It being easier to verify 2 power limits work right than a whole bunch of them is likely at least part of the answer. the other one is “it has always been done like that”.
On the other hand you can just override the power limits the ec sets so you aren’t entirely screwed here.
1W?! that’s just full kneecap and actually uses more power to do the same thing as something more reasonable so if you do that for battery saving you are screwing yourself XD
I wonder if there’s a program that can find a given cpu’s effeciency sweet spot. If not, it may be possible to create a script that helps find it (if it exists). E.g. set power target, run benchmarks, calculate score / power, change power target, rinse and repeat.
I did some mining on my gaming rig back in 2017 and did this myself. I found that just lowering the GPU’s power target as far as it would go was the path to max effeciency.
I did that by hand though I think the cpu reports something like that to the os (lowest non linear frequency or something).
Given how constrained those power sliders tend to be I am not surprised that works, as far as it would go was likely nowhere near as far as 1W is going here. All the way down on a gpu maybe for 450 to 380 or back in 2017 maybe from 250 to 200.
For an ultra power saving mode I’d go for somewhere between 10 and 15 but I am not sure if that is the same for the hs, it is the same die but the different binning (and chip configuration) may make a difference here.