zoukankan      html  css  js  c++  java
  • SimpleThreadPool实践

    前言

    并发(Concurrency)一直谈论java绕不开的一个话题,从移动开发工程师到后端工程师,几乎所有的面试都要涉及到并发/多线程的一些问题。虽然多数时候我们使用线程池,都是已经实现好的框架——jdk7中就有现成的ThreadPoolExecutor供我们使用,不过,自己实现一个简化的线程池,对于帮助我们理解其内部原理还是有一些帮助的。

    设计

    核心思想如下:

    • 线程池实例为单例
    • 线程池实例中保存着一个线程数组,用来分发任务
    • 线程池中通过一个BlockingQueue实例,来实现FIFO的任务队列,这个实例同时被线程数组中的每一个线程拥有
    • 线程通过while循环,不断从队列中取出任务执行(Runnable

    实现

    首先是线程池示例SimpleThreadPool.java,使用单例模式。

    public class SimpleThreadPool {
        private static SimpleThreadPool ourInstance;
        private static final int QUEUE_SIZE = 100; // todo 调整 最优的capacity
        private PoolThread[] threadArray;
        private BlockingQueue<Runnable> blockingQueue = new ArrayBlockingQueue<>(QUEUE_SIZE);
        private boolean isStopped;
    
        public static SimpleThreadPool getInstance() {
            if (ourInstance == null) {
                ourInstance = new SimpleThreadPool();
            }
            return ourInstance;
        }
    
        public SimpleThreadPool initPoolSize(int size) {
            if (size < 1 || size > 10) {
                throw new RuntimeException("size must be 1~10!");
            }
            threadArray = new PoolThread[size];
            for (int i = 0; i < size; i++) {
                threadArray[i] = new PoolThread(blockingQueue);
                threadArray[i].start();
            }
            return this;
        }
    
        public synchronized SimpleThreadPool execute(Runnable runnable) {
            if (isStopped) {
                throw new IllegalStateException("Thread Pool is stopped!");
            }
            blockingQueue.offer(runnable);
            return this;
        }
    
        public synchronized void doStop() {
            for (PoolThread pt : threadArray) {
                pt.doStop();
            }
            isStopped = true;
        }
    
    }

    然后是自定义的线程池线程,PoolThread.java

    是否需要在run、doStop两个方法前声明synchronized,存疑。

    public class PoolThread extends Thread {
        private BlockingQueue<Runnable> blockingQueue;
        private boolean isStopped;
    
        public PoolThread(BlockingQueue<Runnable> blockingQueue) {
            this.blockingQueue = blockingQueue;
        }
    
        @Override
        public void run() {
            while (!isStopped) {
                try {
                    Runnable runnable = blockingQueue.take();
                    System.out.println(getName() + " is running...");
                    runnable.run();
                } catch (InterruptedException e) {
                    // todo: log it
                } catch (Exception e) {
                    throw new RuntimeException(e);
                }
            }
        }
    
        public void doStop() {
            interrupt();
            isStopped = true;
        }
    }

     写个Main函数测试一下。

    public class Main {
        public static void main(String[] args) {
            SimpleThreadPool.getInstance().initPoolSize(10).execute(new Runnable() {
                @Override
                public void run() {
                    System.out.println("1+1=" + (1 + 1));
                }
            }).execute(new Runnable() {
                @Override
                public void run() {
                    System.out.println("2+2=" + (2 + 2));
                }
            }).execute(new Runnable() {
                @Override
                public void run() {
                    System.out.println("3+3=" + (3 + 3));
                }
            }).execute(new Runnable() {
                @Override
                public void run() {
                    System.out.println("4+4=" + (4 + 4));
                }
            }).execute(new Runnable() {
                @Override
                public void run() {
                    System.out.println("5+5=" + (5 + 5));
                }
            })
            ;
            SimpleThreadPool.getInstance().doStop();
        }
    }

    示例输出:

    Thread-9 is running...
    1+1=2
    Thread-9 is running...
    2+2=4
    Thread-9 is running...
    3+3=6
    Thread-3 is running...
    4+4=8
    Thread-4 is running...
    5+5=10
  • 相关阅读:
    【C#进阶系列】29 混合线程同步构造
    【C#进阶系列】28 基元线程同步构造
    【C#进阶系列】27 I/O限制的异步操作
    【C#进阶系列】26 计算限制的异步操作
    快速幂算法
    纪中……结束了……
    洛谷P4526 【模板】自适应辛普森法2
    洛谷P4525 【模板】自适应辛普森法1与2
    洛谷P1056 排座椅
    纪中23日c组T2 2159. 【2017.7.11普及】max 洛谷P1249 最大乘积
  • 原文地址:https://www.cnblogs.com/maozhige/p/5084302.html
Copyright © 2011-2022 走看看