GPU

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.

Deploy GPU Hot with a single Docker command: a lightweight yet polished real-time GPU dashboard. It auto-detects multiple GPUs, tracks processes, and shows live charts with threshold markers. Includes startup commands and comparison benchmarks, ideal for daily monitoring on personal servers and multi-GPU workstations.

This article walks through key terminology and the end‑to‑end execution flow of GPU workloads, then explains PCIe pass‑through and why it is not true virtualization. It compares vCUDA and cGPU approaches for Kubernetes, introduces several open source and commercial GPU virtualization and pooling solutions, and summarizes how different vendors isolate, share, and schedule GPU compute and memory.