转自：http://blog.csdn.net/liweisnake/article/details/12966761
今天看到一篇文章，是关于java中如何等待所有线程都执行结束，文章总结得很好，原文如下http://software.intel.com/zh-cn/blogs/2013/10/15/java-countdownlatchcyclicbarrier/?utm_campaign=CSDN&utm_source=intel.csdn.net&utm_medium=Link&utm_content=others-%20Java
看过之后在想java中有很大的灵活性，应该有更多的方式可以做这件事。
这个事情的场景是这样的：许多线程并行的计算一堆问题，然后每个计算存在一个队列，在主线程要等待所有计算结果完成后排序并展示出来。这样的问题其实很常见。
1. 使用join。这种方式其实并不是那么的优雅，将所有线程启动完之后还需要将所有线程都join，但是每次join都会阻塞，直到被join线程完成，很可能所有被阻塞线程已经完事了，主线程还在不断地join，貌似有点浪费，而且两个循环也不太好看。
public void testThreadSync1() {
final Vector<Integer> list = new Vector<Integer>();
Thread[] threads = new Thread[TEST_THREAD_COUNT];
try {
for (int i = 0; i < TEST_THREAD_COUNT; i++) {
final int num = i;
threads[i] = new Thread(new Runnable() {
public void run() {
try {
Thread.sleep(random.nextInt(100));
} catch (InterruptedException e) {
e.printStackTrace();
}
list.add(num);
System.out.print(num + " add.\t");
}
});
threads[i].start();
}
for (int i = 0; i < threads.length; i++) {
threads[i].join();
System.out.print(i + " end.\t");
}
} catch (InterruptedException ie) {
ie.printStackTrace();
}
printSortedResult(list);
}
9 add. 7 add. 3 add. 5 add. 4 add. 1 add. 0 add. 0 end. 1 end. 8 add. 2 add. 2 end. 3 end. 4 end. 5 end. 6 add. 6 end. 7 end. 8 end. 9 end.
before sort
9 7 3 5 4 1 0 8 2 6
after sort
0 1 2 3 4 5 6 7 8 9
2. 使用wait/notifyAll，这个方式其实跟上面是类似的，只是比较底层些吧(join实际上也是wait)。
@Test
public void testThreadSync2() throws IOException, InterruptedException {
final Object waitObject = new Object();
final AtomicInteger count = new AtomicInteger(TEST_THREAD_COUNT);
final Vector<Integer> list = new Vector<Integer>();
Thread[] threads = new Thread[TEST_THREAD_COUNT];
for (int i = 0; i < TEST_THREAD_COUNT; i++) {
final int num = i;
threads[i] = new Thread(new Runnable() {
public void run() {
try {
Thread.sleep(random.nextInt(100));
} catch (InterruptedException e) {
e.printStackTrace();
}
list.add(num);
System.out.print(num + " add.\t");
synchronized (waitObject) {
int cnt = count.decrementAndGet();
if (cnt == 0) {
waitObject.notifyAll();
}
}
}
});
threads[i].start();
}
synchronized (waitObject) {
while (count.get() != 0) {
waitObject.wait();
}
}
printSortedResult(list);
}
3. 使用CountDownLatch，这其实是最优雅的写法了，每个线程完成后都去将计数器减一，最后完成时再来唤醒。
例1
@Test
public void testThreadSync3() {
final Vector<Integer> list = new Vector<Integer>();
Thread[] threads = new Thread[TEST_THREAD_COUNT];
final CountDownLatch latch = new CountDownLatch(TEST_THREAD_COUNT);
for (int i = 0; i < TEST_THREAD_COUNT; i++) {
final int num = i;
threads[i] = new Thread(new Runnable() {
public void run() {
try {
Thread.sleep(random.nextInt(100));
} catch (InterruptedException e) {
e.printStackTrace();
}
list.add(num);
System.out.print(num + " add.\t");
latch.countDown();
}
});
threads[i].start();
}
try {
latch.await();
} catch (InterruptedException e) {
e.printStackTrace();
}
printSortedResult(list);
}
例2
CountDownLatch 初始化设置count，即等待(await)count个线程或一个线程count次计数，通过工作线程来countDown计数减一，直到计数为0，await阻塞结束。
设置的count不可更改，如需要动态设置计数的线程数，可以使用CyclicBarrier.
