Example analysis of two Java threads adding 1 to a variable

1--Wrong conventional writing

<code>public static int i=0;</code><code>public static void add(){</code><code>    i=i+1;</code><code>    action();</code><code>}</code><code>public static void action(){</code><code>    System.out.println("==>"+Thread.currentThread().getName()+":"+i);</code><code>}</code><code>public static void main(String[] args) throws InterruptedException {</code><code>    Thread t1 = new Thread(SysUserServiceImpl::add,"t1");</code><code>    Thread t2= new Thread(SysUserServiceImpl::add,"t2");</code><code>    t1.start();</code><code>    t2.start();</code><code>}</code><code>运行结果==></code><code>==>t1:1</code><code>==>t2:2</code><code><br></code><code>==>t1:2</code><code>==>t2:1</code><code><br></code><code>==>t1:2</code><code>==>t2:2</code>

The results are inconsistent each time. In a multi-threaded environment, t1 performs a 1 operation on i in the shared memory, but does not refresh the value to the main memory. At this time, t2 happens to also perform a 1 operation on i or 0. Operation, so that the final result i is all 1. In the same way, t1 is 1 after processing, and t2 is 2 after processing. The results of multiple runs are inconsistent.

Improvement method 1 --Synchronization lock

<code>public class ThreadException {</code><code>    public static volatile int i=0;</code><code>    public static void add(){</code><code>        synchronized (ThreadException.class){</code><code>            i=i+1;</code><code>            action();</code><code>        }</code><code>    }</code><code>    public static void action(){</code><code>        System.out.println("==>"+Thread.currentThread().getName()+":"+i);</code><code>    }</code><code>    public static void main(String[] args) throws InterruptedException {</code><code>        Thread t1 = new Thread(ThreadException::add,"t1");</code><code>        Thread t2= new Thread(ThreadException::add,"t2");</code><code>        t1.start();</code><code>        t2.start();</code><code><br></code><code>    }</code><code>}</code>

Advantages: simple implementation

Disadvantages: large locking granularity, low performance, distributed environment, multiple JVM conditions, synchronized failure, synchronized is only a local lock, and the lock is only the current one For objects under jvm, in distributed scenarios, distributed locks should be used

Improvement method 2 AtomicInteger

public class ThreadException {    private static AtomicInteger num = new AtomicInteger(0);    public static void add(){        int i = num.getAndIncrement();        action(i);    }    public static void action(int i){        System.out.println("由"+i+"==>"+Thread.currentThread().getName()+":"+num);    }    public static void main(String[] args) throws InterruptedException {        Thread t1 = new Thread(ThreadException::add,"t1");        Thread t2= new Thread(ThreadException::add,"t2");        t1.start();        t2.start();    }}

Improvement method 3 lock

<code>public class ThreadException {</code><code>    public static volatile int i=0;</code><code>    public static void action(){</code><code>        System.out.println("==>"+Thread.currentThread().getName()+":"+i);</code><code>    }</code><code><br></code><code>    static Lock lock=new ReentrantLock();</code><code>    public static void inc() {</code><code>        lock.lock();</code><code>        try {</code><code>            Thread.sleep(1);</code><code>            i=i+1;</code><code>            action();</code><code>        } catch (InterruptedException e) {</code><code>            e.printStackTrace();</code><code>        } finally {</code><code>            lock.unlock();</code><code>        }</code><code>    }</code><code>    public static void main(String[] args) throws InterruptedException {</code><code>        Thread t1 = new Thread(ThreadException::inc,"t1");</code><code>        Thread t2= new Thread(ThreadException::inc,"t2");</code><code>        t1.start();</code><code>        t2.start();</code><code>    }</code><code>}</code><code><br></code>

Distributed lock: ensure the synchronous execution of multiple nodes
Implementation plan: 1. Based on database, 2. Based on redis cache, 3. Based on zookeeper

The above is the detailed content of Example analysis of two Java threads adding 1 to a variable. For more information, please follow other related articles on the PHP Chinese website!