Short backstory: I’ve been facing an overheating issue on my ASUS Vivobook 14 M1402IA-AM173 laptop running Manjaro (kernel 6.5.3-1-MANJARO, scaling_driver amd-pstate) with 16GB of RAM and a Ryzen 7 4800Hs processor. Every time I launch applications like PHPStorm (project indexing) or run PHP unit tests (Paratest), all CPU cores max out their frequency to the maximum value. According to htop, it reaches 4300 MHz, and even when the processor heats up, it doesn’t reduce the frequency to prevent overheating. Eventually, it hits a critical temperature of 105 degrees, and the laptop shuts down.
I couldn’t find another solution except for what I’ll describe below. Maybe you can help me with this! I’m tired of dealing with it; I need to unplug the laptop from the charger to get my work done.
I found the program TLP, and it does indeed control the governor and sets the maximum value (I talked to the maintainer, and they said the application is functioning correctly). However, nothing changes; it’s as if the processor itself is ignoring these settings and still raising the frequencies to 4300. Here’s a summary of tlp-stat -p:
--- TLP 1.6.0 --------------------------------------------
+++ Processor
CPU model = AMD Ryzen 7 4800HS with Radeon Graphics
/sys/devices/system/cpu/cpu0/cpufreq/scaling_driver = amd-pstate
/sys/devices/system/cpu/cpu0/cpufreq/scaling_governor = ondemand
/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors = conservative ondemand userspace powersave performance schedutil
/sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq = 400000 [kHz]
/sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq = 3900000 [kHz]
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_min_freq = 400000 [kHz]
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq = 4300000 [kHz]
/sys/devices/system/cpu/cpu1..cpu15: omitted for clarity, use -v to show all
/sys/devices/system/cpu/amd_pstate/status = guided
/sys/devices/system/cpu/amd_pstate/cppc_dynamic_boost = (not available)
/sys/devices/system/cpu/cpufreq/boost = 1
/sys/module/workqueue/parameters/power_efficient = Y
/proc/sys/kernel/nmi_watchdog = 0
+++ Platform Profile
/sys/firmware/acpi/platform_profile = (not available)
/sys/firmware/acpi/platform_profile_choices = (not available)
The most interesting part is that there is no such issue on Windows; the problem seems to be in Linux. I’ve tried it on Ubuntu 22.04, and it’s the same story.
If you know how to resolve this or can help, please do. I’m tired of this. The laptop is only a month old.
Your story leaves me wondering one thing: How is the CPU fan controlled? Via BIOS/UEFI or through the OS? Does the fan become louder/faster when the CPU generates heat?
If the fan speed has to be controlled by the OS here, then you have to set it accordingly. The default settings are most likely not good enough for your laptop.
In any case, it is always better to leave the fan control to the BIOS/UEFI if possible. Check the UEFI Settings.
Well, problematic is the turbo boost as I see. Possibly disable it with tlp like that in /etc/tlp.conf to mitigate you problem at the time being:
Thank you for your response!
I was already considering turning off the turbo mode. Most likely, if I don’t find a solution, I will have to do that. But for me, it’s kind of like buying an RTX 4090 but playing on the integrated graphics card :).
After checking the BIOS, I couldn’t find any information about fan control. Could you please advise me on where to find these settings?
I’ve also updated the BIOS to the latest version (v 310).
In most cases there is no documentations about the UEFI, only how you enter it. So sorry I guess you have to share pictures of it so that someone is able to help you.
But if there is none, then most likely it have to be managed by the OS.
Yes, I understood you, thank you for at least trying to help! CoreCtrl doesn’t provide me with any controls and settings for frequency scaling, etc. I can only change the governor as in TLP, and that’s it. Try amdgpu-fan
I tried amdgpu-fan, but unfortunately, the program doesn’t work. Here’s the log it provided me with:
sudo amdgpu-fan
starting amdgpu-fan
Traceback (most recent call last):
File "/usr/bin/amdgpu-fan", line 8, in <module>
sys.exit(main())
^^^^^^
File "/usr/lib/python3.11/site-packages/amdgpu_fan/controller.py", line 95, in main
FanController(config).main()
File "/usr/lib/python3.11/site-packages/amdgpu_fan/controller.py", line 47, in main
logger.debug(f'{name}: Temp {temp}, Setting fan speed to: {speed}, fan speed: {card.fan_speed}, min: {card.fan_min}, max: {card.fan_max}')
^^^^^^^^^^^^
File "/usr/lib/python3.11/site-packages/amdgpu_fan/lib/amdgpu.py", line 64, in fan_min
return int(self.read_endpoint('pwm1_min'))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3.11/site-packages/amdgpu_fan/lib/amdgpu.py", line 36, in read_endpoint
with open(self._endpoints[endpoint], 'r') as e:
~~~~~~~~~~~~~~~^^^^^^^^^^
KeyError: 'pwm1_min'
So you need a cpu fan control… since it is ASUS, take a look at the ArchWiki, there is a section for ASUS: Fan speed control - ArchWiki Probably you get it working with pwmconfig and fancontrol?!
Alright, I started dealing with pwmconfig and fancontrol. I immediately encountered the issue that I don’t have pwm1
Here’s the ls output:
device fan1_input fan1_label name power pwm1_enable subsystem uevent
Here’s the pwd:
/sys/devices/platform/asus-nb-wmi/hwmon/hwmon5
From what works for me:
echo 2 > pwm_enable - works
echo 1 > pwm_enable - says it’s an invalid value
echo 0 > pwm_enable - turns the fans on at full speed. Honestly, I’ve never heard them running so powerfully. However, I think you’re right; this fan has potential, and perhaps they could cool the processor effectively.
I don’t think they conflict but do your own reading, i cannot guarantee for sure. I have it deinstalled now in favor of Corectl (they worked both, i only needed the governor and not the frequency scaling), now i see the tlp service. I guess if they conflict it is automatically dis/enabled on install.
I tried disabling and completely removing TLP. After that, cpupower doesn’t retain its settings after reboot, which is strange but kind of okay. Then I set a limit, but still saw impressive numbers at 4300 MHz in htop.
But, okay, thanks for your help!
Okay… well. Now you have 2 modes: automatic and full speed. You can write a simple script and monitor the temperatur and when it reaches a threshold, it will switch to full speed until it is below the threshold again. That would need little work, but better than nothing when you do intensive tasks.
Many years ago, on a Xubuntu 14.04 i used this.
It is very old and probably needs heavy adapting to work on manjaro today, but as a starting point for a project
Thank you, I’ll try to write a script.
I’m curious, what could be causing this?
Maybe it’s worth opening the laptop and checking the thermal paste? I have some experience with laptop disassembly. But I’m not sure if it’s the right solution.
While I havent done exhaustive testing the above configuration seems to work for me with some fan spin up under load, otherwise quiet, and never hitting a major heat ceiling.
Thank you for the advice, I’ll give this option a try. A couple of questions:
Did your laptop also overheat?
What does “not power-profiles-daemon” mean in this context with TLP (not power-profiles-daemon)?
No overheating that I have observed.
Though it was louder while also being less energy efficient before I started tweaking. tlp conflicts with power-profiles-daemon so one cannot have both. power-profiles-daemon being a newer package that is supposed to integrate with desktop power management (being able to choose ‘performance’ in KDE power widget, etc).
Well … you can have both installed, but then you would need to mask power-profiles-daemon’s service.
I guess you have to share pictures of it so that someone is able to help you.
