今天小编给大家分享一下怎么优化Java线程池的相关知识点，内容详细，逻辑清晰，相信大部分人都还太了解这方面的知识，所以分享这篇文章给大家参考一下，希望大家阅读完这篇文章后有所收获，下面我们一起来了解一下吧。

创新互联建站主营平陆网站建设的网络公司,主营网站建设方案,重庆App定制开发,平陆h5小程序定制开发搭建,平陆网站营销推广欢迎平陆等地区企业咨询

升级版线程池的优化

1：新增了4种拒绝策略。分别为：MyAbortPolicy、MyDiscardPolicy、MyDiscardOldestPolicy、MyCallerRunsPolicy

2：对线程池MyThreadPoolExecutor的构造方法进行优化，增加了参数校验，防止乱传参数现象。

3：这是最重要的一个优化。

• 移除线程池的线程预热功能。因为线程预热会极大的耗费内存，当我们不用线程池时也会一直在运行状态。

• 换来的是在调用execute方法添加任务时通过检查workers线程集合目前的大小与corePoolSize的值去比较，再通过new MyWorker()去创建添加线程到线程池，这样好处就是当我们创建线程池如果不使用的话则对当前内存没有一点影响，当使用了才会创建线程并放入线程池中进行复用。

线程池构造器

public MyThreadPoolExecutor(){
        this(5,new ArrayBlockingQueue<>(10), Executors.defaultThreadFactory(),defaultHandle);
    }
    public MyThreadPoolExecutor(int corePoolSize, BlockingQueue waitingQueue,ThreadFactory threadFactory) {
        this(corePoolSize,waitingQueue,threadFactory,defaultHandle);
    }
    public MyThreadPoolExecutor(int corePoolSize, BlockingQueue waitingQueue,ThreadFactory threadFactory,MyRejectedExecutionHandle handle) {
        this.workers=new HashSet<>(corePoolSize);
        if(corePoolSize>=0&&waitingQueue!=null&&threadFactory!=null&&handle!=null){
            this.corePoolSize=corePoolSize;
            this.waitingQueue=waitingQueue;
            this.threadFactory=threadFactory;
            this.handle=handle;
        }else {
            throw new NullPointerException("线程池参数不合法");
        }
    }

线程池拒绝策略

策略接口：MyRejectedExecutionHandle

package com.springframework.concurrent;

/**
 * 自定义拒绝策略
 * @author 游政杰
 */
public interface MyRejectedExecutionHandle {

    void rejectedExecution(Runnable runnable,MyThreadPoolExecutor threadPoolExecutor);

}

策略内部实现类

/**
     * 实现自定义拒绝策略
     */
    //抛异常策略（默认）
    public static class MyAbortPolicy implements MyRejectedExecutionHandle{
        public MyAbortPolicy(){

        }
        @Override
        public void rejectedExecution(Runnable r, MyThreadPoolExecutor t) {
            throw new MyRejectedExecutionException("任务-> "+r.toString()+"被线程池-> "+t.toString()+" 拒绝");
        }
    }
    //默默丢弃策略
    public static class MyDiscardPolicy implements MyRejectedExecutionHandle{

        public MyDiscardPolicy() {
        }
        @Override
        public void rejectedExecution(Runnable runnable, MyThreadPoolExecutor threadPoolExecutor) {

        }
    }
    //丢弃掉最老的任务策略
    public static class MyDiscardOldestPolicy implements MyRejectedExecutionHandle{
        public MyDiscardOldestPolicy() {
        }
        @Override
        public void rejectedExecution(Runnable runnable, MyThreadPoolExecutor threadPoolExecutor) {
            if(!threadPoolExecutor.isShutdown()){ //如果线程池没被关闭
                threadPoolExecutor.getWaitingQueue().poll();//丢掉最老的任务，此时就有位置当新任务了
                threadPoolExecutor.execute(runnable); //把新任务加入到队列中
            }
        }
    }
    //由调用者调用策略
    public static class MyCallerRunsPolicy implements MyRejectedExecutionHandle{
        public MyCallerRunsPolicy(){

        }
        @Override
        public void rejectedExecution(Runnable runnable, MyThreadPoolExecutor threadPoolExecutor) {
            if(!threadPoolExecutor.isShutdown()){//判断线程池是否被关闭
                runnable.run();
            }
        }
    }

