单例实现方式一，锁机制

public class Singleton {
    private static Singleton singleton=null;

    public Singleton() {
    }
    public static Singleton getIntance(){
        if(singleton==null){
            synchronized (Singleton.class){
                if(singleton==null){
                    singleton=new Singleton();
                }
            }
        }
        return singleton;
    }
}
单例实现方式二，静态内部类

public class Singleton {
    private static class SingletonHolder {
        private static final Singleton singleton = new Singleton();

    }

    public Singleton() {
    }

    public static Singleton getIntance() {
        return SingletonHolder.singleton;
    }
}
静态内部类虽然没有显示调用synchronized，但是借助ClassLoad的线程安全机制，隐式调用了synchronized

单例实现方式三，CAS操作

public class Singleton {
    private static final AtomicReference<Singleton> instance = new AtomicReference<>();

    public Singleton() {
    }

    public static Singleton getIntance() {
        for (; ; ) {
            Singleton singleton = instance.get();
            if (singleton == null) {
                singleton = new Singleton();
                instance.compareAndSet(null, singleton);
            }
            return singleton;
        }
    }

    public static class SingletonHold implements Runnable {

        @Override
        public void run() {
            System.out.println(Thread.currentThread().getName() + ":" + Singleton.getIntance());
        }

    }

    public static void main(String[] args) {
        SingletonHold singletonHold = new SingletonHold();
        Thread thread1 = new Thread(singletonHold);
        Thread thread2 = new Thread(singletonHold);
        Thread thread3 = new Thread(singletonHold);
        thread1.start();
        thread2.start();
        thread3.start();
    }
}
运行结果：

Thread-0:com.rongke.web.Singleton@7249381e
Thread-1:com.rongke.web.Singleton@7249381e
Thread-2:com.rongke.web.Singleton@7249381e

使用cas好处：不需要使用锁来实现线程安全，而是依赖底层硬件实现，减少了因为锁导致线程切换和阻塞的性能开销，可以支持较大的并行度

使用cas缺点：如果线程一直处于for无限循环中，对cpu性能影响很大，如果多个线程执行singleton=new Singleton()，会产出堆内存浪费。