apiVersion: "helm-expt.confighub.com/v1alpha1" kind: "EffectiveValues" metadata: name: "nvidia-gpu-operator-v26.7.1-aicr-eks-training" spec: files: - path: "effective-values-aicr-eks-training.yaml" source: "inline-proof" sha256: "570ef67a52f7fa0c5794fb8add36f93355a7eb6b6c99e814226cee0c0743577e" mergedValuesCaptured: false values: ccManager: enabled: false cdi: default: false enabled: true daemonsets: tolerations: - operator: "Exists" dcgm: enabled: true dcgmExporter: config: create: true data: "# Clocks,,\nDCGM_FI_DEV_SM_CLOCK, gauge, SM clock frequency (in MHz).\nDCGM_FI_DEV_MEM_CLOCK, gauge, Memory clock frequency (in MHz).\n\n# Temperature,,\nDCGM_FI_DEV_MEMORY_TEMP, gauge, Memory temperature (in C).\nDCGM_FI_DEV_GPU_TEMP, gauge, GPU temperature (in C).\n\n# Power,,\nDCGM_FI_DEV_POWER_USAGE, gauge, Power draw (in W).\nDCGM_FI_DEV_TOTAL_ENERGY_CONSUMPTION, counter, Total energy consumption since boot (in mJ).\n\n# PCIe,,\nDCGM_FI_PROF_PCIE_TX_BYTES, counter, Total number of bytes transmitted through PCIe TX (in KB) via NVML.\nDCGM_FI_PROF_PCIE_RX_BYTES, counter, Total number of bytes received through PCIe RX (in KB) via NVML.\nDCGM_FI_DEV_PCIE_REPLAY_COUNTER, counter, Total number of PCIe retries.\n\n# Utilization (the sample period varies depending on the product),,\nDCGM_FI_DEV_GPU_UTIL, gauge, GPU utilization (in %).\nDCGM_FI_DEV_MEM_COPY_UTIL, gauge, Memory utilization (in %).\nDCGM_FI_DEV_ENC_UTIL, gauge, Encoder utilization (in %).\nDCGM_FI_DEV_DEC_UTIL, gauge, Decoder utilization (in %).\n\n# Errors and violations,,\nDCGM_FI_DEV_XID_ERRORS, gauge, Value of the last XID error encountered.\nDCGM_FI_DEV_POWER_VIOLATION, counter, Throttling duration due to power constraints (in us).\nDCGM_FI_DEV_THERMAL_VIOLATION, counter, Throttling duration due to thermal constraints (in us).\nDCGM_FI_DEV_SYNC_BOOST_VIOLATION, counter, Throttling duration due to sync-boost constraints (in us).\nDCGM_FI_DEV_BOARD_LIMIT_VIOLATION, counter, Throttling duration due to board limit constraints (in us).\nDCGM_FI_DEV_LOW_UTIL_VIOLATION, counter, Throttling duration due to low utilization (in us).\nDCGM_FI_DEV_RELIABILITY_VIOLATION, counter, Throttling duration due to reliability constraints (in us).\n\n# Memory usage,,\nDCGM_FI_DEV_FB_FREE, gauge, Framebuffer memory free (in MiB).\nDCGM_FI_DEV_FB_USED, gauge, Framebuffer memory used (in MiB).\n\n# Retired pages,,\nDCGM_FI_DEV_RETIRED_SBE, counter, Total number of retired pages due to single-bit errors.\nDCGM_FI_DEV_RETIRED_DBE, counter, Total number of retired pages due to double-bit errors.\nDCGM_FI_DEV_RETIRED_PENDING, counter, Total number of pages pending retirement.\n\n# NVLink,,\nDCGM_FI_DEV_NVLINK_BANDWIDTH_TOTAL, counter, Total number of NVLink bandwidth counters for all lanes\nDCGM_FI_PROF_NVLINK_TX_BYTES, counter, The rate of data transmitted over NVLink not including protocol headers in bytes per second.