GPU Passthrough Custom OS Template

Follow the steps outlined below when getting your custom template ready for GPU support.

Bear in mind that GPU passthrough generally falls into two main scenarios:

While the Machine Learning case involves less effort (since the display does not rely on the GPU), we will walk through the entire process to cover both situations.

Supported GPU

If you run DaDesktop on premise, or ask for specific cards to be purchased for your dedicated server, here are a few points to keep in mind.

Our testing mainly focuses on the AMD Radeon RX 6000 series (particularly the RX 6400) and Nvidia RTX cards. Models from before 2018 are typically not supported.

Integrated GPUs on Ryzen 7/9 and iGPUs from Intel should work as well, though we don't run automated tests on them.

DD Node side

The dadesktop_npnode_deploy/modules/build-os-template directory contains scripts that adapt a Linux guest to use the GPU as the primary display and toggle between GPU-accelerated machines and CPU-rendered screens.

Guest Settings

Please check that "support_qemu" inside /var/lib/kvm/templatename/sysinfo.json is set to false so you can connect via VNC inside the guest – qemu vnc cannot show the passthrough GPU's screen.

Testing

Method 1: 
vblank_mode=0 glxgears

Method 2:

/apps/dd-guest/check-gpu

Windows guest settings when no physical monitor is attached (most cases)

We rely on IddSampleDriver to bring a display adapter and virtual monitor into the Windows 10 template. Simply install it if your GPU has no monitor plugged in.

Open "Device Manager" → "Display Adapters" → "IddSampleDriver Device" and right-click to enable it. (You can disable the device later when you don't need the virtual display.)

it will tend to be slow, and you can "Disconnect this display" if that option is available.

(Choose Display x, under the "Multiple displays" section, use the drop-down to pick the Disconnect this display option.)

You may also "Disconnect this display" if that option is present.