inverse.c

// gcc -std=c99 -o inverse inverse.c
#include <stdlib.h>
#include <stdio.h>
#include <stdint.h> // use C99
#include <assert.h>

static inline uint32_t f32(uint32_t x, uint32_t y) {
  return  y * ( 2 - y * x );
}

static uint32_t findInverse32(uint32_t x) {
   uint32_t y = x; 
   y = f32(x,y);
   y = f32(x,y);
   y = f32(x,y);
   y = f32(x,y);
   return y;
}


static inline uint64_t f64(uint64_t x, uint64_t y) {
  return  y * ( 2 - y * x );
}

static uint64_t findInverse64(uint64_t x) {
   uint64_t y = x; 
   y = f64(x,y);
   y = f64(x,y);
   y = f64(x,y);
   y = f64(x,y);
   y = f64(x,y);
   return y;
}

int main(int argc, char ** argv) {
  if(argc < 2 ) {
    printf("I expect an integer.\n");
    return EXIT_FAILURE;
  }
  long long val = strtoll(argv[1], NULL, 0);  
  uint64_t val64 = (uint64_t) val; // wrap around negatives
  val64 |= 1; // make it odd if not already odd
  uint64_t inverse64 = findInverse64(val64);
  assert(inverse64 * val64 == 1);
  printf("The 64-bit inverse of %zu is %zu (%zx) \n", (size_t) val64, (size_t) inverse64, (size_t) inverse64);
  uint32_t val32 = (uint32_t) val64;
  uint32_t inverse32 = findInverse32(val32);
  assert(inverse32 * val32 == 1);
  printf("The 32-bit inverse of %u is %u (%x) \n", val32, inverse32, inverse32);
  return EXIT_SUCCESS;
}