{"id":"W4394896940","doi":"10.1109/tpds.2024.3390109","title":"Sampling-Based Multi-Job Placement for Heterogeneous Deep Learning Clusters","year":2024,"lang":"en","type":"article","venue":"IEEE Transactions on Parallel and Distributed Systems","topic":"Stochastic Gradient Optimization Techniques","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Victoria; University of Toronto; Memorial University of Newfoundland","funders":"British Columbia Knowledge Development Fund; Natural Sciences and Engineering Research Council of Canada; Canada Foundation for Innovation","keywords":"Computer science; Workload; Job scheduler; Scheduling (production processes); Distributed computing; Artificial intelligence; Machine learning; Latency (audio); Mathematical optimization; 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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002061195,0.0002151983,0.0002075435,0.0001896374,0.0003106055,0.0003890484,0.0002229711,0.0001041574,0.000004092601],"category_scores_gemma":[0.000008892752,0.000201147,0.0001173649,0.0002745624,0.00003843425,0.0001451694,0.000002870902,0.0001574462,0.000009362842],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001136946,"about_ca_system_score_gemma":0.00004185782,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002315032,"about_ca_topic_score_gemma":0.000007894461,"domain_scores_codex":[0.9986138,0.0000638906,0.0003311036,0.0004969366,0.0002005157,0.0002938029],"domain_scores_gemma":[0.9991931,0.0002854393,0.00005878422,0.0002480082,0.00008081581,0.0001338411],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00003006533,0.00005929685,0.000003939506,0.0001205915,0.00004998829,0.000004692176,0.000133824,0.9969512,0.00005443453,0.000571817,0.00006590482,0.001954222],"study_design_scores_gemma":[0.0006393521,0.0002793653,0.000004072683,0.0001253162,0.00002942976,0.00002948954,0.00005833664,0.9973072,0.0001786484,0.00003772343,0.00108975,0.0002212621],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0002269185,0.0004700498,0.9963192,0.0001478086,0.0009976004,0.0007905937,0.0001337548,0.0009040136,0.00001004117],"genre_scores_gemma":[0.9529324,0.00002089267,0.04618674,0.00004487081,0.00002221943,0.0005919705,0.00004331,0.0000207572,0.0001369004],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9527054,"threshold_uncertainty_score":0.8202538,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03431295490472693,"score_gpt":0.2792646518185682,"score_spread":0.2449516969138412,"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."}}