Merge pull request #4623 from Neelabh94/feature/dcgm-dev-package

Adding "datacenter-gpu-manager-4-dev" as an additional installation in A* YAML files.

Author: Neelabh94

examples/hypercompute_clusters/a3u-slurm-ubuntu-gcs/a3u-slurm-ubuntu-gcs.yaml

@@ -474,6 +474,7 @@ deployment_groups:
               nvidia_packages:
               - cuda-toolkit-12-8
               - datacenter-gpu-manager
+              - datacenter-gpu-manager-4-dev
               - libnvidia-cfg1-570-server
               - libnvidia-nscq-570
               - nvidia-compute-utils-570-server

examples/machine-learning/a3-megagpu-8g/a3mega-slurm-blueprint.yaml

@@ -186,6 +186,7 @@ deployment_groups:
           apt install -y cuda-toolkit-12-8
           apt install -y nvidia-container-toolkit
           apt install -y datacenter-gpu-manager-4-cuda12
+          apt install -y datacenter-gpu-manager-4-dev
       # this duplicates the ulimits configuration of the HPC VM Image
       - type: data
         destination: /etc/security/limits.d/99-unlimited.conf

examples/machine-learning/a3-megagpu-8g/a3mega-slurm-gcsfuse-lssd-blueprint.yaml

@@ -478,7 +478,8 @@ deployment_groups:
               enable_nvidia_dcgm: $(vars.enable_nvidia_dcgm)
               nvidia_packages:
               - cuda-toolkit-12-8
-              - datacenter-gpu-manager
+              - datacenter-gpu-manager-4-cuda12
+              - datacenter-gpu-manager-4-dev
               - libnvidia-cfg1-570-server
               - libnvidia-nscq-570
               - nvidia-compute-utils-570-server

examples/machine-learning/a3-ultragpu-8g/a3ultra-slurm-blueprint.yaml

@@ -144,6 +144,7 @@ deployment_groups:
           add-nvidia-repositories -y
           apt install -y cuda-toolkit-12-8
           apt install -y datacenter-gpu-manager-4-cuda12
+          apt install -y datacenter-gpu-manager-4-dev
       - type: ansible-local
         destination: settings_nvidia_dcgm.yml
         content: |

examples/machine-learning/a4-highgpu-8g/a4high-slurm-blueprint.yaml

@@ -146,6 +146,7 @@ deployment_groups:
           apt update -y
           apt install -y cuda-toolkit-12-8
           apt install -y datacenter-gpu-manager-4-cuda12
+          apt install -y datacenter-gpu-manager-4-dev
       - type: ansible-local
         destination: configure_cuda_dcgm.yml
         content: |

examples/machine-learning/a4x-highgpu-4g/a4xhigh-slurm-blueprint.yaml

@@ -162,6 +162,7 @@ deployment_groups:
               nvidia_packages:
               - cuda-toolkit-12-8
               - datacenter-gpu-manager-4-cuda12
+              - datacenter-gpu-manager-4-dev
             tasks:
             - name: Download NVIDIA repository package
               ansible.builtin.get_url:

examples/machine-learning/build-service-images/common/blueprint.yaml

@@ -101,6 +101,7 @@ deployment_groups:
               nvidia_packages:
               - cuda-toolkit-12-8
               - datacenter-gpu-manager-4-cuda12
+              - datacenter-gpu-manager-4-dev
             tasks:
             - name: Download NVIDIA repository package
               ansible.builtin.get_url: