Moving to a new GPU

Wednesday, 9 November 2022

I've replaced my ageing Nvidia card with a more recent and much faster
AMD GPU. However when I boot the system, the default display still
shows as the built-in Intel chip, even though AMD modules are loaded:

$ sudo lsmod|grep -i amd
amdgpu              10153984  1
iommu_v2               24576  1 amdgpu
gpu_sched              49152  1 amdgpu
drm_ttm_helper         16384  1 amdgpu
drm_buddy              20480  2 amdgpu,i915
drm_display_helper    180224  2 amdgpu,i915
ttm                    94208  3 amdgpu,drm_ttm_helper,i915

(Nvidia modules are not loaded).

Inxi shows:

$ inxi -G
Graphics:
  Device-1: Intel IvyBridge GT2 [HD Graphics 4000] driver: i915 v: kernel
  Device-2: AMD Ellesmere [Radeon RX 470/480/570/570X/580/580X/590]
    driver: amdgpu v: kernel
  Display: x11 server: X.Org v: 1.20.14 with: Xwayland v: 22.1.3 driver: X:
    loaded: amdgpu,modesetting unloaded: fbdev,vesa dri: crocus gpu: i915
    resolution: 1920x1080~60Hz
  OpenGL: renderer: Mesa Intel HD Graphics 4000 (IVB GT2) v: 4.2 Mesa
    22.1.7

Switcherooctl shows both GPUs:

$ switcherooctl list
Device: 0
  Name:        Intel® HD Graphics 4000
  Default:     yes
  Environment: DRI_PRIME=pci-0000_00_02_0

Device: 1
  Name:        Advanced Micro Devices, Inc. [AMD®/ATI] Ellesmere
[Radeon RX 470/480/570/570X/580/580X/590] (Radeon RX 580)
  Default:     no
  Environment: DRI_PRIME=pci-0000_01_00_0

I should mention that both GPUs are connected through an HDMI switch to
a single monitor. I've used this successfully in the past to get PCI
passthrough on Windows VMs for gaming.

I haven't touched anything else. I tried unplugging the i915 and
rebooting (after removing the HDMI switch and connecting the monitor
directly), but simply got a blank screen (in fact three dots at the top
left).

Is there some grub magic I need to do? I assumed this would be
automatic but it seems not.

poc