class RunnableDemo implements Runnable {
private String threadName;
private RunnableDemo(String name) {
this.threadName = name;
System.out.println("creating thread:" + threadName);
}
@Override
public void run() {
System.out.println("Running " + threadName);
try {
for (int i = 0; i < 10; i++) {
System.out.println("Thread:" + threadName + "," + i);
Thread.sleep(50);
}
} catch (InterruptedException e) {
System.out.println("Thread " + threadName + "interrupter");
}
System.out.println("Thread " + threadName + " exiting");
}
// run
public static void main(String[] args) {
RunnableDemo r = new RunnableDemo("MyThread");
r.run();
}
}
复制代码
public class ThreadDemo extends Thread {
@Override
public void run() {
System.out.println("thread" + Thread.currentThread().getId() + " running...");
}
// run 10 thread
public static void main(String[] args) throws InterruptedException {
ThreadDemo[] threadDemos = new ThreadDemo[10];
for (int i = 0; i < threadDemos.length; i++) {
threadDemos[i] = new ThreadDemo();
}
for (ThreadDemo threadDemo : threadDemos) {
threadDemo.start();
}
// wait other thread complete
for (ThreadDemo threadDemo : threadDemos) {
threadDemo.join();
}
System.out.println("completing");
}
}
复制代码
经过FutureTask
包装一个Callable
的实例,再经过Thread
包装FutureTask
的实例,而后调用Thread
的start()
方法java
public class CallableDemo implements Callable {
@Override
public String call() throws Exception {
return "yo!";
}
@Test
public void callUse() throws Exception {
CallableDemo callableDemo = new CallableDemo();
System.out.println(callableDemo.call());
}
@Test
public void threadUse() throws ExecutionException, InterruptedException {
FutureTask futureTask= new FutureTask<>(new CallableDemo());
Thread thread=new Thread(futureTask);
thread.start();
System.out.println(futureTask.get());
}
}
复制代码
FutureTask继承关系git
通常经过ThreadPoolExecutor类来建立线程github
ThreadPoolExecutor poolExecutor = new ThreadPoolExecutor(
int corePoolSize,
int maximumPoolSize,
long keepAliveTime,
TimeUnit unit,
BlockingQueue workQueue,
ThreadFactory threadFactory,
RejectedExecutionHandler handler
)
复制代码
corePoolSize 线程池的基本大小编程
maximumPoolSize 线程池的最大大小多线程
keepAliveTime 空闲线程(超出基本大小的线程)的存活时间并发
unit 空闲线程存活时间的单位(毫秒,秒...)ide
workQueue 任务队列,提交的任务的阻塞队列(BlockingQueue
)。more: Java多线程 - 阻塞队列详解this
threadFactory 线程工产,线程的建立策略,有默认实现,能够经过定制线程工厂来监听线程信息spa
handler 饱和策略,当线程因为任务队列满了,或者某个任务被提交到一个已被关闭的线程的处理方式线程
AbortPolicy 停止策略,默认策略 ,该策略会抛出RejectExecutionException异常,调用者能够根据这个异常编写本身的处理代码
DiscardRunsPolicy 抛弃策略,悄悄抛弃该任务,不抛异常
DiscardOldestPolicy 抛弃最久任务策略 将工做队列中最老的(也就是下一个要执行的)任务抛弃。优先队列将会是优先级最高的
CallerRunsPolicy 调用者执行策略,将线程添加到添加工做队列的线程去执行
ps: 构造器参考下表
public static void main(String[] args) throws InterruptedException, ExecutionException {
ExecutorService pool=Executors.newFixedThreadPool(2);
pool.execute(() -> System.out.println("yo!"));
pool.shutdown();
}
复制代码
经过调用submit方法
在ExecutorService
中提供了重载的submit()
方法,该方法既能够接收Runnable
实例又能接收Callable
实例。对于实现Callable
接口的类,须要覆写call()
方法,而且只能经过ExecutorService
的submit()
方法来启动call()
方法
public static void main(String[] args) throws InterruptedException, ExecutionException {
ExecutorService pool=Executors.newFixedThreadPool(2);
Future future=pool.submit(() -> {
Thread.sleep(100);
return "yo!";
});
System.out.println(future.get());
pool.shutdown();
}
复制代码
定义:延时任务("在100ms后执行的任务") 周期任务("每10ms执行一次的任务")
使用:经过new ScheduledThreadPoolExector()对象
Demo:
public class ScheduleExecutorDemo implements Runnable {
private String name;
public ScheduleExecutorDemo(String name) {
this.name = name;
}
@Override
public void run() {
System.out.println(name + " 运行");
}
public static void main(String[] args) throws InterruptedException {
ScheduledExecutorService executorService1 = Executors.newScheduledThreadPool(2);
// after 10s run
executorService1.schedule(new ScheduleExecutorDemo("task1"), 10, TimeUnit.SECONDS);
executorService1.shutdown();
ScheduledExecutorService executorService2 = Executors.newScheduledThreadPool(2);
// run per 1s
executorService2.scheduleAtFixedRate(new ScheduleExecutorDemo("task1"),
0, 1, TimeUnit.SECONDS);
// run per 2s
executorService2.scheduleWithFixedDelay(new ScheduleExecutorDemo("task2"),
0, 2, TimeUnit.SECONDS);
}
}
复制代码
来源:阿里巴巴Java开发手册
线程资源必须经过线程池提供,不容许在应用中自行显式建立线程。
说明: 使用线程池的好处是减小在建立和销毁线程上所消耗的时间以及系统资源的开销,解决资源不足的问题。若是不使用线程池,有可能形成系统建立大量同类线程而致使消耗完内存或者“过分切换”的问题。
线程池不容许使用 Executors 去建立,而是经过 ThreadPoolExecutor 的方式,这样的处理方式让写的同窗更加明确线程池运行规则,规避资源耗尽的风险。
说明: Executors 返回的线程池对象的弊端以下:
FixedThreadPool 和 SingleThreadPool :容许的请求队列长度为 Integer.MAX_VALUE ,可能会堆积大量的请求,从而致使 OOM 。
CachedThreadPool 和 ScheduledThreadPool :容许的建立线程数量为 Integer.MAX_VALUE ,可能会建立大量的线程,从而致使 OOM 。
多线程并行处理定时任务时, Timer 运行多个 TimeTask 时,只要其中之一没有捕获抛出的异常,其它任务便会自动终止运行,使用 ScheduledExecutorService 则没有这个问题。
JDK1.8.0 源码