Thread.sleep还是TimeUnit.SECONDS.sleep

转http://stevex.blog.51cto.com/4300375/1285767/

刚看到TimeUnit.SECONDS.sleep()方法时觉得挺奇怪的,这里怎么也提供sleep方法?

1
2
3
4
5
6
7
public void sleep(long timeout) throws InterruptedException {
    if (timeout > 0) {
        long ms = toMillis(timeout);
        int ns = excessNanos(timeout, ms);
        Thread.sleep(ms, ns);
    }
}

结果一看源码,原来是对Thread.sleep方法的包装,实现是一样的,只是多了时间单位转换和验证,然而TimeUnit枚举成员的方法却提供更好的可读性,这可能就是当初创建TimeUnit时提供sleep方法的原因吧,大家都知道sleep方法很常用,但经常要使用一个常量保存sleep的时间,比如3秒钟,我们代码通常会这样写:

1
private final int SLEEP_TIME = 3 1000//3 seconds

因为Thread.sleep方法参数接受的毫秒单位的数值,比较下面代码就知道TimeUnit枚举成员的sleep方法更优雅:

1
2
3
4
5
6
TimeUnit.MILLISECONDS.sleep(10);
TimeUnit.SECONDS.sleep(10);
TimeUnit.MINUTES.sleep(10);
Thread.sleep(10);
Thread.sleep(10*1000);
Thread.sleep(10*60*1000);

但使用TimeUnit枚举成员的sleep方法会不会带来性能损失了,毕竟增加了函数调用开销?

测试测试吧:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
import java.util.concurrent.TimeUnit;
public class TestSleep {
    public static void main(String[] args) throws InterruptedException {       
        sleepByTimeunit(10000);
                                                                                                               
        sleepByThread(10000);      
    }
    private static void sleepByTimeunit(int sleepTimes) throws InterruptedException {
        long start = System.currentTimeMillis();
                                                                                                               
        for(int i=0; i<sleepTimes; i++){
            TimeUnit.MILLISECONDS.sleep(10);
        }
                                                                                                               
        long end = System.currentTimeMillis();
                                                                                                               
        System.out.println("Total time consumed by TimeUnit.MILLISECONDS.sleep : " + (end - start));
    }
                                                                                                           
    private static void sleepByThread(int sleepTimes) throws InterruptedException {
        long start = System.currentTimeMillis();
                                                                                                               
        for(int i=0; i<sleepTimes; i++){
            Thread.sleep(10);
        }
                                                                                                               
        long end = System.currentTimeMillis();
                                                                                                               
        System.out.println("Total time consumed by Thread.sleep : " + (end - start));
    }
}

两次测试结果(Win7+4G+JDK7 测试期间计算资源充足):

1
2
3
4
5
6
Total time consumed by TimeUnit.MILLISECONDS.sleep : 100068
Total time consumed by Thread.sleep : 100134
Difference : -- -66
Total time consumed by TimeUnit.MILLISECONDS.sleep : 100222
Total time consumed by Thread.sleep : 100077
Difference : -- +145

从结果可以看出10000次调用差异很小,甚至一次更快,不排除JVM进行了优化,如果忽略性能方面考虑,从可读性方面建议使用TimeUnit枚举成员的sleep方法。

 

另外TimeUnit是枚举实现一个很好的实例,Doug Lea太神了,佩服得五体投地!

 

本文出自 “力量来源于赤诚的爱!” 博客,请务必保留此出处http://stevex.blog.51cto.com/4300375/1285767

posted @ 2015-06-01 14:48  guoliuya  阅读(288)  评论(0编辑  收藏  举报