C#整形转为字符串
起因
群里有人在说int转string,两种方式那种要好一些.当时心理就觉得调用ToString这种方式要一些.至于为什么直接调用类型ToString函数会好一些呢?- 123.ToString()
- 123 + ""
[Benchmark]
public string ToString12345() => 12345.ToString();
[Benchmark]
public string ToString54321() => 12345 + "";测试结果:
根据性能测试,两种用法性能在同一.Net版本性能相差不大,当然在.Net Core 3.1的是int+""性能要好一些.不过也看出在.Net 5性能改进很大.
看c#编译后的代码:
[Benchmark]
public string ToString12345()
{
return this.val1.ToString();
}
[Benchmark]
public string ToString54321()
{
return this.val2.ToString() ?? "";
}看JIT生成的汇编代码:
C.ToString12345()
L0000: push ebp
L0001: mov ebp, esp
L0003: sub esp, 0xc
L0006: xor eax, eax
L0008: mov [ebp-8], eax
L000b: mov [ebp-0xc], eax
L000e: mov [ebp-4], ecx
L0011: cmp dword ptr [0x5fac1a8], 0
L0018: je short L001f
L001a: call 0x04326120
L001f: mov dword ptr [ebp-8], 0x3039
L0026: lea ecx, [ebp-8]
L0029: call System.Int32.ToString()
L002e: mov [ebp-0xc], eax
L0031: mov eax, [ebp-0xc]
L0034: mov esp, ebp
L0036: pop ebp
L0037: ret
C.ToString54321()
L0000: push ebp
L0001: mov ebp, esp
L0003: sub esp, 0x10
L0006: xor eax, eax
L0008: mov [ebp-8], eax
L000b: mov [ebp-0xc], eax
L000e: mov [ebp-0x10], eax
L0011: mov [ebp-4], ecx
L0014: cmp dword ptr [0x5fac1a8], 0 ;判断是否为空
L001b: je short L0022
L001d: call 0x04326120
L0022: mov dword ptr [ebp-8], 0xd431
L0029: lea ecx, [ebp-8]
L002c: call System.Int32.ToString()
L0031: mov [ebp-0xc], eax
L0034: mov eax, [ebp-0xc]
L0037: mov [ebp-0x10], eax
L003a: cmp dword ptr [ebp-0xc], 0
L003e: jne short L0049
L0040: mov eax, [0x19b42038]
L0046: mov [ebp-0x10], eax
L0049: mov eax, [ebp-0x10]
L004c: mov esp, ebp
L004e: pop ebp
L004f: ret从汇编代码看出,int+""生成汇编代码,还是加了判断是否为空,不为空的时候才会调用ToString方法,是比直接调用ToString多一个判断.
.Net 5/6/7 性能基准测试对比(2022.03.26增加)
刚才找了一下Int32的ToString源码,发现2020年后,就没调整了.
秋风
2021-01-09