Part II · Your system

25. GPU

hub

Which graphics chip you have, which driver runs it, and whether programs can use it.

$ atfq gpu
  card0     Intel, driver i915, shows the boot screen
  Screens   eDP-1 1920x1200 · free: DP-1, HDMI-A-1
  Compute   /dev/dri/renderD128

Commands

lspci -k | grep -A3 -E 'VGA|3D'

GPUs and the kernel driver each uses 5.2

glxinfo -B

OpenGL renderer: your GPU, or llvmpipe for software (mesa-utils) 5.2

vulkaninfo --summary

Vulkan devices and drivers

nvidia-smi

NVIDIA: use, memory, processes, driver and CUDA version 5.4

echo $XDG_SESSION_TYPE

wayland or x11 5.3

Paths

/sys/class/drm/

graphics cards and their screen outputs

/dev/dri/

card* to drive screens, renderD* to draw and compute

Also about this

1. 5.1What a GPU does differentlyA CPU has a few powerful cores, built to run one thread fast, branches and all. A GPU has thousands of simple ones, built to run the same instructions on huge amounts of data at once.
2. 5.2GPU drivers: the kernel, Mesa and NVIDIAA GPU driver has two halves: a kernel driver (DRM) that owns the device and its memory, and a user-space library that turns OpenGL or Vulkan calls into GPU commands.
3. 5.3How a picture reaches the screen: KMS, Wayland and X11The kernel sets up the screen itself (KMS: kernel mode setting): resolution, refresh rate, which output shows which image.
4. 5.4Computing on the GPU, and when your program can't see itCompute frameworks talk to the GPU through a vendor stack: CUDA for NVIDIA, ROCm (HIP) for AMD, oneAPI for Intel. Vulkan and OpenCL work across vendors.