By: ValentinHirschi <valentin.hirschi@gmail.com>
Use native rounded Float arithmetic in precision helpers