使用5个线程计算数组之和

前言

之前写过多线程累加计数，原理跟本篇类似，传送门

累加计数比计算数组之和逻辑稍微简单一点，如果对于这块不熟悉的，可以先看下累加计数。

基本思想已经在之前那篇文章里写过了，这里就直接贴代码了。

这里分别通过自己创建线程来实现功能，还有通过线程池来实现功能。思想都差不多。只是代码写法略有不同。仅供参考。

代码一：

五个线程交替累加计算数组之和，这种方法其实不如单线程直接累加快，因为交替累加需要前一个线程计算的结果。

package test;

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

public class FiveThreadCount {
	private int count=0;
	private int[] arr={1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28};
	private int j=0;
	//定义一个任务，关键点所在
	private class MyThread extends Thread{
		@Override
		public void run() {
			super.run();
				while(j<arr.length)
				{
					synchronized (MyThread.class) {
						if(j>=arr.length){
							return;
						}
						count+=arr[j++];
						try {
							Thread.sleep(100);
						} catch (InterruptedException e) {
							// TODO Auto-generated catch block
							e.printStackTrace();
						}
						System.out.println(Thread.currentThread().getName());
					}
				}
		}
	}

	//方法一
	public void test1(){
		for(int i=0;i<5;i++){
			new MyThread().start();
		}
        try {
			Thread.sleep(10000);
		} catch (InterruptedException e) {
			// TODO Auto-generated catch block
			e.printStackTrace();
		}
        System.out.println(count);
	}
	//方法二
	public void test2(){
		Thread myThread=new MyThread();
		for(int i=0;i<5;i++){
			new Thread(myThread).start();
		}
        try {
			Thread.sleep(10000);
		} catch (InterruptedException e) {
			// TODO Auto-generated catch block
			e.printStackTrace();
		}
        System.out.println(count);
	}
	//方法一的线程池实现版
	public void test3(){
		ExecutorService service=Executors.newCachedThreadPool();
		for(int i=0;i<5;i++){
			service.execute(new MyThread());
		}
        try {
			Thread.sleep(10000);
		} catch (InterruptedException e) {
			// TODO Auto-generated catch block
			e.printStackTrace();
		}
        System.out.println(count);
	}
	//方法二的线程池实现版
	public void test4(){
		ExecutorService service=Executors.newCachedThreadPool();
		Thread myThread=new MyThread();
		for(int i=0;i<5;i++){
			service.execute(myThread);
		}
        try {
			Thread.sleep(10000);
		} catch (InterruptedException e) {
			// TODO Auto-generated catch block
			e.printStackTrace();
		}
        System.out.println(count);
	}

}

上边代码中，用到了sleep方法的原因，sleep(100)是为了让其他线程有时间执行任务，如果不sleep的话，有可能一个线程就全部执行完了。 最后的sleep（10000）是为了等所有线程执行完后，打印最后的计算结果。 

代码二：

将数组分为5等分，让每个线程计算自己负责的那份，并发计算，最后汇总结果。这种方式比代码一速度会快些。因为线程独立计算，不依赖其他线程的结果。最后几个线程将总数累加即可。

方式一：

使用Callable，FutureTask方式，来实现代码：

package test;

import java.util.concurrent.Callable;
import java.util.concurrent.ExecutionException;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.FutureTask;

public class FiveThreadCount2 {
	private int[] arr={1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28};
	private int total=0;
	public void test() throws InterruptedException, ExecutionException{
		ExecutorService service=Executors.newFixedThreadPool(5);
		int length=arr.length;
		for(int j=0;j<length;j+=(length/5)){
			FutureTask<Integer> task;
			if( (j+(length/5))<length){
				task=new FutureTask<Integer>(new MyCallable(arr, j, j+(length/5)));
			}else{
				task=new FutureTask<Integer>(new MyCallable(arr, j, length));
			}
			service.execute(task);
			total+=task.get();
		}
		service.shutdown();
		System.out.println(total);

