并发编程式JAVA的一个相对较复杂的模块。纵观各招聘网站的要求,纷纷要求要具有高并发,高性能开发经验。这一块的知识,不得不学啊。学习并发有断断续续,有很长一段时间了,相关并发书籍也看了很多。感受有些收获,特开系列文章,分享下本身的学习收获。html
线程的内容涉及的内容特别多,仅几篇文章不可能透彻阐明,我的又不喜欢弄个大块头文章,因此仍是作个系列文章吧。做为系列的第一篇,本章主要内容:
java
线程VS进程web
并行VS并发数据库
线程的建立和启动
编程
user thread VS daemon thread
tomcat
返回结果类型的线程安全
join操做多线程
进程有独立的地址空间,对应于系统的程序和软件。好比QQ,word java tomcat web应用,一个进程崩溃,不会影响其余进程。属于重量级的资源管理方式。进程切换时,耗费资源较大,效率要差一些。并发
线程ide
而线程只是一个进程中的不一样执行路径。一个进程,能够包含多个线程。线程有本身的堆栈和局部变量,而多个线程共享内存。线程不可以独立执行,必须依存在应用程序中,由应用程序提供多个线程执行控制。通常状况下main()做为应用程序的入口。属于轻量级别的资源管理方式。
2.并行VS并发
并发:一个处理器同时处理多个任务。经过CPU分片执行任务,属于逻辑上的同时发生。好比:一我的同时吃三个馒头。
并行:多个处理器或者是多核的处理器同时处理多个不一样的任务。物理上的同时发生。好比:三我的同时吃三个馒头。
并行的前提:具有多个处理器和多核的处理资源。
并发的前提:在资源CPU能力必定的基础上,区分I/O密集型任务,CPU密集型任务,尽可能减小CPU等待时间,压榨CPU的处理极限。
JAVA领域,涉及的都是并发。
3.线程的建立和启动
JAVA的线程有2中方式
1.实现 java.lang.Runnable接口
2.继承java.lang.Thread类
HeavyWorkRunnable.java,实现 java.lang.Runnable接口
package com.threadexample.base; public class HeavyWorkRunnable implements Runnable { @Override public void run() { System.out.println("Doing heavy processing - START "+Thread.currentThread().getName()); try { Thread.sleep(1000); //Get database connection, delete unused data from DB doDBProcessing(); } catch (InterruptedException e) { e.printStackTrace(); } System.out.println("Doing heavy processing - END "+Thread.currentThread().getName()); } private void doDBProcessing() throws InterruptedException { Thread.sleep(5000); } }
MyThread.java,继承java.lang.Thread类
package com.threadexample.base; public class MyThread extends Thread { public MyThread(String name) { super(name); } @Override public void run() { System.out.println("MyThread - START "+Thread.currentThread().getName()); try { Thread.sleep(1000); //Get database connection, delete unused data from DB doDBProcessing(); } catch (InterruptedException e) { e.printStackTrace(); } System.out.println("MyThread - END "+Thread.currentThread().getName()); } private void doDBProcessing() throws InterruptedException { Thread.sleep(5000); } }
测试代码ThreadRunExample
package com.threadexample.base; import java.util.concurrent.ExecutorService; import java.util.concurrent.Executors; public class ThreadRunExample { public static void main(String[] args){ Thread t1 = new Thread(new HeavyWorkRunnable(), "t1"); Thread t2 = new Thread(new HeavyWorkRunnable(), "t2"); System.out.println("Starting Runnable threads"); t1.start(); t2.start(); System.out.println("Runnable Threads has been started"); Thread t3 = new MyThread("t3"); Thread t4 = new MyThread("t4"); System.out.println("Starting MyThreads"); t3.start(); t4.start(); System.out.println("MyThreads has been started"); ExecutorService executorService = Executors.newCachedThreadPool(); Thread t5 = new Thread(new HeavyWorkRunnable(), "t5"); HeavyWorkRunnable t6 = new HeavyWorkRunnable(); executorService.submit(t5);//启动 executorService.execute(t6);//启动 executorService.shutdown(); } }
