Java 线程池详解

系统启动一个线程的成本是比较高的，因为它涉及到与操作系统的交互，使用线程池的好处是提高性能，当系统中包含大量并发的线程时，会导致系统性能剧烈下降，甚至导致JVM崩溃，而线程池的最大线程数参数可以控制系统中并发线程数不超过次数。

一、Executors 工厂类用来产生线程池，该工厂类包含以下几个静态工厂方法来创建对应的线程池。创建的线程池是一个ExecutorService对象，使用该对象的submit方法或者是execute方法执行相应的Runnable或者是Callable任务。线程池本身在不再需要的时候调用shutdown()方法停止线程池，调用该方法后，该线程池将不再允许任务添加进来，但是会直到已添加的所有任务执行完成后才死亡。

1、newCachedThreadPool()，创建一个具有缓存功能的线程池，提交到该线程池的任务（Runnable或Callable对象）创建的线程，如果执行完成，会被缓存到CachedThreadPool中，供后面需要执行的任务使用。

import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;

public class CacheThreadPool {
  static class Task implements Runnable {
    @Override
    public void run() {
      System.out.println(this + " " + Thread.currentThread().getName() + " AllStackTraces map size: "
          + Thread.currentThread().getAllStackTraces().size());
    }
  }

  public static void main(String[] args) {
    ExecutorService cacheThreadPool = Executors.newCachedThreadPool();
    
    //先添加三个任务到线程池
    for(int i = 0 ; i < 3; i++) {
      cacheThreadPool.execute(new Task());
    }
    
    //等三个线程执行完成后，再次添加三个任务到线程池
    try {
      Thread.sleep(3000);
    } catch (InterruptedException e) {
      e.printStackTrace();
    }
    
    for(int i = 0 ; i < 3; i++) {
      cacheThreadPool.execute(new Task());
    }
  }

}

执行结果如下：

CacheThreadPool$Task@2d312eb9 pool-1-thread-1 AllStackTraces map size: 7
CacheThreadPool$Task@59522b86 pool-1-thread-3 AllStackTraces map size: 7
CacheThreadPool$Task@73dbb89f pool-1-thread-2 AllStackTraces map size: 7
CacheThreadPool$Task@5795cedc pool-1-thread-3 AllStackTraces map size: 7
CacheThreadPool$Task@256d5600 pool-1-thread-1 AllStackTraces map size: 7
CacheThreadPool$Task@7d1c5894 pool-1-thread-2 AllStackTraces map size: 7

线程池中的线程对象进行了缓存，当有新任务执行时进行了复用。但是如果有特别多的并发时，缓存线程池还是会创建很多个线程对象。

2、newFixedThreadPool(int nThreads) 创建一个指定线程个数，线程可复用的线程池。

import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;

public class FixedThreadPool {
  static class Task implements Runnable {
    @Override
    public void run() {
      System.out.println(this + " " + Thread.currentThread().getName() + " AllStackTraces map size: "
          + Thread.currentThread().getAllStackTraces().size());
    }
  }

  public static void main(String[] args) {
    ExecutorService fixedThreadPool = Executors.newFixedThreadPool(3);

    // 先添加三个任务到线程池
    for (int i = 0; i < 5; i++) {
      fixedThreadPool.execute(new Task());
    }

    // 等三个线程执行完成后，再次添加三个任务到线程池
    try {
      Thread.sleep(3);
    } catch (InterruptedException e) {
      e.printStackTrace();
    }

    for (int i = 0; i < 3; i++) {
      fixedThreadPool.execute(new Task());
    }
  }

}

执行结果：

FixedThreadPool$Task@7045c12d pool-1-thread-2 AllStackTraces map size: 7
FixedThreadPool$Task@50fa0bef pool-1-thread-2 AllStackTraces map size: 7
FixedThreadPool$Task@ccb1870 pool-1-thread-2 AllStackTraces map size: 7
FixedThreadPool$Task@7392b4e3 pool-1-thread-1 AllStackTraces map size: 7
FixedThreadPool$Task@5bdeff18 pool-1-thread-2 AllStackTraces map size: 7
FixedThreadPool$Task@7d5554e1 pool-1-thread-1 AllStackTraces map size: 7
FixedThreadPool$Task@24468092 pool-1-thread-3 AllStackTraces map size: 7
FixedThreadPool$Task@fa7b978 pool-1-thread-2 AllStackTraces map size: 7

3、newSingleThreadExecutor(),创建一个只有单线程的线程池，相当于调用newFixedThreadPool(1)

4、newSheduledThreadPool(int corePoolSize),创建指定线程数的线程池，它可以在指定延迟后执行线程。也可以以某一周期重复执行某一线程，知道调用shutdown()关闭线程池。

示例如下：

import java.util.concurrent.Executors;
import java.util.concurrent.ScheduledExecutorService;
import java.util.concurrent.TimeUnit;

public class ScheduledThreadPool {
  static class Task implements Runnable {
    @Override
    public void run() {
      System.out.println("time " + System.currentTimeMillis() + " " + Thread.currentThread().getName() + " AllStackTraces map size: "
          + Thread.currentThread().getAllStackTraces().size());
    }
  }

  public static void main(String[] args) {
    ScheduledExecutorService scheduledExecutorService = Executors.newScheduledThreadPool(3);
    
    scheduledExecutorService.schedule(new Task(), 3, TimeUnit.SECONDS);
    
    scheduledExecutorService.scheduleAtFixedRate(new Task(), 3, 5, TimeUnit.SECONDS);
  
    try {
      Thread.sleep(30 * 1000);
    } catch (InterruptedException e) {
      e.printStackTrace();
    }
    scheduledExecutorService.shutdown();
  }

