Re: HP Omnibook Ultra Flip 14 - power profiles

Re: HP Omnibook Ultra Flip 14 - power profiles

[Armin Wolf]

Am 26.06.25 um 15:20 schrieb Benjamin Hasselgren-Hallén:

> Hi again,
>
> dmesg: https://drive.benis.se/s/2crz7zPzkrzaqXN

The following message intrigues me:

	platform_profile: Failed to get profile for handler hp-wmi

Can you tell me the exact error message that is returned when reading the platform profile over sysfs?

> So I tested thermald again. So here are some results.
>
> Before:
> Running Valheim the power draw is 35w and the fps is 41.
>
> After I have installed thermald:
> Running Valheim the power draw is 44w and the fps is 46
>
> So it's working, but I also noticed that it doesn't matter if I change the power profile. Same performance and power draw. So something is still weird.

The ACPI code seems to suggest that your device support 7 profiles, while the hp-wmi driver only supports 4. Additionally the drivers fails
to properly mask out the platform profile value returned by the firmware, so that could be the reason why it is not working on your device.

Can you test kernel patches?

> Also the bug with the fun still running while the laptop is suspended is an issue.

I CCed the people from the thermal subsystem, maybe they know if this is expected behavior or a bug inside the int340x_thermal driver.

Thanks,
Armin Wolf

>
> Thank you very much for your guidance Armin!
>
>
>
>
> Best regards,
> Benjamin Hasselgren-Hallén
>
>
> On Wednesday, 25 June 2025 at 23:43, Armin Wolf <W_Armin@gmx.de> wrote:
>
>> Am 25.06.25 um 23:22 schrieb Benjamin Hasselgren-Hallén:
>>
>>> The laptop do complain about not being able to load hp-wmi during boot. Same on Fedora 42 as on Debian 13 (with kernel 6.15).
>>
>> Can you share the output of dmesg?
>>
>>> I did install thermald and that might have increased the power budget (which is good) - but introduced some bug that keeps the fans on even while suspended.
>>
>> Please elaborate.
>>
>> Thanks,
>> Armin Wolf
>>
>>> Best regards,
>>> Benjamin Hasselgren-Hallén
>>>
>>> On Wednesday, 25 June 2025 at 15:44, Armin Wolf W_Armin@gmx.de wrote:
>>>
>>>> Am 25.06.25 um 10:06 schrieb Benjamin Hasselgren-Hallén:
>>>>
>>>>> Hi,
>>>>>
>>>>> The acpidump is here: https://drive.benis.se/s/pcKLAL7i8zncX8q
>>>>>
>>>>> Best regards,
>>>>> Benjamin Hasselgren-Hallén
>>>> The hp-wmi driver should be able to control the platform profile on your device. Does this
>>>> driver load and register a platform profile handler?
>>>>
>>>> Additionally you might need to install and setup the intel thermal daemon for the platform
>>>> profile settings to work.
>>>>
>>>> Thanks,
>>>> Armin Wolf
>>>>
>>>>> On Monday, 23 June 2025 at 20:52, Armin Wolf W_Armin@gmx.de wrote:
>>>>>
>>>>>> Am 23.06.25 um 15:52 schrieb Benjamin Hasselgren-Hallén:
>>>>>>
>>>>>>> Hi dear Linux friends,
>>>>>>>
>>>>>>> This is very much a long shot and I understand if no one got any time or motivation for this.However, I am trying to understand how HP Omnibook Ultra Flip (a laptop with Lunar Lake platform) working with power profiles. The reason is that it seems to be very limited while running Linux (to 30 watts to be exact, no matter the power profile, this is for the whole laptop, to compare with something the Omnibook Ultra 14 with AMD Strix draws up to over 70 watts, sure more power hungry platform but still). Also the gpu performance is not as it should be.
>>>>>>> So if anyone got the time or so - let me know where to start digging!
>>>>>>>
>>>>>>> Best regards,
>>>>>>> Benjamin Hasselgren-Hallén
>>>>>>> Can you share the output of "acpidump"?
>>>>>>> Thanks,
>>>>>>> Armin Wolf

Re: HP Omnibook Ultra Flip 14 - power profiles

[Kurt Borja]

Hi all,

On Fri Jun 27, 2025 at 2:10 PM -03, Armin Wolf wrote:
> Am 26.06.25 um 15:20 schrieb Benjamin Hasselgren-Hallén:
>
>> Hi again,
>>
>> dmesg: https://drive.benis.se/s/2crz7zPzkrzaqXN
>
> The following message intrigues me:
>
> 	platform_profile: Failed to get profile for handler hp-wmi

This might be a regression.

This was reported a couple months ago and I completely forgot until I
saw this thread. See [1].

@Benjamin: Can you please check if your power profiles work on Linux
v6.12 (LTS)?

Also try:

$ cat /sys/firmware/acpi/platform_profile


[1] https://bugzilla.kernel.org/show_bug.cgi?id=220008

-- 
 ~ Kurt

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

Hi Kurt,

I do not experience the same error messages as in the bug report - no error message at all. It respons to changing power profile without any errors.
I also tried 6.12 but same behaviour as 6.15.3 (just getting some gpu glitches - problably because of Lunar Lake).

I am trying to understand how power profiles work - I guess on a high level it's controlled by uefi and the profile is set by the OS? Or is it more complicated than that? 



Best regards,
Benjamin Hasselgren-Hallén


On Friday, 27 June 2025 at 21:15, Kurt Borja <kuurtb@gmail.com> wrote:

> Hi all,
> 
> On Fri Jun 27, 2025 at 2:10 PM -03, Armin Wolf wrote:
> 
> > Am 26.06.25 um 15:20 schrieb Benjamin Hasselgren-Hallén:
> > 
> > > Hi again,
> > > 
> > > dmesg: https://drive.benis.se/s/2crz7zPzkrzaqXN
> > 
> > The following message intrigues me:
> > 
> > platform_profile: Failed to get profile for handler hp-wmi
> 
> 
> This might be a regression.
> 
> This was reported a couple months ago and I completely forgot until I
> saw this thread. See [1].
> 
> @Benjamin: Can you please check if your power profiles work on Linux
> v6.12 (LTS)?
> 
> Also try:
> 
> $ cat /sys/firmware/acpi/platform_profile
> 
> 
> [1] https://bugzilla.kernel.org/show_bug.cgi?id=220008
> 
> --
> ~ Kurt

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

One more thing: I noticed this during boot (before entering password for disk encryption)
https://drive.benis.se/s/5wMkEMKs6SELQzt

Thermal trip point bug and other stuff, might be something?



Best regards,
Benjamin Hasselgren-Hallén


On Friday, 27 June 2025 at 22:49, Benjamin Hasselgren-Hallén <benjamin@benis.se> wrote:

> Hi Kurt,
> 
> I do not experience the same error messages as in the bug report - no error message at all. It respons to changing power profile without any errors.
> I also tried 6.12 but same behaviour as 6.15.3 (just getting some gpu glitches - problably because of Lunar Lake).
> 
> I am trying to understand how power profiles work - I guess on a high level it's controlled by uefi and the profile is set by the OS? Or is it more complicated than that?
> 
> 
> 
> Best regards,
> Benjamin Hasselgren-Hallén
> 
> 
> 
> 
> On Friday, 27 June 2025 at 21:15, Kurt Borja kuurtb@gmail.com wrote:
> 
> > Hi all,
> > 
> > On Fri Jun 27, 2025 at 2:10 PM -03, Armin Wolf wrote:
> > 
> > > Am 26.06.25 um 15:20 schrieb Benjamin Hasselgren-Hallén:
> > > 
> > > > Hi again,
> > > > 
> > > > dmesg: https://drive.benis.se/s/2crz7zPzkrzaqXN
> > > 
> > > The following message intrigues me:
> > > 
> > > platform_profile: Failed to get profile for handler hp-wmi
> > 
> > This might be a regression.
> > 
> > This was reported a couple months ago and I completely forgot until I
> > saw this thread. See [1].
> > 
> > @Benjamin: Can you please check if your power profiles work on Linux
> > v6.12 (LTS)?
> > 
> > Also try:
> > 
> > $ cat /sys/firmware/acpi/platform_profile
> > 
> > [1] https://bugzilla.kernel.org/show_bug.cgi?id=220008
> > 
> > --
> > ~ Kurt

Re: HP Omnibook Ultra Flip 14 - power profiles

[Armin Wolf]

Am 27.06.25 um 23:56 schrieb Benjamin Hasselgren-Hallén:

> One more thing: I noticed this during boot (before entering password for disk encryption)
> https://drive.benis.se/s/5wMkEMKs6SELQzt
>
> Thermal trip point bug and other stuff, might be something?

Yes, it seems the trip code AML code try to access an array called CUZO that is only initialize
inside a ACPI control method called _WAK. Since _WAK is only called when the system wakes from
a sleep state CUZO remains initialized and thus causes this error.

Can you check if a BIOS update is available for your device?

