Rob Klein Wrote:
-----------------
'Complex' mathematical routines on the other hand could benefit from a different algorithmic approach, which may be as simple as using integer maths instead of floating point.

can anyone explain how to use integer maths instead of floating point maths?
do i need to make my own multiplication / division myself or is keil good enuf to do it?
My application needs precision of 8 digits, I'm stuck with slow hardware, and the floating point calculations take 2.5 msec to execute. Can it be optimised to take less than 1 msec?
Thanx