[RESPONDED] 12th gen - excessive CPU throttling on Linux - Can't get above 46c and pegs to 400mhz under load

Framework Laptop 13 Linux
1

Similar, but maybe different to this post: Excessive CPU thermal(?) throttling on 12th-gen

My Framework is a 12th gen w/ the i7-1280p running Gentoo Linux w/ a custom 6.2.8 kernel.

I’ve installed and kinda randomly messed with all the tools on that other thread - RAPL settings, thermald, all that, but it hasn’t really changed behavior any.

At rest, my CPU cores sit at 400Hz and 2000MHz (depending on the type). Under load, they all drop to 400Hz and sit there.

Here’s a s-tui screenshot showing this - from idle, to 100% utilization w/ stress test, back to idle again:

stui
stui2222×1248 111 KB

Here’s cpupower frequency-info:

analyzing CPU 0:
  driver: intel_pstate
  CPUs which run at the same hardware frequency: 0
  CPUs which need to have their frequency coordinated by software: 0
  maximum transition latency:  Cannot determine or is not supported.
  hardware limits: 400 MHz - 4.70 GHz
  available cpufreq governors: performance powersave
  current policy: frequency should be within 400 MHz and 4.70 GHz.
                  The governor "performance" may decide which speed to use
                  within this range.
  current CPU frequency: Unable to call hardware
  current CPU frequency: 400 MHz (asserted by call to kernel)
  boost state support:
    Supported: yes
    Active: yes

I’ve never seen my CPU frequencies go over 2GHz, or my CPU temperature go over 49C. I haven’t heard my fans ever either!

What’s going on / how can I debug? I expect that things would get hotter and work harder, but they don’t.

2

Here’s powercap-info:

intel-rapl-mmio
  enabled: 1
  Zone 0
    name: package-0
    enabled: 1
    max_energy_range_uj: 262143328850
    energy_uj: 48534252705
    Constraint 0
      name: long_term
      power_limit_uw: 28000000
      time_window_us: 27983872
      max_power_uw: 28000000
    Constraint 1
      name: short_term
      power_limit_uw: 60000000
      time_window_us: 2440
      max_power_uw: 0
intel-rapl
  enabled: 1
  Zone 0
    name: package-0
    enabled: 1
    max_energy_range_uj: 262143328850
    energy_uj: 48534256306
    Constraint 0
      name: long_term
      power_limit_uw: 44000000
      time_window_us: 27983872
      max_power_uw: 28000000
    Constraint 1
      name: short_term
      power_limit_uw: 64000000
      time_window_us: 2440
      max_power_uw: 0
    Constraint 2
      name: peak_power
      power_limit_uw: 140000000
      time_window_us: 0
      max_power_uw: 0
    Zone 0:0
      name: core
      enabled: 0
      max_energy_range_uj: 262143328850
      energy_uj: 7729278858
      Constraint 0
        name: long_term
        power_limit_uw: 0
        time_window_us: 976
    Zone 0:1
      name: uncore
      enabled: 0
      max_energy_range_uj: 262143328850
      energy_uj: 734345227
      Constraint 0
        name: long_term
        power_limit_uw: 0
        time_window_us: 976
  Zone 1
    name: psys
    enabled: 0
    max_energy_range_uj: 262143328850
    energy_uj: 162928638809
    Constraint 0
      name: long_term
      power_limit_uw: 0
      time_window_us: 27983872
    Constraint 1
      name: short_term
      power_limit_uw: 0
      time_window_us: 976
3

Bases on my search you better contact Framework Support, it most probably is a faulty motherboard.

4

Thanks, @Anachron. I’ve just submitted a ticket, but if anyone in the community has ideas for debugging or that might fix, let me know. I’ll update here with support’s verdict too.

1 Like
5

Hey all - an update!

Support had me apply the beta 3.0.6 firmware … and everything is working great.

Now, when I enable a high load stress test - the fans start, cpu frequency increases, and temperature increase (up to about 100C). As expected.

Things feel so much snappier.

1 Like
6

For those interested, I had and still have the same issue. Updated the BIOS to the alpha version, re-pasted with PTM7950, double checked the pad locations for the VRMs, the whole nine-yards.

I received a replacement mainboard after months of troubleshooting and the thing idles around 80C and will still clock-throttle down to 400 MHz even at 60C. Then it will constantly jump to 100C and trigger PROCHOT again. I repeated the same procedures that I did with the original board, at my own expense again - no dice. There is something wrong, and I’m not the only one with this issue. I just want to swap to a ryzen board at this point because I’ve never had issues with my other ryzen chips - but I will not at my own expense, it’s not acceptable. I have an early batch 12th gen i7-1280p and have not been able to actually use the thing because of this issue.

7

Just out of curiosity what BIOS? There is no alpha version. What distro? and are you using Thermald?

8

I’m going to bump this (sorry folks!) because Intel and Framework really need to look into the throttling and power delivery issues, I’m linking my previous reply on another thread below for visibility here.

I’d like to see feedback from the community so that we can aggregate how many people are seeing this issue.

All I want is the best for the community and company and we are aware that FW are relatively green still, but we need these issues at the top of the queue so that reputation isn’t tarnished. I would expect that it is in the company’s best interest to uphold sustainable and long-term usage, but many of us can’t even get short-term usage OOB.

I understand first-hand the hardships that come with working with (external) partners and vendors, but I feel it’s been a good amount of time to really start a conversation.

And believe me, I have tried everything and gone above and beyond, spent a lot of my own money and provided more than adequate diagnostic information to the FW support team. (I’m sure they would love to fire me as a customer! :laughing:)

But, jokes aside - I get it, the new launch is coming and the pressure is certainly on. It just feels like in certain circumstances it’s a situation of “biting off more than one can chew”, but I can only make assumptions.

I am willing to run as many diagnostics as possible and provide as much data as I can and I encourage the community to do so as well. (I do this to an extent for a living) Just let me know what you would like to see.

My experience thus far: | (Unusable since received October of 2022)

Thank you and best regards,
Asa K.

9

This is copied from another thread:

“Hi there, don’t know if anybody else tried this, but I may have found a workaround to keep this problem from happening:
The only caveats are that it keeps the CPU from using the full speed available, but it keeps the laptop stable, and if you update the BIOS it resets the settings, so make sure to keep an eye on that.
If you go into BIOS and set CPU to Max-Non Turbo Performance and Disable Intel SpeedStep, SpeedShift and Turbo Boost Max, then it should keep the problem from happening.
This is with a Intel 11 i5, 16gb RAM and Windows 11, but it also fixed the problem on a Intel 12 i5 as well.
Hope this helps somebody!”

I tried this but kept intel SpeedStep enabled, and the 400Mhz problem went away but also kept a partial boost so I’m not stuck at 1.7Ghz. So in conclusion lower boost but also lower throttle. I sure hope this helps you in addition to the other information that was copied!

Update:
I re enabled speedshift and turbo boost max again, as for some reason the E-cores will just idle if those settings are disabled. Somehow after running windows with those settings disabled and re-enabling them seems to have solved the 400Mhz issue. Don’t know whats going on or how it somehow worked, but it did.

10

This throttling issue makes playing even old games at old settings unplayable, tf2 and CSGO for example.

11

Just a reminder - if you are having issues, we ask that you open a ticket so we can take a look at each individual case and instance. Thank you

2 Likes
12

I experienced this issue today on my 12th gen i7-1260p Framework Laptop 13 running Fedora 38. I think it was triggered by the laptop waking up from suspend in my bag, failing to suspend again and getting very hot yesterday evening (draining a lot of the battery). After that, I charged it and used it for a bit before suspending it overnight. I’ve been very frustrated today trying to figure out why my laptop had gotten so slow but was able to reproduce the same behaviour using s-tui as in the OP.

I wasn’t able to fix it by restarting (I tried several times) but was able to eventually resolve it by shutting the laptop down, leaving it off for about 10-15 minutes and then turning it on again. I hope it doesn’t come back :crossed_fingers:

2 Likes
13

Looks like I am not the only one. And it seems to be getting worse over time as well for me. I contacted Support today to see if they have a solution for this. If not, then the framework has to go. It is becoming unbearable.

14

Im using Arch, got nearly no issues here. maybe a user configuration issue?

15

@CodeAsm Which packages do I need to have installed to configure this? AFAIK most of this is part of UEFI settings?

One thing to note is that the fan will often stop spinning, but the laptop will not recover from the throttling state.

16

Just a reminder - if you are having issues, we ask that you open a ticket so we can take a look at each individual case and instance. Thank you

17

Thanks for the suggestion. This worked for me!

18

Delighted to hear it! :slight_smile:

19

@Matt_Hartley I got in touch with support a couple of weeks ago and last week I got a replacement motherboard, which fixed the issue for me. :smile:

2 Likes
20

After struggling with this issue for half a year now I’d like to share my experience: Turns out, that taking out the RAM module and plugging it back in helps fix the issue consistently for a few weeks. So my problem is probably, that the module wiggles itself out of a “stable” state where it doesn’t have to throttle.

13th Gen user here.