详解C#实例化对象的三种方式及性能对比
时间：2021-11-29 10:31:40 | 栏目：.NET代码 | 点击：次
前言
做项目过程中有个需求要实例化两万个对象并添加到List
中，这个过程大概需要1min才能加载完（传参较多），于是开启了代码优化之旅，再此记录。
首先想到的是可能实例化比较耗时，于是开始对每种实例化方式进行测试，过程如下
实例化方式
1、用 New 关键字实例化一个类
2、用 Activator 实例化一个类
3、用 Assembly 实例化一个类
代码实现
测试环境：
vs2019 .NET Framework 4.7
Intel Core i7-10510U CPU
首先定义一个类Person
public class Person
{
public Person()
{
}
public Person(string name)
{
Name = name;
}
public string Name { get; set; }
}
我们先在无参的构造函数中实例化，每种方式进行十次，每次实例化十万次，代码如下
static void Main(string[] args)
{
Console.WriteLine("实例化对象的耗时比较（单位：毫秒）");
Console.Write(" ");
for (int i = 1; i <= 10; i++)
Console.Write("{0:G}", i.ToString().PadLeft(5));
Console.Write("\n");
Console.Write("InstanceByNew".PadRight(20));
for (int i = 1; i <= 10; i++)
{
Person person = null;
Stopwatch watch = new Stopwatch();
watch.Start();
for (int j = 0; j < 100000; j++)
person = new Person();
watch.Stop();
Console.Write(watch.ElapsedMilliseconds.ToString().PadLeft(5));
}
Console.Write("\n");
Console.Write("InstanceByActivator".PadRight(20));
for (int i = 1; i <= 10; i++)
{
Type type = Type.GetType("ConsoleApp1.Person");
Person person = null;
Stopwatch watch = new Stopwatch();
watch.Start();
for (int j = 0; j < 100000; j++)
{
object obj = Activator.CreateInstance(type);
person = obj as Person;
}
watch.Stop();
Console.Write(watch.ElapsedMilliseconds.ToString().PadLeft(5));
}
Console.Write("\n");
Console.Write("InstanceByAssembly".PadRight(20));
for (int i = 1; i <= 10; i++)
{
Assembly assembly = Assembly.Load("InstancePerformance");
Person person = null;
Stopwatch watch = new Stopwatch();
watch.Start();
for (int j = 0; j < 100000; j++)
{
object obj = assembly.CreateInstance("ConsoleApp1.Person");
person = obj as Person;
}
watch.Stop();
Console.Write(watch.ElapsedMilliseconds.ToString().PadLeft(5));
}
Console.Write("\n");
Console.ReadKey();
}
执行结果如下：
然后来看下有参构造函数中实例化的代码和结果
static void Main(string[] args)
{
Console.WriteLine("实例化对象的耗时比较（单位：毫秒）");
Console.Write(" ");
for (int i = 1; i <= 10; i++)
Console.Write("{0:G}", i.ToString().PadLeft(5));
Console.Write("\n");
Console.Write("InstanceByNew".PadRight(20));
for (int i = 1; i <= 10; i++)
{
Person person = null;
Stopwatch watch = new Stopwatch();
watch.Start();
for (int j = 0; j < 100000; j++)
person = new Person("Test"+j);
watch.Stop();
Console.Write(watch.ElapsedMilliseconds.ToString().PadLeft(5));
}
Console.Write("\n");
Console.Write("InstanceByActivator".PadRight(20));
for (int i = 1; i <= 10; i++)
{
Type type = Type.GetType("ConsoleApp1.Person");
Person person = null;
Stopwatch watch = new Stopwatch();
watch.Start();
for (int j = 0; j < 100000; j++)
{
object obj = Activator.CreateInstance(type,new object[]{"Test"+j});
person = obj as Person;
}
watch.Stop();
Console.Write(watch.ElapsedMilliseconds.ToString().PadLeft(5));
}
Console.Write("\n");
Console.Write("InstanceByAssembly".PadRight(20));
for (int i = 1; i <= 10; i++)
{
Assembly assembly = Assembly.Load("InstancePerformance");
Person person = null;
Stopwatch watch = new Stopwatch();
watch.Start();
for (int j = 0; j < 100000; j++)
{
object obj = assembly.CreateInstance("ConsoleApp1.Person", true, System.Reflection.BindingFlags.Default, null, new []{"Test"+j}, null, null);
person = obj as Person;
}
watch.Stop();
Console.Write(watch.ElapsedMilliseconds.ToString().PadLeft(5));
}
Console.Write("\n");
Console.ReadKey();
}
执行结果如下：
结论
从上面的执行结果可以看出这三种方式的性能排序为
New > Activator > Assembly
但使用哪种方法还要视情况而定
后续
以上可以看出New性能最高，而我就是使用的New，问题还没解决，把问题指向判断List是否存在某元素上，于是开始测试List中的Contains, Exists, Any，Where。详情请点击.