Thanks,
Armin Wolf

>
> Best regards,
> Benjamin Hasselgren-Hallén
>
>
> On Friday, 27 June 2025 at 22:49, Benjamin Hasselgren-Hallén <benjamin@benis.se> wrote:
>
>> Hi Kurt,
>>
>> I do not experience the same error messages as in the bug report - no error message at all. It respons to changing power profile without any errors.
>> I also tried 6.12 but same behaviour as 6.15.3 (just getting some gpu glitches - problably because of Lunar Lake).
>>
>> I am trying to understand how power profiles work - I guess on a high level it's controlled by uefi and the profile is set by the OS? Or is it more complicated than that?
>>
>>
>>
>> Best regards,
>> Benjamin Hasselgren-Hallén
>>
>>
>>
>>
>> On Friday, 27 June 2025 at 21:15, Kurt Borja kuurtb@gmail.com wrote:
>>
>>> Hi all,
>>>
>>> On Fri Jun 27, 2025 at 2:10 PM -03, Armin Wolf wrote:
>>>
>>>> Am 26.06.25 um 15:20 schrieb Benjamin Hasselgren-Hallén:
>>>>
>>>>> Hi again,
>>>>>
>>>>> dmesg: https://drive.benis.se/s/2crz7zPzkrzaqXN
>>>> The following message intrigues me:
>>>>
>>>> platform_profile: Failed to get profile for handler hp-wmi
>>> This might be a regression.
>>>
>>> This was reported a couple months ago and I completely forgot until I
>>> saw this thread. See [1].
>>>
>>> @Benjamin: Can you please check if your power profiles work on Linux
>>> v6.12 (LTS)?
>>>
>>> Also try:
>>>
>>> $ cat /sys/firmware/acpi/platform_profile
>>>
>>> [1] https://bugzilla.kernel.org/show_bug.cgi?id=220008
>>>
>>> --
>>> ~ Kurt

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

On Saturday, 28 June 2025 at 02:40, Armin Wolf <W_Armin@gmx.de> wrote:

> Am 27.06.25 um 23:56 schrieb Benjamin Hasselgren-Hallén:
> 
> > One more thing: I noticed this during boot (before entering password for disk encryption)
> > https://drive.benis.se/s/5wMkEMKs6SELQzt
> > 
> > Thermal trip point bug and other stuff, might be something?
> 
> 
> Yes, it seems the trip code AML code try to access an array called CUZO that is only initialize
> inside a ACPI control method called _WAK. Since _WAK is only called when the system wakes from
> a sleep state CUZO remains initialized and thus causes this error.
> 
> Can you check if a BIOS update is available for your device?

I have the latest bios for my device.


> 
> Thanks,
> Armin Wolf
> 
> > Best regards,
> > Benjamin Hasselgren-Hallén
> > 
> > On Friday, 27 June 2025 at 22:49, Benjamin Hasselgren-Hallén benjamin@benis.se wrote:
> > 
> > > Hi Kurt,
> > > 
> > > I do not experience the same error messages as in the bug report - no error message at all. It respons to changing power profile without any errors.
> > > I also tried 6.12 but same behaviour as 6.15.3 (just getting some gpu glitches - problably because of Lunar Lake).
> > > 
> > > I am trying to understand how power profiles work - I guess on a high level it's controlled by uefi and the profile is set by the OS? Or is it more complicated than that?
> > > 
> > > Best regards,
> > > Benjamin Hasselgren-Hallén
> > > 
> > > On Friday, 27 June 2025 at 21:15, Kurt Borja kuurtb@gmail.com wrote:
> > > 
> > > > Hi all,
> > > > 
> > > > On Fri Jun 27, 2025 at 2:10 PM -03, Armin Wolf wrote:
> > > > 
> > > > > Am 26.06.25 um 15:20 schrieb Benjamin Hasselgren-Hallén:
> > > > > 
> > > > > > Hi again,
> > > > > > 
> > > > > > dmesg: https://drive.benis.se/s/2crz7zPzkrzaqXN
> > > > > > The following message intrigues me:
> > > > > 
> > > > > platform_profile: Failed to get profile for handler hp-wmi
> > > > > This might be a regression.
> > > > 
> > > > This was reported a couple months ago and I completely forgot until I
> > > > saw this thread. See [1].
> > > > 
> > > > @Benjamin: Can you please check if your power profiles work on Linux
> > > > v6.12 (LTS)?
> > > > 
> > > > Also try:
> > > > 
> > > > $ cat /sys/firmware/acpi/platform_profile
> > > > 
> > > > [1] https://bugzilla.kernel.org/show_bug.cgi?id=220008
> > > > 
> > > > --
> > > > ~ Kurt

Re: HP Omnibook Ultra Flip 14 - power profiles

[Kurt Borja]

On Fri Jun 27, 2025 at 5:49 PM -03, Benjamin Hasselgren-Hallén wrote:
> Hi Kurt,
>
> I do not experience the same error messages as in the bug report - no error message at all. It respons to changing power profile without any errors.
> I also tried 6.12 but same behaviour as 6.15.3 (just getting some gpu glitches - problably because of Lunar Lake).

Can you paste the output of `cat /sys/firmware/acpi/platform_profile` on
Linux v6.12?

>
> I am trying to understand how power profiles work - I guess on a high level it's controlled by uefi and the profile is set by the OS? Or is it more complicated than that? 

Yes, as quick overview. Power profiles are usually a capability of
individual devices. Drivers of these devices expose this profiles to
userspace through some predefined interfaces. Then your DE enviroment
exposes this to you.

KDE and Gnome use power-profiles-daemon. This program controlls mainly
two kernel interfaces: CPUFreq and platform-profile (there may be more).

You can read about CPUFreq here [1], although I don't know much about
it. The platform-profile interface may be used by any platform driver -
in your case this is the hp-wmi driver.

This driver exposes some firmware defined profiles to sysfs trough the
following paths.

	/sys/firmware/acpi/
	/sys/class/platform-profile/

You won't find the latter in the Linux LTS version though. You will find
documentation on these paths here [2] and [3] respectively.

PD: Please, avoid top-posting in kernel mailing lists, interleaved style
is very much preferred [4].


[1] https://docs.kernel.org/admin-guide/pm/cpufreq.html
[2] https://docs.kernel.org/admin-guide/abi-testing-files.html#abi-file-testing-sysfs-platform-profile
[3] https://docs.kernel.org/admin-guide/abi-testing-files.html#abi-file-testing-sysfs-class-platform-profile
[4] https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

-- 
 ~ Kurt

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

Best regards,
Benjamin Hasselgren-Hallén


On Saturday, 28 June 2025 at 02:45, Kurt Borja <kuurtb@gmail.com> wrote:

> On Fri Jun 27, 2025 at 5:49 PM -03, Benjamin Hasselgren-Hallén wrote:
> 
> > Hi Kurt,
> > 
> > I do not experience the same error messages as in the bug report - no error message at all. It respons to changing power profile without any errors.
> > I also tried 6.12 but same behaviour as 6.15.3 (just getting some gpu glitches - problably because of Lunar Lake).
> 
> 
> Can you paste the output of `cat /sys/firmware/acpi/platform_profile` on
> Linux v6.12?

cat /sys/firmware/acpi/platform_profile_choices 
cool quiet balanced performance

cat /sys/firmware/acpi/platform_profile
cat: /sys/firmware/acpi/platform_profile: Invalid argument

> 
> > I am trying to understand how power profiles work - I guess on a high level it's controlled by uefi and the profile is set by the OS? Or is it more complicated than that?
> 
> 
> Yes, as quick overview. Power profiles are usually a capability of
> individual devices. Drivers of these devices expose this profiles to
> userspace through some predefined interfaces. Then your DE enviroment
> exposes this to you.
> 
> KDE and Gnome use power-profiles-daemon. This program controlls mainly
> two kernel interfaces: CPUFreq and platform-profile (there may be more).
> 
> You can read about CPUFreq here [1], although I don't know much about
> it. The platform-profile interface may be used by any platform driver -
> in your case this is the hp-wmi driver.
> 
> This driver exposes some firmware defined profiles to sysfs trough the
> following paths.
> 
> /sys/firmware/acpi/
> /sys/class/platform-profile/
> 
> You won't find the latter in the Linux LTS version though. You will find
> documentation on these paths here [2] and [3] respectively.
> 
> PD: Please, avoid top-posting in kernel mailing lists, interleaved style
> is very much preferred [4].
> 
> 
> [1] https://docs.kernel.org/admin-guide/pm/cpufreq.html
> [2] https://docs.kernel.org/admin-guide/abi-testing-files.html#abi-file-testing-sysfs-platform-profile
> [3] https://docs.kernel.org/admin-guide/abi-testing-files.html#abi-file-testing-sysfs-class-platform-profile
> [4] https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
> 
> --
> ~ Kurt

Re: HP Omnibook Ultra Flip 14 - power profiles

[Armin Wolf]

Am 30.06.25 um 14:32 schrieb Benjamin Hasselgren-Hallén:

>
>
>
> Best regards,
> Benjamin Hasselgren-Hallén
>
>
> On Saturday, 28 June 2025 at 02:45, Kurt Borja <kuurtb@gmail.com> wrote:
>
>> On Fri Jun 27, 2025 at 5:49 PM -03, Benjamin Hasselgren-Hallén wrote:
>>
>>> Hi Kurt,
>>>
>>> I do not experience the same error messages as in the bug report - no error message at all. It respons to changing power profile without any errors.
>>> I also tried 6.12 but same behaviour as 6.15.3 (just getting some gpu glitches - problably because of Lunar Lake).
>>
>> Can you paste the output of `cat /sys/firmware/acpi/platform_profile` on
>> Linux v6.12?
> cat /sys/firmware/acpi/platform_profile_choices
> cool quiet balanced performance
>
> cat /sys/firmware/acpi/platform_profile
> cat: /sys/firmware/acpi/platform_profile: Invalid argument

Good, it seems that the firmware returns additional data when querying the current platform profile.
As the driver does not mask-out this data, he cannot determine what kind of platform profile is currently
selected and thus returns -EINVAL.+

I can provide you with an experimental patch, could you test that?

