Loading config/perlmutter.yaml 0 → 100644 +51 −0 Original line number Diff line number Diff line system: num_cdus: 36 racks_per_cdu: 3 nodes_per_rack: 128 rectifiers_per_rack: 32 chassis_per_rack: 8 nodes_per_blade: 2 switches_per_chassis: 4 nics_per_node: 4 rectifiers_per_chassis: 4 nodes_per_rectifier: 4 missing_racks: [] down_nodes: [] cpus_per_node: 1 gpus_per_node: 4 cpu_peak_flops: 3580000000000.0 gpu_peak_flops: 9700000000000.0 cpu_fp_ratio: 0.667 gpu_fp_ratio: 0.667 power: power_gpu_idle: 88 power_gpu_max: 300 power_cpu_idle: 90 power_cpu_max: 280 power_mem: 74.26 power_nic: 20 power_nvme: 30 power_switch: 250 power_cdu: 8473.47 power_update_freq: 15 rectifier_peak_threshold: 13670 sivoc_loss_constant: 13 sivoc_efficiency: 0.98 rectifier_loss_constant: 17 rectifier_efficiency: 0.96 power_cost: 0.094 scheduler: seed: 42 job_arrival_time: 900 mtbf: 11 trace_quanta: 15 min_wall_time: 3600 max_wall_time: 43200 ui_update_freq: 900 max_nodes_per_job: 3000 job_end_probs: COMPLETED: 0.63 FAILED: 0.13 CANCELLED: 0.12 TIMEOUT: 0.11 NODE_FAIL: 0.01 config/selene.yaml 0 → 100644 +51 −0 Original line number Diff line number Diff line system: num_cdus: 20 racks_per_cdu: 7 nodes_per_rack: 4 rectifiers_per_rack: 32 chassis_per_rack: 4 nodes_per_blade: 2 switches_per_chassis: 4 nics_per_node: 4 rectifiers_per_chassis: 4 nodes_per_rectifier: 4 missing_racks: [] down_nodes: [] cpus_per_node: 2 gpus_per_node: 8 cpu_peak_flops: 3481000000000.0 gpu_peak_flops: 624000000000000.0 # BF8 performance cpu_fp_ratio: 0.667 gpu_fp_ratio: 0.667 power: power_gpu_idle: 88 power_gpu_max: 400 power_cpu_idle: 90 power_cpu_max: 280 power_mem: 74.26 power_nic: 20 power_nvme: 30 power_switch: 250 power_cdu: 8473.47 power_update_freq: 15 rectifier_peak_threshold: 13670 sivoc_loss_constant: 13 sivoc_efficiency: 0.98 rectifier_loss_constant: 17 rectifier_efficiency: 0.96 power_cost: 0.094 scheduler: seed: 42 job_arrival_time: 900 mtbf: 11 trace_quanta: 15 min_wall_time: 3600 max_wall_time: 43200 ui_update_freq: 900 max_nodes_per_job: 3000 job_end_probs: COMPLETED: 0.63 FAILED: 0.13 CANCELLED: 0.12 TIMEOUT: 0.11 NODE_FAIL: 0.01 Loading
config/perlmutter.yaml 0 → 100644 +51 −0 Original line number Diff line number Diff line system: num_cdus: 36 racks_per_cdu: 3 nodes_per_rack: 128 rectifiers_per_rack: 32 chassis_per_rack: 8 nodes_per_blade: 2 switches_per_chassis: 4 nics_per_node: 4 rectifiers_per_chassis: 4 nodes_per_rectifier: 4 missing_racks: [] down_nodes: [] cpus_per_node: 1 gpus_per_node: 4 cpu_peak_flops: 3580000000000.0 gpu_peak_flops: 9700000000000.0 cpu_fp_ratio: 0.667 gpu_fp_ratio: 0.667 power: power_gpu_idle: 88 power_gpu_max: 300 power_cpu_idle: 90 power_cpu_max: 280 power_mem: 74.26 power_nic: 20 power_nvme: 30 power_switch: 250 power_cdu: 8473.47 power_update_freq: 15 rectifier_peak_threshold: 13670 sivoc_loss_constant: 13 sivoc_efficiency: 0.98 rectifier_loss_constant: 17 rectifier_efficiency: 0.96 power_cost: 0.094 scheduler: seed: 42 job_arrival_time: 900 mtbf: 11 trace_quanta: 15 min_wall_time: 3600 max_wall_time: 43200 ui_update_freq: 900 max_nodes_per_job: 3000 job_end_probs: COMPLETED: 0.63 FAILED: 0.13 CANCELLED: 0.12 TIMEOUT: 0.11 NODE_FAIL: 0.01
config/selene.yaml 0 → 100644 +51 −0 Original line number Diff line number Diff line system: num_cdus: 20 racks_per_cdu: 7 nodes_per_rack: 4 rectifiers_per_rack: 32 chassis_per_rack: 4 nodes_per_blade: 2 switches_per_chassis: 4 nics_per_node: 4 rectifiers_per_chassis: 4 nodes_per_rectifier: 4 missing_racks: [] down_nodes: [] cpus_per_node: 2 gpus_per_node: 8 cpu_peak_flops: 3481000000000.0 gpu_peak_flops: 624000000000000.0 # BF8 performance cpu_fp_ratio: 0.667 gpu_fp_ratio: 0.667 power: power_gpu_idle: 88 power_gpu_max: 400 power_cpu_idle: 90 power_cpu_max: 280 power_mem: 74.26 power_nic: 20 power_nvme: 30 power_switch: 250 power_cdu: 8473.47 power_update_freq: 15 rectifier_peak_threshold: 13670 sivoc_loss_constant: 13 sivoc_efficiency: 0.98 rectifier_loss_constant: 17 rectifier_efficiency: 0.96 power_cost: 0.094 scheduler: seed: 42 job_arrival_time: 900 mtbf: 11 trace_quanta: 15 min_wall_time: 3600 max_wall_time: 43200 ui_update_freq: 900 max_nodes_per_job: 3000 job_end_probs: COMPLETED: 0.63 FAILED: 0.13 CANCELLED: 0.12 TIMEOUT: 0.11 NODE_FAIL: 0.01