{"id":"https://openalex.org/W4396814797","doi":"https://doi.org/10.23919/isc.2024.10528931","title":"Accelerating MPI AllReduce Communication with Efficient GPU-Based Compression Schemes on Modern GPU Clusters","display_name":"Accelerating MPI AllReduce Communication with Efficient GPU-Based Compression Schemes on Modern GPU Clusters","publication_year":2024,"publication_date":"2024-05-01","ids":{"openalex":"https://openalex.org/W4396814797","doi":"https://doi.org/10.23919/isc.2024.10528931"},"language":"en","primary_location":{"id":"doi:10.23919/isc.2024.10528931","is_oa":false,"landing_page_url":"https://doi.org/10.23919/isc.2024.10528931","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ISC High Performance 2024 Research Paper Proceedings (39th International Conference)","raw_type":"proceedings-article"},"type":"conference-paper","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/A5028296912","display_name":"Qinghua Zhou","orcid":"https://orcid.org/0009-0001-8490-7990"},"institutions":[{"id":"https://openalex.org/I52357470","display_name":"The Ohio State University","ror":"https://ror.org/00rs6vg23","country_code":"US","type":"education","lineage":["https://openalex.org/I52357470"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Qinghua Zhou","raw_affiliation_strings":["The Ohio State University,Department of Computer Science and Engineering","Department of Computer Science and Engineering, The Ohio State University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Ohio State University,Department of Computer Science and Engineering","institution_ids":["https://openalex.org/I52357470"]},{"raw_affiliation_string":"Department of Computer Science and Engineering, The Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073814270","display_name":"Bharath Ramesh","orcid":"https://orcid.org/0000-0001-8230-3803"},"institutions":[{"id":"https://openalex.org/I52357470","display_name":"The Ohio State University","ror":"https://ror.org/00rs6vg23","country_code":"US","type":"education","lineage":["https://openalex.org/I52357470"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Bharath Ramesh","raw_affiliation_strings":["The Ohio State University,Department of Computer Science and Engineering","Department of Computer Science and Engineering, The Ohio State University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Ohio State University,Department of Computer Science and Engineering","institution_ids":["https://openalex.org/I52357470"]},{"raw_affiliation_string":"Department of Computer Science and Engineering, The Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5078128277","display_name":"Aamir Shafi","orcid":"https://orcid.org/0000-0002-1924-2769"},"institutions":[{"id":"https://openalex.org/I52357470","display_name":"The Ohio State University","ror":"https://ror.org/00rs6vg23","country_code":"US","type":"education","lineage":["https://openalex.org/I52357470"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Aamir Shafi","raw_affiliation_strings":["The Ohio State University,Department of Computer Science and Engineering","Department of Computer Science and Engineering, The Ohio State University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Ohio State University,Department of Computer Science and Engineering","institution_ids":["https://openalex.org/I52357470"]},{"raw_affiliation_string":"Department of Computer Science and Engineering, The Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5032141491","display_name":"Mustafa Abduljabbar","orcid":"https://orcid.org/0000-0003-1280-130X"},"institutions":[{"id":"https://openalex.org/I52357470","display_name":"The Ohio State University","ror":"https://ror.org/00rs6vg23","country_code":"US","type":"education","lineage":["https://openalex.org/I52357470"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Mustafa Abduljabbar","raw_affiliation_strings":["The Ohio State University,Department of Computer Science and Engineering","Department of Computer Science and Engineering, The Ohio State University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Ohio State University,Department of Computer Science and Engineering","institution_ids":["https://openalex.org/I52357470"]},{"raw_affiliation_string":"Department of Computer Science and Engineering, The Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5034293705","display_name":"Hari Subramoni","orcid":"https://orcid.org/0000-0002-1200-2754"},"institutions":[{"id":"https://openalex.org/I52357470","display_name":"The Ohio