Parallel Programming in .NET

Parallel programming lets you take advantage of multiple CPU cores to improve performance and responsiveness. The .NET runtime provides a rich set of APIs that simplify writing concurrent code.

Key Concepts

Task Parallel Library (TPL) – high‑level abstractions for parallelism.
Parallel Class – simple parallel loops and invocations.
PLINQ – Parallel LINQ for data‑parallel queries.
async/await – asynchronous programming with lightweight tasks.
Dataflow – pipeline‑style processing with System.Threading.Tasks.Dataflow.

Parallel Class

The Parallel class provides static methods for parallel loops and invokes.

using System;
using System.Threading.Tasks;

class Example
{
    static void Main()
    {
        Parallel.For(0, 10, i =>
        {
            Console.WriteLine($"Task {Task.CurrentId} processing index {i}");
        });
    }
}

Parallel LINQ (PLINQ)

Transform sequential LINQ queries into parallel ones with AsParallel.

var numbers = Enumerable.Range(1, 1_000_000);
var parallelResult = numbers
    .AsParallel()
    .Where(n => n % 3 == 0)
    .Select(n => n * n)
    .Take(10)
    .ToArray();

Task Parallel Library (TPL)

Use Task for fine‑grained asynchronous work.

Task<int> ComputeAsync(int value)
{
    return Task.Run(() =>
    {
        // Simulate intensive calculation
        Thread.Sleep(500);
        return value * value;
    });
}

Dataflow (Pipeline)

Build robust processing pipelines with BufferBlock and TransformBlock.

using System.Threading.Tasks.Dataflow;

var input = new BufferBlock();
var square = new TransformBlock(x => x * x);
var print = new ActionBlock(x => Console.WriteLine(x));

input.LinkTo(square);
square.LinkTo(print);
for (int i = 1; i <= 5; i++) await input.SendAsync(i);
input.Complete();
await print.Completion;

Best Practices

Avoid blocking calls inside parallel loops.
Prefer async/await for I/O‑bound work.
Limit degree of parallelism for CPU‑bound tasks (ParallelOptions.MaxDegreeOfParallelism).
Use cancellation tokens to support graceful shutdown.

.NET Parallel Programming