{"id":"W4210863326","doi":"10.1109/mm.2022.3148670","title":"Accelerating Deep Learning Using Interconnect-Aware UCX Communication for MPI Collectives","year":2022,"lang":"en","type":"article","venue":"IEEE Micro","topic":"Parallel Computing and Optimization Techniques","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"Queen's University","funders":"Natural Sciences and Engineering Research Council of Canada; Western Canada Research Grid; Compute Canada","keywords":"Computer science; PCI Express; Speedup; Graphics processing unit; Parallel computing; Message Passing Interface; Network topology; Message passing; Bandwidth (computing); Computer architecture; Embedded system; Field-programmable gate array; Computer network","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007071013,0.0009391597,0.0004661781,0.0006444342,0.00123956,0.0009238467,0.002126278,0.0005071062,0.003550918],"category_scores_gemma":[0.002036103,0.0003892369,0.0005618224,0.001009731,0.0006862482,0.001902791,0.00180778,0.001259571,0.000993078],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001447443,"about_ca_system_score_gemma":0.002352171,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00983245,"about_ca_topic_score_gemma":0.01969247,"domain_scores_codex":[0.9994961,0.00007918351,0.00002664185,0.0001057522,0.0001787714,0.0001135822],"domain_scores_gemma":[0.9990082,0.0002058011,0.00006738913,0.0003771657,0.0002440176,0.00009735718],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0008822271,0.0005499939,0.01054982,0.0003538811,0.0002375712,0.0003459239,0.0007106777,0.4153784,0.06372581,0.03667077,0.05517785,0.415417],"study_design_scores_gemma":[0.00006153841,0.000127797,0.0009027605,0.00000792982,0.00002186272,0.00004106354,0.0001009105,0.9594156,0.02248969,0.00898153,0.007825803,0.00002351404],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1566982,0.0004552117,0.8027114,0.001070329,0.0003017619,0.0001547269,0.0003498998,0.02569526,0.01256309],"genre_scores_gemma":[0.6528933,0.0002536186,0.3359193,0.0002620397,0.0001133134,0.0002738962,0.001093574,0.0008298235,0.008361262],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.00983245,"threshold_uncertainty_score":0.01955044,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04396272083031036,"score_gpt":0.297140060445731,"score_spread":0.2531773396154207,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}