Thanks,
Armin Wolf

>>> I am trying to understand how power profiles work - I guess on a high level it's controlled by uefi and the profile is set by the OS? Or is it more complicated than that?
>>
>> Yes, as quick overview. Power profiles are usually a capability of
>> individual devices. Drivers of these devices expose this profiles to
>> userspace through some predefined interfaces. Then your DE enviroment
>> exposes this to you.
>>
>> KDE and Gnome use power-profiles-daemon. This program controlls mainly
>> two kernel interfaces: CPUFreq and platform-profile (there may be more).
>>
>> You can read about CPUFreq here [1], although I don't know much about
>> it. The platform-profile interface may be used by any platform driver -
>> in your case this is the hp-wmi driver.
>>
>> This driver exposes some firmware defined profiles to sysfs trough the
>> following paths.
>>
>> /sys/firmware/acpi/
>> /sys/class/platform-profile/
>>
>> You won't find the latter in the Linux LTS version though. You will find
>> documentation on these paths here [2] and [3] respectively.
>>
>> PD: Please, avoid top-posting in kernel mailing lists, interleaved style
>> is very much preferred [4].
>>
>>
>> [1] https://docs.kernel.org/admin-guide/pm/cpufreq.html
>> [2] https://docs.kernel.org/admin-guide/abi-testing-files.html#abi-file-testing-sysfs-platform-profile
>> [3] https://docs.kernel.org/admin-guide/abi-testing-files.html#abi-file-testing-sysfs-class-platform-profile
>> [4] https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
>>
>> --
>> ~ Kurt

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

On Monday, 30 June 2025 at 14:39, Armin Wolf <W_Armin@gmx.de> wrote:

> Am 30.06.25 um 14:32 schrieb Benjamin Hasselgren-Hallén:
> 
> > Best regards,
> > Benjamin Hasselgren-Hallén
> > 
> > On Saturday, 28 June 2025 at 02:45, Kurt Borja kuurtb@gmail.com wrote:
> > 
> > > On Fri Jun 27, 2025 at 5:49 PM -03, Benjamin Hasselgren-Hallén wrote:
> > > 
> > > > Hi Kurt,
> > > > 
> > > > I do not experience the same error messages as in the bug report - no error message at all. It respons to changing power profile without any errors.
> > > > I also tried 6.12 but same behaviour as 6.15.3 (just getting some gpu glitches - problably because of Lunar Lake).
> > > 
> > > Can you paste the output of `cat /sys/firmware/acpi/platform_profile` on
> > > Linux v6.12?
> > > cat /sys/firmware/acpi/platform_profile_choices
> > > cool quiet balanced performance
> > 
> > cat /sys/firmware/acpi/platform_profile
> > cat: /sys/firmware/acpi/platform_profile: Invalid argument
> 
> 
> Good, it seems that the firmware returns additional data when querying the current platform profile.
> As the driver does not mask-out this data, he cannot determine what kind of platform profile is currently
> selected and thus returns -EINVAL.+
> 
> I can provide you with an experimental patch, could you test that?

Yes! I am trying to compile 6.16-rc3 atm and the first time I am doing this, but I will happily test a patch.

> 
> Thanks,
> Armin Wolf
> 
> > > > I am trying to understand how power profiles work - I guess on a high level it's controlled by uefi and the profile is set by the OS? Or is it more complicated than that?
> > > 
> > > Yes, as quick overview. Power profiles are usually a capability of
> > > individual devices. Drivers of these devices expose this profiles to
> > > userspace through some predefined interfaces. Then your DE enviroment
> > > exposes this to you.
> > > 
> > > KDE and Gnome use power-profiles-daemon. This program controlls mainly
> > > two kernel interfaces: CPUFreq and platform-profile (there may be more).
> > > 
> > > You can read about CPUFreq here [1], although I don't know much about
> > > it. The platform-profile interface may be used by any platform driver -
> > > in your case this is the hp-wmi driver.
> > > 
> > > This driver exposes some firmware defined profiles to sysfs trough the
> > > following paths.
> > > 
> > > /sys/firmware/acpi/
> > > /sys/class/platform-profile/
> > > 
> > > You won't find the latter in the Linux LTS version though. You will find
> > > documentation on these paths here [2] and [3] respectively.
> > > 
> > > PD: Please, avoid top-posting in kernel mailing lists, interleaved style
> > > is very much preferred [4].
> > > 
> > > [1] https://docs.kernel.org/admin-guide/pm/cpufreq.html
> > > [2] https://docs.kernel.org/admin-guide/abi-testing-files.html#abi-file-testing-sysfs-platform-profile
> > > [3] https://docs.kernel.org/admin-guide/abi-testing-files.html#abi-file-testing-sysfs-class-platform-profile
> > > [4] https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
> > > 
> > > --
> > > ~ Kurt

Re: HP Omnibook Ultra Flip 14 - power profiles

[Armin Wolf]

[-- Attachment #1: Type: text/plain, Size: 3504 bytes --]

Am 30.06.25 um 14:46 schrieb Benjamin Hasselgren-Hallén:

>
> On Monday, 30 June 2025 at 14:39, Armin Wolf <W_Armin@gmx.de> wrote:
>
>> Am 30.06.25 um 14:32 schrieb Benjamin Hasselgren-Hallén:
>>
>>> Best regards,
>>> Benjamin Hasselgren-Hallén
>>>
>>> On Saturday, 28 June 2025 at 02:45, Kurt Borja kuurtb@gmail.com wrote:
>>>
>>>> On Fri Jun 27, 2025 at 5:49 PM -03, Benjamin Hasselgren-Hallén wrote:
>>>>
>>>>> Hi Kurt,
>>>>>
>>>>> I do not experience the same error messages as in the bug report - no error message at all. It respons to changing power profile without any errors.
>>>>> I also tried 6.12 but same behaviour as 6.15.3 (just getting some gpu glitches - problably because of Lunar Lake).
>>>> Can you paste the output of `cat /sys/firmware/acpi/platform_profile` on
>>>> Linux v6.12?
>>>> cat /sys/firmware/acpi/platform_profile_choices
>>>> cool quiet balanced performance
>>> cat /sys/firmware/acpi/platform_profile
>>> cat: /sys/firmware/acpi/platform_profile: Invalid argument
>>
>> Good, it seems that the firmware returns additional data when querying the current platform profile.
>> As the driver does not mask-out this data, he cannot determine what kind of platform profile is currently
>> selected and thus returns -EINVAL.+
>>
>> I can provide you with an experimental patch, could you test that?
> Yes! I am trying to compile 6.16-rc3 atm and the first time I am doing this, but I will happily test a patch.

Alright, i attach the experimental patch. Please note that this might not fully solve the problem, because the
driver also needs support for the additional platform profiles available on your machine.

However after setting the platform profile the driver should be able to read the current platform profile setting
without any errors.

Thanks,
Armin Wolf

>> Thanks,
>> Armin Wolf
>>
>>>>> I am trying to understand how power profiles work - I guess on a high level it's controlled by uefi and the profile is set by the OS? Or is it more complicated than that?
>>>> Yes, as quick overview. Power profiles are usually a capability of
>>>> individual devices. Drivers of these devices expose this profiles to
>>>> userspace through some predefined interfaces. Then your DE enviroment
>>>> exposes this to you.
>>>>
>>>> KDE and Gnome use power-profiles-daemon. This program controlls mainly
>>>> two kernel interfaces: CPUFreq and platform-profile (there may be more).
>>>>
>>>> You can read about CPUFreq here [1], although I don't know much about
>>>> it. The platform-profile interface may be used by any platform driver -
>>>> in your case this is the hp-wmi driver.
>>>>
>>>> This driver exposes some firmware defined profiles to sysfs trough the
>>>> following paths.
>>>>
>>>> /sys/firmware/acpi/
>>>> /sys/class/platform-profile/
>>>>
>>>> You won't find the latter in the Linux LTS version though. You will find
>>>> documentation on these paths here [2] and [3] respectively.
>>>>
>>>> PD: Please, avoid top-posting in kernel mailing lists, interleaved style
>>>> is very much preferred [4].
>>>>
>>>> [1] https://docs.kernel.org/admin-guide/pm/cpufreq.html
>>>> [2] https://docs.kernel.org/admin-guide/abi-testing-files.html#abi-file-testing-sysfs-platform-profile
>>>> [3] https://docs.kernel.org/admin-guide/abi-testing-files.html#abi-file-testing-sysfs-class-platform-profile
>>>> [4] https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
>>>>
>>>> --
>>>> ~ Kurt

[-- Attachment #2: 0001-platform-x86-hp-wmi-Use-bitmask-when-accessing-the-t.patch --]
[-- Type: text/x-patch, Size: 2688 bytes --]

From aed503e4deb4b12e05a7f943deea681695c93ca3 Mon Sep 17 00:00:00 2001
From: Armin Wolf <W_Armin@gmx.de>
Date: Mon, 30 Jun 2025 16:35:16 +0200
Subject: [PATCH] platform/x86: hp-wmi: Use bitmask when accessing the thermal
 profile
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

On the HP Omnibook Ultra Flip 14 the second and third byte inside
the result of HPWMI_THERMAL_PROFILE_QUERY contains additional data.
Currently this additional data is not properly masked out, making it
impossible to read the thermal profile on such machines.

Fix this by using a bitmask to mask-out everything but the first byte
when reading the thermal profile. The same is done when writing the
thermal profile for the sake of symmetry.

