{"id":"https://openalex.org/W4405756373","doi":"https://doi.org/10.1109/sc41406.2024.00109","title":"Network-Offloaded Bandwidth-Optimal Broadcast and Allgather for Distributed AI","display_name":"Network-Offloaded Bandwidth-Optimal Broadcast and Allgather for Distributed AI","publication_year":2024,"publication_date":"2024-11-17","ids":{"openalex":"https://openalex.org/W4405756373","doi":"https://doi.org/10.1109/sc41406.2024.00109"},"language":"en","primary_location":{"id":"doi:10.1109/sc41406.2024.00109","is_oa":false,"landing_page_url":"https://doi.org/10.1109/sc41406.2024.00109","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"SC24: International Conference for High Performance Computing, Networking, Storage and Analysis","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5077395612","display_name":"Mikhail Khalilov","orcid":"https://orcid.org/0000-0003-0862-4662"},"institutions":[{"id":"https://openalex.org/I35440088","display_name":"ETH Zurich","ror":"https://ror.org/05a28rw58","country_code":"CH","type":"education","lineage":["https://openalex.org/I2799323385","https://openalex.org/I35440088"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Mikhail Khalilov","raw_affiliation_strings":["ETH Zurich,Department of Computer Science,Zurich,Switzerland"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ETH Zurich,Department of Computer Science,Zurich,Switzerland","institution_ids":["https://openalex.org/I35440088"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5019905167","display_name":"Salvatore Di Girolamo","orcid":"https://orcid.org/0000-0003-2197-8860"},"institutions":[{"id":"https://openalex.org/I3039010084","display_name":"Tamedia (Switzerland)","ror":"https://ror.org/00bfx7n30","country_code":"CH","type":"company","lineage":["https://openalex.org/I3039010084"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Salvatore Di Girolamo","raw_affiliation_strings":["NVIDIA Corporation,Zurich,Switzerland"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NVIDIA Corporation,Zurich,Switzerland","institution_ids":["https://openalex.org/I3039010084"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5001468122","display_name":"Marcin Chrapek","orcid":"https://orcid.org/0009-0009-7654-6038"},"institutions":[{"id":"https://openalex.org/I35440088","display_name":"ETH Zurich","ror":"https://ror.org/05a28rw58","country_code":"CH","type":"education","lineage":["https://openalex.org/I2799323385","https://openalex.org/I35440088"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Marcin Chrapek","raw_affiliation_strings":["ETH Zurich,Department of Computer Science,Zurich,Switzerland"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ETH Zurich,Department of Computer Science,Zurich,Switzerland","institution_ids":["https://openalex.org/I35440088"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5107437411","display_name":"Rami Nudelman","orcid":"https://orcid.org/0009-0002-3601-2786"},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Rami Nudelman","raw_affiliation_strings":["NVIDIA Corporation,Santa Clara,United States"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NVIDIA Corporation,Santa Clara,United States","institution_ids":["https://openalex.org/I4210127875"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048095854","display_name":"Gil Bloch","orcid":"https://orcid.org/0009-0004-6224-9802"},"institutions":[{"id":"https://openalex.org/I80687555","display_name":"Israel Electric (Israel)","ror":"https://ror.org/01p8dnv11","country_code":"IL","type":"company","lineage":["https://openalex.org/I80687555"]}],"countries":["IL"],"is_corresponding":false,"raw_author_name":"Gil Bloch","raw_affiliation_strings":["NVIDIA Corporation,Yokne’am Illit,Israel"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NVIDIA Corporation,Yokne’am Illit,Israel","institution_ids":["https://openalex.org/I80687555"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5026990786","display_name":"Torsten Hoefler","orcid":"https://orcid.org/0000-0002-1333-9797"},"institutions":[{"id":"https://openalex.org/I35440088","display_name":"ETH Zurich","ror":"https://ror.org/05a28rw58","country_code":"CH","type":"education","lineage":["https://openalex.org/I2799323385","https://openalex.org/I35440088"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Torsten Hoefler","raw_affiliation_strings":["ETH Zurich,Department of Computer Science,Zurich,Switzerland"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ETH Zurich,Department of Computer Science,Zurich,Switzerland","institution_ids":["https://openalex.org/I35440088"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":1.7045,"has_fulltext":false,"cited_by_count":6,"citation_normalized_percentile":{"value":0.86008257,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":98,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"17"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10273","display_name":"IoT