Java 線程池詳解

更新時間：2016年03月26日 09:34:44 作者：趙杰A-124

本文給大家總結(jié)了java中的線程池的相關(guān)問題，非常的詳細(xì)也很實用，有需要的小伙伴可以參考下。

系統(tǒng)啟動一個線程的成本是比較高的，因為它涉及到與操作系統(tǒng)的交互，使用線程池的好處是提高性能，當(dāng)系統(tǒng)中包含大量并發(fā)的線程時，會導(dǎo)致系統(tǒng)性能劇烈下降，甚至導(dǎo)致JVM崩潰，而線程池的最大線程數(shù)參數(shù)可以控制系統(tǒng)中并發(fā)線程數(shù)不超過次數(shù)。

一、Executors 工廠類用來產(chǎn)生線程池，該工廠類包含以下幾個靜態(tài)工廠方法來創(chuàng)建對應(yīng)的線程池。創(chuàng)建的線程池是一個ExecutorService對象，使用該對象的submit方法或者是execute方法執(zhí)行相應(yīng)的Runnable或者是Callable任務(wù)。線程池本身在不再需要的時候調(diào)用shutdown()方法停止線程池，調(diào)用該方法后，該線程池將不再允許任務(wù)添加進(jìn)來，但是會直到已添加的所有任務(wù)執(zhí)行完成后才死亡。

1、newCachedThreadPool()，創(chuàng)建一個具有緩存功能的線程池，提交到該線程池的任務(wù)（Runnable或Callable對象）創(chuàng)建的線程，如果執(zhí)行完成，會被緩存到CachedThreadPool中，供后面需要執(zhí)行的任務(wù)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();
    
    //先添加三個任務(wù)到線程池
    for(int i = 0 ; i < 3; i++) {
      cacheThreadPool.execute(new Task());
    }
    
    //等三個線程執(zhí)行完成后，再次添加三個任務(wù)到線程池
    try {
      Thread.sleep(3000);
    } catch (InterruptedException e) {
      e.printStackTrace();
    }
    
    for(int i = 0 ; i < 3; i++) {
      cacheThreadPool.execute(new Task());
    }
  }

}

執(zhí)行結(jié)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

線程池中的線程對象進(jìn)行了緩存，當(dāng)有新任務(wù)執(zhí)行時進(jìn)行了復(fù)用。但是如果有特別多的并發(fā)時，緩存線程池還是會創(chuàng)建很多個線程對象。

2、newFixedThreadPool(int nThreads) 創(chuàng)建一個指定線程個數(shù)，線程可復(fù)用的線程池。

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);

    // 先添加三個任務(wù)到線程池
    for (int i = 0; i < 5; i++) {
      fixedThreadPool.execute(new Task());
    }

    // 等三個線程執(zhí)行完成后，再次添加三個任務(wù)到線程池
    try {
      Thread.sleep(3);
    } catch (InterruptedException e) {
      e.printStackTrace();
    }

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

}

執(zhí)行結(jié)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(),創(chuàng)建一個只有單線程的線程池，相當(dāng)于調(diào)用newFixedThreadPool(1)

4、newSheduledThreadPool(int corePoolSize),創(chuàng)建指定線程數(shù)的線程池，它可以在指定延遲后執(zhí)行線程。也可以以某一周期重復(fù)執(zhí)行某一線程，知道調(diào)用shutdown()關(guān)閉線程池。

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();
  }

}

運行結(jié)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

由運行時間可看出，任務(wù)是按照5秒的周期執(zhí)行的。

5、newSingleThreadScheduledExecutor() 創(chuàng)建一個只有一個線程的線程池，同調(diào)用newScheduledThreadPool(1)。

二、ForkJoinPool和ForkJoinTask

ForkJoinPool是ExecutorService的實現(xiàn)類，支持將一個任務(wù)劃分為多個小任務(wù)并行計算，在把多個小任務(wù)的計算結(jié)果合并成總的計算結(jié)果。它有兩個構(gòu)造函數(shù)

ForkJoinPool(int parallelism)創(chuàng)建一個包含parallelism個并行線程的ForkJoinPool。

ForkJoinPool(),以Runtime.availableProcessors()方法返回值作為parallelism參數(shù)來創(chuàng)建ForkJoinPool。

ForkJoinTask 代表一個可以并行，合并的任務(wù)。它是實現(xiàn)了Future<T>接口的抽象類，它有兩個抽象子類，代表無返回值任務(wù)的RecuriveAction和有返回值的RecursiveTask?？筛鶕?jù)具體需求繼承這兩個抽象類實現(xiàn)自己的對象，然后調(diào)用ForkJoinPool的submit 方法執(zhí)行。

RecuriveAction 示例如下，實現(xiàn)并行輸出0-300的數(shù)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();
  }

}

在拆分小任務(wù)后，調(diào)用任務(wù)的fork（）方法，加入到ForkJoinPool中并行執(zhí)行。

RecursiveTask示例，實現(xiàn)并行計算100個整數(shù)求和。拆分為每20個數(shù)求和后獲取結(jié)果，在最后合并為最后的結(jié)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();
  }

}

執(zhí)行結(jié)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

子任務(wù)執(zhí)行完后，調(diào)用任務(wù)的join（）方法獲取子任務(wù)執(zhí)行結(jié)果，再相加獲得最后的結(jié)果。

您可能感興趣的文章: