Develop Reference

Multi-threading in a foreach loop

I have read a few stackoverflow threads about multi-threading in a foreach loop, but I am not sure I am understanding and using it right.
I have tried multiple scenarios, but I am not seeing much increase in performance.
Here is what I believe runs Asynchronous tasks, but running synchronously in the loop using a single thread:
Stopwatch stopWatch = new Stopwatch();
stopWatch.Start();
foreach (IExchangeAPI selectedApi in selectedApis)
{
if (exchangeSymbols.TryGetValue(selectedApi.Name, out symbol))
{
ticker = await selectedApi.GetTickerAsync(symbol);
}
}
stopWatch.Stop();
Here is what I hoped to be running Asynchronously (still using a single thread) - I would have expected some speed improvement already here:
List<Task<ExchangeTicker>> exchTkrs = new List<Task<ExchangeTicker>>();
stopWatch.Start();
foreach (IExchangeAPI selectedApi in selectedApis)
{
if (exchangeSymbols.TryGetValue(selectedApi.Name, out symbol))
{
exchTkrs.Add(selectedApi.GetTickerAsync(symbol));
}
}
ExchangeTicker[] retTickers = await Task.WhenAll(exchTkrs);
stopWatch.Stop();
Here is what I would have hoped to run Asynchronously in Multi-thread:
stopWatch.Start();
Parallel.ForEach(selectedApis, async (IExchangeAPI selectedApi) =>
{
if (exchangeSymbols.TryGetValue(selectedApi.Name, out symbol))
{
ticker = await selectedApi.GetTickerAsync(symbol);
}
});
stopWatch.Stop();
Stop watch results interpreted as follows:
Console.WriteLine("Time elapsed (ns): {0}", stopWatch.Elapsed.TotalMilliseconds * 1000000);
Console outputs:
Time elapsed (ns): 4183308100
Time elapsed (ns): 4183946299.9999995
Time elapsed (ns): 4188032599.9999995
Now, the speed improvement looks minuscule. Am I doing something wrong or is that more or less what I should be expecting? I suppose writing to files would be a better to check that.
Would you mind also confirming I am interpreting the different use cases correctly?
Finally, using a foreach loop in order to get the ticker from multiple platforms in parallel may not be the best approach. Suggestions on how to improve this would be welcome.
EDIT
Note that I am using the ExchangeSharp code base that you can find here
Here is what the GerTickerAsync() method looks like:
public virtual async Task<ExchangeTicker> GetTickerAsync(string marketSymbol)
{
marketSymbol = NormalizeMarketSymbol(marketSymbol);
return await Cache.CacheMethod(MethodCachePolicy, async () => await OnGetTickerAsync(marketSymbol), nameof(GetTickerAsync), nameof(marketSymbol), marketSymbol);
}
For the Kraken API, you then have:
protected override async Task<ExchangeTicker> OnGetTickerAsync(string marketSymbol)
{
JToken apiTickers = await MakeJsonRequestAsync<JToken>("/0/public/Ticker", null, new Dictionary<string, object> { { "pair", NormalizeMarketSymbol(marketSymbol) } });
JToken ticker = apiTickers[marketSymbol];
return await ConvertToExchangeTickerAsync(marketSymbol, ticker);
}
And the Caching method:
public static async Task<T> CacheMethod<T>(this ICache cache, Dictionary<string, TimeSpan> methodCachePolicy, Func<Task<T>> method, params object?[] arguments) where T : class
{
await new SynchronizationContextRemover();
methodCachePolicy.ThrowIfNull(nameof(methodCachePolicy));
if (arguments.Length % 2 == 0)
{
throw new ArgumentException("Must pass function name and then name and value of each argument");
}
string methodName = (arguments[0] ?? string.Empty).ToStringInvariant();
string cacheKey = methodName;
for (int i = 1; i < arguments.Length;)
{
cacheKey += "|" + (arguments[i++] ?? string.Empty).ToStringInvariant() + "=" + (arguments[i++] ?? string.Empty).ToStringInvariant("(null)");
}
if (methodCachePolicy.TryGetValue(methodName, out TimeSpan cacheTime))
{
return (await cache.Get<T>(cacheKey, async () =>
{
T innerResult = await method();
return new CachedItem<T>(innerResult, CryptoUtility.UtcNow.Add(cacheTime));
})).Value;
}
else
{
return await method();
}
}

