TaskScheduler Class


The .NET API Reference documentation has a new home. Visit the .NET API Browser on docs.microsoft.com to see the new experience.

Represents an object that handles the low-level work of queuing tasks onto threads.

Namespace:   System.Threading.Tasks
Assembly:  mscorlib (in mscorlib.dll)


[HostProtectionAttribute(SecurityAction.LinkDemand, Synchronization = true, 
	ExternalThreading = true)]
[PermissionSetAttribute(SecurityAction.InheritanceDemand, Unrestricted = true)]
public abstract class TaskScheduler


Initializes the TaskScheduler.


Gets the TaskScheduler associated with the currently executing task.


Gets the default TaskScheduler instance that is provided by the .NET Framework.


Gets the unique ID for this TaskScheduler.


Indicates the maximum concurrency level this TaskScheduler is able to support.


Determines whether the specified object is equal to the current object.(Inherited from Object.)


Allows an object to try to free resources and perform other cleanup operations before it is reclaimed by garbage collection.(Inherited from Object.)


Creates a TaskScheduler associated with the current System.Threading.SynchronizationContext.


Serves as the default hash function. (Inherited from Object.)


For debugger support only, generates an enumerable of Task instances currently queued to the scheduler waiting to be executed.


Gets the Type of the current instance.(Inherited from Object.)


Creates a shallow copy of the current Object.(Inherited from Object.)


Queues a Task to the scheduler.


Returns a string that represents the current object.(Inherited from Object.)


Attempts to dequeue a Task that was previously queued to this scheduler.


Attempts to execute the provided Task on this scheduler.

System_CAPS_protmethodTryExecuteTaskInline(Task, Boolean)

Determines whether the provided Task can be executed synchronously in this call, and if it can, executes it.


Occurs when a faulted task's unobserved exception is about to trigger exception escalation policy, which, by default, would terminate the process.

An instance of the TaskScheduler class represents a task scheduler. A task scheduler ensures that the work of a task is eventually executed.

The default task scheduler is based on the .NET Framework 4 thread pool, which provides work-stealing for load-balancing, thread injection/retirement for maximum throughput, and overall good performance. It should be sufficient for most scenarios.

The TaskScheduler class also serves as the extension point for all customizable scheduling logic. This includes mechanisms such as how to schedule a task for execution, and how scheduled tasks should be exposed to debuggers. If you require special functionality, you can create a custom scheduler and enable it for specific tasks or queries.

In this topic:

The default task scheduler and the thread pool

      The global queue vs. local queues
      Work stealing      
      Long-running tasks      
      Task inlining

Specifying a synchronization context      

The default scheduler for the Task Parallel Library and PLINQ uses the .NET Framework thread pool, which is represented by the ThreadPool class, to queue and execute work. The thread pool uses the information that is provided by the Task type to efficiently support the fine-grained parallelism (short-lived units of work) that parallel tasks and queries often represent.

The thread pool maintains a global FIFO (first-in, first-out) work queue for threads in each application domain. Whenever a program calls the ThreadPool.QueueUserWorkItem (or ThreadPool.UnsafeQueueUserWorkItem) method, the work is put on this shared queue and eventually de-queued onto the next thread that becomes available. Starting with the .NET Framework 4, this queue has been improved to use a lock-free algorithm that resembles the ConcurrentQueue<T> class. By using this lock-free implementation, the thread pool spends less time when it queues and de-queues work items. This performance benefit is available to all programs that use the thread pool.

Top-level tasks, which are tasks that are not created in the context of another task, are put on the global queue just like any other work item. However, nested or child tasks, which are created in the context of another task, are handled quite differently. A child or nested task is put on a local queue that is specific to the thread on which the parent task is executing. The parent task may be a top-level task or it also may be the child of another task. When this thread is ready for more work, it first looks in the local queue. If work items are waiting there, they can be accessed quickly. The local queues are accessed in last-in, first-out order (LIFO) to preserve cache locality and reduce contention. For more information about child tasks and nested tasks, see Attached and Detached Child Tasks.

The use of local queues not only reduces pressure on the global queue, but also takes advantage of data locality. Work items in the local queue frequently reference data structures that are physically near one another in memory. In these cases, the data is already in the cache after the first task has run and can be accessed quickly. Both Parallel LINQ (PLINQ) and the Parallel class use nested tasks and child tasks extensively, and achieve significant speedups by using the local work queues.

