Run apps using intel iGPU while nvidia is the main gpu

TheWeirdDev · 2021-03-25 16:10:10

Hi,
I have been using my desktop PC with the dedicated Nvidia gpu for a while now and everything works without any problems.
I found an option in my BIOS settings that lets you enable the iGPU and I can see that it actually enabled my integrated gpu:

$ inxi -GMCS
...
Graphics:
  Device-1: Intel HD Graphics 530 driver: i915 v: kernel 
  Device-2: NVIDIA GM206 [GeForce GTX 960] driver: nvidia v: 460.56 
...

$ lspci | egrep -i '(VGA|Display)'
00:02.0 Display controller: Intel Corporation HD Graphics 530 (rev 06)
01:00.0 VGA compatible controller: NVIDIA Corporation GM206 [GeForce GTX 960] (rev a1)

The i915 kernel module is loaded and intel_gpu_top detects the gpu.

My question is, can I use the iGPU to run a program while xorg and my DE are running on Nvidia? Something like the opposite of prime-run.

Thank you.

Last edited by TheWeirdDev (2021-03-25 17:26:01)

Lone_Wolf · 2021-03-25 20:24:26

That should be possible, but let's verify how situation is first.

post

$ glxinfo -B
$ DRI_PRIME=1 glxinfo -B

(glxinfo is included in mesa-demos)

TheWeirdDev · 2021-03-26 07:37:32

Lone_Wolf wrote:

That should be possible, but let's verify how situation is first.
post
$ glxinfo -B
$ DRI_PRIME=1 glxinfo -B
(glxinfo is included in mesa-demos)

Both of them show the same thing:

name of display: :1
display: :1  screen: 0
direct rendering: Yes
Memory info (GL_NVX_gpu_memory_info):
    Dedicated video memory: 2048 MB
    Total available memory: 2048 MB
    Currently available dedicated video memory: 1579 MB
OpenGL vendor string: NVIDIA Corporation
OpenGL renderer string: GeForce GTX 960/PCIe/SSE2
OpenGL core profile version string: 4.6.0 NVIDIA 460.56
OpenGL core profile shading language version string: 4.60 NVIDIA
OpenGL core profile context flags: (none)
OpenGL core profile profile mask: core profile

OpenGL version string: 4.6.0 NVIDIA 460.56
OpenGL shading language version string: 4.60 NVIDIA
OpenGL context flags: (none)
OpenGL profile mask: (none)

OpenGL ES profile version string: OpenGL ES 3.2 NVIDIA 460.56
OpenGL ES profile shading language version string: OpenGL ES GLSL ES 3.20

seth · 2021-03-26 08:05:52

Try

__GLX_VENDOR_LIBRARY_NAME=mesa glxinfo

Sanity check: what's the plan here?
Prime exist because it's a cool pun on Optimus and Optimus exists because it allows you to run the deskton on the battery saving intel chip and SuperTurboTurkyPuncher 3 on the more powerful dedicated GPU (to unlock the ultra-nightmare 3D smellvision bloodgore mode)

Running on the nvidia chip all the time and unloading some clients to the intel chip has little to no benefit unless you
a) try to workaround an issue w/ nvidia that you should rather address and where offloading the task doesn't necessarily fix it
b) actually want to run a second session on the other chip for some sort of multihead situation (which is not done this way)

TheWeirdDev · 2021-03-26 13:34:57

seth wrote:

Try

__GLX_VENDOR_LIBRARY_NAME=mesa glxinfo

It gives me this error:

__GLX_VENDOR_LIBRARY_NAME=mesa glxinfo
name of display: :1
libGL error: No matching fbConfigs or visuals found
libGL error: failed to load driver: swrast
X Error of failed request:  GLXBadContext
  Major opcode of failed request:  152 (GLX)
  Minor opcode of failed request:  6 (X_GLXIsDirect)
  Serial number of failed request:  62
  Current serial number in output stream:  61

seth wrote:

Sanity check: what's the plan here?
Prime exist because it's a cool pun on Optimus and Optimus exists because it allows you to run the deskton on the battery saving intel chip and SuperTurboTurkyPuncher 3 on the more powerful dedicated GPU (to unlock the ultra-nightmare 3D smellvision bloodgore mode)
Running on the nvidia chip all the time and unloading some clients to the intel chip has little to no benefit unless you
a) try to workaround an issue w/ nvidia that you should rather address and where offloading the task doesn't necessarily fix it
b) actually want to run a second session on the other chip for some sort of multihead situation (which is not done this way)

I'm trying to get hardware accelerated video decoding to work in gstreamer (It supports VAAPI but not VDPAU), so I thought running gstreamer using intel gpu might fix the issue.
Same thing with chromium, It can play accelerated h264 but not VP9. I just want to experiment with a different GPU to see how it performs.

seth · 2021-03-26 13:42:00

Seems a hot topic ;-)
https://lists.freedesktop.org/archives/ … 60655.html

If Aaron says "it doesn't actually work" that close to definitive

TheWeirdDev · 2021-03-26 14:01:41

That's disappointing
It's not a big problem though. I can live without it.

Arch Linux

#1 2021-03-25 16:10:10

Run apps using intel iGPU while nvidia is the main gpu

#2 2021-03-25 20:24:26

Re: Run apps using intel iGPU while nvidia is the main gpu

#3 2021-03-26 07:37:32

Re: Run apps using intel iGPU while nvidia is the main gpu

#4 2021-03-26 08:05:52

Re: Run apps using intel iGPU while nvidia is the main gpu

#5 2021-03-26 13:34:57

Re: Run apps using intel iGPU while nvidia is the main gpu

#6 2021-03-26 13:42:00

Re: Run apps using intel iGPU while nvidia is the main gpu

#7 2021-03-26 14:01:41

Re: Run apps using intel iGPU while nvidia is the main gpu

Board footer