At first it should be pointed out that what you are trying to achieve is performance, not asynchrony. And you are trying to achieve it by running multiple operations concurrently, not in parallel. To keep the explanation simple I'll use a simplified version of your code, and I'll assume that each operation is a direct web request, without an intermediate caching layer, and with no dependencies to values existing in dictionaries.
foreach (var symbol in selectedSymbols)
{
var ticker = await selectedApi.GetTickerAsync(symbol);
}
The above code runs the operations sequentially. Each operation starts after the completion of the previous one.
var tasks = new List<Task<ExchangeTicker>>();
foreach (var symbol in selectedSymbols)
{
tasks.Add(selectedApi.GetTickerAsync(symbol));
}
var tickers = await Task.WhenAll(tasks);
The above code runs the operations concurrently. All operations start at once. The total duration is expected to be the duration of the longest running operation.
Parallel.ForEach(selectedSymbols, async symbol =>
{
var ticker = await selectedApi.GetTickerAsync(symbol);
});
The above code runs the operations concurrently, like the previous version with Task.WhenAll. It offers no advantage, while having the huge disadvantage that you no longer have a way to await the operations to complete. The Parallel.ForEach method will return immediately after launching the operations, because the Parallel class doesn't understand async delegates (it does not accept Func<Task> lambdas). Essentially there are a bunch of async void lambdas in there, that are running out of control, and in case of an exception they will bring down the process.
So the correct way to run the operations concurrently is the second way, using a list of tasks and the Task.WhenAll. Since you've already measured this method and haven't observed any performance improvements, I am assuming that there something else that serializes the concurrent operations. It could be something like a SemaphoreSlim hidden somewhere in your code, or some mechanism on the server side that throttles your requests. You'll have to investigate further to find where and why the throttling happens.

In general, when you do not see an increase by multi threading, it is because your task is not CPU limited or large enough to offset the overhead.
In your example, i.e.:
selectedApi.GetTickerAsync(symbol);
This can hae 2 reasons:
1: Looking up the ticker is brutally fast and it should not be an async to start with. I.e. when you look it up in a dictionary.
2: This is running via a http connection where the runtime is LIMITING THE NUMBER OF CONCURRENT CALLS. Regardless how many tasks you open, it will not use more than 4 at the same time.
Oh, and 3: you think async is using threads. It is not. It is particularly not the case in a codel ike this:
await selectedApi.GetTickerAsync(symbol);
Where you basically IMMEDIATELY WAIT FOR THE RESULT. There is no multi threading involved here at all.
foreach (IExchangeAPI selectedApi in selectedApis) {
if (exchangeSymbols.TryGetValue(selectedApi.Name, out symbol))
{
ticker = await selectedApi.GetTickerAsync(symbol);
} }
This is linear non threaded code using an async interface to not block the current thread while the (likely expensive IO) operation is in place. It starts one, THEN WAITS FOR THE RESULT. No 2 queries ever start at the same time.
If you want a possible (just as example) more scalable way:
In the foreach, do not await but add the task to a list of tasks.
Then start await once all the tasks have started. Likein a 2nd loop.
WAY not perfect, but at least the runtime has a CHANCE to do multiple lookups at the same time. Your await makes sure that you essentially run single threaded code, except async, so your thread goes back into the pool (and is not waiting for results), increasing your scalability - an item possibly not relevant in this case and definitely not measured in your test.

When CPU is not affected of "SemaphoreSlim.WaitAsync" usage and what are alternatives?