下面的例子，所有的工作线程中准备就绪以后，并不是直接运行，而是等待主线程的信号后再执行具体的操作。
package com.example.multithread; import java.util.concurrent.CountDownLatch; class Driver
{
private static final int TOTAL_THREADS = 10;
private final CountDownLatch mStartSignal = new CountDownLatch(1);
private final CountDownLatch mDoneSignal = new CountDownLatch(TOTAL_THREADS); void main()
{
for (int i = 0; i < TOTAL_THREADS; i++)
{
new Thread(new Worker(mStartSignal, mDoneSignal, i)).start();
}
System.out.println("Main Thread Now:" + System.currentTimeMillis());
doPrepareWork();// 准备工作
mStartSignal.countDown();// 计数减一为0，工作线程真正启动具体操作
doSomethingElse();//做点自己的事情
try
{
mDoneSignal.await();// 等待所有工作线程结束
}
catch (InterruptedException e)
{
// TODO Auto-generated catch block
e.printStackTrace();
}
System.out.println("All workers have finished now.");
System.out.println("Main Thread Now:" + System.currentTimeMillis());
} void doPrepareWork()
{
System.out.println("Ready,GO!");
} void doSomethingElse()
{
for (int i = 0; i < 100000; i++)
{
;// delay
}
System.out.println("Main Thread Do something else.");
}
} class Worker implements Runnable
{
private final CountDownLatch mStartSignal;
private final CountDownLatch mDoneSignal;
private final int mThreadIndex; Worker(final CountDownLatch startSignal, final CountDownLatch doneSignal,
final int threadIndex)
{
this.mDoneSignal = doneSignal;
this.mStartSignal = startSignal;
this.mThreadIndex = threadIndex;
} @Override
public void run()
{
// TODO Auto-generated method stub
try
{
mStartSignal.await();// 阻塞，等待mStartSignal计数为0运行后面的代码
// 所有的工作线程都在等待同一个启动的命令
doWork();// 具体操作
System.out.println("Thread " + mThreadIndex + " Done Now:"
+ System.currentTimeMillis());
mDoneSignal.countDown();// 完成以后计数减一
}
catch (InterruptedException e)
{
// TODO Auto-generated catch block
e.printStackTrace();
}
} public void doWork()
{
for (int i = 0; i < 1000000; i++)
{
;// 耗时操作
}
System.out.println("Thread " + mThreadIndex + ":do work");
}
} public class CountDownLatchTest
{
public static void main(String[] args)
{
// TODO Auto-generated method stub
new Driver().main();
} }
通过Executor启动线程：
class CountDownLatchDriver2
{
private static final int TOTAL_THREADS = 10;
private final CountDownLatch mDoneSignal = new CountDownLatch(TOTAL_THREADS); void main()
{
System.out.println("Main Thread Now:" + System.currentTimeMillis());
doPrepareWork();// 准备工作 Executor executor = Executors.newFixedThreadPool(TOTAL_THREADS);
for (int i = 0; i < TOTAL_THREADS; i++)
{
// 通过内建的线程池维护创建的线程
executor.execute(new RunnableWorker(mDoneSignal, i));
}
doSomethingElse();// 做点自己的事情
try
{
mDoneSignal.await();// 等待所有工作线程结束
}
catch (InterruptedException e)
{
// TODO Auto-generated catch block
e.printStackTrace();
}
System.out.println("All workers have finished now.");
System.out.println("Main Thread Now:" + System.currentTimeMillis());
} void doPrepareWork()
{
System.out.println("Ready,GO!");
} void doSomethingElse()
{
for (int i = 0; i < 100000; i++)
{
;// delay
}
System.out.println("Main Thread Do something else.");
}
} class RunnableWorker implements Runnable
{ private final CountDownLatch mDoneSignal;
private final int mThreadIndex; RunnableWorker(final CountDownLatch doneSignal, final int threadIndex)
{
this.mDoneSignal = doneSignal;
this.mThreadIndex = threadIndex;
} @Override
public void run()
{
// TODO Auto-generated method stub doWork();// 具体操作
System.out.println("Thread " + mThreadIndex + " Done Now:"
+ System.currentTimeMillis());
mDoneSignal.countDown();// 完成以后计数减一
// 计数为0时，主线程接触阻塞，继续执行其他任务
try
{
// 可以继续做点其他的事情，与主线程无关了
Thread.sleep(5000);
System.out.println("Thread " + mThreadIndex
+ " Do something else after notifing main thread"); }
catch (InterruptedException e)
{
// TODO Auto-generated catch block
e.printStackTrace();
} } public void doWork()
{
for (int i = 0; i < 1000000; i++)
{
;// 耗时操作
}
System.out.println("Thread " + mThreadIndex + ":do work");
}
}
输出：
Main Thread Now:1359959480786
Ready,GO!
Thread 0:do work
Thread 0 Done Now:1359959480808
Thread 1:do work
Thread 1 Done Now:1359959480811
Thread 2:do work
Thread 2 Done Now:1359959480813
Main Thread Do something else.
Thread 3:do work
Thread 3 Done Now:1359959480825
Thread 5:do work
Thread 5 Done Now:1359959480827
Thread 7:do work
Thread 7 Done Now:1359959480829
Thread 9:do work
Thread 9 Done Now:1359959480831
Thread 4:do work
Thread 4 Done Now:1359959480833
Thread 6:do work
Thread 6 Done Now:1359959480835
Thread 8:do work
Thread 8 Done Now:1359959480837
All workers have finished now.
