继续浏览精彩内容
慕课网APP
程序员的梦工厂
打开
继续
感谢您的支持,我会继续努力的
赞赏金额会直接到老师账户
将二维码发送给自己后长按识别
微信支付
支付宝支付

worker模式

linux快速入门培训
关注TA
已关注
手记 266
粉丝 31
获赞 120

今天来学学,大家也好对线程池有一个更好的理解。

public class Main {          public static void main(String[] args) {                  Channel channel = new Channel(5);   // 工人线程的數量,即线程池内的线程数目                   channel.startWorkers();//启动线程池内的线程                   new ClientThread("Alice", channel).start();//发送请求的线程,相当于向队列加入请求                   new ClientThread("Bobby", channel).start();                  new ClientThread("Chris", channel).start();          }  }

发送请求的client代码:

public class ClientThread extends Thread {          private final Channel channel;//相当于线程池            private static final Random random = new Random();              public ClientThread(String name, Channel channel) {                  super(name);                  this.channel = channel;          }              public void run() {                  try {                          int i = 0;                          Request request = new Request(getName(), i);//生成请求                           channel.putRequest(request);//向队列中放入请求,也即把请求传给线程池                           Thread.sleep(random.nextInt(1000));                  } catch (InterruptedException e) {                  }          }  }

ClientThread建立请求,并把请求传给了channel,下面来看看channel类(相当于线程池类)

public class Channel {          private static final int MAX_REQUEST = 100;          private final Request[] requestQueue;//存放请求的队列           private int tail;  // 下一个putRequest的地方          private int head;  // 下一个takeRequest的地方          private int count; // Request的数量               private final WorkerThread[] threadPool;              public Channel(int threads) {                  this.requestQueue = new Request[MAX_REQUEST];                  this.head = 0;                  this.tail = 0;                  this.count = 0;                      threadPool = new WorkerThread[threads];                  for (int i = 0; i < threadPool.length; i++) {                          threadPool[i] = new WorkerThread("Worker-" + i, this);//生成线程池中的线程                   }          }          public void startWorkers() {                  for (int i = 0; i < threadPool.length; i++) {                          threadPool[i].start();//启动线程池中的线程                   }          }          public synchronized void putRequest(Request request) {//向队列中存入请求                   while (count >= requestQueue.length) {                          try {                                  wait();                          } catch (InterruptedException e) {                          }                  }                  requestQueue[tail] = request;                  tail = (tail + 1) % requestQueue.length;                  count++;                  notifyAll();          }          public synchronized Request takeRequest() {//从队列取出请求                   while (count <= 0) {                          try {                                  wait();                          } catch (InterruptedException e) {                          }                  }                  Request request = requestQueue[head];                  head = (head + 1) % requestQueue.length;                  count--;                  notifyAll();                  return request;          }  }

channel类把传给他的请求放入队列中,等待worker去取请求,下面看看worker(即工作线程,线程池中已经初始话好的线程)

public class WorkerThread extends Thread {          private final Channel channel;          public WorkerThread(String name, Channel channel) {                  super(name);                  this.channel = channel;          }          public void run() {                  while (true) {                          Request request = channel.takeRequest();//取出请求                           request.execute();//处理请求                   }          }  }

在工作线程中会从线程池的队列里取出请求,并对请求进行处理。这里的workerthread相当于背景线程,他一直都在运行,当有请求的时候,他就会进行处理,这里处理请求的线程是已经存在在channel(线程池里的线程),他不会因为请求的增加而增加(这是本例中的情况),不会来一个请求就新建立一个线程,节省了资源。

再看看请求的代码:

public class Request {          private final String name; //  委托者          private final int number;  // 请求编号           private static final Random random = new Random();          public Request(String name, int number) {                  this.name = name;                  this.number = number;          }          public void execute() {//执行请求                   System.out.println(Thread.currentThread().getName() + " executes " + this);                  try {                          Thread.sleep(random.nextInt(1000));                  } catch (InterruptedException e) {                  }          }          public String toString() {                  return "[ Request from " + name + " No." + number + " ]";          }  }

JAVA SDK所写的 ExecutorService,其就相当于channel,即线程池。至于其实现当然要比channel复杂多了,channel只是举个例子。而WorkerThread可不是工作线程,他相当于发送到channel的请求,也就是request,当执行代码:tpes.execute(workers[i]);时,相当于向线程池加入一个请求,而WorkerThread中的run则相当于request中的execute,这也是当执行tpes.execute(workers[i]);时,并不会产生新的线程的原因。ExecutorService中产生的背景线程(相当于本篇的WorkerThread )我们是看不到的。


打开App,阅读手记
0人推荐
发表评论
随时随地看视频慕课网APP