\nDCGM_FI_PROF_NVLINK_RX_BYTES, counter, The rate of data received over NVLink not including protocol headers in bytes per second.\n\n# Add DCP metrics,,\nDCGM_FI_PROF_GR_ENGINE_ACTIVE, gauge, Ratio of time the graphics engine is active (in %).\nDCGM_FI_PROF_SM_ACTIVE, gauge, The ratio of cycles an SM has at least 1 warp assigned (in %).\nDCGM_FI_PROF_SM_OCCUPANCY, gauge, The ratio of number of warps resident on an SM (in %).\nDCGM_FI_PROF_PIPE_TENSOR_ACTIVE, gauge, Ratio of cycles the tensor (HMMA) pipe is active (in %).\nDCGM_FI_PROF_DRAM_ACTIVE, gauge, Ratio of cycles the device memory interface is active sending or receiving data (in %).\nDCGM_FI_PROF_PCIE_TX_BYTES, counter, The number of bytes of active pcie tx data including both header and payload.\nDCGM_FI_PROF_PCIE_RX_BYTES, counter, The number of bytes of active pcie rx data including both header and payload.\n\n# BCP Additional metrics\nDCGM_FI_DEV_GPU_NVLINK_ERRORS, gauge, Identifies a GPU NVLink error type returned by DCGM_FI_DEV_GPU_NVLINK_ERRORS.\n\n# Added RunAI Additional metrics https://docs.run.ai/latest/developer/metrics/metrics-api/#advanced-metrics\n## NVLink\nDCGM_FI_DEV_NVLINK_BANDWIDTH_L0, counter, The number of bytes of active NVLink rx or tx data including both header and payload.\n## VGPU License status\nDCGM_FI_DEV_VGPU_LICENSE_STATUS, gauge, vGPU License status\n## Remapped rows\nDCGM_FI_DEV_UNCORRECTABLE_REMAPPED_ROWS, counter, Number of remapped rows for uncorrectable errors\nDCGM_FI_DEV_CORRECTABLE_REMAPPED_ROWS, counter, Number of remapped rows for correctable errors\nDCGM_FI_DEV_ROW_REMAP_FAILURE, gauge, Whether remapping of rows has failed\n## Static configuration information. These appear as labels on the other metrics\nDCGM_FI_DRIVER_VERSION, label, Driver Version\n## Profiling metrics\nDCGM_FI_PROF_PIPE_FP64_ACTIVE, gauge, Ratio of cycles the fp64 pipes are active (in %).\nDCGM_FI_PROF_PIPE_FP32_ACTIVE, gauge, Ratio of cycles the fp32 pipes are active (in %).\nDCGM_FI_PROF_PIPE_FP16_ACTIVE, gauge, Ratio of cycles the fp16 pipes are active (in %).\n" name: "dcgm-exporter" serviceMonitor: enabled: true honorLabels: true interval: "30s" devicePlugin: enabled: true env: - name: "DP_DISABLE_HEALTHCHECKS" value: "109" - name: "DEVICE_LIST_STRATEGY" value: "cdi-cri,cdi-annotations,volume-mounts" driver: enabled: true kernelModuleType: "auto" maxParallelUpgrades: 5 rdma: enabled: false version: "580.173.02" fullnameOverride: "gpu-operator" gdrcopy: enabled: true gfd: enabled: true hostPaths: driverInstallDir: "/run/nvidia/driver" kataSandboxDevicePlugin: enabled: false migManager: enabled: true nfd: enabled: false operator: nodeSelector: nodeGroup: "system-worker" resources: limits: cpu: "500m" memory: "700Mi" requests: cpu: "200m" memory: "300Mi" tolerations: - operator: "Exists" upgradeCRD: true toolkit: enabled: true validator: plugin: env: - name: "WITH_WORKLOAD" value: "false"