out.c

#include <x86intrin.h>
#include <stdio.h>
#include <unistd.h>
#include <stdlib.h>
#include <inttypes.h>

static __inline__ uint64_t rdtsc(void){
    uint32_t hi, lo;
    __asm__ __volatile__ ("rdtsc" : "=a"(lo), "=d"(hi));
    return ( (uint64_t)lo)|( ((uint64_t)hi)<<32 );
}

#define COUNT 1024

int main(int argc, char ** argv) {
  uint64_t a1;
  uint64_t a2;
  uint64_t cnt;
  int i;
  
  uint64_t *counts = calloc(COUNT, 64);
  
  for (i = 0; i < 1000000000; i ++) {
    a1 = rdtsc();
    a2 = rdtsc();
    cnt = a2 - a1;
    if (cnt >= COUNT || cnt <= 0) {
      if (cnt > 100000) {
        printf("%" PRIu64 "\n", cnt);
      }
    } else {
      counts[(int)cnt]++;
    }
  }

  for (i = 10; i < 100; i++) {
    printf("%d - %ld\n", i, counts[i]);
  }
  return 0;
}

shell out.txt

$ gcc out.c -Wall -Werror -ansi -pedantic -O3 -o bang
$ perf stat  -- ./bang

$ # lots of output

 Performance counter stats for './bang':

         14,705.11 msec task-clock                #    1.000 CPUs utilized          
                44      context-switches          #    0.003 K/sec                  
                 0      cpu-migrations            #    0.000 K/sec                  
                69      page-faults               #    0.005 K/sec                  
    50,144,334,653      cycles                    #    3.410 GHz                    
    16,014,174,095      instructions              #    0.32  insn per cycle         
     2,002,637,169      branches                  #  136.186 M/sec                  
           126,570      branch-misses             #    0.01% of all branches        

      14.706359541 seconds time elapsed

      14.705729000 seconds user
       0.000000000 seconds sys