Thursday
Room 1
13:40 - 14:40
(UTC+02)
Talk (60 min)
Speedrun Into Massive Data: Using SIMD in C#
.NET speeds up every year, partially due to SIMD—Single Instruction Multiple Data.
Starting with the basics of SIMD and why it’s a good tool in your toolbox, we’ll uncover how System.Numerics.Vectors make leveraging SIMD surprisingly accessible, abstracting the complexities of hardware-specific optimizations like SSE2, SSE4, and friends.
With hands-on examples, we will see how SIMD works in .NET, where it shines, and where it does not (of course, with benchmarks). We will also cover things like CUDA and how we can write .NET code that runs on your graphics card!
After that session, you will have a new tool on your belt that can help you squeeze out every ounce of performance when it matters most while knowing when to stick to simpler approaches to keep your code clean and maintainable.
Steven Giesel
As an experienced .NET Engineer and a Microsoft MVP in Developer Technologies in the heart of Zurich, Steven Giesel bring a substantial breadth of knowledge and expertise to each project.
His enthusiasm for technology goes beyond his professional endeavors - he actively maintains various open-source libraries (for example as co-maintainer for bUnit - a unit testing library for Blazor) and frequently speaks at user groups and conferences.
Furthermore, he has authored concise e-Books in .NET related topics, and peer-reviewed a book focused on building applications with Blazor WebAssembly.