C语言7 switch语句为什么高效
switch语句的定义
语法
switch(表达式)
{
case 常亮表达式1:
语句;
break;
case 常亮表达式:
语句;
break;
case 常亮表达式:
语句;
break;
......
default:
语句;
break;
}
需要注意的点
- 表达式结束不能使浮点数
- case后的值不能一样
- case后的值必须是常量
break:break非常重要，当执行到一个分支后 如果没有break就会继续向下执行，遇到break才跳出switch语句default语句与位置无关，但是当default写在其他条件的前面时。如果没有break;，就会向下继续匹配
Switch语句与 if..else语句的区别：
- switch语句只能进行等值判断，而if..else可以进行区间判断
- switch语句的执行效率远远高于if..else 在分支条件比较多的情况下，这种趋势愈发明显
观察switch语句的反汇编，看看switch语句为啥效率高
为啥switch比if..else效率高
游戏中的switch语句
switch(x)
{
case 1:
printf("A \n");
break;
case 2:
printf("B \n");
break;
case 3:
printf("C \n");
break;
default:
printf("default \n");
break;
}
编译后反汇编代码
11: switch(x)
12: {
00401038 8B 45 08 mov eax,dword ptr [ebp+8]
0040103B 89 45 FC mov dword ptr [ebp-4],eax
0040103E 83 7D FC 01 cmp dword ptr [ebp-4],1
00401042 74 0E je MyPrint+32h (00401052)
00401044 83 7D FC 02 cmp dword ptr [ebp-4],2
00401048 74 17 je MyPrint+41h (00401061)
0040104A 83 7D FC 03 cmp dword ptr [ebp-4],3
0040104E 74 20 je MyPrint+50h (00401070)
00401050 EB 2D jmp MyPrint+5Fh (0040107f)
13: case 1:
14: printf("A \n");
00401052 68 30 00 42 00 push offset string "A \n" (00420030)
00401057 E8 B4 00 00 00 call printf (00401110)
0040105C 83 C4 04 add esp,4
15: break;
0040105F EB 2B jmp MyPrint+6Ch (0040108c)
16: case 2:
17: printf("B \n");
00401061 68 2C 00 42 00 push offset string "B \n" (0042002c)
00401066 E8 A5 00 00 00 call printf (00401110)
0040106B 83 C4 04 add esp,4
18: break;
0040106E EB 1C jmp MyPrint+6Ch (0040108c)
19: case 3:
20: printf("C \n");
00401070 68 28 00 42 00 push offset string "C \n" (00420028)
00401075 E8 96 00 00 00 call printf (00401110)
0040107A 83 C4 04 add esp,4
21: break;
0040107D EB 0D jmp MyPrint+6Ch (0040108c)
22: default:
23: printf("default \n");
0040107F 68 1C 00 42 00 push offset string "default \n" (0042001c)
00401084 E8 87 00 00 00 call printf (00401110)
00401089 83 C4 04 add esp,4
24: break;
25: }
26: }
区别
当调教比较少的时候 没啥区别 。但是当条件比较多的时候：
11: switch(x)
12: {
0040B818 8B 45 08 mov eax,dword ptr [ebp+8]
0040B81B 89 45 FC mov dword ptr [ebp-4],eax
0040B81E 8B 4D FC mov ecx,dword ptr [ebp-4]
0040B821 83 E9 01 sub ecx,1
0040B824 89 4D FC mov dword ptr [ebp-4],ecx
0040B827 83 7D FC 03 cmp dword ptr [ebp-4],3
0040B82B 77 46 ja $L539+0Fh (0040b873)
0040B82D 8B 55 FC mov edx,dword ptr [ebp-4]
0040B830 FF 24 95 91 B8 40 00 jmp dword ptr [edx*4+40B891h]
13: case 1:
14: printf("A \n");
0040B837 68 34 0F 42 00 push offset string "A \n" (00420f34)
0040B83C E8 CF 58 FF FF call printf (00401110)
0040B841 83 C4 04 add esp,4
15: break;
0040B844 EB 3A jmp $L539+1Ch (0040b880)
16: case 2:
17: printf("B \n");
0040B846 68 30 00 42 00 push offset string "B \n" (00420030)
0040B84B E8 C0 58 FF FF call printf (00401110)
0040B850 83 C4 04 add esp,4
18: break;
0040B853 EB 2B jmp $L539+1Ch (0040b880)
19: case 3:
20: printf("C \n");
0040B855 68 2C 00 42 00 push offset string "C \n" (0042002c)
0040B85A E8 B1 58 FF FF call printf (00401110)
0040B85F 83 C4 04 add esp,4
21: break;
0040B862 EB 1C jmp $L539+1Ch (0040b880)
22: case 4:
23: printf("D \n");
0040B864 68 28 00 42 00 push offset string "D \n" (00420028)
0040B869 E8 A2 58 FF FF call printf (00401110)
0040B86E 83 C4 04 add esp,4
24: break;
0040B871 EB 0D jmp $L539+1Ch (0040b880)
25: default:
26: printf("default \n");
0040B873 68 1C 00 42 00 push offset string "default \n" (0042001c)
0040B878 E8 93 58 FF FF call printf (00401110)
0040B87D 83 C4 04 add esp,4
27: break;
28: }
29: }
我们发现 生成一张内存表。
switch通过算法，直接一步算出地址 从表中找到需要跳转的地址
大表和小表
当生成的函数跳转地址表 每个成员有4个字节的时候 我们称之为 大表
总结：
-
即没有大表也没有小表：
- case 项小于等于3项
- case最大值和最小值差值>=255
-
有大表没小表：
- case项大于3项并且大表项空隙小于等于6个
-
有大表有小表
- case项大于3项并且大表项空隙大于6个并且case最大最小差值