Reported-by: Benjamin Hasselgren-Hallén <benjamin@benis.se>
Closes: https://lore.kernel.org/platform-driver-x86/X-40AqXfdmQw5shUOk3VSaHSXmwJYWHPmDDMLyGUH6GpMt56ty5SbNg8EVfyI_uC9J07uqZ2TtGJmmpB_x8-xpcVOw29fnKzJZ4n9L0x78A=@benis.se/
Signed-off-by: Armin Wolf <W_Armin@gmx.de>
---
 drivers/platform/x86/hp/hp-wmi.c | 19 ++++++++++++++++---
 1 file changed, 16 insertions(+), 3 deletions(-)

diff --git a/drivers/platform/x86/hp/hp-wmi.c b/drivers/platform/x86/hp/hp-wmi.c
index db5fdee2109c..290d8d4c6f1d 100644
--- a/drivers/platform/x86/hp/hp-wmi.c
+++ b/drivers/platform/x86/hp/hp-wmi.c
@@ -13,6 +13,8 @@
 
 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 
+#include <linux/bits.h>
+#include <linux/bitfield.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/init.h>
@@ -252,6 +254,9 @@ enum hp_thermal_profile {
 	HP_THERMAL_PROFILE_QUIET		= 0x03,
 };
 
+/* The second and third byte contains unknown data, so both need to be masked out. */
+#define HP_THERMAL_PROFILE_MASK GENMASK(7, 0)
+
 #define IS_HWBLOCKED(x) ((x & HPWMI_POWER_FW_OR_HW) != HPWMI_POWER_FW_OR_HW)
 #define IS_SWBLOCKED(x) !(x & HPWMI_POWER_SOFT)
 
@@ -1431,13 +1436,21 @@ static int platform_profile_omen_set(struct device *dev,
 
 static int thermal_profile_get(void)
 {
-	return hp_wmi_read_int(HPWMI_THERMAL_PROFILE_QUERY);
+	int ret;
+
+	ret = hp_wmi_read_int(HPWMI_THERMAL_PROFILE_QUERY);
+	if (ret < 0)
+		return ret;
+
+	return FIELD_GET(HP_THERMAL_PROFILE_MASK, ret);
 }
 
 static int thermal_profile_set(int thermal_profile)
 {
-	return hp_wmi_perform_query(HPWMI_THERMAL_PROFILE_QUERY, HPWMI_WRITE, &thermal_profile,
-							   sizeof(thermal_profile), 0);
+	int value = FIELD_PREP(HP_THERMAL_PROFILE_MASK, thermal_profile);
+
+	return hp_wmi_perform_query(HPWMI_THERMAL_PROFILE_QUERY, HPWMI_WRITE, &value,
+				    sizeof(thermal_profile), 0);
 }
 
 static int hp_wmi_platform_profile_get(struct device *dev,
-- 
2.39.5

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

Best regards,
Benjamin Hasselgren-Hallén


On Monday, 30 June 2025 at 16:50, Armin Wolf <W_Armin@gmx.de> wrote:

> Am 30.06.25 um 14:46 schrieb Benjamin Hasselgren-Hallén:
> 
> > On Monday, 30 June 2025 at 14:39, Armin Wolf W_Armin@gmx.de wrote:
> > 
> > > Am 30.06.25 um 14:32 schrieb Benjamin Hasselgren-Hallén:
> > > 
> > > > Best regards,
> > > > Benjamin Hasselgren-Hallén
> > > > 
> > > > On Saturday, 28 June 2025 at 02:45, Kurt Borja kuurtb@gmail.com wrote:
> > > > 
> > > > > On Fri Jun 27, 2025 at 5:49 PM -03, Benjamin Hasselgren-Hallén wrote:
> > > > > 
> > > > > > Hi Kurt,
> > > > > > 
> > > > > > I do not experience the same error messages as in the bug report - no error message at all. It respons to changing power profile without any errors.
> > > > > > I also tried 6.12 but same behaviour as 6.15.3 (just getting some gpu glitches - problably because of Lunar Lake).
> > > > > > Can you paste the output of `cat /sys/firmware/acpi/platform_profile` on
> > > > > > Linux v6.12?
> > > > > > cat /sys/firmware/acpi/platform_profile_choices
> > > > > > cool quiet balanced performance
> > > > > > cat /sys/firmware/acpi/platform_profile
> > > > > > cat: /sys/firmware/acpi/platform_profile: Invalid argument
> > > 
> > > Good, it seems that the firmware returns additional data when querying the current platform profile.
> > > As the driver does not mask-out this data, he cannot determine what kind of platform profile is currently
> > > selected and thus returns -EINVAL.+
> > > 
> > > I can provide you with an experimental patch, could you test that?
> > > Yes! I am trying to compile 6.16-rc3 atm and the first time I am doing this, but I will happily test a patch.
> 
> 
> Alright, i attach the experimental patch. Please note that this might not fully solve the problem, because the
> driver also needs support for the additional platform profiles available on your machine.

I think I managed to patch a kernel that I installed. No performance change. 

Then I also tried 
cat /sys/firmware/acpi/platform_profile
on a non-patched kernel and I get the result (balanced) on kernel 6.15 (but not on 6.12). 

Dunno what to do with this tho? What have we learned?


> 
> However after setting the platform profile the driver should be able to read the current platform profile setting
> without any errors.
> 
> Thanks,
> Armin Wolf
> 
> > > Thanks,
> > > Armin Wolf
> > > 
> > > > > > I am trying to understand how power profiles work - I guess on a high level it's controlled by uefi and the profile is set by the OS? Or is it more complicated than that?
> > > > > > Yes, as quick overview. Power profiles are usually a capability of
> > > > > > individual devices. Drivers of these devices expose this profiles to
> > > > > > userspace through some predefined interfaces. Then your DE enviroment
> > > > > > exposes this to you.
> > > > > 
> > > > > KDE and Gnome use power-profiles-daemon. This program controlls mainly
> > > > > two kernel interfaces: CPUFreq and platform-profile (there may be more).
> > > > > 
> > > > > You can read about CPUFreq here [1], although I don't know much about
> > > > > it. The platform-profile interface may be used by any platform driver -
> > > > > in your case this is the hp-wmi driver.
> > > > > 
> > > > > This driver exposes some firmware defined profiles to sysfs trough the
> > > > > following paths.
> > > > > 
> > > > > /sys/firmware/acpi/
> > > > > /sys/class/platform-profile/
> > > > > 
> > > > > You won't find the latter in the Linux LTS version though. You will find
> > > > > documentation on these paths here [2] and [3] respectively.
> > > > > 
> > > > > PD: Please, avoid top-posting in kernel mailing lists, interleaved style
> > > > > is very much preferred [4].
> > > > > 
> > > > > [1] https://docs.kernel.org/admin-guide/pm/cpufreq.html
> > > > > [2] https://docs.kernel.org/admin-guide/abi-testing-files.html#abi-file-testing-sysfs-platform-profile
> > > > > [3] https://docs.kernel.org/admin-guide/abi-testing-files.html#abi-file-testing-sysfs-class-platform-profile
> > > > > [4] https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
> > > > > 
> > > > > --
> > > > > ~ Kurt

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

Not sure what sysfs means? 

/sys/firmware/acpi# cat platform_profile_choices 
cool quiet balanced performance

I don't know how to read the error message for not loading hp-wmi, can you guide me?



Best regards,
Benjamin Hasselgren-Hallén


On Friday, 27 June 2025 at 19:10, Armin Wolf <W_Armin@gmx.de> wrote:

> Am 26.06.25 um 15:20 schrieb Benjamin Hasselgren-Hallén:
> 
> > Hi again,
> > 
> > dmesg: https://drive.benis.se/s/2crz7zPzkrzaqXN
> 
> 
> The following message intrigues me:
> 
> platform_profile: Failed to get profile for handler hp-wmi
> 
> Can you tell me the exact error message that is returned when reading the platform profile over sysfs?
> 
> > So I tested thermald again. So here are some results.
> > 
> > Before:
> > Running Valheim the power draw is 35w and the fps is 41.
> > 
> > After I have installed thermald:
> > Running Valheim the power draw is 44w and the fps is 46
> > 
> > So it's working, but I also noticed that it doesn't matter if I change the power profile. Same performance and power draw. So something is still weird.
> 
> 
> The ACPI code seems to suggest that your device support 7 profiles, while the hp-wmi driver only supports 4. Additionally the drivers fails
> to properly mask out the platform profile value returned by the firmware, so that could be the reason why it is not working on your device.
> 
> Can you test kernel patches?
> 
> > Also the bug with the fun still running while the laptop is suspended is an issue.
> 
> 
> I CCed the people from the thermal subsystem, maybe they know if this is expected behavior or a bug inside the int340x_thermal driver.
> 
> Thanks,
> Armin Wolf
> 
> > Thank you very much for your guidance Armin!
> > 
> > Best regards,
> > Benjamin Hasselgren-Hallén
> > 
> > On Wednesday, 25 June 2025 at 23:43, Armin Wolf W_Armin@gmx.de wrote:
> > 
> > > Am 25.06.25 um 23:22 schrieb Benjamin Hasselgren-Hallén:
> > > 
> > > > The laptop do complain about not being able to load hp-wmi during boot. Same on Fedora 42 as on Debian 13 (with kernel 6.15).
> > > 
> > > Can you share the output of dmesg?
> > > 
> > > > I did install thermald and that might have increased the power budget (which is good) - but introduced some bug that keeps the fans on even while suspended.
> > > 
> > > Please elaborate.
> > > 
> > > Thanks,
> > > Armin Wolf
> > > 
> > > > Best regards,
> > > > Benjamin Hasselgren-Hallén
> > > > 
> > > > On Wednesday, 25 June 2025 at 15:44, Armin Wolf W_Armin@gmx.de wrote:
> > > > 
> > > > > Am 25.06.25 um 10:06 schrieb Benjamin Hasselgren-Hallén:
> > > > > 
> > > > > > Hi,
> > > > > > 
> > > > > > The acpidump is here: https://drive.benis.se/s/pcKLAL7i8zncX8q
> > > > > > 
> > > > > > Best regards,
> > > > > > Benjamin Hasselgren-Hallén
> > > > > > The hp-wmi driver should be able to control the platform profile on your device. Does this
> > > > > > driver load and register a platform profile handler?
> > > > > 
> > > > > Additionally you might need to install and setup the intel thermal daemon for the platform
> > > > > profile settings to work.
> > > > > 
> > > > > Thanks,
> > > > > Armin Wolf
> > > > > 
> > > > > > On Monday, 23 June 2025 at 20:52, Armin Wolf W_Armin@gmx.de wrote:
> > > > > > 
> > > > > > > Am 23.06.25 um 15:52 schrieb Benjamin Hasselgren-Hallén:
> > > > > > > 
> > > > > > > > Hi dear Linux friends,
> > > > > > > > 
> > > > > > > > This is very much a long shot and I understand if no one got any time or motivation for this.However, I am trying to understand how HP Omnibook Ultra Flip (a laptop with Lunar Lake platform) working with power profiles. The reason is that it seems to be very limited while running Linux (to 30 watts to be exact, no matter the power profile, this is for the whole laptop, to compare with something the Omnibook Ultra 14 with AMD Strix draws up to over 70 watts, sure more power hungry platform but still). Also the gpu performance is not as it should be.
> > > > > > > > So if anyone got the time or so - let me know where to start digging!
> > > > > > > > 
> > > > > > > > Best regards,
> > > > > > > > Benjamin Hasselgren-Hallén
> > > > > > > > Can you share the output of "acpidump"?
> > > > > > > > Thanks,
> > > > > > > > Armin Wolf

