NVIDIA/deepops

Tools for building GPU clusters

What it solves

DeepOps simplifies the complex process of deploying and managing GPU server clusters. It automates the installation of the necessary software stack—including drivers, container runtimes, and cluster managers—to ensure that hardware like NVIDIA DGX systems is properly configured for AI and high-performance computing workloads.

How it works

DeepOps uses Ansible playbooks to orchestrate the deployment of infrastructure. Depending on the user's needs, it can:

  • Deploy full Kubernetes clusters via Kubespray for containerized application management.
  • Set up Slurm clusters for resource management and batch job scheduling.
  • Install essential GPU components (NVIDIA drivers, Docker, and NVIDIA Container Runtime) on single machines.
  • Deploy the NVIDIA DGX platform software stack on supported operating systems.

Who it’s for

It is designed for system administrators and infrastructure engineers who need to build GPU-enabled clusters on-premises or in data centers, specifically those using NVIDIA DGX systems or generic GPU servers running Ubuntu or RHEL/Rocky Linux.

Highlights

  • Multi-Scheduler Support: Provides automated paths for both Kubernetes and Slurm.
  • End-to-End Automation: Covers everything from OS-level drivers to high-level cluster orchestration.
  • Modular Design: Can be used to set up a full cluster or just specific components like KubeFlow or NFS storage.
  • DGX Optimization: Encapsulates best practices specifically for NVIDIA DGX systems.

Related

  • Project
    NVIDIA/gpu-operator

    An automation tool for Kubernetes that manages the installation and lifecycle of NVIDIA GPU drivers, runtimes, and monitoring tools as containers.

  • Project
    dstackai/dstack

    A unified control plane for GPU provisioning and orchestration that streamlines AI development, training, and inference across any GPU cloud, Kubernetes, or on-prem cluster.

  • Project
    nebius/soperator

    A Kubernetes operator that automates the deployment and management of Slurm clusters for AI training and HPC, featuring automated health checks and GPU optimization.

  • Project
    spark-arena/sparkrun

    A CLI tool for launching and managing LLM inference workloads on NVIDIA DGX Spark systems without needing Slurm or Kubernetes.

  • Project
    Project-HAMi/HAMi

    A Kubernetes-native virtualization middleware that enables GPU sharing, resource isolation, and device-aware scheduling across heterogeneous AI accelerators.