Honestly I’m getting more and more disappointed with cpufrequtils. Doesn’t work with Raspberry Pi 4 and with recent Intel systems there are only two governors left.
I’m used to have the governors userspace, ondemand, conservative, powersave and performance.
However only powersave and performance are left. Why is that.
Arne told us that with the installation of the add-on one of his scripts enables the governor ondemand. That’s not working anymore because that governor does not exist.
Edit: every time I restart the system any changes (governor, max cpu freq) are gone. Why is that?
I edit the script
I changed the governor and did a reboot. Nothing happened. I did a manual execute, getting echo “OK” 3 times for things I don’t know, but still governor performance.
I thought Arne made that script for setting the governor ondemand on boot? But the info that he made the script is gone? Also were is he? Haven’t seen him for years.
I don’t think this feature is getting much love from Kernel developers any more. Modern processors don’t need it and generally do not save any more energy when using cpufrequtils.
Hm I had a look on my linux pc (Manjaro, Kernel 5.18.3, Intel 10600T) and there are also just 2 governors: performance and powersave. However the PC runs @ powersave - don’t know why. I have never set up anything.
IPfire always runs @ performance and max cpu frequency.
powersave is normally what desktop distributions use by default.
In IPFire, we use “performance” which does not sent the CPU into deep sleep because it can take a long time to wake up the CPU and that will increase network latency.
In our benchmark, we could not measure any significant discrepancies in power consumption.
DeepSx is disabled in BIOS but usually I don’t have that option so this is an true argument. I’m less interested in the power consumption of the CPU, but in its temperature. It heats up all nearby components and that’s not got for my SSDs.
However that topic seems to be less related to ipfire, than the kernel power management.