用多线程实现一个数字的自增长到1000000,分别用无锁模式和锁模式来实现代码.
1.使用ReentrantLock.
package test; import java.util.concurrent.CountDownLatch;
import java.util.concurrent.locks.ReentrantLock; public class ThreadWithLock { private static final int THREAD_COUNT=3;
private static volatile int value= 0;
private static final ReentrantLock lock = new ReentrantLock();
private static final CountDownLatch startSignal = new CountDownLatch(1);
private static final CountDownLatch endSignal = new CountDownLatch(1); public static void main(String[] args) throws InterruptedException{ for(int i=0;i<=THREAD_COUNT;i++){
(new Counter()).start();
}
//start to record time.
long start = System.nanoTime();
//send the started signal to all threads.
startSignal.countDown();
//wait for anyone to finish the counting.
endSignal.await();
long end = System.nanoTime();
System.out.println(end-start);
} static class Counter extends Thread { @Override
public void run() {
try {
startSignal.await();
while(true){
try {
lock.lock();
if(value<1000000){
value++;
}
else {
break;
}
}
finally{
lock.unlock();
}
}
endSignal.countDown();
} catch (InterruptedException e) {
e.printStackTrace();
} }
}
}
使用CAS模式，这其实是一种乐观锁模式，它默认是没有竞争的，如果存在竞争，失败了让代码重试.
package test; import java.util.concurrent.CountDownLatch;
import java.util.concurrent.atomic.AtomicInteger; public class ThreadNoLock { private static final int THREAD_COUNT=3;
private static final AtomicInteger value= new AtomicInteger(0);
private static final CountDownLatch startSignal = new CountDownLatch(1);
private static final CountDownLatch endSignal = new CountDownLatch(1); public static void main(String[] args) throws InterruptedException{
for(int i=0;i<THREAD_COUNT;i++){
(new Counter()).start();
}
//start to record time.
long start = System.nanoTime();
//send the started signal to all threads.
startSignal.countDown();
//wait for anyone to finish the counting.
endSignal.await();
long end = System.nanoTime();
System.out.println(end-start);
} static class Counter extends Thread { @Override
public void run() {
try {
startSignal.await();
int current;
while(true){
if((current = value.get()) <1000000){
value.compareAndSet(current, current+1);
}
else{
break;
}
}
endSignal.countDown();
} catch (InterruptedException e) {
e.printStackTrace();
} }
}
}
把Thread_count分别设置为1000,300,30,3 得到的性能数据如下,我们可以发现无锁模式的效率更高.