cpicfixed-point16-bitpic24

32-bit Integer Arithmetic on 16-bit system for PIC

I need to implement a 32-bit Fixed-point on a 16-bit system. My compiler seems to be giving me an "warning: integer overflow in expression". I'm not sure what I'm missing anything :S . I'll provide a dumbed down version of my code below. Any advice would be appreciate :) 

#include <stdint.h>
typedef int32_t Q22;

#define Q22base     512
#define Q22_Convert(_a) ((Q22)(_a*Q22base))

int main (){
    // a bunch of code that does stuff goes here
    Q22 variable1=0;

    variable1 = Q22_Convert(some_value1) - Q22_Convert(some_value1);      
    return 0;
}
Solution

  • Your Q22base is 16-bit on a 16-bit int system, you need to cast it to Q22 so your * expression is done in at least a 32-bit type.

    #define Q22_Convert(_a) ((Q22)(_a * (Q22) Q22base))