I have a program doing long running tasks, which should be started at program startup and it should be possible to be restarted in any moment after that.
My purpose is if "restart" is called, this to happen:
Request for all tasks to be finished
Waiting for all tasks to be completed
All functionalities are started again
I want to use async/await and still lock the process of starting all tasks in order to be sure that not a single restart is proceed until previous start/restart is finished.
As i saw for many reasons async/await does not work with lock statemnts, so I end up using SemaphoreSlim which works great for me. here is my code:
private readonly SemaphoreSlim m_semaphoreSlim;
private CancellationTokenSource m_cancellationTokenSource;
private CancellationToken m_cancellationToken;
public FeedClientService(IList<IFeedConfigurationBuilder> feedConfigs)
{
m_semaphoreSlim = new SemaphoreSlim(1, 1);
m_feedConfigs = feedConfigs;
}
public void Start()
{
Task.Run(() => this.FetchFeeds());
}
public void Restart()
{
if (m_cancellationTokenSource != null) m_cancellationTokenSource.Cancel();
Task.Run(() => this.FetchFeeds());
}
private async Task FetchFeeds()
{
try
{
await m_semaphoreSlim.WaitAsync();
m_cancellationTokenSource = new CancellationTokenSource();
m_cancellationToken = m_cancellationTokenSource.Token;
Task[] tasks = new Task[m_feedConfigs.Count];
for (int i = 0; i < m_feedConfigs.Count; i++)
{
var index = i;
tasks[index] = Task.Run(async () => await this.FetchFeed(index), m_cancellationToken);
}
await Task.WhenAll(tasks);
}
finally
{
m_semaphoreSlim.Release();
}
}
As it is pointed here https://stackoverflow.com/a/4154526/4664866 - "The SemaphoreSlim class represents a lightweight, fast semaphore that can be used for waiting within a single process when wait times are expected to be very short". I haven't found any source where is specified what means "very short" and I'm not sure if my code will not have performance bottlenecks, because the tasks I am starting are for sure not short running ones.
TL;DR;
What means "very short waiting times"?
What are alternative in case waiting times are very long?

It turned out that SemaphoreSlim.WaitAsync is not using spin-wait technique at all(for reference - implementation of SemaphoreSlim). So locking resources, even with long running tasks in them, will not affect CPU consumption.

How to correctly queue up tasks to run in C#

I have an enumeration of items (RunData.Demand), each representing some work involving calling an API over HTTP. It works great if I just foreach through it all and call the API during each iteration. However, each iteration takes a second or two so I'd like to run 2-3 threads and divide up the work between them. Here's what I'm doing:
ThreadPool.SetMaxThreads(2, 5); // Trying to limit the amount of threads
var tasks = RunData.Demand
.Select(service => Task.Run(async delegate
{
var availabilityResponse = await client.QueryAvailability(service);
// Do some other stuff, not really important
}));
await Task.WhenAll(tasks);
The client.QueryAvailability call basically calls an API using the HttpClient class:
public async Task<QueryAvailabilityResponse> QueryAvailability(QueryAvailabilityMultidayRequest request)
{
var response = await client.PostAsJsonAsync("api/queryavailabilitymultiday", request);
if (response.IsSuccessStatusCode)
{
return await response.Content.ReadAsAsync<QueryAvailabilityResponse>();
}
throw new HttpException((int) response.StatusCode, response.ReasonPhrase);
}
This works great for a while, but eventually things start timing out. If I set the HttpClient Timeout to an hour, then I start getting weird internal server errors.
What I started doing was setting a Stopwatch within the QueryAvailability method to see what was going on.
What's happening is all 1200 items in RunData.Demand are being created at once and all 1200 await client.PostAsJsonAsync methods are being called. It appears it then uses the 2 threads to slowly check back on the tasks, so towards the end I have tasks that have been waiting for 9 or 10 minutes.
Here's the behavior I would like:
I'd like to create the 1,200 tasks, then run them 3-4 at a time as threads become available. I do not want to queue up 1,200 HTTP calls immediately.
Is there a good way to go about doing this?

As I always recommend.. what you need is TPL Dataflow (to install: Install-Package System.Threading.Tasks.Dataflow).
You create an ActionBlock with an action to perform on each item. Set MaxDegreeOfParallelism for throttling. Start posting into it and await its completion:
var block = new ActionBlock<QueryAvailabilityMultidayRequest>(async service =>
{
var availabilityResponse = await client.QueryAvailability(service);
// ...
},
new ExecutionDataflowBlockOptions { MaxDegreeOfParallelism = 4 });
foreach (var service in RunData.Demand)
{
block.Post(service);
}
block.Complete();
await block.Completion;

Old question, but I would like to propose an alternative lightweight solution using the SemaphoreSlim class. Just reference System.Threading.
SemaphoreSlim sem = new SemaphoreSlim(4,4);
foreach (var service in RunData.Demand)
{
await sem.WaitAsync();
Task t = Task.Run(async () =>
{
var availabilityResponse = await client.QueryAvailability(serviceCopy));
// do your other stuff here with the result of QueryAvailability
}
t.ContinueWith(sem.Release());
}
The semaphore acts as a locking mechanism. You can only enter the semaphore by calling Wait (WaitAsync) which subtracts one from the count. Calling release adds one to the count.

