After updating to 22.04 from 20.04, I have experienced the same
performance issues. And adding "inter_iommu=off" did not help. The
laptop is with Intel CPU and integrated graphics, so it seemed like the
correct option.

Investigating a bit further, I notices unusually low CPU temperature
(around 36°C idle, while before the update it was usually around 45°C
idle). Installing linux-tools-generic package and using `turboprop` and
`stress` to bring all CPU cores to 100%, it showed this:

```
Core    CPU Avg_MHz Busy%   Bzy_MHz TSC_MHz IPC IRQ SMI POLL    C1  C1E C3  C6  
C7s C8  C9  C10 POLL%   C1% C1E%    C3% C6% C7s%    C8% C9% C10%    CPU%c1  
CPU%c3  CPU%c6CPU%c7    CoreTmp PkgTmp  GFX%rc6 GFXMHz  GFXAMHz Totl%C0 Any%C0  
GFX%C0  CPUGFX% Pkg%pc2 Pkg%pc3 Pkg%pc6 Pkg%pc7 Pkg%pc8 Pkg%pc9 Pk%pc10 SYS%LPI 
PkgWatt CorWatt GFXWatt RAMWatt PKG_%   RAM_%
-   -   399 99.72   400 1800    0.92    22289   0   0   0   0   0   1   0   0   
0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.28    0.00    0.00    0.00    43  43  77.63   300 300 399.82  99.96   20.87   
20.84   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    1.79    
0.93    0.12    0.48    0.00    0.00
0   0   399 99.72   400 1800    0.94    1476    0   0   0   0   0   0   0   0   
0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.28    0.00    0.00    0.00    43  43  77.63   300 300 399.83  99.96   20.87   
20.84   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    1.79    
0.93    0.12    0.48    0.00    0.00
0   4   399 99.72   400 1800    0.92    1514    0   0   0   0   0   0   0   0   
0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.28
1   1   399 99.72   400 1800    0.90    1692    0   0   0   0   0   0   0   0   
0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.28    0.00    0.00    0.00    43
1   5   399 99.70   400 1800    0.85    11087   0   0   0   0   0   1   0   0   
0   0   0.00    0.00    0.00    0.00    0.02    0.00    0.00    0.00    0.00    
0.30
2   2   399 99.72   400 1800    0.93    1488    0   0   0   0   0   0   0   0   
0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.28    0.00    0.00    0.00    43
2   6   399 99.72   400 1800    0.94    1335    0   0   0   0   0   0   0   0   
0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.28
3   3   399 99.72   400 1800    0.95    1518    0   0   0   0   0   0   0   0   
0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.28    0.00    0.00    0.00    42
3   7   399 99.72   400 1800    0.96    2179    0   0   0   0   0   0   0   0   
0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.28
```

All cores running at 400MHz, which should be the idle cpu-state, not the
cpu-state when the cpu is at 100%. Also it must be noted that the
slowness remained after reboot and even on the boot screen (after adding
intel_iommu=off), including the password screen from the BIOS (outside
grub/linux). Restart seems not to help.

However, after going to sleep state (G1 I think), which happens
automatically after the lid is closed, and then waking up, it was fixed.
The turbostat under stress now shows 2605MHz (the maximum allowed
frequency):

Core    CPU Avg_MHz Busy%   Bzy_MHz TSC_MHz IPC IRQ SMI POLL    C1  C1E C3  C6  
C7s C8  C9  C10 POLL%   C1% C1E%    C3% C6% C7s%    C8% C9% C10%    CPU%c1  
CPU%c3  CPU%c6CPU%c7    CoreTmp PkgTmp  GFX%rc6 GFXMHz  GFXAMHz Totl%C0 Any%C0  
GFX%C0  CPUGFX% Pkg%pc2 Pkg%pc3 Pkg%pc6 Pkg%pc7 Pkg%pc8 Pkg%pc9 Pk%pc10 SYS%LPI 
PkgWatt CorWatt GFXWatt RAMWatt PKG_%   RAM_%
-   -   2605    99.52   2618    1800    0.94    33599   0   0   0   0   0   0   
0   0   0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.00    0.48    0.00    0.00    0.00    82  82  85.96   300 300 399.04  99.76   
13.16   13.02   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
14.95   13.77   0.13    0.43    0.00    0.00
0   0   2605    99.52   2618    1800    0.89    22047   0   0   0   0   0   0   
0   0   0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.00    0.48    0.00    0.00    0.00    82  82  85.96   300 300 399.04  99.76   
13.16   13.02   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
14.95   13.77   0.13    0.43    0.00    0.00
0   4   2605    99.52   2618    1800    0.94    1786    0   0   0   0   0   0   
0   0   0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.00    0.48
1   1   2605    99.52   2618    1800    0.94    1920    0   0   0   0   0   0   
0   0   0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.00    0.48    0.00    0.00    0.00    79
1   5   2605    99.52   2618    1800    0.94    2376    0   0   0   0   0   0   
0   0   0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.00    0.48
2   2   2605    99.52   2618    1800    0.94    1379    0   0   0   0   0   0   
0   0   0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.00    0.48    0.00    0.00    0.00    81
2   6   2605    99.52   2618    1800    0.94    1377    0   0   0   0   0   0   
0   0   0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.00    0.48
3   3   2605    99.52   2618    1800    0.94    1382    0   0   0   0   0   0   
0   0   0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.00    0.48    0.00    0.00    0.00    81
3   7   2605    99.52   2618    1800    0.94    1332    0   0   0   0   0   0   
0   0   0   0   0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00    
0.00    0.48

It seems there is some bug with switching CPU power states, which does
not bring the CPU to higher power state after idling, causing the CPU
even under load to operate at really low frequency.

