{"id":"https://openalex.org/W2985247698","doi":"https://doi.org/10.1109/mm.2019.2949986","title":"Communication Profiling and Characterization of Deep-Learning Workloads on Clusters With High-Performance Interconnects","display_name":"Communication Profiling and Characterization of Deep-Learning Workloads on Clusters With High-Performance Interconnects","publication_year":2019,"publication_date":"2019-10-30","ids":{"openalex":"https://openalex.org/W2985247698","doi":"https://doi.org/10.1109/mm.2019.2949986","mag":"2985247698"},"language":"en","primary_location":{"id":"doi:10.1109/mm.2019.2949986","is_oa":false,"landing_page_url":"https://doi.org/10.1109/mm.2019.2949986","pdf_url":null,"source":{"id":"https://openalex.org/S59697426","display_name":"IEEE Micro","issn_l":"0272-1732","issn":["0272-1732","1937-4143"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Micro","raw_type":"journal-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/A5004330728","display_name":"Ammar Ahmad Awan","orcid":"https://orcid.org/0000-0002-6272-3760"},"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":true,"raw_author_name":"Ammar Ahmad Awan","raw_affiliation_strings":["The Ohio State University"],"affiliations":[{"raw_affiliation_string":"The Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057769693","display_name":"Arpan Jain","orcid":"https://orcid.org/0000-0003-2522-8522"},"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":"Arpan Jain","raw_affiliation_strings":["The Ohio State University"],"affiliations":[{"raw_affiliation_string":"The Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5076405830","display_name":"Ching-Hsiang Chu","orcid":"https://orcid.org/0000-0002-6752-3135"},"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":"Ching-Hsiang Chu","raw_affiliation_strings":["The Ohio State University"],"affiliations":[{"raw_affiliation_string":"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"],"affiliations":[{"raw_affiliation_string":"The Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5105345016","display_name":"Dhableswar K. Panda","orcid":null},"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":"Dhableswar K. Panda","raw_affiliation_strings":["The Ohio State University"],"affiliations":[{"raw_affiliation_string":"The Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5004330728"],"corresponding_institution_ids":["https://openalex.org/I52357470"],"apc_list":null,"apc_paid":null,"fwci":1.073,"has_fulltext":false,"cited_by_count":12,"citation_normalized_percentile":{"value":0.783392,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":89,"max":98},"biblio":{"volume":"40","issue":"1","first_page":"35","last_page":"43"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9993000030517578,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9993000030517578,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12808","display_name":"Ferroelectric and Negative Capacitance Devices","score":0.9987999796867371,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9940999746322632,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.9159793257713318},{"id":"https://openalex.org/keywords/infiniband","display_name":"InfiniBand","score":0.8301887512207031},{"id":"https://openalex.org/keywords/xeon-phi","display_name":"Xeon Phi","score":0.6083343625068665},{"id":"https://openalex.org/keywords/remote-direct-memory-access","display_name":"Remote direct memory access","score":0.5464347004890442},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.545728325843811},{"id":"https://openalex.org/keywords/pci-express","display_name":"PCI Express","score":0.5109657049179077},{"id":"https://openalex.org/keywords/message-passing","display_name":"Message passing","score":0.48508182168006897},{"id":"https://openalex.org/keywords/xeon","display_name":"Xeon","score":0.48277655243873596},{"id":"https://openalex.org/keywords/ibm","display_name":"IBM","score":0.46856382489204407},{"id":"https://openalex.org/keywords/low-latency","display_name":"Low latency (capital markets)","score":0.4621700942516327},{"id":"https://openalex.org/keywords/supercomputer","display_name":"Supercomputer","score":0.46039173007011414},{"id":"https://openalex.org/keywords/profiling","display_name":"Profiling (computer programming)","score":0.4586681127548218},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.4321031868457794},{"id":"https://openalex.org/keywords/computer-architecture","display_name":"Computer architecture","score":0.35123109817504883},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.33545058965682983},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.1571742594242096},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.11867770552635193}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.9159793257713318},{"id":"https://openalex.org/C2781030343","wikidata":"https://www.wikidata.org/wiki/Q922437","display_name":"InfiniBand","level":2,"score":0.8301887512207031},{"id":"https://openalex.org/C96972482","wikidata":"https://www.wikidata.org/wiki/Q1049168","display_name":"Xeon Phi","level":2,"score":0.6083343625068665},{"id":"https://openalex.org/C130795937","wikidata":"https://www.wikidata.org/wiki/Q2561570","display_name":"Remote direct memory access","level":2,"score":0.5464347004890442},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.545728325843811},{"id":"https://openalex.org/C64270927","wikidata":"https://www.wikidata.org/wiki/Q206924","display_name":"PCI Express","level":3,"score":0.5109657049179077},{"id":"https://openalex.org/C854659","wikidata":"https://www.wikidata.org/wiki/Q1859284","display_name":"Message