Found it: https://www.kernel.org/doc/html/latest/admin-guide/pm/intel_pstate.html#operation-modes
You can verify whether your intel_pstate driver runs in active mode by checking if the following reports “intel_pstate”:
cat /sys/devices/system/cpu/cpufreq/policy0/scaling_driver
I see the pcc_cpufreq scaling driver module is loaded as well for some reason, should however not be in use. I found a patch which disabled all dynamic scaling governors for this driver when more than 4 CPUs are present: https://patchwork.kernel.org/project/linux-pm/patch/5423012.ZZnfdYddaT@aspire.rjw.lan/
If this was the case for someone, it could be tried to use the more generic acpi_cpufreq scaling driver instead, by blacklisting echo ‘blacklist pcc_cpufreq’ > /etc/modprobe.d/disable_pcc_cpufreq.conf. But if intel_pstate is used (verified above), it’s doesn’t play a role. The module can still be blacklisted to reduce memory usage a little.
So what we need to change is adjusting the governor text in dietpi-config, when intel_pstate runs in active mode, to reflect that both do dynamic frequency scaling, just more aggressive or conservative, respectively.