Thermal throttling not resetting

J

jordi.torrents157f02e2

New member
Joined
Jul 28, 2023
Messages
4
I have an MSI Prestige 14 A12UC with Intel Core i7-1280P, 32GB RAM, 1TB SSD and Nvidia RTX3050 GPU. I have dual boot with Windows 11 but I use Linux (Ubuntu 23.04) for working in AI.

When I bring my laptop to the limit by training AI networks, it heats up and eventually thermal throttling is activated (less power consumption and CPU frequency, slower computer), that's understandable and ok. After cooling down, the computer seems to be recovered (low sensors temperature and fans are back to normal) but the issue is that the throttling never resets. My laptop remains super slow and unusable until a reboot is done. This is what Ubuntu says...

Screenshot from 2023-07-24 15-48-29.png


I've been checking all Ubuntu's thermal settings, installing and tweaking low level stuff, but I couldn't fix it. I'm starting to think it's related with some security feature in the BIOS. That's why I'm posting this here. I tried to update my BIOS to E14C6IMS.118, but the issue remains.

In Windows, the thermal throttling occurs but it is deactivated as soon as the laptop cools down (as expected).
 
Last edited:
I think the problem is obvious. Based on the provided information, the issue only happens to Linux system. MSI laptop does not official support Linux system, so the problem could be related to the compatibility with Linux.
 
jordi.torrents157f02e2 said:
There has to be a way to reset the throttling without rebooting Ubuntu. Could anyone give me some hint on where to look for it?
Click to expand...
It's the "Passive Trip Point" temperature that activates the thermal throttling,
it seems either your sensor has communication issues with Linux or your temperature doesn't go below the threshold.
Why don't you just set it at Balanced?
Do you really need High Performance for AI?
 

Attachments

  • 230720 GP66 BIOS Trip Point.jpg
    230720 GP66 BIOS Trip Point.jpg
    233.5 KB · Views: 643
scorpiitec15e702ef said:
Maybe this will somehow help"

MSI Laptops To Enjoy Better Linux Support Beginning With The 6.4 Kernel - Phoronix

www.phoronix.com www.phoronix.com
Click to expand...

Thanks! Ubuntu 23.10 (to be released next month) will use Linux Kernel 6.5 (source) which will contain MSI-EC Linux driver, so better MSI support!! It's a bit a shame that this driver hasn't been developed by MSI, god save the open-source community <3

I will try it out and post the results as soon as Ubuntu 23.10 is released, wish me luck.
 
drek4 said:
It's the "Passive Trip Point" temperature that activates the thermal throttling,
it seems either your sensor has communication issues with Linux or your temperature doesn't go below the threshold.
Why don't you just set it at Balanced?
Do you really need High Performance for AI?
Click to expand...

Thanks for pointing out the specific parameter involved in this. Tweaking the BIOS is gonna be my second option if Ubuntu's 23.10 update doesn't fix this. Anyway, both Balanced and Performance modes are disabled once temperatures got high. Also, working on Balanced doesn't prevent the thermal throttling to appear.
 
jordi.torrents157f02e2 said:
Thanks for pointing out the specific parameter involved in this. Tweaking the BIOS is gonna be my second option if Ubuntu's 23.10 update doesn't fix this. Anyway, both Balanced and Performance modes are disabled once temperatures got high. Also, working on Balanced doesn't prevent the thermal throttling to appear.
Click to expand...
Then lower the Active trip point 0 & 1 to lower temperatures, to postpone the thermal throttling from occurring.
 
jordi.torrents157f02e2 said:
Thanks! Ubuntu 23.10 (to be released next month) will use Linux Kernel 6.5 (source) which will contain MSI-EC Linux driver, so better MSI support!! It's a bit a shame that this driver hasn't been developed by MSI, god save the open-source community <3

I will try it out and post the results as soon as Ubuntu 23.10 is released, wish me luck.
Click to expand...
Good Luck!!!
 
Hey
I would like to revisit this issues. This is the only time I found someone having a similar issue. Whenever I GPU thermal throttles (at 86°C) it displays the exact same behavior as explained here. It will thermal throttle my machine, full blast all my fans, limit max load even when temps hit 20°C. Only way to "reset" the thermal throttle is by restarting the laptop.

I run windows 11 (upgrade from windows 10, problem happened in both). I have tried everything I can think of: Clean install windows, bios flash, driver updates, laptop full clean + thermal paste replacement. None with too much success.

The laptop is now +-4 years old. Sent it to MSI or vendor (can't remember) to fix this issue (1,5 years ago), under warrenty. Receive the laptop back, with the message: "Nothing wrong". After receiving my laptop back, the problem seemed to be fixed for about a month. However has persisted since.

