异步等待和并行
Posted
技术标签:
【中文标题】异步等待和并行【英文标题】:Async await and parallel 【发布时间】:2016-05-09 15:50:12 【问题描述】:我对 async/await 如何并行工作有点困惑,所以我在这里做了一个测试代码: 我尝试发送 6 个我用列表模拟的任务。 此任务中的每一个都将执行 3 个其他子任务。
您可以复制/粘贴进行测试
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
using System.Threading;
namespace ConsoleApplication1
class Program
static void Main(string[] args)
//job simulation
Func<int, string, Tuple<int, string>> tc = Tuple.Create;
var input = new List<Tuple<int, string>>
tc( 6000, "task 1" ),
tc( 5000, "task 2" ),
tc( 1000, "task 3" ),
tc( 1000, "task 4" ),
tc( 1000, "task 5" ),
tc( 1000, "task 6" )
;
List<Tuple<int, string>> JobsList = new List<Tuple<int, string>>(input);
//paralelism atempt
List<Task> TaskLauncher = new List<Task>();
Parallel.ForEach<Tuple<int, string>>(JobsList, item => JobDispatcher(item.Item1, item.Item2));
Console.ReadLine();
public static async Task JobDispatcher(int time , string query)
List<Task> TList = new List<Task>();
Task<string> T1 = SubTask1(time, query);
Task<string> T2 = SubTask2(time, query);
Task<string> T3 = SubTask3(time, query);
TList.Add(T1);
TList.Add(T2);
TList.Add(T3);
Console.WriteLine("0 Launched ", query);
await Task.WhenAll(TList.ToArray());
Console.WriteLine(T1.Result);
Console.WriteLine(T2.Result);
Console.WriteLine(T3.Result);
public static async Task<string> SubTask1(int time, string query)
//somework
Thread.Sleep(time);
return query + "Finshed SubTask1";
public static async Task<string> SubTask2(int time, string query)
//somework
Thread.Sleep(time);
return query + "Finshed SubTask2";
public static async Task<string> SubTask3(int time, string query)
//somework
Thread.Sleep(time);
return query + "Finshed SubTask3";
理想情况下,我应该在发布时阅读:
task 1 launched
task 2 launched
task 3 launched
task 4 launched
task 5 launched
task 6 launched
那么此时所有任务运行 6*3 = 18 个线程同时运行 但它不是这里发生的事情似乎执行同步。
结果如下:
用 async/await 编写可以将任务和子任务作为 18 个并行线程启动的东西的正确方法是什么?
【问题讨论】:
***.com/a/11565317/2613020 它不是同步运行的,任务4在任务3之前启动,但在任务3之后完成。 首先它应该在控制台中写入,因为我稍后会在函数中等待子任务 看看these articles。 @Zwan:async
/await
是关于异步(并发没有线程); Parallel
是关于并行性(使用 more 线程的并发性)。这些是完全不同的并发方法,很少需要both。或许如果您描述了您实际尝试做的事情,我们可以提出一个更合理的解决方案?
【参考方案1】:
试试这个示例代码。请注意,它在大约 6 秒内完成,这表明所有任务都是异步运行的:
using System;
using System.Diagnostics;
using System.Threading;
using System.Threading.Tasks;
namespace ConsoleApplication1
class Program
static void Main()
// ThreadPool throttling may cause the speed with which
// the threads are launched to be throttled.
// You can avoid that by uncommenting the following line,
// but that is considered bad form:
// ThreadPool.SetMinThreads(20, 20);
var sw = Stopwatch.StartNew();
Console.WriteLine("Waiting for all tasks to complete");
RunWorkers().Wait();
Console.WriteLine("All tasks completed in " + sw.Elapsed);
public static async Task RunWorkers()
await Task.WhenAll(
JobDispatcher(6000, "task 1"),
JobDispatcher(5000, "task 2"),
JobDispatcher(4000, "task 3"),
JobDispatcher(3000, "task 4"),
JobDispatcher(2000, "task 5"),
JobDispatcher(1000, "task 6")
);
public static async Task JobDispatcher(int time, string query)
var results = await Task.WhenAll(
worker(time, query + ": Subtask 1"),
worker(time, query + ": Subtask 2"),
worker(time, query + ": Subtask 3")
);
Console.WriteLine(string.Join("\n", results));
static async Task<string> worker(int time, string query)
return await Task.Run(() =>
Console.WriteLine("Starting worker " + query);
Thread.Sleep(time);
Console.WriteLine("Completed worker " + query);
return query + ": " + time + ", thread id: " + Thread.CurrentThread.ManagedThreadId;
);
在RunWorkers()
中,您将使用一组任务来代替:
public static async Task RunWorkers()
Task[] tasks = new Task[6];
for (int i = 0; i < 6; ++i)
tasks[i] = JobDispatcher(1000 + i*1000, "task " + i);
await Task.WhenAll(tasks);
【讨论】:
构建线程的有趣方式我也可以在静态异步任务Task.Run()
的版本,但我相信它仍然需要worker()
中的Task.Run()
。 (它之前有await Task.Delay()
,但为了让它更像OP的代码,我用Thread.Sleep()
替换了它,这意味着它现在需要在任务中运行。)
我来自未来,我通过 WPF 尝试了代码(将“控制台”与“调试”交换)。我注意到“所有任务在...中完成”没有出现。解决方案:使用await RunWorkers ();
代替RunWorkers().Wait();
并将'async' 添加到Main
方法。以上是关于异步等待和并行的主要内容,如果未能解决你的问题,请参考以下文章