challenges":[71],"in":[72],"synchronization,":[73],"data":[74],"interleaving,":[75],"parallelization":[78],"faced":[79],"using":[81],"for":[87,119,136,140,144,149],"communications.":[89],"Evaluating":[90],"with":[94,173],"TITAN-II":[96],"switch":[98],"six":[100],"Micron":[101],"CZ120":[102],"cards,":[104],"we":[105],"show":[106],"that":[107,127],"CXL-CCL":[108,128],"achieves":[109,129],"highly":[110],"efficient":[111],"hosts,":[115],"demonstrating":[116],"CXL\u2019s":[117],"potential":[118],"scalable,":[120],"memory-centric":[121],"communication.":[123],"evaluation":[125],"demonstrates":[126],"average":[130],"performance":[131],"improvements":[132],"of":[133],"1.34":[134],"\u00d7":[135,139,143,148,170],"AllGather,":[137],"1.84":[138],"Broadcast,":[141],"1.94":[142],"Gather,":[145],"1.07":[147],"Scatter,":[150],"compared":[151,172],"original":[154],"implementation":[156],"over":[157],"200":[158],"Gbps":[159],"InfiniBand.":[160],"In":[161],"addition,":[162],"an":[163],"LLM":[164],"case":[166],"study":[167],"shows":[168],"1.11":[169],"speedup":[171],"InfiniBand":[175],"while":[176],"hardware":[179],"cost":[180],"2.75":[182],"\u00d7.":[183]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2026-07-03T00:00:00"}
