Quick start — test your GPU

1 · Connect over SSH

Use the dedicated IPv6 address from your welcome email (your SSH key is already installed):

ssh root@<your-IPv6>

2 · Check the GPU — one command

Every server ships with a built-in checker:

gpu-check

It runs VAAPI, OpenCL, Vulkan and OpenVINO tests plus a real hardware transcode.

3 · What you should see

[1] VAAPI / QuickSync   Driver version: Intel iHD driver ... 25.2.3
[2] OpenCL              Device Name: Intel(R) UHD Graphics 770   (Type: GPU)
[3] Vulkan              deviceName = Intel(R) UHD Graphics 770 (ADL-S)
[4] OpenVINO            devices: ['CPU', 'GPU']
[5] VAAPI transcode     OK (hardware H.264 encode)

If you see Intel(R) UHD Graphics 770, your GPU is working. ✅

4 · Run each check manually

# VAAPI / QuickSync (video transcoding)
vainfo

# OpenCL (GPGPU compute)
clinfo | grep -E "Device Name|Device Type"

# Vulkan
vulkaninfo --summary

# OpenVINO (AI inference) — should list 'GPU'
python3 -c "from openvino import Core; print(Core().available_devices)"

5 · Example — hardware video transcode

Transcode to H.264 on the GPU with FFmpeg (VAAPI / QuickSync):

ffmpeg -hwaccel vaapi -vaapi_device /dev/dri/renderD128 \
  -i input.mp4 -vf 'format=nv12,hwupload' -c:v h264_vaapi output.mp4

Docker, k3s and OpenVINO are preinstalled — your pods can request the GPU as a resource.