{"id":"https://openalex.org/W4395073435","doi":"https://doi.org/10.1145/3620666.3651362","title":"TCCL: Discovering Better Communication Paths for PCIe GPU Clusters","display_name":"TCCL: Discovering Better Communication Paths for PCIe GPU Clusters","publication_year":2024,"publication_date":"2024-04-24","ids":{"openalex":"https://openalex.org/W4395073435","doi":"https://doi.org/10.1145/3620666.3651362"},"language":"en","primary_location":{"id":"doi:10.1145/3620666.3651362","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3620666.3651362","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 29th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 3","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3620666.3651362","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5057343217","display_name":"Heehoon Kim","orcid":"https://orcid.org/0000-0002-5392-4413"},"institutions":[{"id":"https://openalex.org/I139264467","display_name":"Seoul National University","ror":"https://ror.org/04h9pn542","country_code":"KR","type":"education","lineage":["https://openalex.org/I139264467"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Heehoon Kim","raw_affiliation_strings":["Seoul National University, Seoul, Korea, South ? Republic of Korea"],"raw_orcid":"https://orcid.org/0000-0002-5392-4413","affiliations":[{"raw_affiliation_string":"Seoul National University, Seoul, Korea, South ? Republic of Korea","institution_ids":["https://openalex.org/I139264467"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5032058183","display_name":"Junyeol Ryu","orcid":"https://orcid.org/0009-0004-8788-7405"},"institutions":[{"id":"https://openalex.org/I139264467","display_name":"Seoul National University","ror":"https://ror.org/04h9pn542","country_code":"KR","type":"education","lineage":["https://openalex.org/I139264467"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Junyeol Ryu","raw_affiliation_strings":["Seoul National University, Seoul, Republic of Korea"],"raw_orcid":"https://orcid.org/0009-0004-8788-7405","affiliations":[{"raw_affiliation_string":"Seoul National University, Seoul, Republic of Korea","institution_ids":["https://openalex.org/I139264467"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100767182","display_name":"Jaejin Lee","orcid":"https://orcid.org/0000-0003-4638-8170"},"institutions":[{"id":"https://openalex.org/I139264467","display_name":"Seoul National University","ror":"https://ror.org/04h9pn542","country_code":"KR","type":"education","lineage":["https://openalex.org/I139264467"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jaejin Lee","raw_affiliation_strings":["Seoul National University, Seoul, Korea, South ? Republic of Korea"],"raw_orcid":"https://orcid.org/0000-0003-4638-8170","affiliations":[{"raw_affiliation_string":"Seoul National University, Seoul, Korea, South ? Republic of Korea","institution_ids":["https://openalex.org/I139264467"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I139264467"],"apc_list":null,"apc_paid":null,"fwci":12.6908,"has_fulltext":false,"cited_by_count":25,"citation_normalized_percentile":{"value":0.99351068,"is_in_top_1_percent":true,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":90,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"999","last_page":"1015"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11181","display_name":"Advanced Data Storage Technologies","score":0.9995999932289124,"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/T11181","display_name":"Advanced Data Storage Technologies","score":0.9995999932289124,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9958000183105469,"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/T12923","display_name":"Digital Image Processing Techniques","score":0.9943000078201294,"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/pci-express","display_name":"PCI Express","score":0.9415134787559509},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7568392753601074},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.5494205951690674},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.2770370841026306},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.1442669928073883}],"concepts":[{"id":"https://openalex.org/C64270927","wikidata":"https://www.wikidata.org/wiki/Q206924","display_name":"PCI