CopyOnWriteArrayList简单使用及性能测试

74 阅读 0 评论 49 点赞

我是靠谱客的博主潇洒爆米花，最近开发中收集的这篇文章主要介绍CopyOnWriteArrayList简单使用及性能测试，觉得挺不错的，现在分享给大家，希望可以做个参考。

概述

CopyOnWriteArrayList用于替代同步List,在某些情况下有更好的并发性能，并且在迭代时不需要加锁或复制。在每次修改的时候都会创建并重新发布一个新的容器副本。

下面是源码：


//添加元素，会加锁，创建一个新的对象，原引用指向新对象
public boolean add(E e) {
final ReentrantLock lock = this.lock;
lock.lock();
try {
Object[] elements = getArray();
int len = elements.length;
Object[] newElements = Arrays.copyOf(elements, len + 1);
newElements[len] = e;
setArray(newElements);
return true;
} finally {
lock.unlock();
}
}
//获取元素，不加锁
private E get(Object[] a, int index) {
return (E) a[index];
}
public E get(int index) {
return get(getArray(), index);
}

优点：多个线程可以同时对这个容器进行迭代，而不会相互干扰或者与修改容器的线程互相干扰，即不会抛出ConcurrentModificationException，并且返回的元素与迭代器创建时的元素完全一致，而不必考虑之后修改带来的影响。

缺点：

1.在添加元素时会创建副本，如果副本占用内较大，会增加gc，严重可能会影响相应时间。

2.数据无法保证实时性，可能读到的还是老的数据。

下面贴上测试代码：

import java.util.ArrayList;
import java.util.Collections;
import java.util.List;
import java.util.Random;
import java.util.concurrent.CopyOnWriteArrayList;
import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
/**
* @author wanting
*
*/
public class ListPerformanceTest {
// 线程数
private static int threadNum;
// 初始容量
private static int initCapacity;
// 循环次数
private static int loop;
private static CountDownLatch startLatch;
private static CountDownLatch endLatch;
private static List<String> cowList;
// 耗时
private volatile static double cowListTotalCost;
private static List<String> syncList;
// 耗时
private volatile static double syncListTotalCost;
public ListPerformanceTest(int threadNum, int initCapacity, int loop) {
ListPerformanceTest.initCapacity = initCapacity;
ListPerformanceTest.threadNum = threadNum;
ListPerformanceTest.loop = loop;
startLatch = new CountDownLatch(threadNum);
endLatch = new CountDownLatch(threadNum);
cowList = new CopyOnWriteArrayList<>(new ArrayList<>(initCapacity));
syncList = Collections.synchronizedList(new ArrayList<>(initCapacity));
initList();
}
public void initList() {
for (int i = 0; i < initCapacity; i++) {
cowList.add(i + "");
syncList.add(i + "");
}
}
public static void testPerformance(List<String> list) {
long t1 = System.nanoTime();
Random random = new Random();
for (int i = 0; i < loop; i++) {
int index = random.nextInt(initCapacity);
list.get(index);
}
long t2 = System.nanoTime();
double costTime = (t2 - t1) / 1000000d;
if (list instanceof CopyOnWriteArrayList) {
cowListTotalCost = cowListTotalCost + costTime;
} else {
syncListTotalCost = syncListTotalCost + costTime;
}
// System.out.println(list.getClass() + "耗时：" + costTime + "秒");
}
public static void addEle(String element) {
cowList.add(element);
syncList.add(element);
}
public static void main(String[] args) {
ListPerformanceTest listPerformanceTest = new ListPerformanceTest(500, 100000, 10000);
ExecutorService pool = Executors.newFixedThreadPool(threadNum);
for (int i = 1; i <= threadNum; i++) {
final int flag = i;
pool.submit(new Runnable() {
@Override
public void run() {
try {
startLatch.await();
//if (flag % 2 == 0) {
//addEle(flag + "");
// }
testPerformance(cowList);
testPerformance(syncList);
} catch (Exception e) {
e.printStackTrace();
}
endLatch.countDown();
}
});
startLatch.countDown();
}
try {
endLatch.await();
} catch (InterruptedException e) {
e.printStackTrace();
}
System.out.println("cowz总耗时：" + cowListTotalCost + "微秒，平均耗时：" + (cowListTotalCost / (threadNum * 1.0) + "微秒"));
System.out
.println("sync总耗时：" + syncListTotalCost + "微秒，平均耗时：" + (syncListTotalCost / (threadNum * 1.0) + "微秒"));
pool.shutdown();
}

总结：适合用于迭代操作远远多于修改操作时。随着并发量增加或者迭代次数增加，CopyOnWriteArrayList性能要优于同步List。