Heavy artifacts with hardware acceleration on Meteor Lake Arc iGPU

Per-Astra-Ad-Deum · 2024-11-09 19:31:37

seth wrote:

This error occurs on any Chromium-based browsers as well as any Electron-based apps like FreeTube for example
Same there as well?

per-astra-ad-deum wrote:

I was so surprised by this that I forgot to check if this error occurs with other applications as well.

seth wrote:

Have you tried the behavior w/ the LTS kernel?

Not yet. Will give it a try and report.

Per-Astra-Ad-Deum · 2024-11-10 07:07:36

Here's my report:

The error occurs with Linux 6.6 as well. Tested with Virt-Manager 3D-Acceleration|OpenGL settings.

I'll focus on that error now. Together with the graphics glitch in Doom 3 it is easily reproducible, unlike the artifacts in e.g. FreeTube, which appear to be random.

seth · 2024-11-10 09:21:00

https://bbs.archlinux.org/viewtopic.php … 3#p2204343 ?
Though that should and did affect pretty much everything…

Also

printenv | grep -iE '(mesa|sdl|gl)' # though that makes no sense w/ the fresh system

Per-Astra-Ad-Deum · 2024-11-10 09:27:32

seth wrote:

Also

printenv | grep -iE '(mesa|sdl|gl)' # though that makes no sense w/ the fresh system

Disclaimer: I'm on my own system again with no direct access to my friend's system.

grep -iE '(mesa|sdl|gl)

yields

KGLOBALACCELD_PLATFORM=org.kde.kwin
MESA_VK_DEVICE_SELECT=8086:7d55
VDPAU_DRIVER=va_gl
__EGL_VENDOR_LIBRARY_FILENAMES=/usr/share/glvnd/egl_vendor.d/50_mesa.json
__GLX_VENDOR_LIBRARY_NAME=mesa
LS_COLORS=rs=0:di=01;34:ln=01;36:mh=00:pi=40;33:so=01;35:do=01;35:bd=40;33;01:cd=40;33;01:or=40;31;01:mi=00:su=37;41:sg=30;43:ca=00:tw=30;42:ow=34;42:st=37;44:ex=01;32:*.7z=01;31:*.ace=01;31:*.alz=01;31:*.apk=01;31:*.arc=01;31:*.arj=01;31:*.bz=01;31:*.bz2=01;31:*.cab=01;31:*.cpio=01;31:*.crate=01;31:*.deb=01;31:*.drpm=01;31:*.dwm=01;31:*.dz=01;31:*.ear=01;31:*.egg=01;31:*.esd=01;31:*.gz=01;31:*.jar=01;31:*.lha=01;31:*.lrz=01;31:*.lz=01;31:*.lz4=01;31:*.lzh=01;31:*.lzma=01;31:*.lzo=01;31:*.pyz=01;31:*.rar=01;31:*.rpm=01;31:*.rz=01;31:*.sar=01;31:*.swm=01;31:*.t7z=01;31:*.tar=01;31:*.taz=01;31:*.tbz=01;31:*.tbz2=01;31:*.tgz=01;31:*.tlz=01;31:*.txz=01;31:*.tz=01;31:*.tzo=01;31:*.tzst=01;31:*.udeb=01;31:*.war=01;31:*.whl=01;31:*.wim=01;31:*.xz=01;31:*.z=01;31:*.zip=01;31:*.zoo=01;31:*.zst=01;31:*.avif=01;35:*.jpg=01;35:*.jpeg=01;35:*.mjpg=01;35:*.mjpeg=01;35:*.gif=01;35:*.bmp=01;35:*.pbm=01;35:*.pgm=01;35:*.ppm=01;35:*.tga=01;35:*.xbm=01;35:*.xpm=01;35:*.tif=01;35:*.tiff=01;35:*.png=01;35:*.svg=01;35:*.svgz=01;35:*.mng=01;35:*.pcx=01;35:*.mov=01;35:*.mpg=01;35:*.mpeg=01;35:*.m2v=01;35:*.mkv=01;35:*.webm=01;35:*.webp=01;35:*.ogm=01;35:*.mp4=01;35:*.m4v=01;35:*.mp4v=01;35:*.vob=01;35:*.qt=01;35:*.nuv=01;35:*.wmv=01;35:*.asf=01;35:*.rm=01;35:*.rmvb=01;35:*.flc=01;35:*.avi=01;35:*.fli=01;35:*.flv=01;35:*.gl=01;35:*.dl=01;35:*.xcf=01;35:*.xwd=01;35:*.yuv=01;35:*.cgm=01;35:*.emf=01;35:*.ogv=01;35:*.ogx=01;35:*.aac=00;36:*.au=00;36:*.flac=00;36:*.m4a=00;36:*.mid=00;36:*.midi=00;36:*.mka=00;36:*.mp3=00;36:*.mpc=00;36:*.ogg=00;36:*.ra=00;36:*.wav=00;36:*.oga=00;36:*.opus=00;36:*.spx=00;36:*.xspf=00;36:*~=00;90:*#=00;90:*.bak=00;90:*.crdownload=00;90:*.dpkg-dist=00;90:*.dpkg-new=00;90:*.dpkg-old=00;90:*.dpkg-tmp=00;90:*.old=00;90:*.orig=00;90:*.part=00;90:*.rej=00;90:*.rpmnew=00;90:*.rpmorig=00;90:*.rpmsave=00;90:*.swp=00;90:*.tmp=00;90:*.ucf-dist=00;90:*.ucf-new=00;90:*.ucf-old=00;90:

seth · 2024-11-10 16:40:30

That's from our vulkan tests? But vkcube was unaffected anyway.
Did you see the max bpc thread?