Starting with the .NET Framework 4, the thread pool also features a work-stealing algorithm to help make sure that no threads are sitting idle while others still have work in their queues. When a thread-pool thread is ready for more work, it first looks at the head of its local queue, then in the global queue, and then in the local queues of other threads. If it finds a work item in the local queue of another thread, it first applies heuristics to make sure that it can run the work efficiently. If it can, it de-queues the work item from the tail (in FIFO order). This reduces contention on each local queue and preserves data locality. This architecture helps the thread pool load-balance work more efficiently than past versions did.

You may want to explicitly prevent a task from being put on a local queue. For example, you may know that a particular work item will run for a relatively long time and is likely to block all other work items on the local queue. In this case, you can specify the TaskCreationOptions.LongRunning option, which provides a hint to the scheduler that an additional thread might be required for the task so that it does not block the forward progress of other threads or work items on the local queue. By using this option you avoid the thread pool completely, including the global and local queues.

In some cases when a Task is waited on, it may be executed synchronously on the thread that is performing the wait operation. This enhances performance by preventing the need for an additional thread and instead using the existing thread, which would have blocked otherwise. To prevent errors due to re-entrancy, task inlining only occurs when the wait target is found in the relevant thread's local queue.

You can use the TaskScheduler.FromCurrentSynchronizationContext method to specify that a task should be scheduled to run on a particular thread. This is useful in frameworks such as Windows Forms and Windows Presentation Foundation where access to user interface objects is often restricted to code that is running on the same thread on which the UI object was created. For more information, see How to: Schedule Work on the User Interface (UI) Thread.

The following example uses the TaskScheduler.FromCurrentSynchronizationContext method in a Windows Presentation Foundation (WPF) app to schedule a task on the same thread that the user interface (UI) control was created on. The example creates a mosaic of images that are randomly selected from a specified directory. The WPF objects are used to load and resize the images. The raw pixels are then passed to a task that uses a For loop to write the pixel data into a large single-byte array. No synchronization is required because no two tiles occupy the same array elements. The tiles can also be written in any order because their position is calculated independently of any other tile. The large array is then passed to a task that runs on the UI thread, where the pixel data is loaded into an Image control.

The example moves data off the UI thread, modifies it by using parallel loops and Task objects, and then passes it back to a task that runs on the UI thread. This approach is useful when you have to use the Task Parallel Library to perform operations that either are not supported by the WPF API, or are not sufficiently fast. Another way to create an image mosaic in WPF is to use a System.Windows.Controls.WrapPanel control and add images to it. The WrapPanel handles the work of positioning the tiles. However, this work can only be performed on the UI thread.

using System;
using System.Threading.Tasks;
using System.Windows;
using System.Windows.Media;
using System.Windows.Media.Imaging;

namespace WPF_CS1
    /// <summary>
    /// Interaction logic for MainWindow.xaml
    /// </summary>
    public partial class MainWindow : Window
        private int fileCount;
        int colCount;
        int rowCount;
        private int tilePixelHeight;
        private int tilePixelWidth;
        private int largeImagePixelHeight;
        private int largeImagePixelWidth;
        private int largeImageStride;
        PixelFormat format;
        BitmapPalette palette = null;

        public MainWindow()

            // For this example, values are hard-coded to a mosaic of 8x8 tiles.
            // Each tile is 50 pixels high and 66 pixels wide and 32 bits per pixel.
            colCount = 12;
            rowCount = 8;
            tilePixelHeight = 50;
            tilePixelWidth = 66;
            largeImagePixelHeight = tilePixelHeight * rowCount;
            largeImagePixelWidth = tilePixelWidth * colCount;
            largeImageStride = largeImagePixelWidth * (32 / 8);
            this.Width = largeImagePixelWidth + 40;
            image.Width = largeImagePixelWidth;
            image.Height = largeImagePixelHeight;


        private void button_Click(object sender, RoutedEventArgs e)

            // For best results use 1024 x 768 jpg files at 32bpp.
            string[] files = System.IO.Directory.GetFiles(@"C:\Users\Public\Pictures\Sample Pictures\", "*.jpg");

            fileCount = files.Length;
            Task<byte[]>[] images = new Task<byte[]>[fileCount];
            for (int i = 0; i < fileCount; i++)
                int x = i;
                images[x] = Task.Factory.StartNew(() => LoadImage(files[x]));

            // When they�ve all been loaded, tile them into a single byte array.
            var tiledImage = Task.Factory.ContinueWhenAll(
                images, (i) => TileImages(i));

