Hi Mahmood,

I can not tell, if it can be optimised, since you tell not, which calculation you want.

Most formulas can be optimised before they are implemented in software. E.g. "x = x * 0.1" was faster than "x = x / 10" .

Assembler should not be the right way. I have tested the Keil 16- and 32-bit integer routines and they are 2 ... 10 times faster as the known assembler routines.


Peter