CyclicBarrier处理巨量的List（适用快速批量处理巨量数据 ）

import java.util.ArrayList;import java.util.List;import java.util.concurrent.BrokenBarrierException;import java.util.concurrent.CyclicBarrier;import java.util.concurrent.ExecutorService;import java.util.concurrent.Executors;public class LargeListIntegerSum { private long sum;//存放整数的和 private CyclicBarrier barrier;//障栅集合点(同步器) private List list;//整数集合List private int threadCounts;//使用的线程数 public LargeListIntegerSum(List list,int threadCounts) { this.list=list; this.threadCounts=threadCounts; } /** * 获取List中所有整数的和 * @return */ public long getIntegerSum(){ ExecutorService exec=Executors.newFixedThreadPool(threadCounts); int len=list.size()/threadCounts;//平均分割List //List中的数量没有线程数多（很少存在） if(len==0){ threadCounts=list.size();//采用一个线程处理List中的一个元素 len=list.size()/threadCounts;//重新平均分割List } barrier=new CyclicBarrier(threadCounts+1); for(int i=0;ilist.size()?list.size():len*(i+1)))); } } try { barrier.await();//关键，使该线程在障栅处等待，直到所有的线程都到达障栅处 } catch (InterruptedException e) { System.out.println(Thread.currentThread().getName()+":Interrupted"); } catch (BrokenBarrierException e) { System.out.println(Thread.currentThread().getName()+":BrokenBarrier"); } exec.shutdown(); return sum; } /** * 分割计算List整数和的线程任务 * */ public class SubIntegerSumTask implements Runnable{ private List subList; public SubIntegerSumTask(List subList) { this.subList=subList; } public void run() { long subSum=0L; for (Integer i : subList) { subSum += i; } synchronized(LargeListIntegerSum.this){//在LargeListIntegerSum对象上同步 sum+=subSum; } try { barrier.await();//关键，使该线程在障栅处等待，直到所有的线程都到达障栅处 } catch (InterruptedException e) { System.out.println(Thread.currentThread().getName()+":Interrupted"); } catch (BrokenBarrierException e) { System.out.println(Thread.currentThread().getName()+":BrokenBarrier"); } System.out.println("分配给线程："+Thread.currentThread().getName()+"那一部分List的整数和为：\tSubSum:"+subSum); } } public static void main(String[] args) { List list = new ArrayList(); int threadCounts = 10;//采用的线程数 for (int i = 1; i <= 1000000; i++) { list.add(i); } long start= System.currentTimeMillis(); LargeListIntegerSum countListIntegerSum=new LargeListIntegerSum(list,threadCounts); long sum=countListIntegerSum.getIntegerSum(); System.out.println("List中所有整数的和为:"+sum); long end= System.currentTimeMillis(); System.out.println(end-start); } }