在Java7在。JDK它提供了多线程开发提供了一个非常强大的框架。这是Fork/Join框架。这是原来的Executors更多
进一步,在原来的基础上添加了并行分治计算中的一种Work-stealing策略。就是指的是。
当一个线程正在等待他创建的
子线程执行的时候,当前线程假设完毕了自己的任务后,就会寻找还没有被执行的任务而且执行他们,这样就是和
Executors这个方式最大的差别,更加有效的使用了线程的资源和功能。所以很推荐使用Fork/Join框架。
以下我们以一个样例来说明这个框架怎样使用。主要就是创建一个含有10000个资源的List,分别去改动他的内容。
package com.bird.concursey.charpet8;
/**
* store the name and price of a product
* @author bird 2014年10月7日 下午11:23:14
*/
public class Product {
private String name;
private double price;
public String getName() {
return name;
}
public void setName(String name) {
this.name = name;
}
public double getPrice() {
return price;
}
public void setPrice(double price) {
this.price = price;
}
}
package com.bird.concursey.charpet8;
import java.util.ArrayList;
import java.util.List;
/**
* generate a list of random products
* @author bird
* 2014年10月7日 下午11:24:47
*/
public class ProductListGenerator {
public List generate(int size) {
List list = new ArrayList();
for(int i = 0 ; i < size; i++) {
Product product = new Product();
product.setName("Product" + i);
product.setPrice(10);
list.add(product);
}
return list;
}
}
package com.bird.concursey.charpet8;
import java.util.List;
import java.util.concurrent.ForkJoinPool;
import java.util.concurrent.RecursiveAction;
import java.util.concurrent.TimeUnit;
public class Task extends RecursiveAction {
private static final long serialVersionUID = 1L;
// These attributes will determine the block of products this task has to
// process.
private List products;
private int first;
private int last;
// store the increment of the price of the products
private double increment;
public Task(List products, int first, int last, double increment) {
super();
this.products = products;
this.first = first;
this.last = last;
this.increment = increment;
}
/**
* If the difference between the last and first attributes is greater than
* or equal to 10, create two new Task objects, one to process the first
* half of products and the other to process the second half and execute
* them in ForkJoinPool using the invokeAll() method.
*/
@Override
protected void compute() {
if (last - first < 10) {
updatePrices();
} else {
int middle = (first + last) / 2;
System.out.printf("Task: Pending tasks:%s\n", getQueuedTaskCount());
Task t1 = new Task(products, first, middle + 1, increment);
Task t2 = new Task(products, middle + 1, last, increment);
invokeAll(t1, t2);
}
}
private void updatePrices() {
for (int i = first; i < last; i++) {
Product product = products.get(i);
product.setPrice(product.getPrice() * (1 + increment));
}
}
public static void main(String[] args) {
ProductListGenerator productListGenerator = new ProductListGenerator();
List products = productListGenerator.generate(10000);
Task task = new Task(products, 0, products.size(), 0.2);
ForkJoinPool pool = new ForkJoinPool();
pool.execute(task);
do {
System.out.printf("Main: Thread Count: %d\n",
pool.getActiveThreadCount());
System.out.printf("Main: Thread Steal: %d\n", pool.getStealCount());
System.out.printf("Main: Parallelism: %d\n", pool.getParallelism());
try {
TimeUnit.MILLISECONDS.sleep(5);
} catch (InterruptedException e) {
e.printStackTrace();
}
} while (!task.isDone());
pool.shutdown();
if(task.isCompletedNormally()) {
System.out.printf("Main: The process has completed normally.\n");
}
for(Product product : products) {
if(product.getPrice() != 12) {
System.out.printf("Product %s: %f\n",product.getName(),product.getPrice());
}
}
System.out.println("Main: End of the program.\n");
}
}
In this example, you have created a ForkJoinPool object and a subclass of the
ForkJoinTask class that you execute in the pool. To create the ForkJoinPool object,
you have used the constructor without arguments, so it will be executed with its default
configuration. It creates a pool with a number of threads equal to the number of processors
of the computer. When the ForkJoinPool object is created, those threads are created and
they wait in the pool until some tasks arrive for their execution.
Since the Task class doesn't return a result, it extends the RecursiveAction class. In the
recipe, you have used the recommended structure for the implementation of the task. If the
task has to update more than 10 products, it divides those set of elements into two blocks,
creates two tasks, and assigns a block to each task. You have used the first and last
attributes in the Task class to know the range of positions that this task has to update in the
list of products. You have used the first and last attributes to use only one copy of the
products list and not create different lists for each task.
To execute the subtasks that a task creates, it calls the invokeAll() method. This is a
synchronous call, and the task waits for the finalization of the subtasks before continuing
(potentially finishing) its execution. While the task is waiting for its subtasks, the worker thread
that was executing it takes another task that was waiting for execution and executes it. With
this behavior, the Fork/Join framework offers a more efficient task management than the
Runnable and Callable objects themselves.
The invokeAll() method of the ForkJoinTask class is one of the main differences
between the Executor and the Fork/Join framework. In the Executor framework, all the tasks
have to be sent to the executor, while in this case, the tasks include methods to execute and
control the tasks inside the pool. You have used the invokeAll() method in the Task class,
that extends the RecursiveAction class that extends the ForkJoinTask class.
You have sent a unique task to the pool to update all the list of products using the execute()
method. In this case, it's an asynchronous call, and the main thread continues its execution.
You have used some methods of the ForkJoinPool class to check the status and the
evolution of the tasks that are running. The class includes more methods that can be useful
for this purpose. See the Monitoring a Fork/Join pool recipe for a complete list of
those methods.
Finally, like with the Executor framework, you should finish ForkJoinPool using the
shutdown() method.
版权声明:本文博客原创文章,博客,未经同意,不得转载。