Re: HP Omnibook Ultra Flip 14 - power profiles

[Zhang, Rui]

> > So I tested thermald again. So here are some results.
> > 
> > Before:
> > Running Valheim the power draw is 35w and the fps is 41.
> > 
> > After I have installed thermald:
> > Running Valheim the power draw is 44w and the fps is 46
> > 
> > So it's working,

Good to know that thermald helps.

can you please also attach the turbostat output? say "turbostat -o ts.log
sleep 1"

> >  but I also noticed that it doesn't matter if I change the power
> > profile. Same performance and power draw. So something is still
> > weird.
> 
> The ACPI code seems to suggest that your device support 7 profiles,
> while the hp-wmi driver only supports 4. Additionally the drivers fails
> to properly mask out the platform profile value returned by the
> firmware, so that could be the reason why it is not working on your
> device.
> 
> Can you test kernel patches?
> 
> > Also the bug with the fun still running while the laptop is suspended
> > is an issue.
> 
> I CCed the people from the thermal subsystem, maybe they know if this
> is expected behavior or a bug inside the int340x_thermal driver.
> 
how do you suspend the system? and can you attach the dmesg output after
system "suspended" and resumed?

thanks,
rui

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

Best regards,
Benjamin Hasselgren-Hallén


On Monday, 30 June 2025 at 03:18, Zhang, Rui <rui.zhang@intel.com> wrote:

> > > So I tested thermald again. So here are some results.
> > > 
> > > Before:
> > > Running Valheim the power draw is 35w and the fps is 41.
> > > 
> > > After I have installed thermald:
> > > Running Valheim the power draw is 44w and the fps is 46
> > > 
> > > So it's working,
> 
> 
> Good to know that thermald helps.
> 
> can you please also attach the turbostat output? say "turbostat -o ts.log
> sleep 1"

https://drive.benis.se/s/KtRtCQXi7mS8GMi 
There you go. I am compiling the linux kernel atm so the laptop is working very hard (I am thinking of logging).

> 
> > > but I also noticed that it doesn't matter if I change the power
> > > profile. Same performance and power draw. So something is still
> > > weird.
> > 
> > The ACPI code seems to suggest that your device support 7 profiles,
> > while the hp-wmi driver only supports 4. Additionally the drivers fails
> > to properly mask out the platform profile value returned by the
> > firmware, so that could be the reason why it is not working on your
> > device.
> > 
> > Can you test kernel patches?
> > 
> > > Also the bug with the fun still running while the laptop is suspended
> > > is an issue.
> > 
> > I CCed the people from the thermal subsystem, maybe they know if this
> > is expected behavior or a bug inside the int340x_thermal driver.
> 
> how do you suspend the system? and can you attach the dmesg output after
> system "suspended" and resumed?

Will test later!

> 
> thanks,
> rui

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

Best regards,
Benjamin Hasselgren-Hallén


On Monday, 30 June 2025 at 14:48, Benjamin Hasselgren-Hallén <benjamin@benis.se> wrote:

> 
> 
> 
> 
> Best regards,
> Benjamin Hasselgren-Hallén
> 
> 
> 
> 
> On Monday, 30 June 2025 at 03:18, Zhang, Rui rui.zhang@intel.com wrote:
> 
> > > > So I tested thermald again. So here are some results.
> > > > 
> > > > Before:
> > > > Running Valheim the power draw is 35w and the fps is 41.
> > > > 
> > > > After I have installed thermald:
> > > > Running Valheim the power draw is 44w and the fps is 46
> > > > 
> > > > So it's working,
> > 
> > Good to know that thermald helps.
> > 
> > can you please also attach the turbostat output? say "turbostat -o ts.log
> > sleep 1"
> 
> 
> https://drive.benis.se/s/KtRtCQXi7mS8GMi
> There you go. I am compiling the linux kernel atm so the laptop is working very hard (I am thinking of logging).
> 
> > > > but I also noticed that it doesn't matter if I change the power
> > > > profile. Same performance and power draw. So something is still
> > > > weird.
> > > 
> > > The ACPI code seems to suggest that your device support 7 profiles,
> > > while the hp-wmi driver only supports 4. Additionally the drivers fails
> > > to properly mask out the platform profile value returned by the
> > > firmware, so that could be the reason why it is not working on your
> > > device.
> > > 
> > > Can you test kernel patches?
> > > 
> > > > Also the bug with the fun still running while the laptop is suspended
> > > > is an issue.
> > > 
> > > I CCed the people from the thermal subsystem, maybe they know if this
> > > is expected behavior or a bug inside the int340x_thermal driver.
> > 
> > how do you suspend the system? and can you attach the dmesg output after
> > system "suspended" and resumed?
> 
> 
> Will test later!

https://drive.benis.se/s/sBZSFGAskp7YZy8

I hope this is correct.

> 
> > thanks,
> > rui

Re: HP Omnibook Ultra Flip 14 - power profiles

[Zhang, Rui]

> > > 
> > > how do you suspend the system? and can you attach the dmesg output
> > > after
> > > system "suspended" and resumed?
> > 
> > 
> > Will test later!
> 
> https://drive.benis.se/s/sBZSFGAskp7YZy8
> 
> I hope this is correct.
> 
> 
Yeah, the log shows the system indeed ran a s2idle cycle.
But there is nothing suspicious in the log. The ACPI errors should always
be there no matter if thermald is activated or not.

can you please do the following test?
1. make sure thermald is not activated
2. do a s2idle cycle and make sure that there is no fan spinning
3. launch thermald with --loglevel=debug
4. do a s2idle cycle and make sure that the fan is indeed spinning
5. attach the dmesg output and the output of "journalctl --since=-1h |
grep thermal" and let's see what changes thermald has made.

thanks,
rui

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

Best regards,
Benjamin Hasselgren-Hallén


On Tuesday, 1 July 2025 at 09:02, Zhang, Rui <rui.zhang@intel.com> wrote:

> > > > how do you suspend the system? and can you attach the dmesg output
> > > > after
> > > > system "suspended" and resumed?
> > > 
> > > Will test later!
> > 
> > https://drive.benis.se/s/sBZSFGAskp7YZy8
> > 
> > I hope this is correct.
> 
> Yeah, the log shows the system indeed ran a s2idle cycle.
> But there is nothing suspicious in the log. The ACPI errors should always
> be there no matter if thermald is activated or not.
> 
> can you please do the following test?
> 1. make sure thermald is not activated
> 2. do a s2idle cycle and make sure that there is no fan spinning
> 3. launch thermald with --loglevel=debug
> 4. do a s2idle cycle and make sure that the fan is indeed spinning
> 5. attach the dmesg output and the output of "journalctl --since=-1h |
> grep thermal" and let's see what changes thermald has made.

with thermald
https://drive.benis.se/s/bF5AfDGBw6DFNZt

without thermald
https://drive.benis.se/s/47xJdg33ayHerDF

journald
https://drive.benis.se/s/8JdDJG2bFbHeDmz

I hopefully got everything by your instructions - otherwise let me know! 

> 
> thanks,
> rui

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

Looping in another Omnibook Ultra Flip user.