State University","ror":"https://ror.org/00rs6vg23","country_code":"US","type":"education","lineage":["https://openalex.org/I52357470"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Hari Subramoni","raw_affiliation_strings":["The Ohio State University,Department of Computer Science and Engineering","Department of Computer Science and Engineering, The Ohio State University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Ohio State University,Department of Computer Science and Engineering","institution_ids":["https://openalex.org/I52357470"]},{"raw_affiliation_string":"Department of Computer Science and Engineering, The Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5024879682","display_name":"Dhabaleswar K. Panda","orcid":"https://orcid.org/0000-0002-0356-1781"},"institutions":[{"id":"https://openalex.org/I52357470","display_name":"The Ohio State University","ror":"https://ror.org/00rs6vg23","country_code":"US","type":"education","lineage":["https://openalex.org/I52357470"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Dhabaleswar K. Panda","raw_affiliation_strings":["The Ohio State University,Department of Computer Science and Engineering","Department of Computer Science and Engineering, The Ohio State University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Ohio State University,Department of Computer Science and Engineering","institution_ids":["https://openalex.org/I52357470"]},{"raw_affiliation_string":"Department of Computer Science and Engineering, The Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I52357470"],"apc_list":null,"apc_paid":null,"fwci":0.9645,"has_fulltext":false,"cited_by_count":4,"citation_normalized_percentile":{"value":0.79769733,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":95,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"12"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.9991999864578247,"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"}},"topics":[{"id":"https://openalex.org/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.9991999864578247,"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"}},{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9973999857902527,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.996399998664856,"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/computer-science","display_name":"Computer science","score":0.8931779265403748},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.6637534499168396},{"id":"https://openalex.org/keywords/stochastic-gradient-descent","display_name":"Stochastic gradient descent","score":0.5652514100074768},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.5571585893630981},{"id":"https://openalex.org/keywords/data-parallelism","display_name":"Data parallelism","score":0.531292736530304},{"id":"https://openalex.org/keywords/supercomputer","display_name":"Supercomputer","score":0.4987056255340576},{"id":"https://openalex.org/keywords/gpu-cluster","display_name":"GPU cluster","score":0.46903666853904724},{"id":"https://openalex.org/keywords/parallelism","display_name":"Parallelism (grammar)","score":0.24662601947784424},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.21705135703086853},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.14538425207138062}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8931779265403748},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.6637534499168396},{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.5652514100074768},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.5571585893630981},{"id":"https://openalex.org/C61483411","wikidata":"https://www.wikidata.org/wiki/Q3124522","display_name":"Data parallelism","level":3,"score":0.531292736530304},{"id":"https://openalex.org/C83283714","wikidata":"https://www.wikidata.org/wiki/Q121117","display_name":"Supercomputer","level":2,"score":0.4987056255340576},{"id":"https://openalex.org/C2781335571","wikidata":"https://www.wikidata.org/wiki/Q2633544","display_name":"GPU cluster","level":3,"score":0.46903666853904724},{"id":"https://openalex.org/C2781172179","wikidata":"https://www.wikidata.org/wiki/Q853109","display_name":"Parallelism (grammar)","level":2,"score":0.24662601947784424},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.21705135703086853},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.14538425207138062}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.23919/isc.2024.10528931","is_oa":false,"landing_page_url":"https://doi.org/10.23919/isc.2024.10528931","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ISC