Main Thread Now:1359959480838
Thread 0 Do something else after notifing main thread
Thread 1 Do something else after notifing main thread
Thread 2 Do something else after notifing main thread
Thread 3 Do something else after notifing main thread
Thread 9 Do something else after notifing main thread
Thread 7 Do something else after notifing main thread
Thread 5 Do something else after notifing main thread
Thread 4 Do something else after notifing main thread
Thread 6 Do something else after notifing main thread
Thread 8 Do something else after notifing main thread
4. 使用CyclicBarrier。这里其实类似上面，这个berrier只是在等待完成后自动调用传入CyclicBarrier的Runnable。
例1
@Test
public void testThreadSync4() throws IOException {
final Vector<Integer> list = new Vector<Integer>();
Thread[] threads = new Thread[TEST_THREAD_COUNT];
final CyclicBarrier barrier = new CyclicBarrier(TEST_THREAD_COUNT,
new Runnable() {
public void run() {
printSortedResult(list);
}
});
for (int i = 0; i < TEST_THREAD_COUNT; i++) {
final int num = i;
threads[i] = new Thread(new Runnable() {
public void run() {
try {
Thread.sleep(random.nextInt(100));
} catch (InterruptedException e) {
e.printStackTrace();
}
list.add(num);
System.out.print(num + " add.\t");
try {
barrier.await();
} catch (InterruptedException e) {
e.printStackTrace();
} catch (BrokenBarrierException e) {
e.printStackTrace();
}
}
});
threads[i].start();
}
System.in.read();
}
例2
class WalkTarget
{
private final int mCount = 5;
private final CyclicBarrier mBarrier;
ExecutorService mExecutor; class BarrierAction implements Runnable
{
@Override
public void run()
{
// TODO Auto-generated method stub
System.out.println("所有线程都已经完成任务,计数达到预设值");
//mBarrier.reset();//恢复到初始化状态 }
} WalkTarget()
{
//初始化CyclicBarrier
mBarrier = new CyclicBarrier(mCount, new BarrierAction());
mExecutor = Executors.newFixedThreadPool(mCount); for (int i = 0; i < mCount; i++)
{
//启动工作线程
mExecutor.execute(new Walker(mBarrier, i));
}
}
} //工作线程
class Walker implements Runnable
{
private final CyclicBarrier mBarrier;
private final int mThreadIndex; Walker(final CyclicBarrier barrier, final int threadIndex) {
mBarrier = barrier;
mThreadIndex = threadIndex;
} @Override
public void run()
{
// TODO Auto-generated method stub
System.out.println("Thread " + mThreadIndex + " is running...");
// 执行任务
try
{
TimeUnit.MILLISECONDS.sleep(5000);
// do task
}
catch (InterruptedException e)
{
// TODO Auto-generated catch block
e.printStackTrace();
} // 完成任务以后，等待其他线程完成任务
try
{
mBarrier.await();
}
catch (InterruptedException e)
{
// TODO Auto-generated catch block
e.printStackTrace();
}
catch (BrokenBarrierException e)
{
// TODO Auto-generated catch block
e.printStackTrace();
}
// 其他线程任务都完成以后，阻塞解除，可以继续接下来的任务
System.out.println("Thread " + mThreadIndex + " do something else");
} } public class CountDownLatchTest
{
public static void main(String[] args)
{
// TODO Auto-generated method stub
//new CountDownLatchDriver2().main();
new WalkTarget();
} }
输出(注意，只有所有的线程barrier.await之后才能继续执行其他的操作）：
Thread 0 is running... Thread 2 is running... Thread 3 is running... Thread 1 is running... Thread 4 is running... 所有线程都已经完成任务,计数达到预设值 Thread 4 do something else Thread 0 do something else Thread 2 do something else Thread 3 do something else Thread 1 do something else5、
CountDownLatch和CyclicBarrier简单比较：
|
CountDownLatch |
CyclicBarrier |
|
|---|---|---|
|
软件包 |
java.util.concurrent |
java.util.concurrent |
|
适用情景 |
主线程等待多个工作线程结束 |
多个线程之间互相等待，直到所有线程达到一个障碍点(Barrier point) |
|
主要方法 |
CountDownLatch(int count) （主线程调用） 初始化计数 CountDownLatch.await （主线程调用） 阻塞，直到等待计数为0解除阻塞 CountDownLatch.countDown 计数减一（工作线程调用） |
CyclicBarrier(int parties, Runnable barrierAction) //初始化参与者数量和障碍点执行Action，Action可选。由主线程初始化 CyclicBarrier.await() //由参与者调用 阻塞，直到所有线程达到屏障点 |
|
等待结束 |
各线程之间不再互相影响，可以继续做自己的事情。不再执行下一个目标工作。 |
在屏障点达到后，允许所有线程继续执行，达到下一个目标。可以重复使用CyclicBarrier |
|
异常 |
如果其中一个线程由于中断，错误，或超时导致永久离开屏障点，其他线程也将抛出异常。 |
|
|
其他 |
如果BarrierAction不依赖于任何Party中的所有线程，那么在任何party中的一个线程被释放的时候，可以直接运行这个Action。 If(barrier.await()==2) { //do action } |