Best regards,
Benjamin Hasselgren-Hallén


On Monday, 7 July 2025 at 21:55, Benjamin Hasselgren-Hallén <benjamin@benis.se> wrote:

> 
> 
> 
> 
> Best regards,
> Benjamin Hasselgren-Hallén
> 
> 
> 
> 
> On Wednesday, 2 July 2025 at 10:00, Zhang, Rui rui.zhang@intel.com wrote:
> 
> > Remove the list as I want to grab more details.
> > 
> > On Tue, 2025-07-01 at 07:44 +0000, Benjamin Hasselgren-Hallén wrote:
> > 
> > > with thermald
> > > https://drive.benis.se/s/bF5AfDGBw6DFNZt
> > > 
> > > without thermald
> > > https://drive.benis.se/s/47xJdg33ayHerDF
> > 
> > There is no much difference, which is expected.
> > 
> > > journald
> > > https://drive.benis.se/s/8JdDJG2bFbHeDmz
> > 
> > The log doesn't have anything useful.
> > It shows your last launch of thermald failed
> > 
> > Jul 01 09:38:36 computer thermald[10950]: Couldn't get lock file 10950
> > Jul 01 09:38:36 computer thermald[10950]: An instance of thermald is
> > already running, exiting ...
> > 
> > BTW, when do you start to hear the fan spinning? upon changing platform
> > profile? upon launching thermald?
> > is there anyway to figure out what fan device brings this noise?
> 
> 
> I can get the fans starting by activating performance and do any work more or less.
> 
> I have thermald enabled all the time now. I think that as long as I don't have the fans active while suspending - it's fine.
> 
> > -rui

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

Best regards,
Benjamin Hasselgren-Hallén


On Thursday, 17 July 2025 at 03:15, Zhang, Rui <rui.zhang@intel.com> wrote:

> On Mon, 2025-07-07 at 19:55 +0000, Benjamin Hasselgren-Hallén wrote:
> 
> > Best regards,
> > Benjamin Hasselgren-Hallén
> > 
> > On Wednesday, 2 July 2025 at 10:00, Zhang, Rui rui.zhang@intel.com
> > wrote:
> > 
> > > Remove the list as I want to grab more details.
> > > 
> > > On Tue, 2025-07-01 at 07:44 +0000, Benjamin Hasselgren-Hallén wrote:
> > > 
> > > > with thermald
> > > > https://drive.benis.se/s/bF5AfDGBw6DFNZt
> > > > 
> > > > without thermald
> > > > https://drive.benis.se/s/47xJdg33ayHerDF
> > > 
> > > There is no much difference, which is expected.
> > > 
> > > > journald
> > > > https://drive.benis.se/s/8JdDJG2bFbHeDmz
> > > 
> > > The log doesn't have anything useful.
> > > It shows your last launch of thermald failed
> > > 
> > > Jul 01 09:38:36 computer thermald[10950]: Couldn't get lock file
> > > 10950
> > > Jul 01 09:38:36 computer thermald[10950]: An instance of thermald is
> > > already running, exiting ...
> > > 
> > > BTW, when do you start to hear the fan spinning? upon changing
> > > platform
> > > profile? upon launching thermald?
> > > is there anyway to figure out what fan device brings this noise?
> > 
> > I can get the fans starting by activating performance and do any work
> > more or less.
> > 
> > I have thermald enabled all the time now. I think that as long as I
> > don't have the fans active while suspending - it's fine.
> 
> 
> Now I'm confused.
> The fan issue is related via platform profile or thermald?
> 
> say, after a refresh boot, does the fan spin during suspend when
> 1. thermald is NOT activated and platform profile is NOT changed?
> 2. thermald is NOT activated and platform profile is changed?
> 3. thermald is activated and platform profile is NOT changed?
> 4. thermald is activated and platform profile is changed?
> 
> Or, what change does it make by switching platform profile only?
> what change does it make by activating thermald only?

The combination of thermald + running the fan while suspending keeps the fan running. Dunno for how long (if it's for ever or just for a while).

However, thermald helps a bit but the system still under performance quite a lot. 

I am regulary trying latest kernel + latest linux-firmware but no changes. 

Anything else we can do? 

> 
> 
> thanks,
> rui

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

Best regards,
Benjamin Hasselgren-Hallén


On Saturday, 19 July 2025 at 22:19, Benjamin Hasselgren-Hallén <benjamin@benis.se> wrote:

> 
> 
> 
> 
> Best regards,
> Benjamin Hasselgren-Hallén
> 
> 
> 
> 
> On Thursday, 17 July 2025 at 03:15, Zhang, Rui rui.zhang@intel.com wrote:
> 
> > On Mon, 2025-07-07 at 19:55 +0000, Benjamin Hasselgren-Hallén wrote:
> > 
> > > Best regards,
> > > Benjamin Hasselgren-Hallén
> > > 
> > > On Wednesday, 2 July 2025 at 10:00, Zhang, Rui rui.zhang@intel.com
> > > wrote:
> > > 
> > > > Remove the list as I want to grab more details.
> > > > 
> > > > On Tue, 2025-07-01 at 07:44 +0000, Benjamin Hasselgren-Hallén wrote:
> > > > 
> > > > > with thermald
> > > > > https://drive.benis.se/s/bF5AfDGBw6DFNZt
> > > > > 
> > > > > without thermald
> > > > > https://drive.benis.se/s/47xJdg33ayHerDF
> > > > 
> > > > There is no much difference, which is expected.
> > > > 
> > > > > journald
> > > > > https://drive.benis.se/s/8JdDJG2bFbHeDmz
> > > > 
> > > > The log doesn't have anything useful.
> > > > It shows your last launch of thermald failed
> > > > 
> > > > Jul 01 09:38:36 computer thermald[10950]: Couldn't get lock file
> > > > 10950
> > > > Jul 01 09:38:36 computer thermald[10950]: An instance of thermald is
> > > > already running, exiting ...
> > > > 
> > > > BTW, when do you start to hear the fan spinning? upon changing
> > > > platform
> > > > profile? upon launching thermald?
> > > > is there anyway to figure out what fan device brings this noise?
> > > 
> > > I can get the fans starting by activating performance and do any work
> > > more or less.
> > > 
> > > I have thermald enabled all the time now. I think that as long as I
> > > don't have the fans active while suspending - it's fine.
> > 
> > Now I'm confused.
> > The fan issue is related via platform profile or thermald?
> > 
> > say, after a refresh boot, does the fan spin during suspend when
> > 1. thermald is NOT activated and platform profile is NOT changed?
> > 2. thermald is NOT activated and platform profile is changed?
> > 3. thermald is activated and platform profile is NOT changed?
> > 4. thermald is activated and platform profile is changed?
> > 
> > Or, what change does it make by switching platform profile only?
> > what change does it make by activating thermald only?
> 
> 
> The combination of thermald + running the fan while suspending keeps the fan running. Dunno for how long (if it's for ever or just for a while).
> 
> However, thermald helps a bit but the system still under performance quite a lot.
> 
> I am regulary trying latest kernel + latest linux-firmware but no changes.
> 
> Anything else we can do?
I can also add that when thermald is running - the fan i still running when the cpu is 40c and the power profile is balance. It should not run now.

> 
> > thanks,
> > rui

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

I got tips from Eddie (on CC) about https://github.com/horshack-dpreview/setPL and it seems to get the cpu to boost higher, I think. However, the gpu still downclocks way more on Linux than Windows.



Best regards,
Benjamin Hasselgren-Hallén


On Saturday, 19 July 2025 at 22:38, Benjamin Hasselgren-Hallén <benjamin@benis.se> wrote:

> 
> 
> 
> 
> Best regards,
> Benjamin Hasselgren-Hallén
> 
> 
> 
> 
> On Saturday, 19 July 2025 at 22:19, Benjamin Hasselgren-Hallén benjamin@benis.se wrote:
> 
> > Best regards,
> > Benjamin Hasselgren-Hallén
> > 
> > On Thursday, 17 July 2025 at 03:15, Zhang, Rui rui.zhang@intel.com wrote:
> > 
> > > On Mon, 2025-07-07 at 19:55 +0000, Benjamin Hasselgren-Hallén wrote:
> > > 
> > > > Best regards,
> > > > Benjamin Hasselgren-Hallén
> > > > 
> > > > On Wednesday, 2 July 2025 at 10:00, Zhang, Rui rui.zhang@intel.com
> > > > wrote:
> > > > 
> > > > > Remove the list as I want to grab more details.
> > > > > 
> > > > > On Tue, 2025-07-01 at 07:44 +0000, Benjamin Hasselgren-Hallén wrote:
> > > > > 
> > > > > > with thermald
> > > > > > https://drive.benis.se/s/bF5AfDGBw6DFNZt
> > > > > > 
> > > > > > without thermald
> > > > > > https://drive.benis.se/s/47xJdg33ayHerDF
> > > > > 
> > > > > There is no much difference, which is expected.
> > > > > 
> > > > > > journald
> > > > > > https://drive.benis.se/s/8JdDJG2bFbHeDmz
> > > > > 
> > > > > The log doesn't have anything useful.
> > > > > It shows your last launch of thermald failed
> > > > > 
> > > > > Jul 01 09:38:36 computer thermald[10950]: Couldn't get lock file
> > > > > 10950
> > > > > Jul 01 09:38:36 computer thermald[10950]: An instance of thermald is
> > > > > already running, exiting ...
> > > > > 
> > > > > BTW, when do you start to hear the fan spinning? upon changing
> > > > > platform
> > > > > profile? upon launching thermald?
> > > > > is there anyway to figure out what fan device brings this noise?
> > > > 
> > > > I can get the fans starting by activating performance and do any work
> > > > more or less.
> > > > 
> > > > I have thermald enabled all the time now. I think that as long as I
> > > > don't have the fans active while suspending - it's fine.
> > > 
> > > Now I'm confused.
> > > The fan issue is related via platform profile or thermald?
> > > 
> > > say, after a refresh boot, does the fan spin during suspend when
> > > 1. thermald is NOT activated and platform profile is NOT changed?
> > > 2. thermald is NOT activated and platform profile is changed?
> > > 3. thermald is activated and platform profile is NOT changed?
> > > 4. thermald is activated and platform profile is changed?
> > > 
> > > Or, what change does it make by switching platform profile only?
> > > what change does it make by activating thermald only?
> > 
> > The combination of thermald + running the fan while suspending keeps the fan running. Dunno for how long (if it's for ever or just for a while).
> > 
> > However, thermald helps a bit but the system still under performance quite a lot.
> > 
> > I am regulary trying latest kernel + latest linux-firmware but no changes.
> > 
> > Anything else we can do?
> 
> I can also add that when thermald is running - the fan i still running when the cpu is 40c and the power profile is balance. It should not run now.
> 
> > > thanks,
> > > rui

Re: HP Omnibook Ultra Flip 14 - power profiles

[Zhang, Rui]

