这里我就不给大家去研究这两上下文关键字的il实现了,感兴趣的朋友可以自己去用ildasm分析一下。
我写这篇文章只是想强调微软官方文档上写的一点:
async和await并没有开辟新的线程,我们来分别执行一下io密集型和cpu密集型代码
:
static void Main(string[] args){ Stopwatch stopwatch = new Stopwatch(); stopwatch.Start(); GetUrlContentLengthAsync(); stopwatch.Stop(); Console.WriteLine("async,await时间: " + stopwatch.ElapsedMilliseconds); Stopwatch stopwatch1 = new Stopwatch(); stopwatch1.Start(); GetUrlContentLength(); stopwatch1.Stop(); Console.WriteLine("非async,await时间: " + stopwatch1.ElapsedMilliseconds); } public async static Task<int> GetUrlContentLength() { var client = new HttpClient(); string getStringTask = client.GetStringAsync("https://learn.microsoft.com/dotnet").ToString(); DoIndependentWork(); string contents = getStringTask; return contents.Length; } public async static Task<int> GetUrlContentLengthAsync() { var client = new HttpClient(); Task<string> getStringTask = client.GetStringAsync("https://learn.microsoft.com/dotnet"); DoIndependentWork(); string contents = await getStringTask; return contents.Length; } static void DoIndependentWork() { Console.WriteLine("Working..."); }
运行结果:
这是io密集型,就是微软官方文档的示例代码
为什么用了这个反而更慢?我理解的是因为存在上下文状态的切换。
可能我这样说不是很直观,我推荐阅读微软原文档:
https://learn.microsoft.com/zh-cn/dotnet/csharp/asynchronous-programming/task-asynchronous-programming-model?source=recommendations
我们再来看一个cpu密集型操作:
static void Main(string[] args){ Stopwatch stopwatch = new Stopwatch(); stopwatch.Start(); GetnumAsync(); stopwatch.Stop(); Console.WriteLine("async,await时间: " + stopwatch.ElapsedMilliseconds); Stopwatch stopwatch1 = new Stopwatch(); stopwatch1.Start(); Getnum(); stopwatch1.Stop(); Console.WriteLine("主线程时间: " + stopwatch1.ElapsedMilliseconds); } public async static Task<int> GetnumAsync() { int num = 0; await Task.Run(() => { for (int i = 0; i < 100000000; i++) { num += i; } }); return num; } public static int Getnum() { Stopwatch stopwatch1 = new Stopwatch(); stopwatch1.Start(); int num = 0; for (int i = 0; i < 100000000; i++) { num += i; } stopwatch1.Stop(); Console.WriteLine("子线程时间: " + stopwatch1.ElapsedMilliseconds); return num; }
运行结果如下:
此时,反而开辟了新线程,用时会更长,为什么呢?很简单,因为线程间的切换是需要资源的。
希望大家亲手实践一下我提供的这两例子,加深自己的映像。若有不对,请指正。