(Java 基础知识) Java线程池

淩亂°似流年 2023-10-10 10:48 2阅读 0赞

**ExecutorService 建立多线程的步骤：**

<table style="width:100%"> 
 <tbody> 
  <tr> 
   <td>1。定义线程类</td> 
   <td>class Handler implements Runnable{ <br> }</td> 
  </tr> 
  <tr> 
   <td>2。建立ExecutorService线程池</td> 
   <td>ExecutorService executorService = Executors.newCachedThreadPool();<br> <br> 或者<br> <br> int cpuNums = Runtime.getRuntime().availableProcessors();<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; //获取当前系统的CPU 数目<br> ExecutorService executorService =Executors.newFixedThreadPool(cpuNums * POOL_SIZE);<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; //ExecutorService通常根据系统资源情况灵活定义线程池大小</td> 
  </tr> 
  <tr> 
   <td>3。调用线程池操作</td> 
   <td>循环操作，成为daemon,把新实例放入Executor池中<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp; while(true){ <br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; executorService.execute(new Handler(socket));&nbsp;<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; // class Handler implements Runnable{ <br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 或者<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; executorService.execute(createTask(i));<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; //private static Runnable createTask(final int taskID)<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp; }<br> <br> execute(Runnable对象)方法<br> 其实就是对Runnable对象调用start()方法<br> （当然还有一些其他后台动作，比如队列，优先级，IDLE timeout，active激活等）</td> 
  </tr> 
 </tbody> 
</table>

**几种不同的ExecutorService线程池对象**

<table style="width:100%"> 
 <tbody> 
  <tr> 
   <td>1.newCachedThreadPool()&nbsp;</td> 
   <td>-缓存型池子，先查看池中有没有以前建立的线程，如果有，就reuse.如果没有，就建一个新的线程加入池中<br> -缓存型池子通常用于执行一些<span style="font-weight:bold">生存期很短的异步型任务</span><br> &nbsp;因此在一些面向连接的daemon型SERVER中用得不多。<br> -能reuse的线程，必须是timeout IDLE内的池中线程，<span style="font-weight:bold">缺省timeout是60s</span>,超过这个IDLE时长，线程实例将被终止及移出池。<br> &nbsp; 注意，放入CachedThreadPool的线程不必担心其结束，<span style="font-weight:bold">超过TIMEOUT不活动，其会自动被终止。</span></td> 
  </tr> 
  <tr> 
   <td>2. newFixedThreadPool</td> 
   <td>-newFixedThreadPool与cacheThreadPool差不多，也是能reuse就用，但不能随时建新的线程<br> -其独特之处:任意时间点，最多只能有固定数目的活动线程存在，<span style="font-weight:bold">此时如果有新的线程要建立，只能放在另外的队列中等待</span>，<span style="font-weight:bold">直到当前的线程中某个线程终止直接被移出池子</span><br> -和cacheThreadPool不同，FixedThreadPool没有IDLE机制（可能也有，但既然文档没提，肯定非常长，类似依赖上层的TCP或UDP IDLE机制之类的），所以<span style="font-weight:bold">FixedThreadPool多数针对一些很稳定很固定的正规并发线程，多用于服务器</span><br> -<span style="color:#ff0000">从方法的源代码看，cache池和fixed 池调用的是同一个底层池，只不过参数不同:</span><br style="color:rgb(255,0,0)"> <span style="color:#ff0000; font-weight:bold">fixed池线程数固定，并且是0秒IDLE（无IDLE）</span><br style="color:rgb(255,0,0)"> <span style="font-weight:bold"><span style="color:#ff0000">cache池线程数支持0-Integer.MAX_VALUE(显然完全没考虑主机的资源承受能力），60秒IDLE</span>&nbsp;</span>&nbsp;</td> 
  </tr> 
  <tr> 
   <td>3.ScheduledThreadPool</td> 
   <td>-调度型线程池<br> -这个池子里的线程可以按schedule依次delay执行，或周期执行</td> 
  </tr> 
  <tr> 
   <td>4.SingleThreadExecutor</td> 
   <td>-单例线程，任意时间池中只能有一个线程<br> -<span style="font-weight:bold">用的是和cache池和fixed池相同的底层池，但线程数目是1-1,0秒IDLE（无IDLE）</span></td> 
  </tr> 
 </tbody> 
</table>

**上面四种线程池，都使用Executor的缺省线程工厂建立线程，也可单独定义自己的线程工厂**  
下面是缺省线程工厂代码:

<table style="width:100%"> 
 <tbody> 
  <tr> 
   <td>&nbsp;&nbsp;&nbsp; static class DefaultThreadFactory implements ThreadFactory { <br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; static final AtomicInteger poolNumber = new AtomicInteger(1);<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; final ThreadGroup group;<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; final AtomicInteger threadNumber = new AtomicInteger(1);<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; final String namePrefix;<br> <br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; DefaultThreadFactory() { <br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; SecurityManager s = System.getSecurityManager();<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; group = (s != null)? s.getThreadGroup() :Thread.currentThread().getThreadGroup();<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; namePrefix = "pool-" + poolNumber.getAndIncrement() + "-thread-";<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; }<br> <br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; public Thread newThread(Runnable r) { <br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Thread t = new Thread(group, r,namePrefix + threadNumber.getAndIncrement(),0);<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; if (t.isDaemon())<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; t.setDaemon(false);<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; if (t.getPriority() != Thread.NORM_PRIORITY)<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; t.setPriority(Thread.NORM_PRIORITY);<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; return t;<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; }<br> &nbsp;&nbsp;&nbsp; }</td> 
  </tr> 
 </tbody> 
</table>

也可自己定义ThreadFactory，加入建立池的参数中

<table style="width:100%"> 
 <tbody> 
  <tr> 
   <td>&nbsp;public static ExecutorService newCachedThreadPool<span style="font-weight:bold">(ThreadFactory threadFactory</span>) { </td> 
  </tr> 
 </tbody> 
</table>

**Executor的execute()方法**  
execute() 方法将Runnable实例加入pool中,并进行一些pool size计算和优先级处理  
execute() 方法本身在Executor接口中定义,有多个实现类都定义了不同的execute()方法  
如ThreadPoolExecutor类（cache,fiexed,single三种池子都是调用它）的execute方法如下：

<table style="width:100%"> 
 <tbody> 
  <tr> 
   <td>&nbsp;&nbsp;&nbsp; public void execute(Runnable command) { <br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; if (command == null)<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; throw new NullPointerException();<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; if (poolSize &gt;= corePoolSize || !addIfUnderCorePoolSize(command)) { <br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; if (runState == RUNNING &amp;&amp; workQueue.offer(command)) { <br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; if (runState != RUNNING || poolSize == 0)<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ensureQueuedTaskHandled(command);<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; }<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; else if (!addIfUnderMaximumPoolSize(command))<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; reject(command); // is shutdown or saturated<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; }<br> &nbsp;&nbsp;&nbsp; }</td> 
  </tr> 
 </tbody> 
</table>

*  java.util.concurrent.[**Executor**][Executor]
    
     *  java.util.concurrent.[**ExecutorService**][ExecutorService]
        
         *  java.util.concurrent.[**ScheduledExecutorService**][ScheduledExecutorService]

[Executor]: http://blog.csdn.net/java/util/concurrent/Executor.html
[ExecutorService]: http://blog.csdn.net/java/util/concurrent/ExecutorService.html
[ScheduledExecutorService]: http://blog.csdn.net/java/util/concurrent/ScheduledExecutorService.html