Trouble with batteries
I finally managed to make Tony's dyn-tick patch to work... It did interfere badly with cpu frequency scaling... kind of understendable.
Now I'm searching for a way to measure power consuption... evo N620c internal ampermeter seems to be only updated when I [un]plug the cord. nx5000's internal ampermeter does not update under CPU load [or something like that]. Arima's ampermeter seems to work, but athlon64 machine is not exactly ideal for playing with power consumption... and then I have external ampermeter with only 2W accuracy :-(. I *was* able to measure ~0.6 W power difference between HZ=100 and HZ=1000 long time ago, so dyn-tick could help quite a lot.