cpu-cache-effects.c

#include <bits/time.h>
#include <stdint.h>
#include <stdlib.h>
#include <stdio.h>
#include <sys/types.h>
#include <time.h>

// https://igoro.com/archive/gallery-of-processor-cache-effects/

u_int64_t now_ns()
{
    struct timespec ts;
    clock_gettime(CLOCK_MONOTONIC_RAW, &ts);
    return ts.tv_sec * (uint64_t)1e9 + ts.tv_nsec;
}

__inline__ uint64_t rdtsc(void)
{
  uint32_t lo, hi;
  __asm__ __volatile__ (      // serialize
  "xorl %%eax,%%eax \n        cpuid"
  ::: "%rax", "%rbx", "%rcx", "%rdx");
  /* We cannot use "=A", since this would use %rax on x86_64 and return only the lower 32bits of the TSC */
  __asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi));
  return (uint64_t)hi << 32 | lo;
}

#define N (1024*1024*64)
int main()
{
    int * array = calloc(sizeof(int), N);
    for (size_t k = 0; k < 4; ++k)
    {
        uint64_t start = rdtsc();
        for (size_t i = 0; i < N; i += 1)
        {
            array[i] *= 3;
        }
        uint64_t end = rdtsc();
        printf("elapsed: %lu\n", end - start);
    }
    free(array);
    return 0;
}