Book Image

Hands-On Parallel Programming with C# 8 and .NET Core 3

By : Shakti Tanwar
Book Image

Hands-On Parallel Programming with C# 8 and .NET Core 3

By: Shakti Tanwar

Overview of this book

In today’s world, every CPU has a multi-core processor. However, unless your application has implemented parallel programming, it will fail to utilize the hardware’s full processing capacity. This book will show you how to write modern software on the optimized and high-performing .NET Core 3 framework using C# 8. Hands-On Parallel Programming with C# 8 and .NET Core 3 covers how to build multithreaded, concurrent, and optimized applications that harness the power of multi-core processors. Once you’ve understood the fundamentals of threading and concurrency, you’ll gain insights into the data structure in .NET Core that supports parallelism. The book will then help you perform asynchronous programming in C# and diagnose and debug parallel code effectively. You’ll also get to grips with the new Kestrel server and understand the difference between the IIS and Kestrel operating models. Finally, you’ll learn best practices such as test-driven development, and run unit tests on your parallel code. By the end of the book, you’ll have developed a deep understanding of the core concepts of concurrency and asynchrony to create responsive applications that are not CPU-intensive.
Table of Contents (22 chapters)
Free Chapter
1
Section 1: Fundamentals of Threading, Multitasking, and Asynchrony
6
Section 2: Data Structures that Support Parallelism in .NET Core
10
Section 3: Asynchronous Programming Using C#
13
Section 4: Debugging, Diagnostics, and Unit Testing for Async Code
16
Section 5: Parallel Programming Feature Additions to .NET Core

Introduction to distributed systems

We have already discussed how distributed computing works in this book. In this section, we will try to understand distributed computing with a small example that works on arrays.

Let's say we have an array of 1,040 elements and we would like to find the sum of all the numbers:

a = [1,2,3, 4...., n]

If the total time that's taken to add numbers is x (let's say all of the numbers are large) and we want to compute them all as fast as possible, we can take advantage of distributed computing. We would divide the array into multiple arrays (let's say, four arrays), each with 25% of the original number of elements, and send each array to a different processor to calculate the sum, as follows:

In this arrangement, the total time that's taken to add all the numbers is reduced to (x/4 + d) or (x/number of processors +d), where...