java – 动态编程实现不好或HashMap慢？

参见英文答案 > How do I write a correct micro-benchmark in Java?????????????????????????????????????11个
我有两种我正在测试的Fibonacci方法.两者都应该是线性的.我不明白memoization或HashMap查找比我想象的要慢.

我知道递归函数不应该通过添加DP来加速,但是因为它是静态的,所以它只需要计算一次(永远不会超出范围).第一次执行后,它从HashMap中检索答案.

我这样做是为了准备实现O(log n)斐波那契函数和基准测试,但是当我看到这个时,我会略微绕道而行. (即使对迭代方法添加memoization也会减慢它的速度).

请赐教,我此刻感觉很傻.

迭代方法：

public static int linearIterativeFib(int x) {
    if (x <= 2) {
        return 1;
    } else {
        int prev = 0;
        int result = 1;
        for (int i = 2; i <= x; i++) {
            int temp = result;
            result += prev;
            prev = temp;
        }
        return result;
    }       
}

递归方法：

static Map<Integer,Integer> memo = new HashMap<Integer,Integer>();

public static int linearRecursiveFib(int x) {
    if (x <= 2) {
        return 1;
    } else if (memo.containsKey(x)) {
        return memo.get(x);
    } else {
        int result = linearRecursiveFib(x - 1) 
                + linearRecursiveFib(x - 2);
        memo.put(x,result);
        return result;
    }
}

并进行单元测试：

@Test
public void testSpeed() {
    int input = 35;
    long start = System.currentTimeMillis();
    for (int i = 0; i < 10000000; i++) {
        fib.linearIterativeFib(input);
    }
    System.out.println("Iterative approach took " 
            + (System.currentTimeMillis() - start) + "ms");

    start = System.currentTimeMillis();
    for (int i = 0; i < 10000000; i++) {
        fib.linearRecursiveFib(input);
    }
    System.out.println("Recursive approach took " 
            + (System.currentTimeMillis() - start) + "ms");
}

返回：

Iterative approach took 6ms
Recursive approach took 132ms