You're using async HTTP calls, so limiting the number of threads will not help (nor will ParallelOptions.MaxDegreeOfParallelism in Parallel.ForEach as one of the answers suggests). Even a single thread can initiate all requests and process the results as they arrive.
One way to solve it is to use TPL Dataflow.
Another nice solution is to divide the source IEnumerable into partitions and process items in each partition sequentially as described in this blog post:
public static Task ForEachAsync<T>(this IEnumerable<T> source, int dop, Func<T, Task> body)
{
return Task.WhenAll(
from partition in Partitioner.Create(source).GetPartitions(dop)
select Task.Run(async delegate
{
using (partition)
while (partition.MoveNext())
await body(partition.Current);
}));
}

While the Dataflow library is great, I think it's a bit heavy when not using block composition. I would tend to use something like the extension method below.
Also, unlike the Partitioner method, this runs the async methods on the calling context - the caveat being that if your code is not truly async, or takes a 'fast path', then it will effectively run synchronously since no threads are explicitly created.
public static async Task RunParallelAsync<T>(this IEnumerable<T> items, Func<T, Task> asyncAction, int maxParallel)
{
var tasks = new List<Task>();
foreach (var item in items)
{
tasks.Add(asyncAction(item));
if (tasks.Count < maxParallel)
continue;
var notCompleted = tasks.Where(t => !t.IsCompleted).ToList();
if (notCompleted.Count >= maxParallel)
await Task.WhenAny(notCompleted);
}
await Task.WhenAll(tasks);
}

Random tasks from Task.Factory.StartNew never finishes

I am using Async await with Task.Factory method.
public async Task<JobDto> ProcessJob(JobDto jobTask)
{
try
{
var T = Task.Factory.StartNew(() =>
{
JobWorker jobWorker = new JobWorker();
jobWorker.Execute(jobTask);
});
await T;
}
This method I am calling inside a loop like this
for(int i=0; i < jobList.Count(); i++)
{
tasks[i] = ProcessJob(jobList[i]);
}
What I notice is that new tasks opens up inside Process explorer and they also start working (based on log file). however out of 10 sometimes 8 or sometimes 7 finishes. Rest of them just never come back.
why would that be happening ?
Are they timing out ? Where can I set timeout for my tasks ?
UPDATE
Basically above, I would like each Task to start running as soon as they are called and wait for the response on AWAIT T keyword. I am assuming here that once they finish each of them will come back at Await T and do the next action. I am alraedy seeing this result for 7 out of 10 tasks but 3 of them are not coming back.
Thanks

It is hard to say what the issues is without the rest if the code, but you code can be simplified by making ProcessJob synchronous and then calling Task.Run with it.
public JobDto ProcessJob(JobDto jobTask)
{
JobWorker jobWorker = new JobWorker();
return jobWorker.Execute(jobTask);
}
Start tasks and wait for all tasks to finish. Prefer using Task.Run rather than Task.Factory.StartNew as it provides more favourable defaults for pushing work to the background. See here.
for(int i=0; i < jobList.Count(); i++)
{
tasks[i] = Task.Run(() => ProcessJob(jobList[i]));
}
try
{
await Task.WhenAll(tasks);
}
catch(Exception ex)
{
// handle exception
}