High Performance 2024 Research Paper Proceedings (39th International Conference)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G202197749","display_name":"Computation for the Endless Frontier","funder_award_id":"1818253","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":31,"referenced_works":["https://openalex.org/W1558115001","https://openalex.org/W1902814702","https://openalex.org/W1964981582","https://openalex.org/W2052440657","https://openalex.org/W2193413348","https://openalex.org/W2194775991","https://openalex.org/W2271840356","https://openalex.org/W2486202470","https://openalex.org/W2488497244","https://openalex.org/W2769644379","https://openalex.org/W2899771611","https://openalex.org/W2953328958","https://openalex.org/W2964137095","https://openalex.org/W2973727699","https://openalex.org/W3037565727","https://openalex.org/W3039165326","https://openalex.org/W3043058721","https://openalex.org/W3043494828","https://openalex.org/W3086105743","https://openalex.org/W3106011734","https://openalex.org/W3130395060","https://openalex.org/W3132977829","https://openalex.org/W3177445289","https://openalex.org/W3209029431","https://openalex.org/W4221008227","https://openalex.org/W4285105236","https://openalex.org/W4312443924","https://openalex.org/W4367147749","https://openalex.org/W4399282167","https://openalex.org/W6687566353","https://openalex.org/W6694517276"],"related_works":["https://openalex.org/W2056717482","https://openalex.org/W1966243865","https://openalex.org/W2017587301","https://openalex.org/W2030707850","https://openalex.org/W2170611190","https://openalex.org/W2566934642","https://openalex.org/W1582436825","https://openalex.org/W2092007952","https://openalex.org/W2963113252","https://openalex.org/W2548246577"],"abstract_inverted_index":{"With":[0],"the":[1,16,67,75,90,95,107,160,163,183,217],"increasing":[2],"scale":[3],"of":[4,19,74,100,174],"High-Performance":[5],"Computing":[6],"(HPC)":[7],"and":[8,47,142,166,178,185,187,212],"Deep":[9],"Learning":[10],"(DL)":[11],"applications":[12,46],"through":[13],"GPU":[14,83,96,197],"adaptation,":[15],"seamless":[17],"communication":[18,37],"data":[20,84,133],"stored":[21],"on":[22,195],"GPUs":[23,62],"has":[24],"become":[25],"a":[26,36,57,72,154],"critical":[27],"factor":[28],"in":[29,44,153],"enhancing":[30],"overall":[31],"application":[32],"performance.":[33],"AllReduce":[34,79,105,144],"is":[35,41,56,128],"collective":[38],"operation":[39,80],"that":[40,146],"commonly":[42],"used":[43,64],"HPC":[45],"distributed":[48,200],"DL":[49,76,201],"training,":[50],"especially":[51],"Data":[52,54],"Parallelism.":[53],"Parallelism":[55],"common":[58],"strategy":[59],"where":[60],"parallel":[61],"are":[63],"to":[65,89,130,176,182,191,210],"process":[66],"partitioned":[68],"training":[69,117,202,215],"dataset":[70],"using":[71],"replica":[73],"model.":[77],"However,":[78],"for":[81],"large":[82],"still":[85],"performs":[86],"poorly":[87],"due":[88],"limited":[91],"interconnect":[92],"bandwidth":[93],"between":[94],"nodes.":[97],"Some":[98],"strategies":[99],"Gradient":[101,109],"Quantization":[102],"or":[103],"Sparse":[104],"modifying":[106],"Stochastic":[108],"Descent":[110],"(SGD)":[111],"algorithms":[112,145,168],"may":[113],"not":[114],"support":[115,171],"different":[116],"scenarios.":[118],"Recent":[119],"research":[120],"shows":[121],"integrating":[122],"GPU-based":[123,150],"compression":[124,151,170,194],"into":[125],"MPI":[126,157],"libraries":[127],"efficient":[129,148],"achieve":[131,172],"faster":[132,214],"transmission.":[134],"In":[135],"this":[136],"paper,":[137],"we":[138],"propose":[139],"optimized":[140],"Recursive-Doubling":[141,165],"Ring":[143,167],"encompass":[147],"collective-level":[149],"schemes":[152],"state-of-the-art":[155],"GPU-Aware":[156],"library.":[158],"At":[159],"microbenchmark":[161],"level,":[162],"proposed":[164],"with":[169,203],"benefits":[173],"up":[175,209],"75.3%":[177],"85.5%":[179],"respectively":[180,189],"compared":[181,190],"baseline,":[184,218],"24.8%":[186],"66.1%":[188],"naive":[192],"point-to-point":[193],"modern":[196],"clusters.":[198],"For":[199],"PyTorch-DDP,":[204],"these":[205],"two":[206],"approaches":[207],"yield":[208],"32.3%":[211],"35.7%":[213],"than":[216],"while":[219],"maintaining":[220],"similar":[221],"accuracy.":[222]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":2}],"updated_date":"2026-08-18T07:49:30.821534","created_date":"2025-10-10T00:00:00"}