Merge branch 'esterel26' into 'master'

Esterel26

See merge request !1159

data/grid5000/sites/sophia/clusters/esterel26/esterel26.json

@@ -49,12 +49,39 @@
         "port": 9100,
         "protocol": "prometheus"
       }
+    },
+    {
+      "description": "Default subset of metrics from Prometheus Nvidia DCGM Exporter",
+      "name": "prom_nvgpu_default_metrics",
+      "optional_period": 15000,
+      "period": 0,
+      "source": {
+        "id": [
+          "DCGM_FI_DEV_SM_CLOCK",
+          "DCGM_FI_DEV_MEM_CLOCK",
+          "DCGM_FI_DEV_GPU_TEMP",
+          "DCGM_FI_DEV_POWER_USAGE",
+          "DCGM_FI_DEV_MEM_COPY_UTIL"
+        ],
+        "port": 9400,
+        "protocol": "prometheus"
+      }
+    },
+    {
+      "description": "All metrics from Prometheus Nvidia DCGM Exporter",
+      "name": "prom_nvgpu_all_metrics",
+      "optional_period": 15000,
+      "period": 0,
+      "source": {
+        "port": 9400,
+        "protocol": "prometheus"
+      }
     }
   ],
   "model": "Dell PowerEdge T640",
   "nodes_count": 1,
-  "nodes_description": "2 CPUs Intel Xeon Silver 4216, 16 cores/CPU, 384GB RAM, 558GB HDD, 3575GB HDD, 1 x 1Gb Ethernet, 1 x 40Gb InfiniBand",
-  "priority": 202010,
+  "nodes_description": "2 CPUs Intel Xeon Silver 4216, 16 cores/CPU, 4 GPUs Quadro RTX 8000, 384GB RAM, 558GB HDD, 3575GB HDD, 1 x 1Gb Ethernet, 1 x 40Gb InfiniBand",
+  "priority": 202110,
   "queues": [
     "admin",
     "testing"

data/grid5000/sites/sophia/clusters/esterel26/nodes/esterel26-1.json

@@ -20,6 +20,76 @@
     "warranty_end": "2027-10-31"
   },
   "exotic": false,
+  "gpu_devices": {
+    "nvidia0": {
+      "compute_capability": "7.5",
+      "cores": 4608,
+      "cpu_affinity": 0,
+      "device": "/dev/nvidia0",
+      "memory": 48318382080,
+      "microarchitecture": "Turing",
+      "model": "Quadro RTX 8000",
+      "performance": {
+        "fp-16": 32620000000000,
+        "fp-32": 16310000000000,
+        "fp-64": 510000000000
+      },
+      "power_default_limit": "250.00 W",
+      "vbios_version": "90.02.4E.00.03",
+      "vendor": "Nvidia"
+    },
+    "nvidia1": {
+      "compute_capability": "7.5",
+      "cores": 4608,
+      "cpu_affinity": 0,
+      "device": "/dev/nvidia1",
+      "memory": 48318382080,
+      "microarchitecture": "Turing",
+      "model": "Quadro RTX 8000",
+      "performance": {
+        "fp-16": 32620000000000,
+        "fp-32": 16310000000000,
+        "fp-64": 510000000000
+      },
+      "power_default_limit": "250.00 W",
+      "vbios_version": "90.02.4E.00.03",
+      "vendor": "Nvidia"
+    },
+    "nvidia2": {
+      "compute_capability": "7.5",
+      "cores": 4608,
+      "cpu_affinity": 1,
+      "device": "/dev/nvidia2",
+      "memory": 48318382080,
+      "microarchitecture": "Turing",
+      "model": "Quadro RTX 8000",
+      "performance": {
+        "fp-16": 32620000000000,
+        "fp-32": 16310000000000,
+        "fp-64": 510000000000
+      },
+      "power_default_limit": "250.00 W",
+      "vbios_version": "90.02.4E.00.03",
+      "vendor": "Nvidia"
+    },
+    "nvidia3": {
+      "compute_capability": "7.5",
+      "cores": 4608,
+      "cpu_affinity": 1,
+      "device": "/dev/nvidia3",
+      "memory": 48318382080,
+      "microarchitecture": "Turing",
+      "model": "Quadro RTX 8000",
+      "performance": {
+        "fp-16": 32620000000000,
+        "fp-32": 16310000000000,
+        "fp-64": 510000000000
+      },
+      "power_default_limit": "250.00 W",
+      "vbios_version": "90.02.4E.00.03",
+      "vendor": "Nvidia"
+    }
+  },
   "main_memory": {
     "ram_size": 412316860416
   },

input/grid5000/sites/sophia/clusters/esterel26/esterel26_metrics.yaml

@@ -38,3 +38,24 @@ metrics:
     source:
       protocol: prometheus
       port: 9100
+
+  - name: prom_nvgpu_default_metrics
+    description: Default subset of metrics from Prometheus Nvidia DCGM Exporter
+    period: 0
+    optional_period: 15000
+    source:
+      protocol: prometheus
+      port: 9400
+      id:
+        - DCGM_FI_DEV_SM_CLOCK
+        - DCGM_FI_DEV_MEM_CLOCK
+        - DCGM_FI_DEV_GPU_TEMP
+        - DCGM_FI_DEV_POWER_USAGE
+        - DCGM_FI_DEV_MEM_COPY_UTIL
+  - name: prom_nvgpu_all_metrics
+    description: All metrics from Prometheus Nvidia DCGM Exporter
+    period: 0
+    optional_period: 15000
+    source:
+      protocol: prometheus
+      port: 9400

input/grid5000/sites/sophia/clusters/esterel26/nodes/esterel26-1.yaml

@@ -16,6 +16,39 @@ esterel26-1:
     manufacturer: Dell Inc.
     name: PowerEdge T640
     serial: 61GCH73
+  gpu_devices:
+    nvidia0:
+      cpu_affinity: 0
+      device: "/dev/nvidia0"
+      memory: 48318382080
+      model: Quadro RTX 8000
+      power_default_limit: 250.00 W
+      vbios_version: 90.02.4E.00.03
+      vendor: Nvidia
+    nvidia1:
+      cpu_affinity: 0
+      device: "/dev/nvidia1"
+      memory: 48318382080
+      model: Quadro RTX 8000
+      power_default_limit: 250.00 W
+      vbios_version: 90.02.4E.00.03
+      vendor: Nvidia
+    nvidia2:
+      cpu_affinity: 1
+      device: "/dev/nvidia2"
+      memory: 48318382080
+      model: Quadro RTX 8000
+      power_default_limit: 250.00 W
+      vbios_version: 90.02.4E.00.03
+      vendor: Nvidia
+    nvidia3:
+      cpu_affinity: 1
+      device: "/dev/nvidia3"
+      memory: 48318382080
+      model: Quadro RTX 8000
+      power_default_limit: 250.00 W
+      vbios_version: 90.02.4E.00.03
+      vendor: Nvidia
   main_memory:
     ram_size: 412316860416
   memory_devices: