Java多線程系列六——Map實現類

參考資料：

https://crunchify.com/hashmap-vs-concurrenthashmap-vs-synchronizedmap-how-a-hashmap-can-be-synchronized-in-java/

https://stackoverflow.com/questions/35534906/java-hashmap-getobject-infinite-loop

Map的一些實現類有及其特性

類	線程安全	特性
Hashtable	是	Key不能爲null
HashMap	否	讀寫效率最高，但在Java6多線程環境下使用不當可能陷入死循環，進而致使CPU使用率太高(原理可參見：http://coolshell.cn/articles/9606.html)
Collections.synchronizedMap	是	Collections.SynchronizedMap在Map全部方法基礎上加鎖，效率與HashTable至關
ConcurrentHashMap	是	採用分段鎖，get通常不加鎖，put分段鎖，key/value不能爲null，效率僅次於HashMap

如下代碼測試各種的讀寫效率

import java.util.Collections;
import java.util.HashMap;
import java.util.Hashtable;
import java.util.Map;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.TimeUnit;

/**
 * @Description: 測試map
 */
public class ThreadMapTest {
    public static void main(String[] args) throws InterruptedException {
        Map<Integer, Integer> hashtable = new Hashtable<>();
        Map<Integer, Integer> hashmap = new HashMap<>();
        Map<Integer, Integer> synchronizedHashMap = Collections.synchronizedMap(new HashMap<>());
        Map<Integer, Integer> concurrentHashMap = new ConcurrentHashMap<>();

        test(hashtable);
        test(hashmap);
        test(synchronizedHashMap);
        test(concurrentHashMap);
    }

    private static void test(Map<Integer, Integer> map) throws InterruptedException {
        int testTimes = 5;
        long totalTimeMillis = 0;
        for (int k = 0; k < testTimes; k++) {
            totalTimeMillis += costTimeMillis(map);
        }
        System.out.println("Test " + map.getClass() + " average time " + (totalTimeMillis / testTimes));
    }

    private static long costTimeMillis(Map<Integer, Integer> map) throws InterruptedException {
        int count = 5;
        ExecutorService executorService = Executors.newFixedThreadPool(count);
        long startMillis = System.currentTimeMillis();
        for (int i = 0; i < count; i++) {
            executorService.execute(new Runnable() {
                @Override
                public void run() {
                    for (int j = 0; j < 500000; j++) {
                        map.put(0, 0);
                        map.get(0);
                    }
                }
            });
        }
        executorService.shutdown();
        executorService.awaitTermination(Long.MAX_VALUE, TimeUnit.DAYS);
        return System.currentTimeMillis() - startMillis;
    }
}

輸出結果以下：

Test class java.util.Hashtable average time 267
Test class java.util.HashMap average time 67
Test class java.util.Collections$SynchronizedMap average time 262
Test class java.util.concurrent.ConcurrentHashMap average time 167