            // We are currently on the UI thread. Save the sync context and pass it to
            // the next task so that it can access the UI control "image".
            var UISyncContext = TaskScheduler.FromCurrentSynchronizationContext();

            //  On the UI thread, put the bytes into a bitmap and
            // and display it in the Image control.
            var t3 = tiledImage.ContinueWith((antecedent) =>
                // Get System DPI.
                Matrix m = PresentationSource.FromVisual(Application.Current.MainWindow)
                double dpiX = m.M11;
                double dpiY = m.M22;

                BitmapSource bms = BitmapSource.Create(largeImagePixelWidth,
                    palette, //use default palette
                image.Source = bms;
            }, UISyncContext);

        byte[] LoadImage(string filename)
            // Use the WPF BitmapImage class to load and 
            // resize the bitmap. NOTE: Only 32bpp formats are supported correctly.
            // Support for additional color formats is left as an exercise
            // for the reader. For more information, see documentation for ColorConvertedBitmap.

            BitmapImage bitmapImage = new BitmapImage();
            bitmapImage.UriSource = new Uri(filename);
            bitmapImage.DecodePixelHeight = tilePixelHeight;
            bitmapImage.DecodePixelWidth = tilePixelWidth;

            format = bitmapImage.Format;
            int size = (int)(bitmapImage.Height * bitmapImage.Width);
            int stride = (int)bitmapImage.Width * 4;
            byte[] dest = new byte[stride * tilePixelHeight];

            bitmapImage.CopyPixels(dest, stride, 0);

            return dest;

        int Stride(int pixelWidth, int bitsPerPixel)
            return (((pixelWidth * bitsPerPixel + 31) / 32) * 4);

        // Map the individual image tiles to the large image
        // in parallel. Any kind of raw image manipulation can be
        // done here because we are not attempting to access any 
        // WPF controls from multiple threads.
        byte[] TileImages(Task<byte[]>[] sourceImages)
            byte[] largeImage = new byte[largeImagePixelHeight * largeImageStride];
            int tileImageStride = tilePixelWidth * 4; // hard coded to 32bpp

            Random rand = new Random();
            Parallel.For(0, rowCount * colCount, (i) =>
                // Pick one of the images at random for this tile.
                int cur = rand.Next(0, sourceImages.Length);
                byte[] pixels = sourceImages[cur].Result;

                // Get the starting index for this tile.
                int row = i / colCount;
                int col = (int)(i % colCount);
                int idx = ((row * (largeImageStride * tilePixelHeight)) + (col * tileImageStride));

                // Write the pixels for the current tile. The pixels are not contiguous
                // in the array, therefore we have to advance the index by the image stride
                // (minus the stride of the tile) for each scanline of the tile.
                int tileImageIndex = 0;
                for (int j = 0; j < tilePixelHeight; j++)
                    // Write the next scanline for this tile.
                    for (int k = 0; k < tileImageStride; k++)
                        largeImage[idx++] = pixels[tileImageIndex++];
                    // Advance to the beginning of the next scanline.
                    idx += largeImageStride - tileImageStride;
            return largeImage;

To create the example, crate a WPF application project in Visual Studio and assign it a name of your choice. Then do the following:

  1. In design view, drag an Image control from the Toolbox to the design surface. In XAML view, specify the horizontal alignment as "Left." The size does not matter because the control is be dynamically resized at run time. Accept the default name, "image".

  2. Drag a Button control from the Toolbox to the lower left part of the application window. Double-click the button to add a Click event handler. In XAML view, specify the Content property of the button as "Make a Mosaic" and specify its horizontal alignment as "Left". Accept the default name, "button".

  3. Replace the entire contents of the MainWindow.xaml.cs or MainWindow.xaml.vb file with the code from this example. Make sure that the name of the workspace matches the project name.

  4. The example reads JPEG images from a directory named C:\Users\Public\Pictures\Sample Pictures\. Either create the directory and place some images in it, or change the path to refer to some other directory that contains images.

This example has some limitations. For example, only 32-bits-per-pixel images are supported; images in other formats are corrupted by the BitmapImage object during the resizing operation. Also, the source images must all be larger than the tile size. As a further exercise, you can add functionality to handle multiple pixel formats and file sizes.

