Public bug reported:

I'm having some thermal trouble since I've started using kernel >4.0.x
with my new laptop. According to psensor and sensord, CPU temperature
jumps above the 90C, stays there for around 1second, and goes back to
50C. All transitions happen in around 100ms. I've cleaned and re-applied
thermal paste and checked the connections twice. More interestingly,
this doesn't happen when the CPU is under heavy-load, but when it is
waiting idly. Sometimes sensord reports the CPU temp has gone over the
critical threshold of the CPU and device shuts down itself immediately.

There is also Windows 10 installed as a dual boot and didn't see this
problem on it when idle or under heavy load.

Dmesg output is full of following warnings;

[ 1282.296247] CPU0: Core temperature above threshold, cpu clock throttled 
(total events = 109)[ 1282.296267] CPU4: Core temperature above threshold, cpu 
clock throttled (total events = 109)
[ 1282.296269] CPU5: Package temperature above threshold, cpu clock throttled 
(total events = 220)
[ 1282.296269] CPU1: Package temperature above threshold, cpu clock throttled 
(total events = 220)
[ 1282.296271] CPU4: Package temperature above threshold, cpu clock throttled 
(total events = 220)
[ 1282.296272] CPU6: Package temperature above threshold, cpu clock throttled 
(total events = 220)
[ 1282.296273] CPU2: Package temperature above threshold, cpu clock throttled 
(total events = 220)
[ 1282.296274] CPU7: Package temperature above threshold, cpu clock throttled 
(total events = 220)
[ 1282.296275] CPU3: Package temperature above threshold, cpu clock throttled 
(total events = 220)
[ 1282.296281] CPU0: Package temperature above threshold, cpu clock throttled 
(total events = 220)
[ 1282.297226] CPU4: Core temperature/speed normal
[ 1282.297227] CPU5: Package temperature/speed normal
[ 1282.297228] CPU0: Core temperature/speed normal
[ 1282.297229] CPU1: Package temperature/speed normal
[ 1282.297229] CPU0: Package temperature/speed normal
[ 1282.297230] CPU4: Package temperature/speed normal
[ 1282.297233] CPU3: Package temperature/speed normal
[ 1282.297233] CPU7: Package temperature/speed normal
[ 1282.297269] CPU2: Package temperature/speed normal
[ 1282.297269] CPU6: Package temperature/speed normal

Because of wrong temperature reporting, kernel throttles the CPU and
reduces overall performance, which results in frustrating user
experience.

I've tried followings and find a temporary solution;
 - intel_pstate enabled, turbo-boost enabled > problem exists
 - intel_pstate enabled, turbo-boost disabled > problem frequency reduced
 - intel_pstate disabled, governor set to anythig other than powersave > 
problem exists
 - intel_pstate disabled, governor set to powersave > problem frequency reduced

I'm suspicious about a kernel bug such as;
- https://bugzilla.redhat.com/show_bug.cgi?id=924570
- https://bugzilla.redhat.com/show_bug.cgi?id=1317190

** Affects: linux (Ubuntu)
     Importance: Undecided
         Status: New

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1781924

Title:
  Kernel wrong temperature reporting

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1781924/+subscriptions

-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to