c - 为什么n个线程的平均速度不如C中一个单线程的速度？

Question

我写了一个程序，有 2 个线程做同样的事情，但我发现每个线程的吞吐量比我只生成一个线程的吞吐量要慢。然后我编写了这个简单的测试，看看这是我的问题还是系统的问题。

#include <stdio.h>
#include <stdlib.h>
#include <pthread.h>
#include <time.h>


/*
 * Function: run_add
 * -----------------------
 * Do addition operation for iteration ^ 3 times
 *
 * returns: void
 */
void *run_add(void *ptr) {
  clock_t t1, t2;
  t1 = clock();

  int sum = 0;
  int i = 0, j = 0, k = 0;
  int iteration = 1000;
  long total = iteration * iteration * iteration;
  for (i = 0; i < iteration; i++) {
    for (j = 0; j < iteration; j++) {
      for (k = 0; k < iteration; k++) {
        sum++;
      }
    }
  }

  t2 = clock();
  float diff = ((float)(t2 - t1) / 1000000.0F );
  printf("thread id = %d\n", (int)(pthread_self()));
  printf("Total addtions: %ld\n", total);
  printf("Total time: %f second\n", diff);
  printf("Addition per second: %f\n", total / diff);
  printf("\n");

  return NULL;
}


void run_test(int num_thread) {
  pthread_t pth_arr[num_thread];
  int i = 0;
  for (i = 0; i < num_thread; i++) {
    pthread_create(&pth_arr[i], NULL, run_add, NULL);
  }

  for (i = 0; i < num_thread; i++) {
    pthread_join(pth_arr[i], NULL);
  }
}

int main() {
  int num_thread = 5;
  int i = 0;
  for (i = 1; i < num_thread; i++) {
    printf("Running SUM with %d threads. \n\n", i);
    run_test(i);
  }
  return 0;
}

结果仍然显示 n 个线程的平均速度比单个线程慢。我拥有的线程越多，每个线程就越慢。

结果如下:

Running SUM with 1 threads.

thread id = 528384,Total addtions: 1000000000,Total time: 1.441257 second,Addition per second: 693838784.000000

Running SUM with 2 threads.

thread id = 528384,Total addtions: 1000000000,Total time: 2.970870 second,Addition per second: 336601728.000000

thread id = 1064960,Total addtions: 1000000000,Total time: 2.972992 second,Addition per second: 336361504.000000

Running SUM with 3 threads.

thread id = 1064960,Total addtions: 1000000000,Total time: 4.434701 second,Addition per second: 225494352.000000

thread id = 1601536,Total addtions: 1000000000,Total time: 4.449250 second,Addition per second: 224756976.000000

thread id = 528384,Total addtions: 1000000000,Total time: 4.454826 second,Addition per second: 224475664.000000

Running SUM with 4 threads.

thread id = 528384,Total addtions: 1000000000,Total time: 6.261967 second,Addition per second: 159694224.000000

thread id = 1064960,Total addtions: 1000000000,Total time: 6.293107 second,Addition per second: 158904016.000000

thread id = 2138112,Total addtions: 1000000000,Total time: 6.295047 second,Addition per second: 158855056.000000

thread id = 1601536,Total addtions: 1000000000,Total time: 6.306261 second,Addition per second: 158572560.000000

我有一个 4 核 CPU，我的系统监视器显示每次我运行 n 个线程时，n 个 CPU 内核都被 100% 使用。 n 个线程(<= 我的 CPU 内核)的运行速度应该是一个线程的 n 倍，这是真的吗？为什么这里不是这种情况？

Answer 1

clock() 测量 CPU 时间而不是“墙”时间。它还测量所有线程的总时间..

CPU 时间是处理器执行代码的时间，wall time 是真实世界的运行时间(就像墙上的时钟显示的那样)

使用/usr/bin/time 为你的程序计时，看看到底发生了什么。或者使用像 time()、gettimeofday() 或 clock_gettime() 这样的挂钟函数

clock_gettime() 可以测量此线程、此进程或挂钟时间的 CPU 时间。 - 这可能是进行此类实验的最佳方式。