[mobad]

In order to understand how doing a one benchmark app it could and would be compared with one by competition in auto racing?

One Golf GTI in Nuremberg on the racing circuit

- driver is not always the same
- Weather Conditions rain fall winds blow in different directions
- temperature asphalt
- how many in the circuit participants
Auto
- new motor
- new oil
- new tires
- new brakes
- quality gasoline
- car's weight

One benchmark is, as same as a stopwatch in the circut.
- How has been started android
- filled correctly all the drivers
- in the background is not running anything
- no application running to backup make on SD card
- what's your`s kernel
- what's your`s ROM
- what`s your`s governor
- what is your init.d folder

In the above mentioned cases,to get the result always the same all this parameter must always be the same?
Otherwise you will never get the same result?
With, it does not matter of is the stopwatch in circuit in Nuremberg incorrect or has a mistake.
The same conclusion benchmark has itself a mistake of not.
Stopwatch in Nuremberg-circuit carry his mistake always, it is no matter how much time you changed a engine, tires, brakes etc.
Applications like benchmarks carry his mistakes always no matter how much time you changed ROMs, kernel, software on the same Device?
Benchmarks count with their mistake in itself
always some the result is approximate exact.


Nothing is wrong with the benchmarks there are something wrong in the head of the way of thinking at xda members and staff!