CPU frequency issues after upgrading to Core Update 159

After update 159 —>>> CPU frequency 1 core max

Hi,

welcome to the IPFire community. :slight_smile:

Since your issue differs, I have moved your post into a new thread, so things don’t get mixed up.

Actually, this is harmless - it just means Core Update 159 (and the Linux kernel 5.10.x inside) changed the behaviour of controlling your CPU’s frequency.

Please provide us further information:

  • Which CPU are you using? (A link to your Fireinfo profile is helpful, see here.)
  • Do you have the cpufrequtils add-on installed?

Thanks, and best regards,
Peter Müller

1 Like

Have you leaved the WebGui open? The ajax speeddisplay currently make a lot of cpu load.

1 Like

Only upgraded a few to 159 so far, but majority reduced frequency in my case.
Core 2 Duo E8400 down


Core 2 6600 up

OK Thanks you

My INFO fireinfo.ipfire.org - Profile cc218309030bb945f12776144f53ed25ea9e754b

EDIT: mod added missing [/quote]

For completeness of information…

Intel(R) Pentium(R) D CPU 3.00GHz x2

Intel(R) Core™2 Duo CPU E8500 @ 3.16GHz x2

Intel(R) Core™2 Duo CPU E7500 @ 2.93GHz x2


Intel(R) Core™2 Duo CPU E8400 @ 3.00GHz x2

Intel(R) Celeron(R) J4105 CPU @ 1.50GHz x4

the situation readjusted itself after a few days

CPU Frequency stuck around minimum after update to core 159
Intel Celeron CPU 847 @ 1.10GHz
Fireinfo

Problem solved:
So far I had set the CPU governor to powersave. This worked perfectly under the old kernel, CPU clock between 800 MHz and 1000 MHz depending on the load. With the new kernel, powersave apparently sets the clock to the minimum of 800MHz. I have now set the governor to conservative and now the CPU speed is adjusted to the system load again.

Same problem here…
One core run’s always at full speed (2415MHz).
But this govenor-settings doesnt help…
On core 158, all cores were working at 1333MHz - so there is maybe more load or something else…

Since Core Update 159 with the new kernel my four cores are much more balanced in terms of frequency than they were before. I also tend to see a more balanced activity percentage per core using htop than before, when two cores tended to do the majority of the work and the other two cores were lower until the first two cores were nearly at max. This also matches with the frequency effects with two cores typically at lower frequency and the other two at higher, although the difference is not huge anyway, ranging from 1.4MHz to 1.75MHz