intel_pstate has too aggressive frequency selection

Bug #1598394 reported by chives
14
This bug affects 3 people
Affects Status Importance Assigned to Milestone
linux (Ubuntu)
Confirmed
Medium
Unassigned

Bug Description

I was recently upgrading from Ubuntu Gnome 15.10 to 16.04 on my ThinkPad. Before the frequency-scaling chose the minimum possible frequency when the system was on low load. Now, when I allow the CPU to choose anything between the lowest and highest possible frequency, it sticks to the highest frequency for 90% of time and only occasionally drops to the low frequency.

I can confirm (htop) that my system load is practically zero. Also the CPU Power Manager Applet from Gnome Shell works well, because I can force low frequency by reducing the maximum allowed frequency and can also switch off Turbo Boost.

ProblemType: Bug
DistroRelease: Ubuntu 16.04
Package: linux-generic 4.4.0.28.30
ProcVersionSignature: Ubuntu 4.4.0-28.47-generic 4.4.13
Uname: Linux 4.4.0-28-generic x86_64
ApportVersion: 2.20.1-0ubuntu2.1
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: timeshifter 1768 F.... pulseaudio
 /dev/snd/controlC1: timeshifter 1768 F.... pulseaudio
CurrentDesktop: GNOME
Date: Sat Jul 2 12:11:42 2016
EcryptfsInUse: Yes
HibernationDevice: RESUME=UUID=87cca434-274f-4464-9598-97fd7a1338fe
InstallationDate: Installed on 2016-03-19 (104 days ago)
InstallationMedia: Ubuntu-GNOME 15.10 "Wily Werewolf" - Release amd64 (20151021)
MachineType: LENOVO 20C0003SFR
ProcEnviron:
 TERM=xterm-256color
 PATH=(custom, no user)
 XDG_RUNTIME_DIR=<set>
 LANG=de_DE.UTF-8
 SHELL=/usr/bin/zsh
ProcFB: 0 inteldrmfb
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-4.4.0-28-generic root=UUID=e52c99ea-9646-4e8a-b8ad-c47b175ac5a3 ro quiet splash vt.handoff=7
RelatedPackageVersions:
 linux-restricted-modules-4.4.0-28-generic N/A
 linux-backports-modules-4.4.0-28-generic N/A
 linux-firmware 1.157.1
SourcePackage: linux
UpgradeStatus: Upgraded to xenial on 2016-07-02 (0 days ago)
dmi.bios.date: 11/07/2014
dmi.bios.vendor: LENOVO
dmi.bios.version: B0ET24WW (1.11 )
dmi.board.asset.tag: Not Available
dmi.board.name: 20C0003SFR
dmi.board.vendor: LENOVO
dmi.board.version: SDK0E50510 WIN
dmi.chassis.asset.tag: No Asset Information
dmi.chassis.type: 10
dmi.chassis.vendor: LENOVO
dmi.chassis.version: Not Available
dmi.modalias: dmi:bvnLENOVO:bvrB0ET24WW(1.11):bd11/07/2014:svnLENOVO:pn20C0003SFR:pvrThinkPadS1Yoga:rvnLENOVO:rn20C0003SFR:rvrSDK0E50510WIN:cvnLENOVO:ct10:cvrNotAvailable:
dmi.product.name: 20C0003SFR
dmi.product.version: ThinkPad S1 Yoga
dmi.sys.vendor: LENOVO

Revision history for this message
chives (mra+ubuntu) wrote :
Revision history for this message
chives (mra+ubuntu) wrote :

Added attachment

Revision history for this message
chives (mra+ubuntu) wrote :
Revision history for this message
Brad Figg (brad-figg) wrote : Status changed to Confirmed

This change was made by a bot.

Changed in linux (Ubuntu):
status: New → Confirmed
Revision history for this message
Joseph Salisbury (jsalisbury) wrote :

Would it be possible for you to test the latest upstream kernel? Refer to https://wiki.ubuntu.com/KernelMainlineBuilds . Please test the latest v4.7 kernel[0].

If this bug is fixed in the mainline kernel, please add the following tag 'kernel-fixed-upstream'.

If the mainline kernel does not fix this bug, please add the tag: 'kernel-bug-exists-upstream'.

Once testing of the upstream kernel is complete, please mark this bug as "Confirmed".

Thanks in advance.

[0] http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.7-rc6-yakkety/

Changed in linux (Ubuntu):
importance: Undecided → Medium
status: Confirmed → Incomplete
Revision history for this message
chives (mra+ubuntu) wrote :

The bug exists in the upstream kernel 4.7-rc6.

Changed in linux (Ubuntu):
status: Incomplete → Confirmed
tags: added: kernel-bug-exists-upstream
Revision history for this message
Jean-Pierre van Riel (jpvr) wrote :

Notable upstream bugs with related issues

alegidly patched: https://bugzilla.kernel.org/show_bug.cgi?id=115771
maybe not patched yet: https://bugzilla.kernel.org/show_bug.cgi?id=93521

Workaround (to help other who find this bug)

Add `intel_pstate=disable` to `GRUB_CMDLINE_LINUX_DEFAULT` in /etc/default/grub and update-grub. This will revert to the acpi-cpufreq driver and ondemand (default) policy.

check with `cpupower frequency-info`

I've observed more sane scaling of CPU frequencies, and better fan speed and temperature on Haswell. E.g. acpi-cpufreq vs pstate (while relatively idle)

* temp: ~52 vs 70 degrees
* fan: ~3800 rpm vs 4600 rpm

pstate will very often turbo boost many cores and seldom drop below standard max frequency even when lightly loaded per core. When loaded, temp would bounce up to 89 degrees quickly. I suspect the CPU kept needing to throttle back often.

Revision history for this message
dimahetman (dimahetman) wrote :

I recently bought a new Lenovo IdeaPad 520-15IKB 81BF005JRK laptop and installed Ubuntu (18.10 downloaded a few days ago, during install, updates were installed). But I cannot work with it as immediately after turning on the laptop, the processor (i8520u) on all cores operates at maximum frequencies (over 3 GHz) and goes into hard trotting with pauses of 20-30 seconds. I have some time open the console, but I can not execute any command. It's owful.
I cann't install cpufrequtils, cpupower or other software

 I tried to boot from a live-usb (there are no such problems) and edit /boot/config-* :
#CONFIG_CPU_FREQ_GOV_PERFORMANCE
CONFIG_CPU_FREQ_GOV_POWERSAVE=y

the governor was established to POWERSAVE but the laptop still works at the maximum frequency.
may be I can Add `intel_pstate=disable` to `GRUB_CMDLINE_LINUX_DEFAULT` in /etc/default/grub
but I don't know how update-grub at GPT UEFI partition.

It even sometimes seems to me that with each new start in such a boost it slows down more and more, I'm afraid, as it were, that it does not kill the processor on a new computer. After the first installation, I didn’t immediately notice trotting and left the computer for the night. And now it's just not possible to work with him

Brad Figg (brad-figg)
tags: added: cscc
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.