{"id":"W4411689634","doi":"10.1016/j.jpdc.2025.105138","title":"Topology-aware GPU job scheduling with deep reinforcement learning and heuristics","year":2025,"lang":"en","type":"article","venue":"Journal of Parallel and Distributed Computing","topic":"Distributed and Parallel Computing Systems","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":false,"ca_institutions":"IBM (Canada); York University","funders":"IBM Canada; Natural Sciences and Engineering Research Council of Canada; Canadian Anesthesiologists' Society","keywords":"Computer science; Heuristics; Reinforcement learning; Scheduling (production processes); Parallel computing; Job shop scheduling; General-purpose computing on graphics processing units; Artificial intelligence; Mathematical optimization; Computer graphics (images); Operating system; Graphics","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":[0.0006984663,0.0002546787,0.0005051593,0.000182872,0.0004769038,0.0003471911,0.0004478747,0.000107135,0.000001934229],"category_scores_gemma":[0.0001098272,0.0002045866,0.00007533145,0.0004507096,0.0001280092,0.0002136861,0.0003810258,0.0005283427,0.000001139307],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005647651,"about_ca_system_score_gemma":0.0001388281,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001530416,"about_ca_topic_score_gemma":0.000001911756,"domain_scores_codex":[0.9980808,0.0001456695,0.0007493629,0.0003156159,0.0002913217,0.0004172479],"domain_scores_gemma":[0.9983293,0.0003329817,0.0006010459,0.0001845518,0.0003634456,0.0001886338],"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":[0.0001217576,0.00006970196,0.05507837,0.0002279936,0.0003217008,0.0002400819,0.0008668828,0.8986087,0.00004325789,0.03002679,0.0005734976,0.01382124],"study_design_scores_gemma":[0.001996174,0.0004723595,0.008181538,0.0006096387,0.000065899,0.0007169681,0.0007999075,0.9814429,0.00001330828,0.001346424,0.004028522,0.0003263524],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06896501,0.001492443,0.9278748,0.000761638,0.0003118845,0.00009736962,0.000001384337,0.00006669213,0.0004287707],"genre_scores_gemma":[0.9729449,0.0000740616,0.02663336,0.0001165756,0.0001467116,7.822813e-7,0.000009737694,0.000007005524,0.0000668994],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9039798,"threshold_uncertainty_score":0.8342801,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.009154864165384203,"score_gpt":0.2529626781795075,"score_spread":0.2438078140141233,"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."}}