On Mon, 2025-06-30 at 12:48 +0000, Benjamin Hasselgren-Hallén wrote:
> 
> 
> 
> 
> Best regards,
> Benjamin Hasselgren-Hallén
> 
> 
> On Monday, 30 June 2025 at 03:18, Zhang, Rui <rui.zhang@intel.com>
> wrote:
> 
> > > > So I tested thermald again. So here are some results.
> > > > 
> > > > Before:
> > > > Running Valheim the power draw is 35w and the fps is 41.
> > > > 
> > > > After I have installed thermald:
> > > > Running Valheim the power draw is 44w and the fps is 46
> > > > 
> > > > So it's working,
> > 
> > 
> > Good to know that thermald helps.
> > 
> > can you please also attach the turbostat output? say "turbostat -o
> > ts.log
> > sleep 1"
> 
> https://drive.benis.se/s/KtRtCQXi7mS8GMi 
> There you go. I am compiling the linux kernel atm so the laptop is
> working very hard (I am thinking of logging).
> 
is this done with or without thermald activated?
I see the TDP is 17W, not sure if it has already been set by thermald or
not.
can you also paste the lscpu output?

thanks,
rui

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

Best regards,
Benjamin Hasselgren-Hallén


On Tuesday, 1 July 2025 at 08:55, Zhang, Rui <rui.zhang@intel.com> wrote:

> On Mon, 2025-06-30 at 12:48 +0000, Benjamin Hasselgren-Hallén wrote:
> 
> > Best regards,
> > Benjamin Hasselgren-Hallén
> > 
> > On Monday, 30 June 2025 at 03:18, Zhang, Rui rui.zhang@intel.com
> > wrote:
> > 
> > > > > So I tested thermald again. So here are some results.
> > > > > 
> > > > > Before:
> > > > > Running Valheim the power draw is 35w and the fps is 41.
> > > > > 
> > > > > After I have installed thermald:
> > > > > Running Valheim the power draw is 44w and the fps is 46
> > > > > 
> > > > > So it's working,
> > > 
> > > Good to know that thermald helps.
> > > 
> > > can you please also attach the turbostat output? say "turbostat -o
> > > ts.log
> > > sleep 1"
> > 
> > https://drive.benis.se/s/KtRtCQXi7mS8GMi
> > There you go. I am compiling the linux kernel atm so the laptop is
> > working very hard (I am thinking of logging).
> 
> is this done with or without thermald activated?

Thermald is active. 

> I see the TDP is 17W, not sure if it has already been set by thermald or
> not.

I can do a test with a different load if that matters?

> can you also paste the lscpu output?

Architecture:                x86_64
  CPU op-mode(s):            32-bit, 64-bit
  Address sizes:             42 bits physical, 48 bits virtual
  Byte Order:                Little Endian
CPU(s):                      8
  On-line CPU(s) list:       0-7
Vendor ID:                   GenuineIntel
  Model name:                Intel(R) Core(TM) Ultra 7 258V
    CPU family:              6
    Model:                   189
    Thread(s) per core:      1
    Core(s) per socket:      8
    Socket(s):               1
    Stepping:                1
    CPU(s) scaling MHz:      12%
    CPU max MHz:             4800.0000
    CPU min MHz:             400.0000
    BogoMIPS:                6604.80
    Flags:                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pg
                             e mca cmov pat pse36 clflush dts acpi mmx fxsr sse 
                             sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm cons
                             tant_tsc art arch_perfmon pebs bts rep_good nopl xt
                             opology nonstop_tsc cpuid aperfmperf tsc_known_freq
                              pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm
                             2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 
                             x2apic movbe popcnt tsc_deadline_timer aes xsave av
                             x f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault
                              epb ssbd ibrs ibpb stibp ibrs_enhanced tpr_shadow 
                             flexpriority ept vpid ept_ad fsgsbase tsc_adjust bm
                             i1 avx2 smep bmi2 erms invpcid rdt_a rdseed adx sma
                             p clflushopt clwb intel_pt sha_ni xsaveopt xsavec x
                             getbv1 xsaves split_lock_detect user_shstk avx_vnni
                              lam wbnoinvd dtherm ida arat pln pts hwp hwp_notif
                             y hwp_act_window hwp_epp hwp_pkg_req hfi vnmi umip 
                             pku ospke waitpkg gfni vaes vpclmulqdq rdpid bus_lo
                             ck_detect movdiri movdir64b fsrm md_clear serialize
                              pconfig arch_lbr ibt flush_l1d arch_capabilities
Virtualization features:     
  Virtualization:            VT-x
Caches (sum of all):         
  L1d:                       320 KiB (8 instances)
  L1i:                       512 KiB (8 instances)
  L2:                        14 MiB (5 instances)
  L3:                        12 MiB (1 instance)
NUMA:                        
  NUMA node(s):              1
  NUMA node0 CPU(s):         0-7
Vulnerabilities:             
  Gather data sampling:      Not affected
  Ghostwrite:                Not affected
  Indirect target selection: Not affected
  Itlb multihit:             Not affected
  L1tf:                      Not affected
  Mds:                       Not affected
  Meltdown:                  Not affected
  Mmio stale data:           Not affected
  Reg file data sampling:    Not affected
  Retbleed:                  Not affected
  Spec rstack overflow:      Not affected
  Spec store bypass:         Mitigation; Speculative Store Bypass disabled via p
                             rctl
  Spectre v1:                Mitigation; usercopy/swapgs barriers and __user poi
                             nter sanitization
  Spectre v2:                Mitigation; Enhanced / Automatic IBRS; IBPB conditi
                             onal; PBRSB-eIBRS Not affected; BHI BHI_DIS_S
  Srbds:                     Not affected
  Tsx async abort:           Not affected


> 
> thanks,
> rui

Re: HP Omnibook Ultra Flip 14 - power profiles

[Zhang, Rui]

On Tue, 2025-07-01 at 07:04 +0000, Benjamin Hasselgren-Hallén wrote:
> 
> 
> 
> 
> Best regards,
> Benjamin Hasselgren-Hallén
> 
> 
> On Tuesday, 1 July 2025 at 08:55, Zhang, Rui <rui.zhang@intel.com>
> wrote:
> 
> > On Mon, 2025-06-30 at 12:48 +0000, Benjamin Hasselgren-Hallén wrote:
> > 
> > > Best regards,
> > > Benjamin Hasselgren-Hallén
> > > 
> > > On Monday, 30 June 2025 at 03:18, Zhang, Rui rui.zhang@intel.com
> > > wrote:
> > > 
> > > > > > So I tested thermald again. So here are some results.
> > > > > > 
> > > > > > Before:
> > > > > > Running Valheim the power draw is 35w and the fps is 41.
> > > > > > 
> > > > > > After I have installed thermald:
> > > > > > Running Valheim the power draw is 44w and the fps is 46
> > > > > > 
> > > > > > So it's working,
> > > > 
> > > > Good to know that thermald helps.
> > > > 
> > > > can you please also attach the turbostat output? say "turbostat -
> > > > o
> > > > ts.log
> > > > sleep 1"
> > > 
> > > https://drive.benis.se/s/KtRtCQXi7mS8GMi
> > > There you go. I am compiling the linux kernel atm so the laptop is
> > > working very hard (I am thinking of logging).
> > 
> > is this done with or without thermald activated?
> 
> Thermald is active. 
> 
> > I see the TDP is 17W, not sure if it has already been set by thermald
> > or
> > not.
> 
> I can do a test with a different load if that matters?
> 
> > can you also paste the lscpu output?
> 
> Architecture:                x86_64
>   CPU op-mode(s):            32-bit, 64-bit
>   Address sizes:             42 bits physical, 48 bits virtual
>   Byte Order:                Little Endian
> CPU(s):                      8
>   On-line CPU(s) list:       0-7
> Vendor ID:                   GenuineIntel
>   Model name:                Intel(R) Core(TM) Ultra 7 258V

https://www.intel.com/content/www/us/en/products/sku/240957/intel-core-ultra-7-processor-258v-12m-cache-up-to-4-80-ghz/specifications.html
so the TDP is indeed 17W.

And the turbostat output you attached shows that the processors are
running at around 17W, so the processor power is not capped in your
previous test (with thermald activated).

I suspect that the firmware caps the power via MMIO RAPL, and thermald
fixes that.
can you please do the following test
1. do a fresh boot
2. run "grep . /sys/class/powercap/*/* > powercap-thermald-off.log"
3. run "sudo turbostat -o ts-thermald-off.log stress -c 8 -t 10"
4. activate thermald
5. run "sudo turbostat -o ts-thermald-on.log stress -c 8 -t 10"
6. run "grep . /sys/class/powercap/*/* > powercap-thermald-on.log"
7. attach the four log files