代码展现了2种建立线程的方式,以及多种启动方式。须要注意的是ExecutorService API,ExecutorService 的设计是为了尽量实现线程的复用,由于线程也会占用资源,线程复用能够下降资源占用,提高性能,是并发优化的一个途径,固然使用ThreadFactory的工厂,为线程管理提供了便利。
Runnable vs Thread
1.不只仅提供执行任务,还提供额外功能,实现 java.lang.Runnable接口
2.若是类的惟一功能是做为线程执行,能够继承java.lang.Thread
须要注意,thread 提供的方法没有返回值。
------------------------------
执行结果
Starting Runnable threads
Runnable Threads has been started
Starting MyThreads
MyThreads has been started
MyThread - START t3
Doing heavy processing - START t2
MyThread - START t4
Doing heavy processing - START t1
Doing heavy processing - START pool-1-thread-1
Doing heavy processing - START pool-1-thread-2
MyThread - END t3
Doing heavy processing - END t1
MyThread - END t4
Doing heavy processing - END t2
Doing heavy processing - END pool-1-thread-1
Doing heavy processing - END pool-1-thread-2
经过结果分析:线程实际执行结果,和程序编写顺序不一致。并且屡次运行执行的结果也不一样。这就是多线程,执行结果的不肯定性。
4.user thread VS daemon thread
对上面的ThreadRunExample ,进行DEBUG跟踪。
threads有些是本身建立的,有些不是本身建立的,如Finalizer线程,负责垃圾收集线程,是虚拟机自动建立的线程。
导出thread信息,部分信息截图
经过两个线程对比,Finalizer是daemon线程,优先级是8. 属于 “user thread"
咱们建立的线程不是daemon线程,优先级是5.属于"daemon thread".
所谓守护线程(daemon thread),是指在程序运行的时候在后台提供一种通用服务的线程,好比垃圾回收线程就是一个很称职的守护者,而且这种线程并不属于程序中不可或缺的部分。因 此,当全部的非守护线程结束时,程序也就终止了,同时会杀死进程中的全部守护线程。反过来讲,只要任何非守护线程还在运行,程序就不会终止。
用户线程和守护线程二者几乎没有区别,惟一的不一样之处就在于退出时机的差异:若是用户线程已经所有退出运行了,只剩下守护线程存在了,虚拟机也就退出了。 由于没有了被守护者,守护线程也就没有工做可作了,也就没有继续运行程序的必要了。
将线程转换为守护线程能够经过调用Thread对象的setDaemon(true)方法来实现。在使用守护线程时须要注意一下几点:
(1) thread.setDaemon(true)必须在thread.start()以前设置,不然会跑出一个IllegalThreadStateException异常。你不能把正在运行的常规线程设置为守护线程。
(2) 在Daemon线程中产生的新线程也是Daemon的。
(3) 守护线程应该永远不去访问固有资源,如文件、数据库,由于它会在任什么时候候甚至在一个操做的中间发生中断。
JavaDaemonThread.java代码
package com.threadexample.base; public class JavaDaemonThread { public static void main(String[] args) throws InterruptedException { Thread dt = new Thread(new DaemonThread(), "dt"); dt.setDaemon(true); dt.start(); //continue program Thread.sleep(30000); System.out.println("Finishing program"); } } class DaemonThread implements Runnable{ //此处是个死循环(若是不是daemon线程,他会一直执行下去,永不退出; //对于daemon线程,它会随user thread一并退出 @Override public void run() { while(true){ processSomething(); } } private void processSomething() { try { System.out.println("Processing daemon thread"); Thread.sleep(5000); } catch (InterruptedException e) { e.printStackTrace(); } } }
执行结果
Processing daemon thread
Processing daemon thread
Processing daemon thread
Processing daemon thread
Processing daemon thread
Processing daemon thread
Finishing program
5.返回结果类型的线程(Callable Future)
下面是Runnable 接口.执行方法没有返回值。
public interface Runnable { public abstract void run(); }
若是在面对执行计算任务是,须要有返回值怎么办?
