ThreadPoolExecutor corepoolsize 理解

 

在用jdk自带的jvm查看工具（bin/jvisualvm.exe） jvisualvm查看项目的运行状况的时候，发现Thread一只在增加（Thread dump），故进一步了解了一下ThreadPoolExecutor。

 

ThreadPoolExecutor 的 corePoolSize 的理解：

 

根据ThreadPoolExecutor的官方doc文档说明：When a new task is submitted in method execute(java.lang.Runnable), and fewer than corePoolSize threads are running, a new thread is created to handle the request, even if other worker threads are idle.

 

可以看出这个corePoolSize 的意思是说：当外部来了一个execute的请求，如果这时候正在运行状态的线程数 < corePoolSize 的时候，就会创建一个新的Thread去处理。但是需要注意的是，一个ThreadPoolExecutor 在创建之初，是不会立即初始化corePoolSize数量的Thread的，而是通过外部request来一个一个的创建，当达到corePoolSize数目之后，就会维持至少corePoolSize数目的Thread在pool中，哪怕他们都处于空闲状态（idle）.

 

通过例子更好理解一点，下面用ThreadPoolExecutor的子类ScheduledThreadPoolExecutor做例子：

 

 

import java.util.concurrent.ScheduledThreadPoolExecutor;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.atomic.AtomicInteger;

public class MyTest {
    
    private static AtomicInteger sequenceNumGenerator = new AtomicInteger();
    
    public static void main(String[] args) throws Exception {
        //pool with corePoolSize 10
        ScheduledThreadPoolExecutor pool = new ScheduledThreadPoolExecutor(10);
        
        //print the active thread and total thread num when pool created
        System.out.println(pool.getActiveCount());
        System.out.println(pool.getPoolSize());
        
        //execute 2 task
        pool.execute(getTask());
        pool.execute(getTask());
        
        //print the active thread and total thread num when tasks not finish
        System.out.println(pool.getActiveCount());
        System.out.println(pool.getPoolSize());
        
        //wait to ensure all tasks finish
        Thread.sleep(3000);
        
        //print the active thread and total thread num when all tasks finish
        System.out.println(pool.getActiveCount());
        System.out.println(pool.getPoolSize());
        
        //shutdown thread pool
        pool.shutdown();
    }
    
    private static Runnable getTask(){
        return new Runnable(){
            @Override
            public void run() {
                int sequenceNum = sequenceNumGenerator.incrementAndGet();
                System.out.println("begin task " + sequenceNum);
                try {
                    TimeUnit.SECONDS.sleep(1);
                } catch (InterruptedException e) {
                    e.printStackTrace();
                }
                System.out.println("finish task " + sequenceNum);
            }
        };
    }
}

 

 

 

 

 

根据结果可以看出：

 

ThreadPool在初始化的时候，active的thread 和 pool中的thread总数量都是0.也就是说没有初始化任何的thread

 

执行2个task，并且在task结束之前， active的thread 和 pool中的thread总数量都是2

 

在执行了2个task之后，并且在task都结束了之后， active的thread 和 pool中的thread总数量分别是0和2

 

这样就很清楚了，ThreadPoolExecutor的corePoolSize 数量的thread并不是在ThreadPoolExecutor创建的时候就理解初始化的，而是慢慢的通过外部的调用（例如调用execute）来一个一个的创建，最终达到corePoolSize 数量。

 

下面修改一下corePoolSize,让执行的task数量大于corePoolSize：

 

 

import java.util.concurrent.ScheduledThreadPoolExecutor;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.atomic.AtomicInteger;

public class MyTest {
    
    private static AtomicInteger sequenceNumGenerator = new AtomicInteger();
    
    public static void main(String[] args) throws Exception {
        //pool with corePoolSize 10
        ScheduledThreadPoolExecutor pool = new ScheduledThreadPoolExecutor(3);
        
        //print the active thread and total thread num when pool created
        System.out.println(pool.getActiveCount());
        System.out.println(pool.getPoolSize());
        
        //execute 2 task
        pool.execute(getTask());
        pool.execute(getTask());
        pool.execute(getTask());
        pool.execute(getTask());
        
        //print the active thread and total thread num when tasks not finish
        System.out.println(pool.getActiveCount());
        System.out.println(pool.getPoolSize());
        
        //wait to ensure all tasks finish
        Thread.sleep(4000);
        
        //print the active thread and total thread num when all tasks finish
        System.out.println(pool.getActiveCount());
        System.out.println(pool.getPoolSize());
        
        //shutdown thread pool
        pool.shutdown();
    }
    
    private static Runnable getTask(){
        return new Runnable(){
            @Override
            public void run() {
                int sequenceNum = sequenceNumGenerator.incrementAndGet();
                System.out.println("begin task " + sequenceNum);
                try {
                    TimeUnit.SECONDS.sleep(1);
                } catch (InterruptedException e) {
                    e.printStackTrace();
                }
                System.out.println("finish task " + sequenceNum);
            }
        };
    }
}

 

 

就会发现ThreadPoolExecutor在所以task完成之后，会维持corePoolSize数量的thread在threadpool中，哪怕这些task已经完成工作，处于空闲状态（idle）