and Edge/Fog Computing","score":0.951200008392334,"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/T10273","display_name":"IoT and Edge/Fog Computing","score":0.951200008392334,"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":0.7380815744400024},{"id":"https://openalex.org/keywords/bandwidth","display_name":"Bandwidth (computing)","score":0.7279503345489502},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.6075126528739929}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7380815744400024},{"id":"https://openalex.org/C2776257435","wikidata":"https://www.wikidata.org/wiki/Q1576430","display_name":"Bandwidth (computing)","level":2,"score":0.7279503345489502},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.6075126528739929}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/sc41406.2024.00109","is_oa":false,"landing_page_url":"https://doi.org/10.1109/sc41406.2024.00109","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"SC24: International Conference for High Performance Computing, Networking, Storage and Analysis","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/7","display_name":"Affordable and clean energy","score":0.5099999904632568}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":42,"referenced_works":["https://openalex.org/W1558115001","https://openalex.org/W1999985771","https://openalex.org/W2000882268","https://openalex.org/W2002037063","https://openalex.org/W2042120213","https://openalex.org/W2062764889","https://openalex.org/W2068020713","https://openalex.org/W2130531694","https://openalex.org/W2132777761","https://openalex.org/W2151194859","https://openalex.org/W2170796499","https://openalex.org/W2745034467","https://openalex.org/W2761028280","https://openalex.org/W2784240842","https://openalex.org/W2889463335","https://openalex.org/W2926767350","https://openalex.org/W2963198032","https://openalex.org/W2972087877","https://openalex.org/W3081168214","https://openalex.org/W3086105743","https://openalex.org/W3129251660","https://openalex.org/W3129831491","https://openalex.org/W3155879802","https://openalex.org/W3165940308","https://openalex.org/W3193181769","https://openalex.org/W3195540550","https://openalex.org/W3205377733","https://openalex.org/W4242547806","https://openalex.org/W4242634435","https://openalex.org/W4247434968","https://openalex.org/W4308995130","https://openalex.org/W4321636619","https://openalex.org/W4382931639","https://openalex.org/W4384705383","https://openalex.org/W4384705453","https://openalex.org/W4386768656","https://openalex.org/W4401176521","https://openalex.org/W6684980599","https://openalex.org/W6716975455","https://openalex.org/W6743330237","https://openalex.org/W6764726765","https://openalex.org/W6861513570"],"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":{"In":[0,19],"the":[1,16,32,89,94,102,117,133],"Fully":[2],"Sharded":[3],"Data":[4],"Parallel":[5],"(FSDP)":[6],"training":[7],"pipeline,":[8],"collective":[9,49,127],"operations":[10,23],"can":[11,29,130],"be":[12],"interleaved":[13],"to":[14,58,111,132],"maximize":[15],"communication/computation":[17],"overlap.":[18],"this":[20,41],"scenario,":[21],"outstanding":[22],"such":[24],"as":[25],"Allgather":[26,48,72,75,106],"and":[27,35,108],"Reduce-Scatter":[28],"compete":[30],"for":[31,68,115],"injection":[33],"bandwidth":[34],"create":[36],"pipeline":[37],"bubbles.":[38],"To":[39,87],"address":[40],"problem,":[42],"we":[43,96],"propose":[44],"a":[45,60,65,83,112],"novel":[46],"bandwidth-optimal":[47],"algorithm":[50,76,107],"that":[51,124],"leverages":[52],"hardware":[53],"multicast.":[54],"We":[55,100,122],"use":[56],"multicast":[57],"build":[59],"constant-time":[61],"reliable":[62],"Broadcast":[63],"protocol,":[64,95],"building":[66],"block":[67],"constructing":[69],"an":[70],"optimal":[71],"schedule.":[73],"Our":[74],"achieves":[77],"$2":[78],"\\times$":[79],"traffic":[80],"reduction":[81],"on":[82],"188":[84],"-node":[85],"testbed.":[86],"free":[88],"host":[90],"side":[91],"from":[92],"running":[93],"employ":[97],"SmartNIC":[98,113],"offloading.":[99],"extract":[101],"parallelism":[103],"in":[104],"our":[105,125],"map":[109],"it":[110],"specialized":[114],"hiding":[116],"cost":[118],"of":[119,136],"data":[120],"movement.":[121],"show":[123],"SmartNIC-offloaded":[126],"progress":[128],"engine":[129],"scale":[131],"next":[134],"generation":[135],"1.6":[137],"Tbit/s":[138],"links.":[139]},"counts_by_year":[{"year":2025,"cited_by_count":6}],"updated_date":"2025-12-21T01:58:51.020947","created_date":"2025-10-10T00:00:00"}