封装拒绝方法

protected final void reject(Runnable runnable){
        this.handle.rejectedExecution(runnable, this);
    }

    protected final void reject(Runnable runnable,MyThreadPoolExecutor threadPoolExecutor){
        this.handle.rejectedExecution(runnable, threadPoolExecutor);
    }

execute方法

@Override
    public boolean execute(Runnable runnable)
    {
        if (!this.waitingQueue.offer(runnable)) {
            this.reject(runnable);
            return false;
        }
        else {
            if(this.workers!=null&&this.workers.size()
可以看出只有当往线程池放任务时才会创建线程对象。
手写线程池源码
MyExecutorService
package com.springframework.concurrent;

import java.util.concurrent.BlockingQueue;

/**
 * 自定义线程池业务接口
 * @author 游政杰
 */
public interface MyExecutorService {

    boolean execute(Runnable runnable);

    void shutdown();

    void shutdownNow();

    boolean isShutdown();

    BlockingQueue getWaitingQueue();

}
MyRejectedExecutionException
package com.springframework.concurrent;

/**
 * 自定义拒绝异常
 */
public class MyRejectedExecutionException extends RuntimeException {

    public MyRejectedExecutionException() {
    }
    public MyRejectedExecutionException(String message) {
        super(message);
    }

    public MyRejectedExecutionException(String message, Throwable cause) {
        super(message, cause);
    }

    public MyRejectedExecutionException(Throwable cause) {
        super(cause);
    }

}
MyRejectedExecutionHandle
package com.springframework.concurrent;

/**
 * 自定义拒绝策略
 * @author 游政杰
 */
public interface MyRejectedExecutionHandle {

    void rejectedExecution(Runnable runnable,MyThreadPoolExecutor threadPoolExecutor);

}
核心类MyThreadPoolExecutor
package com.springframework.concurrent;

import java.util.HashSet;
import java.util.Set;
import java.util.concurrent.ArrayBlockingQueue;
import java.util.concurrent.BlockingQueue;
import java.util.concurrent.Executors;
import java.util.concurrent.ThreadFactory;
import java.util.concurrent.atomic.AtomicInteger;

/**
 * 纯手撸线程池框架
 * @author 游政杰
 */
public class MyThreadPoolExecutor implements MyExecutorService{

    private static final AtomicInteger taskcount=new AtomicInteger(0);//执行任务次数
    private static final AtomicInteger threadNumber=new AtomicInteger(0); //线程编号
    private static volatile int corePoolSize; //核心线程数
    private final HashSet workers; //工作线程
    private final BlockingQueue waitingQueue; //等待队列
    private static final String THREADPOOL_NAME="MyThread-Pool-";//线程名称
    private volatile boolean isRunning=true; //是否运行
    private volatile boolean STOPNOW=false; //是否立刻停止
    private volatile ThreadFactory threadFactory; //线程工厂
    private static final MyRejectedExecutionHandle defaultHandle=new MyThreadPoolExecutor.MyAbortPolicy();//默认拒绝策略
    private volatile MyRejectedExecutionHandle handle; //拒绝紫略

    public MyThreadPoolExecutor(){
        this(5,new ArrayBlockingQueue<>(10), Executors.defaultThreadFactory(),defaultHandle);
    }
    public MyThreadPoolExecutor(int corePoolSize, BlockingQueue waitingQueue,ThreadFactory threadFactory) {
        this(corePoolSize,waitingQueue,threadFactory,defaultHandle);
    }
    public MyThreadPoolExecutor(int corePoolSize, BlockingQueue waitingQueue,ThreadFactory threadFactory,MyRejectedExecutionHandle handle) {
        this.workers=new HashSet<>(corePoolSize);
        if(corePoolSize>=0&&waitingQueue!=null&&threadFactory!=null&&handle!=null){
            this.corePoolSize=corePoolSize;
            this.waitingQueue=waitingQueue;
            this.threadFactory=threadFactory;
            this.handle=handle;
        }else {
            throw new NullPointerException("线程池参数不合法");
        }
    }
    /**
     * 实现自定义拒绝策略
     */
    //抛异常策略（默认）
    public static class MyAbortPolicy implements MyRejectedExecutionHandle{
        public MyAbortPolicy(){

        }
        @Override
        public void rejectedExecution(Runnable r, MyThreadPoolExecutor t) {
            throw new MyRejectedExecutionException("任务-> "+r.toString()+"被线程池-> "+t.toString()+" 拒绝");
        }
    }
    //默默丢弃策略
    public static class MyDiscardPolicy implements MyRejectedExecutionHandle{

        public MyDiscardPolicy() {
        }
        @Override
        public void rejectedExecution(Runnable runnable, MyThreadPoolExecutor threadPoolExecutor) {