Doom3 is the open source linux client, not some windows binary?

Per-Astra-Ad-Deum · 2024-11-10 16:44:41

seth wrote:

That's from our vulkan tests? But vkcube was unaffected anyway.

Whuat. What are you talking about now? If you're referring to my latest codeblock, that was the output of

printenv | grep -iE '(mesa|sdl|gl)'

seth wrote:

Did you see the max bpc thread?

The one you've linked? Yes, but I don't think these issues are related :-(

seth wrote:

Doom3 is the open source linux client, not some windows binary?

Yes, open source client.

seth · 2024-11-10 16:52:38

What are you talking about now?

MESA_VK_DEVICE_SELECT=8086:7d55
VDPAU_DRIVER=va_gl
__EGL_VENDOR_LIBRARY_FILENAMES=/usr/share/glvnd/egl_vendor.d/50_mesa.json
__GLX_VENDOR_LIBRARY_NAME=mesa

They're probably left overs from earlier efforts to fix this?
If not, remove/unet those and try again.

Yes, but I don't think these issues are related :-(

I don't really either, but we're starting to run out of reasonable explanations…
Did you use the ame monitor on both systems?

Last edited by seth (2024-11-10 16:52:53)

Per-Astra-Ad-Deum · 2024-11-10 16:58:03

seth wrote:

They're probably left overs from earlier efforts to fix this?
If not, remove/unet those and try again.

Not quite. They are there to prevent certain processes from running on the NVIDIA eGPU. But the issue at hand precedes these settings, so I know for sure they are not related.

seth wrote:

I don't really either, but we're starting to run out of reasonable explanations…
Did you use the ame monitor on both systems?

Nope, as I've said before, no hardware or software or files or configs or peripherals were linked to my system when the same error occurred on my friend's desktop.

You're right though, this is getting tedious. Say, do you know where I can report qemu/virt-manager/libvirt/virtio bugs and errors? Do they have a Github of some sorts? I couldn't find a definite place to report this.

seth · 2024-11-10 19:48:45

Is the eGPU attached when this happens?
Does doom3 or qemu run ok on the nvidia GPU?

https://libvirt.org/bugs.html but w/ the Doom3 issues (did we test https://archlinux.org/packages/extra/x86_64/xonotic/ or https://archlinux.org/packages/extra/x86_64/warsow/ or so?) there must be some common element/wider factor (SDL?)

The default mesa dri definitions include

        <application name="Doom 3: BFG" executable="Doom3BFG.exe">
            <option name="allow_glsl_builtin_variable_redeclaration" value="true" />
            <option name="force_glsl_extensions_warn" value="true" />
        </application>

which will likely not meet your doom executable - what if you run it sth. like

allow_glsl_builtin_variable_redeclaration=true force_glsl_extensions_warn=true doom3

?

Per-Astra-Ad-Deum · 2024-11-13 08:54:24

seth wrote:

Is the eGPU attached when this happens?

As I've said, the issue predates my eGPU setup.

seth wrote:

Does doom3 or qemu run ok on the nvidia GPU?

I've tried switching to NVIDIA for Renderer in OpenGL settings in Qemu, but it didn't work. Doom3 ran flawlessly on the eGPU I had before the NVIDIA (Intel Arch A770 LE)

seth wrote:

https://libvirt.org/bugs.html but w/ the Doom3 issues (did we test https://archlinux.org/packages/extra/x86_64/xonotic/ or https://archlinux.org/packages/extra/x86_64/warsow/ or so?) there must be some common element/wider factor (SDL?)

Tested with Xonotic (SDL and GLX), ran flawlessly.

seth wrote:

The default mesa dri definitions include

        <application name="Doom 3: BFG" executable="Doom3BFG.exe">
            <option name="allow_glsl_builtin_variable_redeclaration" value="true" />
            <option name="force_glsl_extensions_warn" value="true" />
        </application>

which will likely not meet your doom executable - what if you run it sth. like

allow_glsl_builtin_variable_redeclaration=true force_glsl_extensions_warn=true doom3

?

I've banned Doom III to my VM :-D
I'd have to re-install it and re-install the native port/mod.

But I made another discovery that may help us. I've been trying to pinpoint the issue with Chromium-based applications and I think I found the culprit:

This is when running Brave with "--ozone-platform-hint=auto --enable-features=VaapiVideoDecodeLinuxGL". Hardware acceleration is confirmed to be active with intel_gpu_top.

This is the same scenario, but this time the flag "--ozone-platform-hint=auto". Hardware acceleration is confirmed to be active with intel_gpu_top.