thanks,
rui

>     CPU family:              6
>     Model:                   189
>     Thread(s) per core:      1
>     Core(s) per socket:      8
>     Socket(s):               1
>     Stepping:                1
>     CPU(s) scaling MHz:      12%
>     CPU max MHz:             4800.0000
>     CPU min MHz:             400.0000
>     BogoMIPS:                6604.80
>     Flags:                   fpu vme de pse tsc msr pae mce cx8 apic
> sep mtrr pg
>                              e mca cmov pat pse36 clflush dts acpi mmx
> fxsr sse 
>                              sse2 ss ht tm pbe syscall nx pdpe1gb
> rdtscp lm cons
>                              tant_tsc art arch_perfmon pebs bts
> rep_good nopl xt
>                              opology nonstop_tsc cpuid aperfmperf
> tsc_known_freq
>                               pni pclmulqdq dtes64 monitor ds_cpl vmx
> smx est tm
>                              2 ssse3 sdbg fma cx16 xtpr pdcm pcid
> sse4_1 sse4_2 
>                              x2apic movbe popcnt tsc_deadline_timer aes
> xsave av
>                              x f16c rdrand lahf_lm abm 3dnowprefetch
> cpuid_fault
>                               epb ssbd ibrs ibpb stibp ibrs_enhanced
> tpr_shadow 
>                              flexpriority ept vpid ept_ad fsgsbase
> tsc_adjust bm
>                              i1 avx2 smep bmi2 erms invpcid rdt_a
> rdseed adx sma
>                              p clflushopt clwb intel_pt sha_ni xsaveopt
> xsavec x
>                              getbv1 xsaves split_lock_detect user_shstk
> avx_vnni
>                               lam wbnoinvd dtherm ida arat pln pts hwp
> hwp_notif
>                              y hwp_act_window hwp_epp hwp_pkg_req hfi
> vnmi umip 
>                              pku ospke waitpkg gfni vaes vpclmulqdq
> rdpid bus_lo
>                              ck_detect movdiri movdir64b fsrm md_clear
> serialize
>                               pconfig arch_lbr ibt flush_l1d
> arch_capabilities
> Virtualization features:     
>   Virtualization:            VT-x
> Caches (sum of all):         
>   L1d:                       320 KiB (8 instances)
>   L1i:                       512 KiB (8 instances)
>   L2:                        14 MiB (5 instances)
>   L3:                        12 MiB (1 instance)
> NUMA:                        
>   NUMA node(s):              1
>   NUMA node0 CPU(s):         0-7
> Vulnerabilities:             
>   Gather data sampling:      Not affected
>   Ghostwrite:                Not affected
>   Indirect target selection: Not affected
>   Itlb multihit:             Not affected
>   L1tf:                      Not affected
>   Mds:                       Not affected
>   Meltdown:                  Not affected
>   Mmio stale data:           Not affected
>   Reg file data sampling:    Not affected
>   Retbleed:                  Not affected
>   Spec rstack overflow:      Not affected
>   Spec store bypass:         Mitigation; Speculative Store Bypass
> disabled via p
>                              rctl
>   Spectre v1:                Mitigation; usercopy/swapgs barriers and
> __user poi
>                              nter sanitization
>   Spectre v2:                Mitigation; Enhanced / Automatic IBRS;
> IBPB conditi
>                              onal; PBRSB-eIBRS Not affected; BHI
> BHI_DIS_S
>   Srbds:                     Not affected
>   Tsx async abort:           Not affected
> 
> 
> > 
> > thanks,
> > rui

Re: HP Omnibook Ultra Flip 14 - power profiles

[Benjamin Hasselgren-Hallén]

Best regards,
Benjamin Hasselgren-Hallén


On Wednesday, 2 July 2025 at 03:14, Zhang, Rui <rui.zhang@intel.com> wrote:

> On Tue, 2025-07-01 at 07:04 +0000, Benjamin Hasselgren-Hallén wrote:
> 
> > Best regards,
> > Benjamin Hasselgren-Hallén
> > 
> > On Tuesday, 1 July 2025 at 08:55, Zhang, Rui rui.zhang@intel.com
> > wrote:
> > 
> > > On Mon, 2025-06-30 at 12:48 +0000, Benjamin Hasselgren-Hallén wrote:
> > > 
> > > > Best regards,
> > > > Benjamin Hasselgren-Hallén
> > > > 
> > > > On Monday, 30 June 2025 at 03:18, Zhang, Rui rui.zhang@intel.com
> > > > wrote:
> > > > 
> > > > > > > So I tested thermald again. So here are some results.
> > > > > > > 
> > > > > > > Before:
> > > > > > > Running Valheim the power draw is 35w and the fps is 41.
> > > > > > > 
> > > > > > > After I have installed thermald:
> > > > > > > Running Valheim the power draw is 44w and the fps is 46
> > > > > > > 
> > > > > > > So it's working,
> > > > > 
> > > > > Good to know that thermald helps.
> > > > > 
> > > > > can you please also attach the turbostat output? say "turbostat -
> > > > > o
> > > > > ts.log
> > > > > sleep 1"
> > > > 
> > > > https://drive.benis.se/s/KtRtCQXi7mS8GMi
> > > > There you go. I am compiling the linux kernel atm so the laptop is
> > > > working very hard (I am thinking of logging).
> > > 
> > > is this done with or without thermald activated?
> > 
> > Thermald is active.
> > 
> > > I see the TDP is 17W, not sure if it has already been set by thermald
> > > or
> > > not.
> > 
> > I can do a test with a different load if that matters?
> > 
> > > can you also paste the lscpu output?
> > 
> > Architecture: x86_64
> > CPU op-mode(s): 32-bit, 64-bit
> > Address sizes: 42 bits physical, 48 bits virtual
> > Byte Order: Little Endian
> > CPU(s): 8
> > On-line CPU(s) list: 0-7
> > Vendor ID: GenuineIntel
> > Model name: Intel(R) Core(TM) Ultra 7 258V
> 
> 
> https://www.intel.com/content/www/us/en/products/sku/240957/intel-core-ultra-7-processor-258v-12m-cache-up-to-4-80-ghz/specifications.html
> so the TDP is indeed 17W.
> 
> And the turbostat output you attached shows that the processors are
> running at around 17W, so the processor power is not capped in your
> previous test (with thermald activated).
> 
> I suspect that the firmware caps the power via MMIO RAPL, and thermald
> fixes that.
> can you please do the following test
> 1. do a fresh boot
> 2. run "grep . /sys/class/powercap// > powercap-thermald-off.log"
> 
> 3. run "sudo turbostat -o ts-thermald-off.log stress -c 8 -t 10"
> 4. activate thermald
> 5. run "sudo turbostat -o ts-thermald-on.log stress -c 8 -t 10"
> 6. run "grep . /sys/class/powercap// > powercap-thermald-on.log"
> 
> 7. attach the four log files

Logs
https://drive.benis.se/s/o75BQ9WfPEWNcrn
https://drive.benis.se/s/pFBLr7NDLJkyr83
https://drive.benis.se/s/E3e8mEbLsrmmoHJ
https://drive.benis.se/s/qSYAJtLEwkSAzQT

I have noticed that if I only run CPU or GPU - the performance is good. So when running Valheim (game) that do not require any cpu - the performance is on par with Windows, but when benchmarking Company of Heroes 3 or Cyberpunk 2077, the performance is like 50% of Windows (I might have state this before). I can also see how the fps are dropping very fast after starting the game. I do not see the gpu mhz in MangoHud and when I try intel_gpu_top I get:
sudo intel_gpu_top
No device filter specified and no discrete/integrated i915 devices found
It's not supported?

//Benjamin


> 
> thanks,
> rui
> 
> > CPU family: 6
> > Model: 189
> > Thread(s) per core: 1
> > Core(s) per socket: 8
> > Socket(s): 1
> > Stepping: 1
> > CPU(s) scaling MHz: 12%
> > CPU max MHz: 4800.0000
> > CPU min MHz: 400.0000
> > BogoMIPS: 6604.80
> > Flags: fpu vme de pse tsc msr pae mce cx8 apic
> > sep mtrr pg
> > e mca cmov pat pse36 clflush dts acpi mmx
> > fxsr sse
> > sse2 ss ht tm pbe syscall nx pdpe1gb
> > rdtscp lm cons
> > tant_tsc art arch_perfmon pebs bts
> > rep_good nopl xt
> > opology nonstop_tsc cpuid aperfmperf
> > tsc_known_freq
> > pni pclmulqdq dtes64 monitor ds_cpl vmx
> > smx est tm
> > 2 ssse3 sdbg fma cx16 xtpr pdcm pcid
> > sse4_1 sse4_2
> > x2apic movbe popcnt tsc_deadline_timer aes
> > xsave av
> > x f16c rdrand lahf_lm abm 3dnowprefetch
> > cpuid_fault
> > epb ssbd ibrs ibpb stibp ibrs_enhanced
> > tpr_shadow
> > flexpriority ept vpid ept_ad fsgsbase
> > tsc_adjust bm
> > i1 avx2 smep bmi2 erms invpcid rdt_a
> > rdseed adx sma
> > p clflushopt clwb intel_pt sha_ni xsaveopt
> > xsavec x
> > getbv1 xsaves split_lock_detect user_shstk
> > avx_vnni
> > lam wbnoinvd dtherm ida arat pln pts hwp
> > hwp_notif
> > y hwp_act_window hwp_epp hwp_pkg_req hfi
> > vnmi umip
> > pku ospke waitpkg gfni vaes vpclmulqdq
> > rdpid bus_lo
> > ck_detect movdiri movdir64b fsrm md_clear
> > serialize
> > pconfig arch_lbr ibt flush_l1d
> > arch_capabilities
> > Virtualization features:
> > Virtualization: VT-x
> > Caches (sum of all):
> > L1d: 320 KiB (8 instances)
> > L1i: 512 KiB (8 instances)
> > L2: 14 MiB (5 instances)
> > L3: 12 MiB (1 instance)
> > NUMA:
> > NUMA node(s): 1
> > NUMA node0 CPU(s): 0-7
> > Vulnerabilities:
> > Gather data sampling: Not affected
> > Ghostwrite: Not affected
> > Indirect target selection: Not affected
> > Itlb multihit: Not affected
> > L1tf: Not affected
> > Mds: Not affected
> > Meltdown: Not affected
> > Mmio stale data: Not affected
> > Reg file data sampling: Not affected
> > Retbleed: Not affected
> > Spec rstack overflow: Not affected
> > Spec store bypass: Mitigation; Speculative Store Bypass
> > disabled via p
> > rctl
> > Spectre v1: Mitigation; usercopy/swapgs barriers and
> > __user poi
> > nter sanitization
> > Spectre v2: Mitigation; Enhanced / Automatic IBRS;
> > IBPB conditi
> > onal; PBRSB-eIBRS Not affected; BHI
> > BHI_DIS_S
> > Srbds: Not affected
> > Tsx async abort: Not affected
> > 
> > > thanks,
> > > rui