也是能够的,之前面的例子稍加改造。
HeavyWorkRunnableWithResult.java
package com.threadexample.base; public class HeavyWorkRunnableWithResult implements Runnable { private String result; public String getResult(){ return result; } @Override public void run() { System.out.println("Doing heavy processing - START "+Thread.currentThread().getName()); try { Thread.sleep(1000); //Get database connection, delete unused data from DB doDBProcessing(); this.result="执行结果="+Thread.currentThread().getName(); } catch (InterruptedException e) { e.printStackTrace(); } System.out.println("Doing heavy processing - END "+Thread.currentThread().getName()); } private void doDBProcessing() throws InterruptedException { Thread.sleep(5000); } }
测试代码ThreadRunExampleWithResult.java
package com.threadexample.base; public class ThreadRunExampleWithResult { public static void main(String[] args){ System.out.println("Runnable Threads has been started"); HeavyWorkRunnableWithResult runnableWithResult = new HeavyWorkRunnableWithResult(); Thread t1 = new Thread(runnableWithResult,"t1"); t1.start(); System.out.println("输出结果:"+runnableWithResult.getResult()); } }
打印结果
------------------------
Runnable Threads has been started
输出结果:null
Doing heavy processing - START t1
Doing heavy processing - END t1
--------------------------
经测试结果不是咱们想要的结果,返回值为空。这是由于在打印输出结果时,线程尚未执行完。
怎么修复这个问题呢。
升级后的HeavyWorkRunnableWithResult.java,添加了个完成标记字段。
package com.threadexample.base; public class HeavyWorkRunnableWithResult implements Runnable { private String result; private volatile boolean done=false; public String getResult(){ return result; } public boolean isDone(){ return this.done; } @Override public void run() { System.out.println("Doing heavy processing - START "+Thread.currentThread().getName()); try { Thread.sleep(1000); //Get database connection, delete unused data from DB doDBProcessing(); this.result="执行结果="+Thread.currentThread().getName(); this.done=true; } catch (InterruptedException e) { e.printStackTrace(); } System.out.println("Doing heavy processing - END "+Thread.currentThread().getName()); } private void doDBProcessing() throws InterruptedException { Thread.sleep(5000); } }
注意必定要将done 设置为volatile 类型,不然main线程永远不会读取不到更新的值。关于volatile
,又涉及到底层的JMM模型了,不在此涉及了。
改造后的ThreadRunExampleWithResult.java,加个循环,保证有结果以后打打印。
package com.threadexample.base; public class ThreadRunExampleWithResult { public static void main(String[] args){ System.out.println("Runnable Threads has been started"); HeavyWorkRunnableWithResult runnableWithResult = new HeavyWorkRunnableWithResult(); Thread t1 = new Thread(runnableWithResult,"t1"); t1.start(); while(!runnableWithResult.isDone()){ //空循环 } System.out.println("输出结果:"+runnableWithResult.getResult()); } }
执行结果
-----------------------------
Runnable Threads has been started
Doing heavy processing - START t1
Doing heavy processing - END t1
输出结果:执行结果=t1
-----------------------------
仔细想一想,咱们作了好多无用的事,好比完成判断,循环等待。JAVA专为解决此类问题,提供了Callable接口。
public interface Callable<V> { /** * Computes a result, or throws an exception if unable to do so. * * @return computed result * @throws Exception if unable to compute a result */ V call() throws Exception; }
HeavyWorkCallable.java 做为HeavyWorkRunnableWithResult 的Callable版本
package com.threadexample.base; import java.util.concurrent.Callable; public class HeavyWorkCallable implements Callable<String> { @Override public String call() { System.out.println("Doing heavy processing - START "+Thread.currentThread().getName()); try { Thread.sleep(1000); //Get database connection, delete unused data from DB doDBProcessing(); return "执行结果="+Thread.currentThread().getName(); } catch (InterruptedException e) { e.printStackTrace(); } return null; } private void doDBProcessing() throws InterruptedException { Thread.sleep(5000); } }