On a side note, what about the pictures? I understand posting pictures is generally discouraged here, but this is a graphical issue, so I guess it is appropriate to post graphics for illustration?

seth · 2024-11-14 08:26:14

https://bbs.archlinux.org/viewtopic.php?id=244031&p=33

Browser video HW accel is a special (and sad) case.
https://wiki.archlinux.org/title/Hardwa … tion#Intel

pacman -Qs intel

https://wiki.archlinux.org/title/Hardwa … ing_VA-API

The browser artifacts are then limited to video playback only?
So the remaining offenders are doom3 and libvirt?

On a side note, what about the pictures?

You're supposed to link them and not embed anything > 250x250, regardless of the nature of the thread to keep loading fast and the mouse wheel cool.

Per-Astra-Ad-Deum · 2024-11-24 10:33:01

Hey there, me again.

Kernel 6.12 is out, and sadly, the error still persists.

seth wrote:

The browser artifacts are then limited to video playback only?

Actually, haven't seen any of the random artifacts for quite some time now. And browser playback is fixed with the addition of "--ozone-platform-hint=auto". So this part is solved practically.

seth wrote:

So the remaining offenders are doom3 and libvirt?

I don't really care about Doom III since I've banned all Gaming to a VM meanwhile. But libvirt, damn, that's really bothering me!

And yes, it still is an issue and it is as reproducible as ever.

In the meantime I've reported the error to both libvirt as well as qemu gitlabs:

Libvirt

Qemu-Project

Also, coincidentally I've configured an older Dell Precision 7530 for a friend of mine. I've installed Debian and set-up a GPU-passthrough for the laptop's dGPU (NVIDIA Quadro P2000). Now, would you believe it: no artifacts/glitching/whatever on this machine! Settings are the same: 3D acceleration box ticked for Video Virtio submenu, OpenGL box ticked for Spice Graphics submenu!
So on Arch and Tumbleweed this is an issue, but not on Debian. I assume mesa/qemu/libvirt/intel-media-driver/spice ... broke somewhere along the road between the stale Debian package versions and the bleeding-edge Arch and Tumbleweed versions.

seth wrote:

pacman -Qs intel

local/intel-gmmlib 22.5.2-1
    Intel Graphics Memory Management Library
local/intel-gpu-tools 1.29-1
    Tools for development and testing of the Intel DRM driver
local/intel-media-driver 24.4.2-1
    Intel Media Driver for VAAPI — Broadwell+ iGPUs
local/intel-ucode 20241112-1
    Microcode update files for Intel CPUs
local/libmfx 23.2.2-3
    Intel Media SDK dispatcher library
local/libva-utils 2.22.0-1
    Intel VA-API Media Applications and Scripts for libva
local/libvpl 2.13.0-1
    Intel Video Processing Library
local/onetbb 2022.0.0-1
    High level abstract threading library (oneAPI Threading Building Blocks)
local/vulkan-intel 1:24.2.7-1
    Open-source Vulkan driver for Intel GPUs

seth · 2024-11-25 09:18:22

I've configured an older Dell Precision 7530 for a friend of mine. I've installed Debian and set-up a GPU-passthrough for the laptop's dGPU (NVIDIA Quadro P2000)

That's completely different HW, though and notably no arc IGP.

WRT https://bbs.archlinux.org/viewtopic.php … 0#p2202120 and some journal bits from your post, you're not even running on xe but i915
Maybe we give that a shot: https://wiki.archlinux.org/title/Intel_ … _Xe_driver

Arch Linux

#26 2024-11-09 19:31:37

Re: Heavy artifacts with hardware acceleration on Meteor Lake Arc iGPU

#27 2024-11-10 07:07:36

Re: Heavy artifacts with hardware acceleration on Meteor Lake Arc iGPU

#28 2024-11-10 09:21:00

Re: Heavy artifacts with hardware acceleration on Meteor Lake Arc iGPU

#29 2024-11-10 09:27:32

Re: Heavy artifacts with hardware acceleration on Meteor Lake Arc iGPU

#30 2024-11-10 16:40:30

Re: Heavy artifacts with hardware acceleration on Meteor Lake Arc iGPU

#31 2024-11-10 16:44:41

Re: Heavy artifacts with hardware acceleration on Meteor Lake Arc iGPU

#32 2024-11-10 16:52:38

Re: Heavy artifacts with hardware acceleration on Meteor Lake Arc iGPU

#33 2024-11-10 16:58:03

Re: Heavy artifacts with hardware acceleration on Meteor Lake Arc iGPU

#34 2024-11-10 19:48:45

Re: Heavy artifacts with hardware acceleration on Meteor Lake Arc iGPU

#35 2024-11-13 08:54:24

Re: Heavy artifacts with hardware acceleration on Meteor Lake Arc iGPU

#36 2024-11-14 08:26:14

Re: Heavy artifacts with hardware acceleration on Meteor Lake Arc iGPU

#37 2024-11-24 10:33:01

Re: Heavy artifacts with hardware acceleration on Meteor Lake Arc iGPU

#38 2024-11-25 09:18:22

Re: Heavy artifacts with hardware acceleration on Meteor Lake Arc iGPU

Board footer