CUDA

This guide walks through building a full NVIDIA GPU compute node on Ubuntu 20.04/22.04, targeting HGX A800 and Tesla data center GPUs. It covers disabling Nouveau, installing the NVIDIA driver and CUDA, configuring Infiniband with MLNX OFED, enabling key kernel modules, building OpenMPI, setting up NCCL and FabricManager, and validating with single-node and multi-node NCCL+MPI benchmarks.

This guide walks through installing NVIDIA GPU drivers with CUDA support on Ubuntu 22.04, including prerequisite packages, Vulkan libraries, CUDA Toolkit, and Clang. It also covers setting up Docker and NVIDIA GPU support for containers via nvidia-docker2, and verifying that the GPU, CUDA environment, and GPU-enabled Docker runtime are working correctly.