NlogN比N^2要慢？
最大和序列问题有多种解法，据说动态规划的要比两重循环快。（详见附件）
algo_cmp_windows.cpp在一台P4 3.0G Hz上的结果是（尽管显示的不对），
CPU frequency is 3579545 Hz
Cubic time:
8
7
7
8
9
9
10
11
12
13
15
17
19
Quadratic time:
7
6
6
6
7
7
8
7
8
8
9
9
9
N log N time:
7
7
8
8
8
9
9
9
10
10
11
11
12
Linear time:
6
6
6
6
6
6
6
6
6
6
6
6
6
（在其他机器上CPU频率和计数都是正确的，这个可以忽略）
Linux版的algo_cmp_linux.cpp（编译时需要-l rt）在一台SMP上的结果是，
cpu MHz : 2994.204
cpu MHz : 2994.204
Cubic time:
Size: 3, Maximun sum is 10, Time: 1000 nano seconds.
Size: 4, Maximun sum is 19, Time: 1000 nano seconds.
Size: 5, Maximun sum is 19, Time: 1000 nano seconds.
Size: 6, Maximun sum is 19, Time: 1000 nano seconds.
Size: 7, Maximun sum is 35, Time: 1000 nano seconds.
Size: 8, Maximun sum is 35, Time: 1000 nano seconds.
Size: 9, Maximun sum is 35, Time: 2000 nano seconds.
Size: 10, Maximun sum is 35, Time: 2000 nano seconds.
Size: 11, Maximun sum is 35, Time: 3000 nano seconds.
Size: 12, Maximun sum is 129, Time: 3000 nano seconds.
Size: 13, Maximun sum is 129, Time: 4000 nano seconds.
Size: 14, Maximun sum is 129, Time: 5000 nano seconds.
Size: 15, Maximun sum is 129, Time: 5000 nano seconds.
Quadratic time:
Size: 3, Maximun sum is 10, Time: 1000 nano seconds.
Size: 4, Maximun sum is 19, Time: 1000 nano seconds.
Size: 5, Maximun sum is 19, Time: 1000 nano seconds.
Size: 6, Maximun sum is 19, Time: 1000 nano seconds.
Size: 7, Maximun sum is 35, Time: 1000 nano seconds.
Size: 8, Maximun sum is 35, Time: 1000 nano seconds.
Size: 9, Maximun sum is 35, Time: 1000 nano seconds.
Size: 10, Maximun sum is 35, Time: 2000 nano seconds.
Size: 11, Maximun sum is 35, Time: 2000 nano seconds.
Size: 12, Maximun sum is 129, Time: 1000 nano seconds.
Size: 13, Maximun sum is 129, Time: 2000 nano seconds.
Size: 14, Maximun sum is 129, Time: 2000 nano seconds.
Size: 15, Maximun sum is 129, Time: 1000 nano seconds.
N log N time:
Size: 3, Maximun sum is 10, Time: 2000 nano seconds.
Size: 4, Maximun sum is 19, Time: 1000 nano seconds.
Size: 5, Maximun sum is 19, Time: 1000 nano seconds.
Size: 6, Maximun sum is 19, Time: 1000 nano seconds.
Size: 7, Maximun sum is 35, Time: 1000 nano seconds.
Size: 8, Maximun sum is 35, Time: 2000 nano seconds.
Size: 9, Maximun sum is 35, Time: 2000 nano seconds.
Size: 10, Maximun sum is 35, Time: 2000 nano seconds.
Size: 11, Maximun sum is 35, Time: 2000 nano seconds.
Size: 12, Maximun sum is 129, Time: 2000 nano seconds.
Size: 13, Maximun sum is 129, Time: 2000 nano seconds.
Size: 14, Maximun sum is 129, Time: 2000 nano seconds.
Size: 15, Maximun sum is 129, Time: 2000 nano seconds.
Linear time:
Size: 3, Maximun sum is 10, Time: 1000 nano seconds.
Size: 4, Maximun sum is 19, Time: 1000 nano seconds.
Size: 5, Maximun sum is 19, Time: 1000 nano seconds.
Size: 6, Maximun sum is 19, Time: 1000 nano seconds.
Size: 7, Maximun sum is 35, Time: 1000 nano seconds.
Size: 8, Maximun sum is 35, Time: 1000 nano seconds.
Size: 9, Maximun sum is 35, Time: 1000 nano seconds.
Size: 10, Maximun sum is 35, Time: 0 nano seconds.
Size: 11, Maximun sum is 35, Time: 0 nano seconds.
Size: 12, Maximun sum is 129, Time: 1000 nano seconds.
Size: 13, Maximun sum is 129, Time: 0 nano seconds.
Size: 14, Maximun sum is 129, Time: 1000 nano seconds.
Size: 15, Maximun sum is 129, Time: 1000 nano seconds.
总之maxSubSum3比maxSubSum2要慢？
[解决办法]
数据规模太小了，一次函数调用的代价都差不多同你计算部分的代价差不多了。
至少也得用上数百数千的数据测试一下吧
[解决办法]
是呀，历史上还出现过，对于同一问题，对于较大的数据，多项式算法不如指数算法的情况。
