4.3.3 thread对性能有何帮助
public class ThreadLocalDemo {
public static final int GE_COUNT = 10000000;
public static final int THREAD_COUT = 4;
static ExecutorService executorService = Executors.newFixedThreadPool(THREAD_COUT);
public static Random random = new Random(123);
public static ThreadLocal<Random> randomThreadLocal = new ThreadLocal<Random>() {
protected Random initialValue() {
return new Random(123);
}
};
public static class RandomTask implements Callable<Long> {
private int mode = 0;
public RandomTask(int mode) {
this.mode = mode;
}
public Random getRandom() {
if (mode == 0) {
return random;
} else if (mode == 1) {
return randomThreadLocal.get();
} else {
return null;
}
}
public Long call() throws Exception {
long b = System.currentTimeMillis();
for (int i = 0; i < GE_COUNT; i++) {
getRandom().nextInt();
}
long e = System.currentTimeMillis();
System.out.println(Thread.currentThread().getName() + " spend" + (e - b) + "ms");
return e - b;
}
}
public static void main(String args[]) throws ExecutionException, InterruptedException {
Future<Long>[] futures = new Future[THREAD_COUT];
for (int i = 0; i < THREAD_COUT; i++) {
futures[i] = executorService.submit(new RandomTask(0));
}
long totalTime = 0;
for (int i = 0; i < THREAD_COUT; i++) {
totalTime += futures[i].get();
}
System.out.println("多线程访问同一个Random实例:" + totalTime + "ms");
executorService.shutdown();
executorService=Executors.newFixedThreadPool(4);
//ThreadLocal的情况
for (int i = 0; i < THREAD_COUT; i++) {
futures[i] = executorService.submit(new RandomTask(1));
}
totalTime = 0;
for (int i = 0; i < THREAD_COUT; i++) {
totalTime += futures[i].get();
}
executorService.shutdown();
System.out.println("使用ThreadLocal包装Random实例:" + totalTime + "ms");
}
运行结果:
pool-1-thread-3 spend1353ms
pool-1-thread-1 spend1542ms
pool-1-thread-2 spend1573ms
pool-1-thread-4 spend1593ms
多线程访问同一个Random实例:6061ms
pool-2-thread-1 spend531ms
pool-2-thread-2 spend512ms
pool-2-thread-3 spend450ms
pool-2-thread-4 spend414ms
使用ThreadLocal包装Random实例:1907ms