ThreadRunExampleCallable.java 测试类
package com.threadexample.base; import java.util.ArrayList; import java.util.Date; import java.util.List; import java.util.concurrent.*; public class ThreadRunExampleCallable { public static void main(String[] args){ System.out.println("Runnable Threads has been started"); //Get ExecutorService from Executors utility class, thread pool size is 10 ExecutorService executor = Executors.newFixedThreadPool(5); //create a list to hold the Future object associated with Callable List<Future<String>> list = new ArrayList<Future<String>>(); //Create MyCallable instance Callable callable = new HeavyWorkCallable(); for(int i=0; i< 10; i++){ //submit Callable tasks to be executed by thread pool Future<String> future = executor.submit(callable); //add Future to the list, we can get return value using Future list.add(future); } for(Future<String> fut : list){ try { //print the return value of Future, notice the output delay in console // because Future.get() waits for task to get completed System.out.println(new Date()+ "::"+fut.get()); } catch (InterruptedException e) { e.printStackTrace(); } catch (ExecutionException e) { e.printStackTrace(); } } //shut down the executor service now executor.shutdown(); // System.out.println("输出结果:"+runnableWithResult.getResult()); } }
执行结果以下
-----------------------
Runnable Threads has been started
Doing heavy processing - START pool-1-thread-1
Doing heavy processing - START pool-1-thread-2
Doing heavy processing - START pool-1-thread-3
Doing heavy processing - START pool-1-thread-5
Doing heavy processing - START pool-1-thread-4
Doing heavy processing - START pool-1-thread-3
Sun Jun 26 17:41:51 CST 2016::执行结果=pool-1-thread-1
Doing heavy processing - START pool-1-thread-5
Sun Jun 26 17:41:57 CST 2016::执行结果=pool-1-thread-2
Doing heavy processing - START pool-1-thread-1
Sun Jun 26 17:41:57 CST 2016::执行结果=pool-1-thread-3
Doing heavy processing - START pool-1-thread-4
Doing heavy processing - START pool-1-thread-2
Sun Jun 26 17:41:57 CST 2016::执行结果=pool-1-thread-4
Sun Jun 26 17:41:57 CST 2016::执行结果=pool-1-thread-5
Sun Jun 26 17:41:57 CST 2016::执行结果=pool-1-thread-3
Sun Jun 26 17:42:03 CST 2016::执行结果=pool-1-thread-1
Sun Jun 26 17:42:03 CST 2016::执行结果=pool-1-thread-5
Sun Jun 26 17:42:03 CST 2016::执行结果=pool-1-thread-2
Sun Jun 26 17:42:03 CST 2016::执行结果=pool-1-thread-4
Process finished with exit code 0
-----------------------
6.join操做
Java Thread join方法能够用于直到指定线程死亡时中止当前线程。有三个重载方法
public final void join()
public final synchronized void join(long millis): 中止当前线程直到线程死时或者超过millis时间。
public final synchronized void join(long millis, int nanos)
咱们能够对5节中的ThreadRunExampleWithResult,进行改造下,达到想要的效果
package com.threadexample.base; public class ThreadRunExampleJoin { public static void main(String[] args){ System.out.println("Runnable Threads has been started"); HeavyWorkRunnableWithResult runnableWithResult = new HeavyWorkRunnableWithResult(); Thread t1 = new Thread(runnableWithResult,"t1"); t1.start(); // while(!runnableWithResult.isDone()){ // //空循环 // } try { t1.join(); } catch (InterruptedException e) { e.printStackTrace(); } System.out.println("输出结果:"+runnableWithResult.getResult()); } }
join方法通常使用在“直到....才”的场景。和自带的CountDownLatch工具类有重合的部分。
输出结果
----------
Runnable Threads has been started
Doing heavy processing - START t1
Doing heavy processing - END t1
输出结果:执行结果=t1
---------
引用资源
http://www.journaldev.com/1024/java-thread-join-example
http://www.cnblogs.com/luochengor/archive/2011/08/11/2134818.html
http://www.cnblogs.com/lmule/archive/2010/08/18/1802774.html
最后
本文主要讲了线程的几点用法,比较简单。
Runnable ->Thread->Daemon ->Callable-->Join。捎带着提了下violatile关键字。
因为线程间没有共享变量,因此不涉及到一丁点线程安全的东西。后面专门抽出几篇讲解下。
附件为代码