Express","level":3,"score":0.9415134787559509},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7568392753601074},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.5494205951690674},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.2770370841026306},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.1442669928073883}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3620666.3651362","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3620666.3651362","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 29th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 3","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3620666.3651362","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3620666.3651362","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 29th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 3","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G2045000886","display_name":null,"funder_award_id":"RS-2023-00222663","funder_id":"https://openalex.org/F4320322120","funder_display_name":"National Research Foundation of Korea"}],"funders":[{"id":"https://openalex.org/F4320322120","display_name":"National Research Foundation of Korea","ror":"https://ror.org/013aysd81"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":36,"referenced_works":["https://openalex.org/W60432105","https://openalex.org/W95608104","https://openalex.org/W1518038022","https://openalex.org/W1637731592","https://openalex.org/W1825216778","https://openalex.org/W2059300917","https://openalex.org/W2131613942","https://openalex.org/W2160054705","https://openalex.org/W2734941459","https://openalex.org/W2753239506","https://openalex.org/W2867345499","https://openalex.org/W2901299405","https://openalex.org/W2910100551","https://openalex.org/W2913954081","https://openalex.org/W2922527104","https://openalex.org/W2926767350","https://openalex.org/W2930869794","https://openalex.org/W2970971581","https://openalex.org/W3072623287","https://openalex.org/W3101708369","https://openalex.org/W3129488589","https://openalex.org/W3129831491","https://openalex.org/W3137135510","https://openalex.org/W3188065709","https://openalex.org/W3193985311","https://openalex.org/W3204998121","https://openalex.org/W3206314279","https://openalex.org/W4280611847","https://openalex.org/W4285071004","https://openalex.org/W4285092328","https://openalex.org/W4308995130","https://openalex.org/W4312060029","https://openalex.org/W4360831831","https://openalex.org/W4393406951","https://openalex.org/W6794216301","https://openalex.org/W7067822191"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W4385894176","https://openalex.org/W2347371119","https://openalex.org/W2612768808","https://openalex.org/W3131402800","https://openalex.org/W2388965158","https://openalex.org/W2366201982","https://openalex.org/W2808480228"],"abstract_inverted_index":{"Exploiting":[0],"parallelism":[1,139],"to":[2,9,64,75,90,118],"train":[3],"deep":[4,95],"learning":[5],"models":[6],"requires":[7],"GPUs":[8,30],"cooperate":[10],"through":[11,33],"collective":[12,44,85],"communication":[13,45,67,86,92,122],"primitives.":[14],"While":[15],"systems":[16,28],"like":[17],"DGX,":[18],"equipped":[19],"with":[20,131,135],"proprietary":[21],"interconnects,":[22],"have":[23,35],"been":[24],"extensively":[25],"studied,":[26],"the":[27,77,120],"where":[29],"mainly":[31],"communicate":[32],"PCIe":[34],"received":[36],"limited":[37],"attention.":[38],"This":[39],"paper":[40],"introduces":[41],"TCCL,":[42],"a":[43,56,62,70],"library":[46],"designed":[47],"explicitly":[48],"for":[49,58],"such":[50,97],"systems.":[51],"TCCL":[52,106,115,130],"has":[53],"three":[54,108],"components:":[55],"profiler":[57],"multi-transfer":[59],"performance":[60],"measurement,":[61],"pathfinder":[63],"discover":[65],"optimal":[66],"paths,":[68],"and":[69,100,125],"modified":[71],"runtime":[72],"of":[73,105,138],"NCCL":[74,124],"utilize":[76],"identified":[78],"paths.":[79],"The":[80,102],"focus":[81],"is":[82],"on":[83,107],"ring-based":[84],"algorithms":[87],"that":[88,114],"apply":[89],"popular":[91],"operations":[93],"in":[94],"learning,":[96],"as":[98],"AllReduce":[99],"AllGather.":[101],"evaluation":[103],"results":[104],"different":[109],"PCIe-dependent":[110],"GPU":[111],"clusters":[112],"show":[113],"outperforms":[116],"(up":[117],"\u00d72.07)":[119],"state-of-the-art":[121],"libraries,":[123],"MSCCL.":[126],"We":[127],"also":[128],"evaluate":[129],"DL":[132],"training":[133],"workloads":[134],"various":[136],"combinations":[137],"types.":[140]},"counts_by_year":[{"year":2026,"cited_by_count":16},{"year":2025,"cited_by_count":8},{"year":2024,"cited_by_count":1}],"updated_date":"2026-08-22T07:34:49.880490","created_date":"2025-10-10T00:00:00"}