$ GPU, driver, CUDA and ROCm news sources

On Linux, GPU support depends on several moving parts: the kernel driver, the userspace stack (Mesa for open drivers, the proprietary NVIDIA driver, or AMD's ROCm for compute), and the framework on top. A mismatch between any of them is a common reason for a setup that suddenly stops working.

To stay ahead of this, follow release notes for the compute stack you use (CUDA or ROCm), a news site that tracks Mesa and kernel graphics work, and the vendor engineering blogs, which usually explain what new hardware and software versions mean for running models locally.

Sources for gpu and driver news (6)

Phoronixlinuxgpu-driverskernel

Linux hardware and open-source news: GPU drivers, Mesa, kernel features and benchmarks.

NVIDIA Technical Bloggpu-driverslocal-ai

CUDA, inference optimisation and GPU-accelerated AI deep dives.

AMD ROCm Blogsgpu-driverslocal-ai

AMD's ROCm engineering blog: running AI on Radeon and Instinct GPUs.

ROCm Releasesgpu-drivers

Release notes for the ROCm GPU compute stack.

vLLM Releaseslocal-aigpu-drivers

Release notes for the vLLM high-throughput inference server.

Hacker News: Linuxlinuxgpu-drivers

Hacker News stories about Linux, kernels, distros and GPU stacks with 100+ points.

FAQ

Do I need ROCm or CUDA for local AI?
It depends on your GPU: CUDA is for NVIDIA cards and ROCm for supported AMD cards. CPU-only inference also works, just slower.

See what these sources published in the last 24 hours on the latest checkup.

Related

Linux newsLinux kernel newsLinux distribution newsLocal AI newsOpen models newsAI agents newsVibe coding & dev tooling news

Best Linux news sources (2026)Best local AI sources (2026)all sources