        }
    }
    //丢弃掉最老的任务策略
    public static class MyDiscardOldestPolicy implements MyRejectedExecutionHandle{
        public MyDiscardOldestPolicy() {
        }
        @Override
        public void rejectedExecution(Runnable runnable, MyThreadPoolExecutor threadPoolExecutor) {
            if(!threadPoolExecutor.isShutdown()){ //如果线程池没被关闭
                threadPoolExecutor.getWaitingQueue().poll();//丢掉最老的任务，此时就有位置当新任务了
                threadPoolExecutor.execute(runnable); //把新任务加入到队列中
            }
        }
    }
    //由调用者调用策略
    public static class MyCallerRunsPolicy implements MyRejectedExecutionHandle{
        public MyCallerRunsPolicy(){

        }
        @Override
        public void rejectedExecution(Runnable runnable, MyThreadPoolExecutor threadPoolExecutor) {
            if(!threadPoolExecutor.isShutdown()){//判断线程池是否被关闭
                runnable.run();
            }
        }
    }
    //call拒绝方法
    protected final void reject(Runnable runnable){
        this.handle.rejectedExecution(runnable, this);
    }

    protected final void reject(Runnable runnable,MyThreadPoolExecutor threadPoolExecutor){
        this.handle.rejectedExecution(runnable, threadPoolExecutor);
    }

    /**
     * MyWorker就是我们每一个线程对象
     */
    private final class MyWorker implements Runnable{

        final Thread thread; //为每个MyWorker

        MyWorker(){
            Thread td = threadFactory.newThread(this);
            td.setName(THREADPOOL_NAME+threadNumber.getAndIncrement());
            this.thread=td;
            this.thread.start();
            workers.add(this);
        }

        //执行任务
        @Override
        public void run() {
            //循环接收任务
                while (true)
                {
                    //循环退出条件：
                    //1：当isRunning为false并且waitingQueue的队列大小为0（也就是无任务了），会优雅的退出。
                    //2：当STOPNOW为true，则说明调用了shutdownNow方法进行暴力退出。
                    if((!isRunning&&waitingQueue.size()==0)||STOPNOW)
                    {
                        break;
                    }else {
                        //不断取任务，当任务！=null时则调用run方法处理任务
                        Runnable runnable = waitingQueue.poll();
                        if(runnable!=null){
                            runnable.run();
                            System.out.println("task==>"+taskcount.incrementAndGet());
                        }
                    }
                }
        }
    }

    //往线程池中放任务
    @Override
    public boolean execute(Runnable runnable)
    {
        if (!this.waitingQueue.offer(runnable)) {
            this.reject(runnable);
            return false;
        }
        else {
            if(this.workers!=null&&this.workers.size() getWaitingQueue() {
        return this.waitingQueue;
    }
}
线程池测试类
package com.springframework.test;

import com.springframework.concurrent.MyThreadPoolExecutor;

import java.util.concurrent.ArrayBlockingQueue;
import java.util.concurrent.Executors;

public class ThreadPoolTest {

  public static void main(String[] args) {


//      MyThreadPoolExecutor myThreadPoolExecutor = new MyThreadPoolExecutor
//              (5,new ArrayBlockingQueue<>(6), Executors.defaultThreadFactory(),new MyThreadPoolExecutor.MyAbortPolicy());

//      MyThreadPoolExecutor myThreadPoolExecutor = new MyThreadPoolExecutor
//              (5,new ArrayBlockingQueue<>(6), Executors.defaultThreadFactory(),new MyThreadPoolExecutor.MyDiscardPolicy());

//      MyThreadPoolExecutor myThreadPoolExecutor = new MyThreadPoolExecutor
//              (5,new ArrayBlockingQueue<>(6), Executors.defaultThreadFactory(),new MyThreadPoolExecutor.MyDiscardOldestPolicy());

      MyThreadPoolExecutor myThreadPoolExecutor = new MyThreadPoolExecutor
              (5,new ArrayBlockingQueue<>(6), Executors.defaultThreadFactory(),new MyThreadPoolExecutor.MyCallerRunsPolicy());


      for(int i=0;i<11;i++){

          int finalI = i;
          myThreadPoolExecutor.execute(()->{
              System.out.println(Thread.currentThread().getName()+">>>>"+ finalI);
          });

      }

      myThreadPoolExecutor.shutdown();

//      myThreadPoolExecutor.shutdownNow();




  }
}
好了升级版线程池就优化到这了，后面可能还会出完善版，不断进行优化。
以上就是“怎么优化Java线程池”这篇文章的所有内容，感谢各位的阅读！相信大家阅读完这篇文章都有很大的收获，小编每天都会为大家更新不同的知识，如果还想学习更多的知识，请关注创新互联行业资讯频道。
            
            
                            

                分享名称：怎么优化Java线程池                

                转载来源：http://scyanting.com/article/pgiies.html