Measuring FLOPS with a simple C routine

Question

I am trying to measure FLOPS in rather simple way:

clock_gettime(CLOCK_REALTIME, &start);
num1 + num2;
clock_gettime(CLOCK_REALTIME, &end);
ns += end.tv_nsec - start.tv_nsec;

I run this in a loop and then compute how many nanoseconds on average does it take to do this operation.

I am obtaining results that I was not expecting based on the published performance numbers of my CPU.

After further reading my guess is that I am erroneously equating the C statement of adding two floating point numbers with a Floating Point Operation.

My question is: How exactly is a FLOP measured? Is it purely based properties of the CPU such as their frequency?

My complete code:

#include <stdio.h>
#include <stdlib.h>
#include <time.h>
#include <unistd.h>
int main(int argc, char *argv[])
{
    if (argc != 2) return -1;
    int n = atoi(argv[1]);
    if (n <= 0) return -1;

    float num1;
    float num2;
    struct timespec start, end, res;
    float ns = 0;
    clock_getres(CLOCK_REALTIME, &res);
    fprintf(stderr, "CLOCK resolution: %ld nanosecond(s).\n", res.tv_nsec);

    for (int i = 0; i < n; i++) {
        num1 = ((float)rand()/(float)(RAND_MAX));
        num2 = ((float)rand()/(float)(RAND_MAX));
        clock_gettime(CLOCK_REALTIME, &start);
        num1 + num2;
        clock_gettime(CLOCK_REALTIME, &end);
        ns += end.tv_nsec - start.tv_nsec;
    }
    fprintf(stderr, "Average time per operation: %.4f\n", ns/n);
}