First, let's make a reproducible version of your code. This is NOT the best way to achieve what you are doing, but to show you what is happening in your code!
I'll keep the code almost same as your code, except I'll use simple int rather than your JobDto and on completion of the job Execute() I'll write in a file that we can verify later. Here's the code
public class SomeMainClass
{
public void StartProcessing()
{
var jobList = Enumerable.Range(1, 10).ToArray();
var tasks = new Task[10];
//[1] start 10 jobs, one-by-one
for (int i = 0; i < jobList.Count(); i++)
{
tasks[i] = ProcessJob(jobList[i]);
}
//[4] here we have 10 awaitable Task in tasks
//[5] do all other unrelated operations
Thread.Sleep(1500); //assume it works for 1.5 sec
// Task.WaitAll(tasks); //[6] wait for tasks to complete
// The PROCESS IS COMPLETE here
}
public async Task ProcessJob(int jobTask)
{
try
{
//[2] start job in a ThreadPool, Background thread
var T = Task.Factory.StartNew(() =>
{
JobWorker jobWorker = new JobWorker();
jobWorker.Execute(jobTask);
});
//[3] await here will keep context of calling thread
await T; //... and release the calling thread
}
catch (Exception) { /*handle*/ }
}
}
public class JobWorker
{
static object locker = new object();
const string _file = #"C:\YourDirectory\out.txt";
public void Execute(int jobTask) //on complete, writes in file
{
Thread.Sleep(500); //let's assume does something for 0.5 sec
lock(locker)
{
File.AppendAllText(_file,
Environment.NewLine + "Writing the value-" + jobTask);
}
}
}
After running just the StartProcessing(), this is what I get in the file
Writing the value-4
Writing the value-2
Writing the value-3
Writing the value-1
Writing the value-6
Writing the value-7
Writing the value-8
Writing the value-5
So, 8/10 jobs has completed. Obviously, every time you run this, the number and order might change. But, the point is, all the jobs did not complete!
Now, if I un-comment the step [6] Task.WaitAll(tasks);, this is what I get in my file
Writing the value-2
Writing the value-3
Writing the value-4
Writing the value-1
Writing the value-5
Writing the value-7
Writing the value-8
Writing the value-6
Writing the value-9
Writing the value-10
So, all my jobs completed here!
Why the code is behaving like this, is already explained in the code-comments. The main thing to note is, your tasks run in ThreadPool based Background threads. So, if you do not wait for them, they will be killed when the MAIN process ends and the main thread exits!!
If you still don't want to await the tasks there, you can return the list of tasks from this first method and await the tasks at the very end of the process, something like this
public Task[] StartProcessing()
{
...
for (int i = 0; i < jobList.Count(); i++)
{
tasks[i] = ProcessJob(jobList[i]);
}
...
return tasks;
}
//in the MAIN METHOD of your application/process
var tasks = new SomeMainClass().StartProcessing();
// do all other stuffs here, and just at the end of process
Task.WaitAll(tasks);
Hope this clears all confusion.

It's possible your code is swallowing exceptions. I would add a ContineWith call to the end of the part of the code that starts the new task. Something like this untested code:
var T = Task.Factory.StartNew(() =>
{
JobWorker jobWorker = new JobWorker();
jobWorker.Execute(jobTask);
}).ContinueWith(tsk =>
{
var flattenedException = tsk.Exception.Flatten();
Console.Log("Exception! " + flattenedException);
return true;
});
},TaskContinuationOptions.OnlyOnFaulted); //Only call if task is faulted
Another possibility is that something in one of the tasks is timing out (like you mentioned) or deadlocking. To track down whether a timeout (or maybe deadlock) is the root cause, you could add some timeout logic (as described in this SO answer):
int timeout = 1000; //set to something much greater than the time it should take your task to complete (at least for testing)
var task = TheMethodWhichWrapsYourAsyncLogic(cancellationToken);
if (await Task.WhenAny(task, Task.Delay(timeout, cancellationToken)) == task)
{
// Task completed within timeout.
// Consider that the task may have faulted or been canceled.
// We re-await the task so that any exceptions/cancellation is rethrown.
await task;
}
else
{
// timeout/cancellation logic
}
Check out the documentation on exception handling in the TPL on MSDN.

How do I convert this to an async task?

