How to use Nvidia PRIME offload to run the X server on the integrated board

But your GPU/Intel architecure is Turing/Coffee Lake or newer, right?