...turns out PDE is slower than expected; doing the same thing in c99 (-O3'ed) is an order of magnitude faster?

csrt.c

// c99

#include <stdio.h>
#include <stdlib.h>
#include <stdint.h>
#include <string.h>
#include <time.h>

#define COUNT 200000000
#define RND_RANGE 200000000
//#define OUTPUT

int32_t ip[COUNT];
int32_t op[RND_RANGE];


int32_t main()
{
    srand( time(NULL) );
    printf("generating PRN array[%i], range: %i\n", COUNT, RND_RANGE);

    for( int32_t i = 0; i < COUNT; i++ )
    {
        ip[i] = rand() % RND_RANGE;
    }

    memset(op, 0, sizeof(op) );

    printf("sorting PRN array...\n");

    for( int32_t i = 0; i < COUNT; i++ )
    {
        op[ ip[i] ]++;
    }

#ifdef OUTPUT
    printf("output: \n");

    for( int32_t i = 0; i < RND_RANGE; i++ )
    {
        if (op[i] > 0) printf("%i(%i)\t", i, op[i] );
    }
#endif // OUTPUT

    return 0;
}