Given the following code...
static void DoSomething(int id) {
Thread.Sleep(50);
Console.WriteLine(#"DidSomething({0})", id);
}
I know I can convert this to an async task as follows...
static async Task DoSomethingAsync(int id) {
await Task.Delay(50);
Console.WriteLine(#"DidSomethingAsync({0})", id);
}
And that by doing so if I am calling multiple times (Task.WhenAll) everything will be faster and more efficient than perhaps using Parallel.Foreach or even calling from within a loop.
But for a minute, lets pretend that Task.Delay() does not exist and I actually have to use Thread.Sleep(); I know in reality this is not the case, but this is concept code and where the Delay/Sleep is would normally be an IO operation where there is no async option (such as early EF).
I have tried the following...
static async Task DoSomethingAsync2(int id) {
await Task.Run(() => {
Thread.Sleep(50);
Console.WriteLine(#"DidSomethingAsync({0})", id);
});
}
But, though it runs without error, according to Lucien Wischik this is in fact bad practice as it is merely spinning up threads from the pool to complete each task (it is also slower using the following console application - if you swap between DoSomethingAsync and DoSomethingAsync2 call you can see a significant difference in the time that it takes to complete)...
static void Main(string[] args) {
MainAsync(args).Wait();
}
static async Task MainAsync(String[] args) {
List<Task> tasks = new List<Task>();
for (int i = 1; i <= 1000; i++)
tasks.Add(DoSomethingAsync2(i)); // Can replace with any version
await Task.WhenAll(tasks);
}
I then tried the following...
static async Task DoSomethingAsync3(int id) {
await new Task(() => {
Thread.Sleep(50);
Console.WriteLine(#"DidSomethingAsync({0})", id);
});
}
Transplanting this in place of the original DoSomethingAsync, the test never completes and nothing is shown on screen!
I have also tried multiple other variations that either do not compile or do not complete!
So, given the constraint that you cannot call any existing asynchronous methods and must complete both the Thread.Sleep and the Console.WriteLine in an asynchronous task, how do you do it in a manner that is as efficient as the original code?
The objective here for those of you who are interested is to give me a better understanding of how to create my own async methods where I am not calling anybody elses. Despite many searches, this seems to be the one area where examples are really lacking - whilst there are many thousands of examples of calling async methods that call other async methods in turn I cannot find any that convert an existing void method to an async task where there is no call to a further async task other than those that use the Task.Run(() => {} ) method.

There are two kinds of tasks: those that execute code (e.g., Task.Run and friends), and those that respond to some external event (e.g., TaskCompletionSource<T> and friends).
What you're looking for is TaskCompletionSource<T>. There are various "shorthand" forms for common situations so you don't always have to use TaskCompletionSource<T> directly. For example, Task.FromResult or TaskFactory.FromAsync. FromAsync is most commonly used if you have an existing *Begin/*End implementation of your I/O; otherwise, you can use TaskCompletionSource<T> directly.
For more information, see the "I/O-bound Tasks" section of Implementing the Task-based Asynchronous Pattern.
The Task constructor is (unfortunately) a holdover from Task-based parallelism, and should not be used in asynchronous code. It can only be used to create a code-based task, not an external event task.
So, given the constraint that you cannot call any existing asynchronous methods and must complete both the Thread.Sleep and the Console.WriteLine in an asynchronous task, how do you do it in a manner that is as efficient as the original code?
I would use a timer of some kind and have it complete a TaskCompletionSource<T> when the timer fires. I'm almost positive that's what the actual Task.Delay implementation does anyway.