The following example is taken from the Samples for Parallel Programming with the .NET Framework 4 on the MSDN Code Gallery Web site. It creates a custom task scheduler that limits the number of threads used by the app. It then launches two sets of tasks and displays information about the task and the thread on which the task is executing.

using System;
using System.Collections.Generic;
using System.Threading;
using System.Threading.Tasks;

class Example
   static void Main()
       // Create a scheduler that uses two threads. 
       LimitedConcurrencyLevelTaskScheduler lcts = new LimitedConcurrencyLevelTaskScheduler(2);
       List<Task> tasks = new List<Task>();

       // Create a TaskFactory and pass it our custom scheduler. 
       TaskFactory factory = new TaskFactory(lcts);
       CancellationTokenSource cts = new CancellationTokenSource();

       // Use our factory to run a set of tasks. 
       Object lockObj = new Object();
       int outputItem = 0;

       for (int tCtr = 0; tCtr <= 4; tCtr++) {
          int iteration = tCtr;
          Task t = factory.StartNew(() => {
                                       for (int i = 0; i < 1000; i++) {
                                          lock (lockObj) {
                                             Console.Write("{0} in task t-{1} on thread {2}   ", 
                                                           i, iteration, Thread.CurrentThread.ManagedThreadId);
                                             if (outputItem % 3 == 0)
                                    }, cts.Token);
      // Use it to run a second set of tasks.                       
      for (int tCtr = 0; tCtr <= 4; tCtr++) {
         int iteration = tCtr;
         Task t1 = factory.StartNew(() => {
                                       for (int outer = 0; outer <= 10; outer++) {
                                          for (int i = 0x21; i <= 0x7E; i++) {
                                             lock (lockObj) {
                                                Console.Write("'{0}' in task t1-{1} on thread {2}   ", 
                                                              Convert.ToChar(i), iteration, Thread.CurrentThread.ManagedThreadId);
                                                if (outputItem % 3 == 0)
                                    }, cts.Token);           

      // Wait for the tasks to complete before displaying a completion message.
      Console.WriteLine("\n\nSuccessful completion.");

// Provides a task scheduler that ensures a maximum concurrency level while 
// running on top of the thread pool.
public class LimitedConcurrencyLevelTaskScheduler : TaskScheduler
   // Indicates whether the current thread is processing work items.
   private static bool _currentThreadIsProcessingItems;

  // The list of tasks to be executed 
   private readonly LinkedList<Task> _tasks = new LinkedList<Task>(); // protected by lock(_tasks)

   // The maximum concurrency level allowed by this scheduler. 
   private readonly int _maxDegreeOfParallelism;

   // Indicates whether the scheduler is currently processing work items. 
   private int _delegatesQueuedOrRunning = 0;

   // Creates a new instance with the specified degree of parallelism. 
   public LimitedConcurrencyLevelTaskScheduler(int maxDegreeOfParallelism)
       if (maxDegreeOfParallelism < 1) throw new ArgumentOutOfRangeException("maxDegreeOfParallelism");
       _maxDegreeOfParallelism = maxDegreeOfParallelism;

   // Queues a task to the scheduler. 
   protected sealed override void QueueTask(Task task)
      // Add the task to the list of tasks to be processed.  If there aren't enough 
      // delegates currently queued or running to process tasks, schedule another. 
       lock (_tasks)
           if (_delegatesQueuedOrRunning < _maxDegreeOfParallelism)

   // Inform the ThreadPool that there's work to be executed for this scheduler. 
   private void NotifyThreadPoolOfPendingWork()
       ThreadPool.UnsafeQueueUserWorkItem(_ =>
           // Note that the current thread is now processing work items.
           // This is necessary to enable inlining of tasks into this thread.
           _currentThreadIsProcessingItems = true;
               // Process all available items in the queue.
               while (true)
                   Task item;
                   lock (_tasks)
                       // When there are no more items to be processed,
                       // note that we're done processing, and get out.
                       if (_tasks.Count == 0)

                       // Get the next item from the queue
                       item = _tasks.First.Value;

                   // Execute the task we pulled out of the queue
           // We're done processing items on the current thread
           finally { _currentThreadIsProcessingItems = false; }
       }, null);

   // Attempts to execute the specified task on the current thread. 
   protected sealed override bool TryExecuteTaskInline(Task task, bool taskWasPreviouslyQueued)
       // If this thread isn't already processing a task, we don't support inlining
       if (!_currentThreadIsProcessingItems) return false;

       // If the task was previously queued, remove it from the queue
       if (taskWasPreviouslyQueued) 
          // Try to run the task. 
          if (TryDequeue(task)) 
            return base.TryExecuteTask(task);
             return false; 
          return base.TryExecuteTask(task);

