You are not logged in.
Hello everyone!
Distro: arch(last update was at sept 23)
Kernel: zen 6.16.7
Bios version: IFGP6.03.05
Hardware confit: Ryzen 7840hs with 128GB of crucial 5600MHz ram
The issue the laptop after aproximately 5 minutes of gaming hits the apu skin temperature(50C) with a very aggresive fan curve(fans at 100%), and the fps gets cut in half, basically from 60 to 30 in most cases. And at 30 it creates a squarewave looking frametime( for a few milliseconds it outpus 45 fps and after for a few milliseconds it outputs 30fps), because of this it looks 3ven more stuttery than it is.
I tried to adjust the apu_skin_temp_limit with ryzenadj, but the system straigh up ignores it.
The interesting part is that the gpu stays at 2700MHz and consumes the sam 35-40W of power.
As you can see the frame time is very spikey and has that straight piece in the middle.
Does anybody have this issue? Or I made something wrong during the installation?
The laptop is doing this since I got it( late july), but until noe I didn’t know the cause of it.
Thank you foe your response!
Offline
Isn't there a bios setting for thermals?
Can you check the actual cpu temperature? Within sensors it should be the k10temp. Around 50 it really should not be throttling.
Offline
Isn't there a bios setting for thermals?
Can you check the actual cpu temperature? Within sensors it should be the k10temp. Around 50 it really should not be throttling.
I cannot find it! It isn't the cpu what is throttling it, it is the apu_skin_temperature which is the tempreature of the case of the laptop.
Offline
As you can see the frame time is very spikey and has that straight piece in the middle.
Can we? Where?
after aproximately 5 minutes of gaming
What "gaming" specifically? Blobby Volley?
Sanity check:
glxinfo -B
Offline
As you can see the frame time is very spikey and has that straight piece in the middle.
Can we? Where?
https://us1.discourse-cdn.com/flex001/u … 4f134.jpeg
after aproximately 5 minutes of gaming
What "gaming" specifically? Blobby Volley?
Snowrunner, Beamng, Deathloop
Sanity check:
glxinfo -B
What should I look for?
Offline
Whether you're running on software (llvmpipe) but not according to your screenshot.
https://community.frame.work/t/fw16-apu … ture/75619
The die temperatures don't look that much off,
https://wiki.archlinux.org/title/AMDGPU#Monitoring
https://wiki.archlinux.org/title/AMDGPU#Overclocking
https://wiki.archlinux.org/title/AMDGPU … nce_levels
What happens if you underclock the GPU or force it to a lower performance level/power profile?
Offline
Whether you're running on software (llvmpipe) but not according to your screenshot.
https://community.frame.work/t/fw16-apu … ture/75619The die temperatures don't look that much off,
https://wiki.archlinux.org/title/AMDGPU#Monitoring
https://wiki.archlinux.org/title/AMDGPU#Overclocking
https://wiki.archlinux.org/title/AMDGPU … nce_levels
What happens if you underclock the GPU or force it to a lower performance level/power profile?
I tried lowering the clocks with corectrl but even after that, if the skin temperature reached 50C. The fps drops.
Offline
That seems to be expectable and the idea would be to prevent the skin temperature from rising by controlling the APU temperature - which apparently isn't driven by the GPU?
Is one of the cores (even if they shift through) constantly at 100%
Maybe better post the output of
glxinfo -B
On a formal note, please avoid bloating the thread with pointless full quotes of previous posts.
Offline
Is one of the cores (even if they shift through) constantly at 100%
Maybe better post the output ofglxinfo -B
None of the cores are reaching 100% utilization while playing snowrunner.
Last edited by Bgf12 (2025-09-29 04:32:18)
Offline
Doesn't look like software rendering.
So is the CPU driving the skin temperature?
https://wiki.archlinux.org/title/CPU_fr … amd_pstate - try the powersaving governor.
Are the games just overly demanding?
Does any of https://aur.archlinux.org/packages?O=0&K=unigine cause the same (probably eventually, but how fast)
Last but not least the thermal management of the device might be broken or simply not up to the task t run the APU anywhere near its full capacity for very long…
Offline
In my experience laptops do not like high temperatures at all. Life expectancy is reduced when pushed. However, you can try to change that skin temperature limit, RyzenAdj does that.
Offline
However, you can try to change that skin temperature limit, RyzenAdj does that.
I tried that it didn't helped. It cha ges the limit but the laptop straight up ignores it.
Offline
You can probably lower it, but 50°C sounds like a sane cap - focus on why it's getting there itfp.
Can you disable PBO?
The turbo on intel CPUs seems to drive the die temperature more than the general frequency range, but idk whether this is true for AMD or relevant to your situation (though I'll say that one of the cores in your screenshot /is/ at 100%)
Offline
LuxFerre wrote:However, you can try to change that skin temperature limit, RyzenAdj does that.
I tried that it didn't helped. It cha ges the limit but the laptop straight up ignores it.
Maybe ask for help in the Ryzenadj github, they'd probably know better...
Offline