值类型和引用类型在hashtable里面性能比较分析1 -GetHashCode()

上篇文章中，测试了一下值类型数据和引用类型数据在hashtable中插入和读取的性能，测试结果和本人预期也有一些出入，msdn有一篇文章介绍在box,unbox的时候，性能关系为:class>interface>int,原文： Open the Box! Quick，进一步分析了上篇测试，发现其实影响测试性能还有其它几个方面的因素，本篇就针对不同数据类型在GetHashCode()上面的消耗
测试程序如下;

 1  System.Diagnostics.Stopwatch stopWatch  =   new  System.Diagnostics.Stopwatch();
 2             stopWatch.Start();
 3              for  ( int  i  =   0 ; i  <  CompareCount; i ++ )
 4              {
 5                int hashCode = i.GetHashCode();
 6            }
 7             stopWatch.Stop();
 8             System.Diagnostics.Stopwatch stopWatch2  =   new  System.Diagnostics.Stopwatch();
 9             stopWatch2.Start();
10              for  ( int  i  =   0 ; i  <  CompareCount; i ++ )
11              {
12                int hashCode = i.ToString().GetHashCode();
13            }
14             stopWatch2.Stop();
15             System.Diagnostics.Stopwatch stopWatch3  =   new  System.Diagnostics.Stopwatch();
16             stopWatch3.Start();
17              for  ( int  i  =   0 ; i  <  CompareCount; i ++ )
18              {
19                int hashCode = new Test().GetHashCode();
20            }
21             stopWatch3.Stop();
22              long  t1  =  stopWatch.ElapsedTicks;
23              long  t2  =  stopWatch2.ElapsedTicks;
24              long  t3  =  stopWatch3.ElapsedTicks;
25             Console.WriteLine(t1.ToString());
26             Console.WriteLine(t2.ToString());
27             Console.WriteLine(t3.ToString());
28             Console.WriteLine(( double )(t2 / t1));
29             ArrayList list  =   new  ArrayList();
30         }

结果表，我暂且不画在这上面了，因为比较多，而且不好排版，我只说一下结果，int类型的数据比字符串类型的在GetHashCode()的效率上要高50多倍,比Class的也高50-70倍
察看了GetHashCode()的实现
Int32
     

1 public   override   int  GetHashCode()
2 {
3      return this;
4}
5
6

String

 1 [ReliabilityContract(Consistency.WillNotCorruptState, Cer.MayFail)]
 2 public   override   unsafe   int  GetHashCode()
 3 {
 4      fixed (char* text1 = ((char*) this))
 5      {
 6            char* chPtr1 = text1;
 7            int num1 = 0x15051505;
 8            int num2 = num1;
 9            int* numPtr1 = (int*) chPtr1;
10            for (int num3 = this.Length; num3 > 0; num3 -= 4)
11            {
12                  num1 = (((num1 << 5) + num1) + (num1 >> 0x1b)) ^ numPtr1[0];
13                  if (num3 <= 2)
14                  {
15                        break;
16                  }
17                  num2 = (((num2 << 5) + num2) + (num2 >> 0x1b)) ^ numPtr1[1];
18                  numPtr1 += 2;
19            }
20            return (num1 + (num2 * 0x5d588b65));
21      }
22}
23
24  
25

而object的实现方法，我 不太理解：

1 public   virtual   int  GetHashCode()
2 {
3      return object.InternalGetHashCode(this);
4}
5
6  
7 [MethodImpl(MethodImplOptions.InternalCall)]
8 internal   static   extern   int  InternalGetHashCode( object  obj);
9

 object的具体处理方法
而从int,string类型算法可以轻易看出效率差别，虽然string类型还有unsafe .

而且，值类型必须重写GetHashCode。
从这里可以看出，在上次测试中，GetHashCode帮了int类型不少忙呀！