   // Attempt to remove a previously scheduled task from the scheduler. 
   protected sealed override bool TryDequeue(Task task)
       lock (_tasks) return _tasks.Remove(task);

   // Gets the maximum concurrency level supported by this scheduler. 
   public sealed override int MaximumConcurrencyLevel { get { return _maxDegreeOfParallelism; } }

   // Gets an enumerable of the tasks currently scheduled on this scheduler. 
   protected sealed override IEnumerable<Task> GetScheduledTasks()
       bool lockTaken = false;
           Monitor.TryEnter(_tasks, ref lockTaken);
           if (lockTaken) return _tasks;
           else throw new NotSupportedException();
           if (lockTaken) Monitor.Exit(_tasks);
// The following is a portion of the output from a single run of the example:
//    'T' in task t1-4 on thread 3   'U' in task t1-4 on thread 3   'V' in task t1-4 on thread 3   
//    'W' in task t1-4 on thread 3   'X' in task t1-4 on thread 3   'Y' in task t1-4 on thread 3   
//    'Z' in task t1-4 on thread 3   '[' in task t1-4 on thread 3   '\' in task t1-4 on thread 3   
//    ']' in task t1-4 on thread 3   '^' in task t1-4 on thread 3   '_' in task t1-4 on thread 3   
//    '`' in task t1-4 on thread 3   'a' in task t1-4 on thread 3   'b' in task t1-4 on thread 3   
//    'c' in task t1-4 on thread 3   'd' in task t1-4 on thread 3   'e' in task t1-4 on thread 3   
//    'f' in task t1-4 on thread 3   'g' in task t1-4 on thread 3   'h' in task t1-4 on thread 3   
//    'i' in task t1-4 on thread 3   'j' in task t1-4 on thread 3   'k' in task t1-4 on thread 3   
//    'l' in task t1-4 on thread 3   'm' in task t1-4 on thread 3   'n' in task t1-4 on thread 3   
//    'o' in task t1-4 on thread 3   'p' in task t1-4 on thread 3   ']' in task t1-2 on thread 4   
//    '^' in task t1-2 on thread 4   '_' in task t1-2 on thread 4   '`' in task t1-2 on thread 4   
//    'a' in task t1-2 on thread 4   'b' in task t1-2 on thread 4   'c' in task t1-2 on thread 4   
//    'd' in task t1-2 on thread 4   'e' in task t1-2 on thread 4   'f' in task t1-2 on thread 4   
//    'g' in task t1-2 on thread 4   'h' in task t1-2 on thread 4   'i' in task t1-2 on thread 4   
//    'j' in task t1-2 on thread 4   'k' in task t1-2 on thread 4   'l' in task t1-2 on thread 4   
//    'm' in task t1-2 on thread 4   'n' in task t1-2 on thread 4   'o' in task t1-2 on thread 4   
//    'p' in task t1-2 on thread 4   'q' in task t1-2 on thread 4   'r' in task t1-2 on thread 4   
//    's' in task t1-2 on thread 4   't' in task t1-2 on thread 4   'u' in task t1-2 on thread 4   
//    'v' in task t1-2 on thread 4   'w' in task t1-2 on thread 4   'x' in task t1-2 on thread 4   
//    'y' in task t1-2 on thread 4   'z' in task t1-2 on thread 4   '{' in task t1-2 on thread 4   
//    '|' in task t1-2 on thread 4   '}' in task t1-2 on thread 4   '~' in task t1-2 on thread 4   
//    'q' in task t1-4 on thread 3   'r' in task t1-4 on thread 3   's' in task t1-4 on thread 3   
//    't' in task t1-4 on thread 3   'u' in task t1-4 on thread 3   'v' in task t1-4 on thread 3   
//    'w' in task t1-4 on thread 3   'x' in task t1-4 on thread 3   'y' in task t1-4 on thread 3   
//    'z' in task t1-4 on thread 3   '{' in task t1-4 on thread 3   '|' in task t1-4 on thread 3  

In addition, several sample task schedulers are available on Code Gallery: Samples for Parallel Programming with the .NET Framework 4.

Universal Windows Platform
Available since 8
.NET Framework
Available since 4.0
Portable Class Library
Supported in: portable .NET platforms
Available since 5.0
Windows Phone Silverlight
Available since 8.0
Windows Phone
Available since 8.1

All members of the abstract TaskScheduler type are thread-safe and may be used from multiple threads concurrently.

Return to top