So the solution seems is to force changing of CPU power state either by
powering off and bringing the computer to sleep. Since I did that after
adding the "intel_iommu=off", I have no idea of this is required or not.

Also it must be noted, that it seems both ubuntu 20.04 and 22.04 tend to
operate on higher CPU power-state that I would expect when idling. A CPU
with less than 1% operates on average 1GHz (average out from changing of
powerstates), where the highest CPU power-state seems to be 400MHz for
this CPU, an interesting fact I learned from this bug.

-- 
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux in Ubuntu.
https://bugs.launchpad.net/bugs/1973434

Title:
  massive performance issues since 22.04 upgrade

Status in linux package in Ubuntu:
  Confirmed

Bug description:
  Hi,
  After upgrading to 22.04 i had to fight with massive performance issues.

  Browsers appeared to hang every other minute, youtube videos being
  laggy and hang in between, applications in a virtualbox VM where slow
  and also hanging every other minute to a level of not being useable.
  On a pretty recent and powerful system just 2 years old.

  I noticed CPU jumps in top, but also somehow thought it could be a graphics 
issue so invested some time installing nvidia drivers properly.
  Also I wondered if it might be the lowlatency kernel I normally use because I 
do audio stuff, and switched to generic. But nothing helped.

  ThenI had the idea it could be a kernel/scheduler issue because the
  system wasn't always slow, but it appeared certain things kept hanging
  when other processed had a lot of cpu for a few seconds.

  So I got a recent mainline kernel, configured it with my last running
  config from 21.10 before the update, made the debs and installed them,
  and now can tell that a mainline kernel 5.17.7 with all the dkms
  modules that i had before which got compiled automatically at
  installation brings back a "normal" performance.

  I can browse the web, run multiple youtube vids at once, even in
  another browser, have thunderbird running, and a virtualbox machine
  open with another browser for some web app testing and everything runs
  fine and smooth, no lagging.

  Not sure yet what the real reason is - either the kernel version, or a
  patch in the ubuntu version, or the 22.04 kernel config so far, or
  some configuration made in 21.10 that isn't good with 22.04 and it's
  kernel anymore.

  I will go ahead tomorrow and see if I can build a vanilla kernel with
  the config from the ubuntu 22.04 kernel and "make oldconfig", then I
  will be able to tell if only the config is making the difference.

  Please let me know of there is anything I should test to further
  analyze this issue, or any ideas I can try to solve it without having
  to run a mainline manually installed kernel.

  Thanks.

  ProblemType: Bug
  DistroRelease: Ubuntu 22.04
  Package: linux-image-generic 5.15.0.30.33
  ProcVersionSignature: Ubuntu 5.15.0-30.31-generic 5.15.30
  Uname: Linux 5.15.0-30-generic x86_64
  NonfreeKernelModules: nvidia_modeset nvidia
  ApportVersion: 2.20.11-0ubuntu82
  Architecture: amd64
  AudioDevicesInUse:
   USER        PID ACCESS COMMAND
   /dev/snd/controlC0:  henning    6198 F.... pulseaudio
   /dev/snd/controlC1:  henning    6198 F.... pulseaudio
  CasperMD5CheckResult: unknown
  CurrentDesktop: ubuntu:GNOME
  Date: Sat May 14 23:02:38 2022
  InstallationDate: Installed on 2020-04-12 (761 days ago)
  InstallationMedia: Ubuntu 19.10 "Eoan Ermine" - Release amd64 (20191017)
  MachineType: LENOVO 20QV00CEGE
  ProcFB: 0 i915drmfb
  ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-5.15.0-30-generic 
root=/dev/mapper/vgubuntu-root ro
  RelatedPackageVersions:
   linux-restricted-modules-5.15.0-30-generic N/A
   linux-backports-modules-5.15.0-30-generic  N/A
   linux-firmware                             20220329.git681281e4-0ubuntu3
  SourcePackage: linux
  UpgradeStatus: Upgraded to jammy on 2022-04-28 (15 days ago)
  dmi.bios.date: 12/06/2021
  dmi.bios.release: 1.42
  dmi.bios.vendor: LENOVO
  dmi.bios.version: N2OET55W (1.42 )
  dmi.board.asset.tag: Not Available
  dmi.board.name: 20QV00CEGE
  dmi.board.vendor: LENOVO
  dmi.board.version: SDK0T08861 WIN
  dmi.chassis.asset.tag: No Asset Information
  dmi.chassis.type: 10
  dmi.chassis.vendor: LENOVO
  dmi.chassis.version: None
  dmi.ec.firmware.release: 1.23
  dmi.modalias: 
dmi:bvnLENOVO:bvrN2OET55W(1.42):bd12/06/2021:br1.42:efr1.23:svnLENOVO:pn20QV00CEGE:pvrThinkPadX1Extreme2nd:rvnLENOVO:rn20QV00CEGE:rvrSDK0T08861WIN:cvnLENOVO:ct10:cvrNone:skuLENOVO_MT_20QV_BU_Think_FM_ThinkPadX1Extreme2nd:
  dmi.product.family: ThinkPad X1 Extreme 2nd
  dmi.product.name: 20QV00CEGE
  dmi.product.sku: LENOVO_MT_20QV_BU_Think_FM_ThinkPad X1 Extreme 2nd
  dmi.product.version: ThinkPad X1 Extreme 2nd
  dmi.sys.vendor: LENOVO

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1973434/+subscriptions


-- 
Mailing list: https://launchpad.net/~kernel-packages
Post to     : kernel-packages@lists.launchpad.net
Unsubscribe : https://launchpad.net/~kernel-packages
More help   : https://help.launchpad.net/ListHelp

Reply via email to