История изменений

What Landed: Linux Kernel’s Per-Process GPU Scheduling

In Linux 6.11, the kernel officially introduced per-process GPU scheduling and memory accounting — essentially letting the system:

    Assign GPU resources to specific processes.
    Isolate GPU memory per application.
    Dynamically prioritize workloads without freezing the desktop.

It’s paired with updates in Mesa and Wayland compositors that let applications request GPU acceleration in a standardized way — no vendor hacks, no environment variables, no guessing games.