Anyone know of anything I can try to solved this issue, I'm starting to believe I have receive a faulty GPU sensor, one that doesn't reset it throttle warning.

Thanks in advance
 
urepthor.vannijle155b02de said:
Hey
I would like to revisit this issues. This is the only time I found someone having a similar issue. Whenever I GPU thermal throttles (at 86°C) it displays the exact same behavior as explained here. It will thermal throttle my machine, full blast all my fans, limit max load even when temps hit 20°C. Only way to "reset" the thermal throttle is by restarting the laptop.
Click to expand...
The behavior is not normal.
Do you try to perform EC reset by long pressing the power button over 25 seconds? If not, I will suggest you do so.
You can use Furmark to test GPU performance.
1. Run Furmark and monitor the maximum power of GPU.
2. Let the laptop to cool down.
3. Run Furmark again, then check if you get the similar maximum power of GPU.
If the problem still persist, it is suggested to provide mode detailed information to figure out the problem.
1. Which model do you have?
2. How you trigger the issue?
 
I'm having the same kind of problem.

Laptop Bravo 15 c7ve (7735hs-4050), when CPU reach 95c automatically drops wattage up to 10W, fans go to turbo, then the CPU will get chill down to 40ish, but the wattage draw for the cpu keeps at 10W max, and the fans can't be turn off or just slow down, it gets stuck in turbo cool.

This happens in windows 11 (was shipped with this, I've already performed a lot of clean installs for different troubleshooting), also I've already tried the EC reset, I though would work but not at all. If anyone has any idea how to "unstuck" the cpu power will be really helpful.

I was having this issue so I just installed all windows updates that were pending... this updated my bios to the latest available but is the same as before.
 
I’ve already installed it… 2 days before “launch” date, after windows updated my bios I checked on msi website and it was already listed but at the moment with a future release date, unfortunately the same behavior keeps happening. I can “avoid it” but avoiding playing and reaching 95.6 c, but I believe is better to know how to revert this behavior of wattage drop than avoid “using the laptop as normal”
 
Somebody find a solution to this issue? I have the same problem the only solution to return to normal processor speed is rebooting system.
 
kikebeer said:
Somebody find a solution to this issue? I have the same problem the only solution to return to normal processor speed is rebooting system.
Click to expand...
Which model do you have? What is your problem? There are a few different situations from above, please at least descibe what temperature, wattage, and what are you doing when the issue happens.
 
kikebeer said:
Somebody find a solution to this issue? I have the same problem the only solution to return to normal processor speed is rebooting system.
Click to expand...
What I did in my scenario was to setup a thermal limit in my bios to 92c, so the laptop draws max power then throttle just a little and keep working fine.
 
Edric said:
Which model do you have? What is your problem? There are a few different situations from above, please at least descibe what temperature, wattage, and what are you doing when the issue happens.
Click to expand...

Bravo 15 C7VF , when i play any game (DIablo, 7 days to die, minecraft with shaders....) , any game that rise temperature of cpu. I check my bios but i dont see any limits like Alex said. The only solution that i find is before i launch a game uo my fans to maximum speed manually...but sometimes i forgot it and i need restart my computer to work normally...
 
kikebeer said:
Bravo 15 C7VF , when i play any game (DIablo, 7 days to die, minecraft with shaders....) , any game that rise temperature of cpu. I check my bios but i dont see any limits like Alex said. The only solution that i find is before i launch a game uo my fans to maximum speed manually...but sometimes i forgot it and i need restart my computer to work normally...
Click to expand...
I have the same model, and those settings are on the advanced bios, you can google how to go in but you need to be careful, the settings are not that well labeled, a couple of weeks ago there was an update on MSI center that make my laptop better, but I still leave my custom limits.
 
kikebeer said:
Bravo 15 C7VF , when i play any game (DIablo, 7 days to die, minecraft with shaders....) , any game that rise temperature of cpu. I check my bios but i dont see any limits like Alex said. The only solution that i find is before i launch a game uo my fans to maximum speed manually...but sometimes i forgot it and i need restart my computer to work normally...
Click to expand...
Check if you select extreme performance while gaming and make share the fan is spining.
For Bravo 15 C7VF, the CPU temperature should be around 85 under heavy loading.
If the temperature on your laptop is higher than that, please try to clean install the graphics driver to the latest version first.
Follow the FAQ to clean install AMD graphics driver: https://www.msi.com/faq/faq-3725
Disconnect the internet and use DDU to clean install Nvidia graphics driver.
If the laptop is still over heating, I would recommend to clean the laptop and repaste the thermal pad.
 
Back
Top