What kind of battery life are we getting on this bad boy?
So if benchmarks are laughable, how do we quantify if X change improves performance or not? Seat of the pants "I just can tell" measurements are not the way to go.
I wanted a baseline for the current cost of my kernel being on but otherwise idle. For kicks I also ran the same test with the stock kernel as a comparison.
My simple test was as follows. I ran my 1.1.5 kernel at 192MHz/1.83GHz with ondemand and compared it with DSB's stock write (because that was the easiest thing to flash on Viper). I turned the brightness down to 0, screen timeout to 30 minutes and left the phone untouched for 30 minutes. To quantify the battery use, I assume that the screen and radio battery consumption will be constant and looked at the ratio between these as an indicator of the relative battery consumption.
Stock used 1.7x more battery when otherwise idle but turned on.