View previous topic :: View next topic |
Author |
Message |
logrusx Advocate
Joined: 22 Feb 2018 Posts: 2541
|
Posted: Mon Dec 23, 2024 4:22 pm Post subject: Gnome 46/Wayland + Firefox + YouTube stutterng |
|
|
Today I updated and noticed the update included Gnome 46. However now I observe stuttering in video playback in YouTube.
By chance in my YouTube feed there was a video about this Gnome feature that has been requested for so long and it turned out to be variable refresh rate, coming in exactly v46. By playing with the refresh I found out the stuttering was worse at the highest rate. I didn't try anything else other than the highest and the lowest, but what made me curious is why I have two 60.01Hz entries for my internal display. The external one has only one entry for 60.02Hz which is expected as it is 17 years old Eizo.
I also tested it with Xorg and on the dGPU which is Nvidia and it looked like on the dGPU the stuttering was significantly decreased but still noticeable, both under Wayland and Xorg and there were only two entries for my internal display - 60.01 and 165 Hz.
However I know my iGPU which is AMD Vega 10 or something (Ryzen 7 5800H) is perfectly capable of playing 1440p and the stuttering is noticeable on 1080p as well. I also don't want to use the dGPU as the stupid binary driver dims the display due to inactivity too soon, the GPU consumes power and spins fans harder and more often due to the higher thermal output.
Does anyone else experience this?
Best Regards,
Georgi |
|
Back to top |
|
|
gorg86 Guru
Joined: 20 May 2011 Posts: 320
|
Posted: Tue Dec 24, 2024 12:50 am Post subject: |
|
|
I'm not using Gnome but I noticed since I switched to Wayland the power saving function of AMDGPU was bugged. My whole desktop and especially browsers were stuttery all the time because the GPU runs at super low frequencies all the time. Same thing happens on an Intel machine with integrated graphics.
On my AMD machine I have to set this on every start: Code: | echo performance > /sys/class/drm/card0/device/power_dpm_state
echo high > /sys/class/drm/card0/device/power_dpm_force_performance_level
echo 1 > /sys/class/drm/card0/device/pp_power_profile_mode |
|
|
Back to top |
|
|
logrusx Advocate
Joined: 22 Feb 2018 Posts: 2541
|
Posted: Tue Dec 24, 2024 5:41 am Post subject: |
|
|
gorg86 wrote: | I'm not using Gnome but I noticed since I switched to Wayland the power saving function of AMDGPU was bugged. |
Wayland has nothing to do with the power saving of the AMDGPU. That's a kernel thing. Also DPM stands for display power management, which shouldn't affect the performance of your graphics card.
However I gave it the benefit of the doubt and... nothing changed. If it works for you, it must be some kind of placebo effect.
Maybe only the last line may have some effect, but I doubt it.
Best Regards,
Georgi
p.s. did you upgrade/change your kernel?
Last edited by logrusx on Tue Dec 24, 2024 7:10 am; edited 1 time in total |
|
Back to top |
|
|
djselbeck n00b
Joined: 10 Oct 2005 Posts: 33 Location: Germany
|
Posted: Tue Dec 24, 2024 5:55 am Post subject: |
|
|
Quote: |
Also DPM stands for display power management, which shouldn't affect the performance of your graphics card.
|
Not in this case, it is documented here https://docs.kernel.org/gpu/amdgpu/thermal.html#gpu-sysfs-power-state-interfaces
FWIW I can also observe some lost frames on a 780m (7840u processor) for 1080p@60Hz video in firefox. I think it might be related to slowness in Firefox's compositing or so (downloaded in Kodi the video plays smooth). Also happened on Gnome 45 and now on 46. |
|
Back to top |
|
|
logrusx Advocate
Joined: 22 Feb 2018 Posts: 2541
|
Posted: Tue Dec 24, 2024 6:12 am Post subject: |
|
|
It started happening on 46 for me, that's what brought my attention to Gnome.
And i don't think the above-mentioned interfaces have anything to do with Wayland. As I already said, they are a kernel thing and I haven't changed the kernel a bit.
Unfortunately I can't test it on Chromium, unless I want to spend 2-3 hours compiling. Maybe I need to recompile Firefox. I'll try.
Best Regards,
Georgi |
|
Back to top |
|
|
Anon-E-moose Watchman
Joined: 23 May 2008 Posts: 6182 Location: Dallas area
|
Posted: Tue Dec 24, 2024 11:56 am Post subject: |
|
|
Yes setting the different parameters can affect the speed at which the gpu is allowed to do things.
Locked at low performance settings the gpu will only run to the limit of what you are allowing.
These are facts, like it or not.
They may not be your issue, but if you are going to pretend to be a know-it-all (ie. you already know all the answers) then why even ask questions?
You might just run "about:support" in firefox, scroll down to the gpu stuff and look at the various settings.
