Java中对AtomicInteger和int值在多线程下递增操作的测试

java针对多线程下的数值安全计数器设计了一些类,这些类叫做原子类,其中一部分如下:

java.util.concurrent.atomic.atomicboolean;
java.util.concurrent.atomic.atomicinteger;
java.util.concurrent.atomic.atomiclong;
java.util.concurrent.atomic.atomicreference;

下面是一个对比  atomicinteger 与 普通 int 值在多线程下的递增测试,使用的是 junit4;

完整代码:

package test.java;

import java.util.concurrent.countdownlatch;
import java.util.concurrent.atomic.atomicinteger;

import org.junit.assert;
import org.junit.before;
import org.junit.test;

/**
 * 测试atomicinteger与普通int值在多线程下的递增操作
 */
public class testatomic {

 // 原子integer递增对象
 public static atomicinteger counter_integer;// = new atomicinteger(0);
 // 一个int类型的变量
 public static int count_int = 0;

 @before
 public void setup() {
 // 所有测试开始之前执行初始设置工作
 counter_integer = new atomicinteger(0);
 }

 @test
 public void testatomic() throws interruptedexception {
 // 创建的线程数量
 int threadcount = 100;
 // 其他附属线程内部循环多少次
 int loopcount = 10000600;
 // 控制附属线程的辅助对象;(其他await的线程先等着主线程喊开始)
 countdownlatch latch_1 = new countdownlatch(1);
 // 控制主线程的辅助对象;(主线程等着所有附属线程都运行完毕再继续)
 countdownlatch latch_n = new countdownlatch(threadcount);
 // 创建并启动其他附属线程
 for (int i = 0; i < threadcount; i++) {
  thread thread = new atomicintegerthread(latch_1, latch_n, loopcount);
  thread.start();
 }
 long startnano = system.nanotime();
 // 让其他等待的线程统一开始
 latch_1.countdown();
 // 等待其他线程执行完
 latch_n.await();
 //

 long endnano = system.nanotime();
 int sum = counter_integer.get();
 //
 assert.assertequals("sum 不等于 threadcount * loopcount,测试失败",
  sum, threadcount * loopcount);
 system.out.println("--------testatomic(); 预期两者相等------------");
 system.out.println("耗时: " + ((endnano - startnano) / (1000 * 1000)) + "ms");
 system.out.println("threadcount = " + (threadcount) + ";");
 system.out.println("loopcount = " + (loopcount) + ";");
 system.out.println("sum = " + (sum) + ";");
 }

 @test
 public void testintadd() throws interruptedexception {
 // 创建的线程数量
 int threadcount = 100;
 // 其他附属线程内部循环多少次
 int loopcount = 10000600;
 // 控制附属线程的辅助对象;(其他await的线程先等着主线程喊开始)
 countdownlatch latch_1 = new countdownlatch(1);
 // 控制主线程的辅助对象;(主线程等着所有附属线程都运行完毕再继续)
 countdownlatch latch_n = new countdownlatch(threadcount);
 // 创建并启动其他附属线程
 for (int i = 0; i < threadcount; i++) {
  thread thread = new integerthread(latch_1, latch_n, loopcount);
  thread.start();
 }
 long startnano = system.nanotime();
 // 让其他等待的线程统一开始
 latch_1.countdown();
 // 等待其他线程执行完
 latch_n.await();
 //
 long endnano = system.nanotime();
 int sum = count_int;
 //
 assert.assertnotequals(
  "sum 等于 threadcount * loopcount,testintadd()测试失败", 
  sum, threadcount * loopcount);
 system.out.println("-------testintadd(); 预期两者不相等---------");
 system.out.println("耗时: " + ((endnano - startnano) / (1000*1000))+ "ms");
 system.out.println("threadcount = " + (threadcount) + ";");
 system.out.println("loopcount = " + (loopcount) + ";");
 system.out.println("sum = " + (sum) + ";");
 }

 // 线程
 class atomicintegerthread extends thread {
 private countdownlatch latch = null;
 private countdownlatch latchdown = null;
 private int loopcount;

 public atomicintegerthread(countdownlatch latch,
  countdownlatch latchdown, int loopcount) {
  this.latch = latch;
  this.latchdown = latchdown;
  this.loopcount = loopcount;
 }

 @override
 public void run() {
  // 等待信号同步
  try {
  this.latch.await();
  } catch (interruptedexception e) {
  e.printstacktrace();
  }
  //
  for (int i = 0; i < loopcount; i++) {
  counter_integer.getandincrement();
  }
  // 通知递减1次
  latchdown.countdown();
 }
 }

 // 线程
 class integerthread extends thread {
 private countdownlatch latch = null;
 private countdownlatch latchdown = null;
 private int loopcount;

 public integerthread(countdownlatch latch, 
  countdownlatch latchdown, int loopcount) {
  this.latch = latch;
  this.latchdown = latchdown;
  this.loopcount = loopcount;
 }

 @override
 public void run() {
  // 等待信号同步
  try {
  this.latch.await();
  } catch (interruptedexception e) {
  e.printstacktrace();
  }
  //
  for (int i = 0; i < loopcount; i++) {
  count_int++;
  }
  // 通知递减1次
  latchdown.countdown();
 }
 }
}

普通pc机上的执行结果类似如下:

--------------testatomic(); 预期两者相等-------------------
耗时: 85366ms
threadcount = 100;
loopcount = 10000600;
sum = 1000060000;
--------------testintadd(); 预期两者不相等-------------------
耗时: 1406ms
threadcount = 100;
loopcount = 10000600;
sum = 119428988;

从中可以看出, atomicinteger操作 与 int操作的效率大致相差在50-80倍上下,当然,int很不消耗时间,这个对比只是提供一个参照。

如果确定是单线程执行,那应该使用 int; 而int在多线程下的操作执行的效率还是蛮高的, 10亿次只花了1.5秒钟；

 (假设cpu是 2ghz，双核4线程,理论最大8ghz，则每秒理论上有80亿个时钟周期,

 10亿次java的int增加消耗了1.5秒,即 120亿次运算, 算下来每次循环消耗cpu周期 12个;

个人觉得效率不错, c 语言也应该需要4个以上的时钟周期(判断,执行内部代码,自增判断,跳转)

 前提是: jvm和cpu没有进行激进优化.

)

而 atomicinteger 效率其实也不低,10亿次消耗了80秒, 那100万次大约也就是千分之一,80毫秒的样子.