}

运行结果如下：

time 1458921795240 pool-1-thread-1 AllStackTraces map size: 6
time 1458921795241 pool-1-thread-2 AllStackTraces map size: 6
time 1458921800240 pool-1-thread-1 AllStackTraces map size: 7
time 1458921805240 pool-1-thread-1 AllStackTraces map size: 7
time 1458921810240 pool-1-thread-1 AllStackTraces map size: 7
time 1458921815240 pool-1-thread-1 AllStackTraces map size: 7
time 1458921820240 pool-1-thread-1 AllStackTraces map size: 7

由运行时间可看出，任务是按照5秒的周期执行的。

5、newSingleThreadScheduledExecutor() 创建一个只有一个线程的线程池，同调用newScheduledThreadPool(1)。

二、ForkJoinPool和ForkJoinTask

ForkJoinPool是ExecutorService的实现类，支持将一个任务划分为多个小任务并行计算，在把多个小任务的计算结果合并成总的计算结果。它有两个构造函数

ForkJoinPool(int parallelism)创建一个包含parallelism个并行线程的ForkJoinPool。

ForkJoinPool(),以Runtime.availableProcessors()方法返回值作为parallelism参数来创建ForkJoinPool。

ForkJoinTask 代表一个可以并行，合并的任务。它是实现了Future<T>接口的抽象类，它有两个抽象子类，代表无返回值任务的RecuriveAction和有返回值的RecursiveTask。可根据具体需求继承这两个抽象类实现自己的对象，然后调用ForkJoinPool的submit 方法执行。

RecuriveAction 示例如下，实现并行输出0-300的数字。

import java.util.concurrent.ForkJoinPool;
import java.util.concurrent.RecursiveAction;
import java.util.concurrent.TimeUnit;

public class ActionForkJoinTask {
  static class PrintTask extends RecursiveAction {
    private static final int THRESHOLD = 50;
    private int start;
    private int end;

    public PrintTask(int start, int end) {
      this.start = start;
      this.end = end;
    }

    @Override
    protected void compute() {
      if (end - start < THRESHOLD) {
        for(int i = start; i < end; i++) {
          System.out.println(Thread.currentThread().getName() + " " + i);
        }
      } else {
        int middle = (start + end) / 2;
        PrintTask left = new PrintTask(start, middle);
        PrintTask right = new PrintTask(middle, end);
        left.fork();
        right.fork();
      }
    }

  }

  public static void main(String[] args) {
    ForkJoinPool pool = new ForkJoinPool();
    
    pool.submit(new PrintTask(0, 300));
    try {
      pool.awaitTermination(2, TimeUnit.SECONDS);
    } catch (InterruptedException e) {
      e.printStackTrace();
    }
    
    pool.shutdown();
  }

}

在拆分小任务后，调用任务的fork（）方法，加入到ForkJoinPool中并行执行。

RecursiveTask示例，实现并行计算100个整数求和。拆分为每20个数求和后获取结果，在最后合并为最后的结果。

import java.util.Random;
import java.util.concurrent.ExecutionException;
import java.util.concurrent.ForkJoinPool;
import java.util.concurrent.Future;
import java.util.concurrent.RecursiveTask;

public class TaskForkJoinTask {
  static class CalTask extends RecursiveTask<Integer> {
    private static final int THRESHOLD = 20;

    private int arr[];
    private int start;
    private int end;

    public CalTask(int[] arr, int start, int end) {
      this.arr = arr;
      this.start = start;
      this.end = end;
    }

    @Override
    protected Integer compute() {
      int sum = 0;

      if (end - start < THRESHOLD) {
        for (int i = start; i < end; i++) {
          sum += arr[i];
        }
        System.out.println(Thread.currentThread().getName() + " sum:" + sum);
        return sum;
      } else {
        int middle = (start + end) / 2;
        CalTask left = new CalTask(arr, start, middle);
        CalTask right = new CalTask(arr, middle, end);

        left.fork();
        right.fork();

        return left.join() + right.join();
      }
    }

  }

  public static void main(String[] args) {
    int arr[] = new int[100];
    Random random = new Random();
    int total = 0;

    for (int i = 0; i < arr.length; i++) {
      int tmp = random.nextInt(20);
      total += (arr[i] = tmp);
    }
    System.out.println("total " + total);

    ForkJoinPool pool = new ForkJoinPool(4);

    Future<Integer> future = pool.submit(new CalTask(arr, 0, arr.length));
    try {
      System.out.println("cal result: " + future.get());
    } catch (InterruptedException e) {
      e.printStackTrace();
    } catch (ExecutionException e) {
      e.printStackTrace();
    }
    pool.shutdown();
  }

}

执行结果如下：

total 912
ForkJoinPool-1-worker-2 sum:82
ForkJoinPool-1-worker-2 sum:123
ForkJoinPool-1-worker-2 sum:144
ForkJoinPool-1-worker-3 sum:119
ForkJoinPool-1-worker-2 sum:106
ForkJoinPool-1-worker-2 sum:128
ForkJoinPool-1-worker-2 sum:121
ForkJoinPool-1-worker-3 sum:89
cal result: 912

子任务执行完后，调用任务的join（）方法获取子任务执行结果，再相加获得最后的结果。