I doubt wayland or the kernel is at fault though, maybe gnome is setting the gpu to some weird specs. Again up to you to look. _________________ UM780, 6.12 zen kernel, gcc 13, openrc, wayland |
|
Back to top |
|
|
logrusx Advocate
Joined: 22 Feb 2018 Posts: 2541
|
Posted: Tue Dec 24, 2024 12:07 pm Post subject: |
|
|
I appreciate you taking the time to post, but your post is completely useless. Could you please refrain from doing it further?
Best Regards,
Georgi |
|
Back to top |
|
|
pietinger Moderator
Joined: 17 Oct 2006 Posts: 5271 Location: Bavaria
|
Posted: Tue Dec 24, 2024 1:02 pm Post subject: |
|
|
logrusx wrote: | I appreciate you taking the time to post, but your post is completely useless. Could you please refrain from doing it further? |
Giving someone support can sometimes be unsuccessful. You can react to this in different ways. One option is to simply ignore a post. However, your reply will be read by many other users. When I read something like that, I think to myself that I will be very careful in future to try to help you, because a reply like that would hurt me. Maybe other users feel the same way ...
@Anon-E-moose has successfully helped many users - including me. You also often try to help others. Have you always been right? Support is sometimes trial and error ... and sometimes you need a few mistakes before you are successful. Quitting after the first mistake will certainly never lead to success. I try to help with problems with the kernel configuration and yes, there are also very long threads ... because we didn't find the right option right away. But I was also never told to stop helping ... so that in the end the result was successful ... or we have found out that it is simply a bug in the kernel ... this sometimes takes a few iterations ... and the exclusion of an incorrect configuration. _________________ https://wiki.gentoo.org/wiki/User:Pietinger |
|
Back to top |
|
|
logrusx Advocate
Joined: 22 Feb 2018 Posts: 2541
|
Posted: Tue Dec 24, 2024 2:38 pm Post subject: |
|
|
pietinger wrote: |
@Anon-E-moose has successfully helped many users - including me. |
Me too. I'm very surprised by such a replay from him. That's why I decided to push back a little bit.
But let me elaborate on why I reject his response. Thanks to djselbeck, I now know the following:
Quote: | GPU sysfs Power State Interfaces
GPU power controls are exposed via sysfs files.
power_dpm_state
The power_dpm_state file is a legacy interface and is only provided for backwards compatibility. The amdgpu driver provides a sysfs API for adjusting certain power related parameters. The file power_dpm_state is used for this. It accepts the following arguments:
battery
balanced
performance
battery
On older GPUs, the vbios provided a special power state for battery operation. Selecting battery switched to this state. This is no longer provided on newer GPUs so the option does nothing in that case.
balanced
On older GPUs, the vbios provided a special power state for balanced operation. Selecting balanced switched to this state. This is no longer provided on newer GPUs so the option does nothing in that case.
performance
On older GPUs, the vbios provided a special power state for performance operation. Selecting performance switched to this state. This is no longer provided on newer GPUs so the option does nothing in that case.
|
The power_dpm_force_performance_level key makes no difference between auto and high, which means auto does its job correctly.
pp_power_profile_mode does nothing, because:
Quote: | To select a profile or create a custom profile, first select manual using power_dpm_force_performance_level. |
The profile was initially set to video, which should have sufficed even if what was set there had any effect.
Also I'm confident nothing in my graphics subsystem changed. Even under Xorg using the dGPU, even though there are no dropped frames reported in the YT video statistics, the stuttering is still noticeable. This leads me to believe the problem is in Gnome, moreover they made some serious changes to enable variable refresh rate.
What I'm asking is if someone else experiences this. Only one user has responded to this question so far and they claim they've solved it in a way it doesn't really do anything.
Best Regards,
Georgi |
|
Back to top |
|
|
rab0171610 Guru
Joined: 24 Dec 2022 Posts: 445
|
Posted: Tue Dec 24, 2024 3:09 pm Post subject: |
|
|
Quote: | Does anyone else experience this?
Best Regards,
Georgi |
No. Best Regards. |
|
Back to top |
|
|
logrusx Advocate
Joined: 22 Feb 2018 Posts: 2541
|
Posted: Tue Dec 24, 2024 3:45 pm Post subject: |
|
|
Firefox is recompiled, no difference.
I switched to manual performance level and tested video and 3d_full_screen profiles. The difference is only in the number of dropped frames reported, still stuttering. But then last night after playing with different refresh rates I had the same impression only to wake up today to the same situation as yesterday.
I'm starting to consider moving to a different DE.
Best Regards,
Georgi
Last edited by logrusx on Tue Dec 24, 2024 6:09 pm; edited 1 time in total |
|
Back to top |
|
|
logrusx Advocate
Joined: 22 Feb 2018 Posts: 2541
|
Posted: Tue Dec 24, 2024 4:39 pm Post subject: |
|
|
rab0171610 wrote: | Quote: | Does anyone else experience this?
Best Regards,
Georgi |
No. Best Regards. |
What's your setup?
Kernel, video adapters, Gnome version, browser, graphics session?
Best Regards,
Georgi |
|
Back to top |
|
|
|