So, given the constraint that you cannot call any existing
asynchronous methods and must complete both the Thread.Sleep and the
Console.WriteLine in an asynchronous task, how do you do it in a
manner that is as efficient as the original code?
IMO, this is a very synthetic constraint that you really need to stick with Thread.Sleep. Under this constraint, you still can slightly improve your Thread.Sleep-based code. Instead of this:
static async Task DoSomethingAsync2(int id) {
await Task.Run(() => {
Thread.Sleep(50);
Console.WriteLine(#"DidSomethingAsync({0})", id);
});
}
You could do this:
static Task DoSomethingAsync2(int id) {
return Task.Run(() => {
Thread.Sleep(50);
Console.WriteLine(#"DidSomethingAsync({0})", id);
});
}
This way, you'd avoid an overhead of the compiler-generated state machine class. There is a subtle difference between these two code fragments, in how exceptions are propagated.
Anyhow, this is not where the bottleneck of the slowdown is.
(it is also slower using the following console application - if you
swap between DoSomethingAsync and DoSomethingAsync2 call you can see a
significant difference in the time that it takes to complete)
Let's look one more time at your main loop code:
static async Task MainAsync(String[] args) {
List<Task> tasks = new List<Task>();
for (int i = 1; i <= 1000; i++)
tasks.Add(DoSomethingAsync2(i)); // Can replace with any version
await Task.WhenAll(tasks);
}
Technically, it requests 1000 tasks to be run in parallel, each supposedly to run on its own thread. In an ideal universe, you'd expect to execute Thread.Sleep(50) 1000 times in parallel and complete the whole thing in about 50ms.
However, this request is never satisfied by the TPL's default task scheduler, for a good reason: thread is a precious and expensive resource. Moreover, the actual number of concurrent operations is limited to the number of CPUs/cores. So in reality, with the default size of ThreadPool, I'm getting 21 pool threads (at peak) serving this operation in parallel. That is why DoSomethingAsync2 / Thread.Sleep takes so much longer than DoSomethingAsync / Task.Delay. DoSomethingAsync doesn't block a pool thread, it only requests one upon the completion of the time-out. Thus, more DoSomethingAsync tasks can actually run in parallel, than DoSomethingAsync2 those.
The test (a console app):
// https://stackoverflow.com/q/21800450/1768303
using System;
using System.Collections.Generic;
using System.Diagnostics;
using System.Threading;
using System.Threading.Tasks;
namespace Console_21800450
{
public class Program
{
static async Task DoSomethingAsync(int id)
{
await Task.Delay(50);
UpdateMaxThreads();
Console.WriteLine(#"DidSomethingAsync({0})", id);
}
static async Task DoSomethingAsync2(int id)
{
await Task.Run(() =>
{
Thread.Sleep(50);
UpdateMaxThreads();
Console.WriteLine(#"DidSomethingAsync2({0})", id);
});
}
static async Task MainAsync(Func<int, Task> tester)
{
List<Task> tasks = new List<Task>();
for (int i = 1; i <= 1000; i++)
tasks.Add(tester(i)); // Can replace with any version
await Task.WhenAll(tasks);
}
volatile static int s_maxThreads = 0;
static void UpdateMaxThreads()
{
var threads = Process.GetCurrentProcess().Threads.Count;
// not using locks for simplicity
if (s_maxThreads < threads)
s_maxThreads = threads;
}
static void TestAsync(Func<int, Task> tester)
{
s_maxThreads = 0;
var stopwatch = new Stopwatch();
stopwatch.Start();
MainAsync(tester).Wait();
Console.WriteLine(
"time, ms: " + stopwatch.ElapsedMilliseconds +
", threads at peak: " + s_maxThreads);
}
static void Main()
{
Console.WriteLine("Press enter to test with Task.Delay ...");
Console.ReadLine();
TestAsync(DoSomethingAsync);
Console.ReadLine();
Console.WriteLine("Press enter to test with Thread.Sleep ...");
Console.ReadLine();
TestAsync(DoSomethingAsync2);
Console.ReadLine();
}
}
}
Output:
Press enter to test with Task.Delay ...
...
time, ms: 1077, threads at peak: 13
Press enter to test with Thread.Sleep ...
...
time, ms: 8684, threads at peak: 21
Is it possible to improve the timing figure for the Thread.Sleep-based DoSomethingAsync2? The only way I can think of is to use TaskCreationOptions.LongRunning with Task.Factory.StartNew:
You should think twice before doing this in any real-life application:
static async Task DoSomethingAsync2(int id)
{
await Task.Factory.StartNew(() =>
{
Thread.Sleep(50);
UpdateMaxThreads();
Console.WriteLine(#"DidSomethingAsync2({0})", id);
}, TaskCreationOptions.LongRunning | TaskCreationOptions.PreferFairness);
}
// ...
static void Main()
{
Console.WriteLine("Press enter to test with Task.Delay ...");
Console.ReadLine();
TestAsync(DoSomethingAsync);
Console.ReadLine();
Console.WriteLine("Press enter to test with Thread.Sleep ...");
Console.ReadLine();
TestAsync(DoSomethingAsync2);
Console.ReadLine();
}
Output:
Press enter to test with Thread.Sleep ...
...
time, ms: 3600, threads at peak: 163
The timing gets better, but the price for this is high. This code asks the task scheduler to create a new thread for each new task. Do not expect this thread to come from the pool:
Task.Factory.StartNew(() =>
{
Thread.Sleep(1000);
Console.WriteLine("Thread pool: " +
Thread.CurrentThread.IsThreadPoolThread); // false!
}, TaskCreationOptions.LongRunning).Wait();