© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
0e1f616
(Get the code!)
I have some ~different behavior, but I was still able to achieve bug 1.
Initially I saw some intel_pstate/
long_term_power would fluctuate from very low values all the way up to the nominal 65. Previously when triggering the bug it would slowly go mostly down toward 0.125, getting stuck between 0.125 and ~4.0, with the cpu clocked at 400Mhz. Now it will go up, seemingly reset up to 65, and allow the max frequencies.
I was able to get something like bug 1 to happen twice.
long_term_power got stuck at 19.5 (unusual, might have been 18.5 the first time)
long_term_time got stuck at 28.0 (typical, its usually this or 32.0)
This happened when I toggled between one and two pegged cores to keep the temperature at a maximum. If I killed the stress commands at the right moment while the fan was spun up, while long_term_power was in the high teens, if I then waited for the spans to spin down - I would get into a situation where I could not sustain maximum clocks long enough to reach temperatures to trigger thermald to.... poke whatever it is that resets long_term_power back up to 65.
I was able to do this twice.
On my third attempt I accidentally got a long_term_power value 1f 12.5. I waited for fans to spin down, started stressing again, and it pretty immediately dropped to 0.125 with cpu speeds locked to 400Mhz. I also noticed at the end that intelpstate/
Finally after resetting the long_power_mode to 65 I noticed that I still couldn't get the fan up because intelpstate/
(Likely irrelevant: even though I see there is that intel wireless driver deb package (I installed all of the packages) in that 6.2-26 kernel folder you shared with me, my intel wifi card does not work when booted into it.)
this was with the thermald build in adaptive mode, the full command was:
sudo ./thermald --no-daemon --adaptive --loglevel=debug