passing","level":2,"score":0.48508182168006897},{"id":"https://openalex.org/C145108525","wikidata":"https://www.wikidata.org/wiki/Q656154","display_name":"Xeon","level":2,"score":0.48277655243873596},{"id":"https://openalex.org/C70388272","wikidata":"https://www.wikidata.org/wiki/Q5968558","display_name":"IBM","level":2,"score":0.46856382489204407},{"id":"https://openalex.org/C46637626","wikidata":"https://www.wikidata.org/wiki/Q6693015","display_name":"Low latency (capital markets)","level":2,"score":0.4621700942516327},{"id":"https://openalex.org/C83283714","wikidata":"https://www.wikidata.org/wiki/Q121117","display_name":"Supercomputer","level":2,"score":0.46039173007011414},{"id":"https://openalex.org/C187191949","wikidata":"https://www.wikidata.org/wiki/Q1138496","display_name":"Profiling (computer programming)","level":2,"score":0.4586681127548218},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.4321031868457794},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.35123109817504883},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.33545058965682983},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.1571742594242096},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.11867770552635193},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0},{"id":"https://openalex.org/C192562407","wikidata":"https://www.wikidata.org/wiki/Q228736","display_name":"Materials science","level":0,"score":0.0},{"id":"https://openalex.org/C171250308","wikidata":"https://www.wikidata.org/wiki/Q11468","display_name":"Nanotechnology","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/mm.2019.2949986","is_oa":false,"landing_page_url":"https://doi.org/10.1109/mm.2019.2949986","pdf_url":null,"source":{"id":"https://openalex.org/S59697426","display_name":"IEEE Micro","issn_l":"0272-1732","issn":["0272-1732","1937-4143"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Micro","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G4368463115","display_name":null,"funder_award_id":"#ACI-1450440","funder_id":"https://openalex.org/F4320335353","funder_display_name":"National Science Foundation of Sri Lanka"},{"id":"https://openalex.org/G5644214236","display_name":null,"funder_award_id":"#CCF-1565414","funder_id":"https://openalex.org/F4320335353","funder_display_name":"National Science Foundation of Sri Lanka"},{"id":"https://openalex.org/G841161129","display_name":null,"funder_award_id":"#CNS-1513120","funder_id":"https://openalex.org/F4320335353","funder_display_name":"National Science Foundation of Sri Lanka"},{"id":"https://openalex.org/G8973714196","display_name":null,"funder_award_id":"#ACI-1664137","funder_id":"https://openalex.org/F4320335353","funder_display_name":"National Science Foundation of Sri Lanka"}],"funders":[{"id":"https://openalex.org/F4320335353","display_name":"National Science Foundation of Sri Lanka","ror":"https://ror.org/010xaa060"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":14,"referenced_works":["https://openalex.org/W1817724761","https://openalex.org/W2766362889","https://openalex.org/W2787998955","https://openalex.org/W2898319404","https://openalex.org/W2904180680","https://openalex.org/W2912208425","https://openalex.org/W2914698571","https://openalex.org/W2963674387","https://openalex.org/W2985247698","https://openalex.org/W2992165038","https://openalex.org/W3017432630","https://openalex.org/W3103417206","https://openalex.org/W4301239768","https://openalex.org/W6748645090"],"related_works":["https://openalex.org/W2981664121","https://openalex.org/W1949642813","https://openalex.org/W2056476515","https://openalex.org/W2043143850","https://openalex.org/W3146239682","https://openalex.org/W2146390972","https://openalex.org/W1964981582","https://openalex.org/W2136638258","https://openalex.org/W1974037879","https://openalex.org/W2107463933"],"abstract_inverted_index":{"Heterogeneous":[0],"high-performance":[1,9],"computing":[2],"systems":[3,70],"with":[4,8,71,93],"GPUs":[5,79],"are":[6],"equipped":[7],"interconnects":[10,30,85],"like":[11,73,80],"InfiniBand,":[12],"Omni-Path,":[13],"PCIe,":[14],"and":[15,48,56,76,83,105,141],"NVLink.":[16],"However,":[17],"little":[18],"exists":[19],"in":[20,58],"the":[21,25,88],"literature":[22],"that":[23],"captures":[24],"performance":[26,116],"impact":[27],"of":[28,69,102],"these":[29],"on":[31,122],"distributed":[32,43],"deep":[33],"learning":[34],"(DL).":[35],"In":[36],"this":[37],"article,":[38],"we":[39,128,145],"choose":[40],"Horovod,":[41,134],"a":[42,66,130],"training":[44,52],"middleware,":[45],"to":[46,60,86],"analyze":[47,87],"profile":[49],"various":[50,84],"DNN":[51],"workloads":[53],"using":[54],"TensorFlow":[55],"PyTorch":[57],"addition":[59],"standard":[61],"MPI":[62],"microbenchmarks.":[63],"We":[64,113],"use":[65],"wide":[67],"variety":[68],"CPUs":[72],"Intel":[74],"Xeon":[75],"IBM":[77],"POWER9,":[78],"Volta":[81],"V100,":[82],"following":[89],"metrics:":[90],"1)":[91],"message-size":[92,97],"Horovod's":[94],"tensor-fusion;":[95,99],"2)":[96],"without":[98],"3)":[100],"number":[101],"MPI/NCCL":[103,111],"calls;":[104],"4)":[106],"time":[107],"taken":[108],"by":[109],"each":[110],"call.":[112],"observed":[114,146],"extreme":[115],"variations":[117],"for":[118,133,148],"non-power-of-two":[119],"message":[120],"sizes":[121],"different":[123],"platforms.":[124],"To":[125],"address":[126],"this,":[127],"design":[129],"message-padding":[131],"scheme":[132],"illustrate":[135],"significantly":[136],"smoother":[137],"allreduce":[138],"latency":[139],"profiles,":[140],"report":[142],"cases":[143],"where":[144],"improvement":[147],"end-to-end":[149],"training.":[150]},"counts_by_year":[{"year":2025,"cited_by_count":2},{"year":2023,"cited_by_count":1},{"year":2022,"cited_by_count":1},{"year":2021,"cited_by_count":6},{"year":2020,"cited_by_count":1},{"year":2019,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}