	}
	
	public class MyCallable implements Callable<Integer>{
		int[] arr;
		int startIndex;
		int endIndex;
		public MyCallable(int[] arr,int startIndex,int endIndex){
			this.arr=arr;
			this.startIndex=startIndex;
			this.endIndex=endIndex;
		}
		@Override
		public Integer call() throws Exception {
			int sum=0;
			for(int i=startIndex;i<endIndex;i++){
				sum+=arr[i];
			}
			System.out.println(Thread.currentThread().getName());
			return sum;
		}
	}
	
}

这个方式有一个缺点，看似5个线程异步执行，其实是顺序执行，因为 task.get是要等待线程执行完毕才会执行下边的代码。所以效率不会高,可能换种写法可以解决这个问题，这里就不深入研究。

方式二：

通过java工具类CountDownLatch实现并发计算

package test;

import java.util.concurrent.CountDownLatch;

public class FiveThreadCount3 {
	private int[] arr={1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28};
	private int total=0;
	public void test() throws InterruptedException{
		int length=arr.length;
		CountDownLatch latch=new CountDownLatch(length%5==0?5:6);
		System.out.println(length);
		for(int j=0;j<length;j+=(length/5)){
			MyThread task;
			if( (j+(length/5))<=length){
				task=new MyThread(arr, j, j+(length/5), latch);
			}else{
				task=new MyThread(arr, j, length, latch);
			}
			new Thread(task).start();
		}
		latch.await();
		System.out.println(total);
	}
	
	private class MyThread implements Runnable{
		int[] arr;
		int startIndex;
		int endIndex;
		CountDownLatch latch;
		public MyThread(int[] arr,int startIndex,int endIndex,CountDownLatch latch){
			this.arr=arr;
			this.startIndex=startIndex;
			this.endIndex=endIndex;
			this.latch=latch;
		}
		@Override
		public void run() {
			int sum=0;
			for(int i=startIndex;i<endIndex;i++){
				sum+=arr[i];
			}
			synchronized (MyThread.class) {
				total+=sum;
			}

			System.out.println(Thread.currentThread().getName());
			latch.countDown();

		}
		
	}
}

对于CountDownLatch不熟悉的可以搜索下用法。 

方式三：

通过java工具类 CyclicBarrier实现并发计算。

package test;

import java.util.concurrent.BrokenBarrierException;
import java.util.concurrent.CyclicBarrier;


public class FiveThreadCount1 {
	private int[] arr={1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28};
	private int total=0;
	public void test() throws InterruptedException, BrokenBarrierException{
		int length=arr.length;
		CyclicBarrier barrier=new CyclicBarrier((length%5==0?5:6)+1);
		System.out.println(length);
		for(int j=0;j<length;j+=(length/5)){
			MyThread task;
			if( (j+(length/5))<=length){
				task=new MyThread(arr, j, j+(length/5), barrier);
			}else{
				task=new MyThread(arr, j, length, barrier);
			}
			new Thread(task).start();
		}
		barrier.await();
		System.out.println(total);
	}
	
	private class MyThread implements Runnable{
		int[] arr;
		int startIndex;
		int endIndex;
		CyclicBarrier barrier;
		public MyThread(int[] arr,int startIndex,int endIndex,CyclicBarrier barrier){
			this.arr=arr;
			this.startIndex=startIndex;
			this.endIndex=endIndex;
			this.barrier=barrier;
		}
		@Override
		public void run() {
			int sum=0;
			for(int i=startIndex;i<endIndex;i++){
				sum+=arr[i];
			}
			synchronized (MyThread.class) {
				total+=sum;
			}
			
			try {
				System.out.println(Thread.currentThread().getName());
				barrier.await();
			} catch (InterruptedException e) {
				// TODO Auto-generated catch block
				e.printStackTrace();
			} catch (BrokenBarrierException e) {
				// TODO Auto-generated catch block
				e.printStackTrace();
			}
		}
		
	}
}

总结

总体来说，代码二的方式二、三，效率会高一点。以上代码都是通过main方法调用示例代码的test方法，输出结果到控制台。