5

I've got a frustrating mystery and I'm hoping that someone can point me in the right direction.

Problem:

  • Under load, my system doesn't increase CPU frequency to meet demand after some period of time passes after boot.
  • I first observed this during video calls where the audio degrades and my laptop generally becomes sluggish.
  • I mostly assess frequency and load in my manual tests with s-tui and stress.
  • This actually happens under windows and linux but I'm running linux right now and have no intention of going back to windows, and I assume the tools and options are better on linux.

enter image description here

Properties of my system:

  • lscpu:
    • Intel(R) Core(TM) i7-10610U CPU @ 1.80GHz
    • CPU max MHz: 2301.0000
    • CPU min MHz: 400.0000
  • Lenovo P15s ThinkPad Gen 1 . The listing claimed gen2 but I guess it was a lie.
  • Ubuntu 22.04,
  • Bios version 1.27. Apparently the latest available.
  • uname -a Linux 5.15.0-60-generic #66-Ubuntu SMP Fri Jan 20 14:29:49 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
  • lsmod | grep intel
    • snd_soc_intel_hda_dsp_commonintel_tcc_coolingintel_rapl_msrintel_powerclampsnd_sof_pci_intel_cnlsnd_sof_intel_hda_commonsoundwire_intelsoundwire_generic_allocationsoundwire_cadencesnd_sof_intel_hdasnd_sof_pcisnd_sof_xtensa_dspsnd_sofsnd_soc_hdac_hdasnd_hda_ext_coresnd_soc_acpi_intel_matchsnd_soc_acpisoundwire_bussnd_soc_corekvm_intelkvmsnd_hda_intelsnd_intel_dspcfgbtintelsnd_intel_sdw_acpibluetoothsnd_hda_codecintel_cstateintel_wmi_thunderboltsnd_hda_coresnd_pcmintel_rapl_commonintel_soc_dts_iosfintel_pch_thermalsndintel_hidsparse_keymapghash_clmulni_intelaesni_intelcrypto_simdcryptdwmi
  • lsmod | grep acpi
    • snd_soc_acpi_intel_matchsnd_soc_acpisnd_intel_sdw_acpithinkpad_acpinvramledtrig_audioucsi_acpitypec_ucsisndplatform_profileacpi_thermal_relacpi_padvideo
  • Output of sudo cpupower frequency-info
analyzing CPU 0:
  driver: acpi-cpufreq
  CPUs which run at the same hardware frequency: 0
  CPUs which need to have their frequency coordinated by software: 0
  maximum transition latency: 10.0 us
  hardware limits: 400 MHz - 2.30 GHz
  available frequency steps:  2.30 GHz, 2.30 GHz, 2.20 GHz, 2.00 GHz, 1.80 GHz, 1.70 GHz, 1.60 GHz, 1.50 GHz, 1.30 GHz, 1.20 GHz, 1.10 GHz, 900 MHz, 800 MHz, 700 MHz, 500 MHz, 400 MHz
  available cpufreq governors: conservative ondemand userspace powersave performance schedutil
  current policy: frequency should be within 400 MHz and 1.80 GHz.
                  The governor "ondemand" may decide which speed to use
                  within this range.
  current CPU frequency: 1.80 GHz (asserted by call to hardware)
  boost state support:
    Supported: yes
    Active: yes

Thinks I've tried so far:

  • Enabling/Disabling intel pstate driver by adding intel_pstate=disbale to grub. I don't think my system even uses it though.
  • ondemand and performance governor. Nothing changes.
  • Enabling/Disabling all of the power/CPU related BIOS options (intel speed step, etc)
  • Manually setting min/max freq and governor using cpupower frequency-set --governor and cpufreq-set -r --min=0.8GHz --max=2.3GHz
  • Using tools like (throttlestop)[https://github.com/agoose77/throttlestop] to see if it was a thermal issue, even though the tempts never get higher than 48 degrees.
  • Manually unplugging my battery and holding the power button for 30 seconds (I got desperate).
  • Enabling acpi driver with sudo apt install acpi-support acpid acpi. I've tried removing these as well. freq-info claims I'm using acpi even if I remove these so I probably had no effect.

Things I don't understand:

  • The min/max frequencies reported by cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq and cpupower frequency-info don't seem to be true. I can see the frequencies go as high as 4.2ghz in s-tui before this bad state kicks in.
  • If I reboot and run s-tui and start stress testing I actually see what I want, all cores go up to 3.5ghz -4.2ghz. At some point after they only go to .8ghz - 1.1ghz. They actually get higher just being idle before the stress test, then lock in at the lower freq.
  • The amazon listing actually claims an 11th gen cpu and I just noticed lscpu is claiming a 10th gen, so I guess I got ripped off.
  • Setting frequencies and governors doesn't actually change anything after I get into this bad state, even though the change are reflected in cpupower frequency-info. The CPU freq under load won't go up.
  • While in this bad state, if I start a load test then the freq does momentarily jump to 3.5-4, but then drops down to the capped lower freq.

Any pointers would be greatly appreciated.

EDIT:

I gave up and bought a new laptop. No more thinkpads for a while.

17
  • regarding lscpu is claiming a 10th gen - what does the BIOS say? Feb 28 at 22:06
  • @JaromandaX lscpu agrees with the bios. Feb 28 at 22:28
  • I actually am pretty out of date. sudo dmidecode -s bios-version says N2XET37W (1.27 ). I'll have to find out how to do the update from linux and see if that does anything. @Freddy Feb 28 at 23:16
  • 1
    These helped me sort out (at least improve) a similar issue on my P14s: forums.lenovo.com/topic/findpost/1306/5087833/5530806 and github.com/daniel-kristjansson/smart-fancontrol/blob/main/…. Have a look at those, try installing thermald and see if that helps. I spent ages getting mine to work and have kept notes, so feel free to ping me in chat and I might be able to help. If it looks like this is a promising avenue to explore, I'll post a write up of what ended up working for my machine as an answer.
    – terdon
    Mar 1 at 0:02
  • 1
    I put my README file here: gist.github.com/terdon/e1dc9cdec58d0351a0f6bbeb0051474b It is messy since it has all the things I tried and I wrote it as notes for myself, and don't have time to make it nice and clean, but it might help you figure out what's going on.
    – terdon
    Mar 1 at 18:35

1 Answer 1

0

I had the same issue where the CPU throttled shortly after the boot up. I tried many different tweaks and utilities until the OEM eventually had replaced the mother board on my laptop. I'm pretty sure it was caused by a faulty thermal sensor reporting wrong temperature causing the CPU to throttle.

1
  • Your answer could be improved with additional supporting information. Please edit to add further details, such as citations or documentation, so that others can confirm that your answer is correct. You can find more information on how to write good answers in the help center.
    – Community Bot
    Mar 6 at 20:49

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .