Math.Net Numerics Beta 2, plus a competition for a fast matrix multiply

  • Article

The contributors to Math.NET have just released their final beta for Math.NET Numerics. This wonderful package includes both NuGet packages and F# integration! From the Math.NET Numerics site:

Download the MathNet.Numerics.dll assembly, add a reference to it to your project and you're done. To make this even easier we also publish binary releases to the NuGet Gallery as package MathNet.Numerics (or MathNet.Numerics.FSharp for F# integration), so you can add a managed reference directly in Visual Studio. More details on the quick start page.

They are also running a competition for a fast matrix multiply routine - you can win $1500... Here's the announcement:

Two years ago, the people who developed Math.Net Iridium and dnAnalytics, both open source numerical libraries for .NET, came together with the idea of creating one great numerical library for .NET. Based on our combined experiences, we designed a new library and called it Math.Net Numerics.

During the past two years, with financial support from Microsoft Research and together with some great new developers on the team, we are on the final leg of delivering Math.Net Numerics 1.0: the last release of Math.Net Numerics beta.

The .NET platform is a good place for scientific computing. Modulo a few essential things we'd like to see in the future, e.g. integration with SIMD instructions, the .NET runtime allows us to write fairly efficient numerical code. Nonetheless, there are places where improvements should be possible and we'd like to invite you to help us out.

Starting from today, we're running a contest on writing the most efficient matrix multiplication routine in .NET. We will incorporate the best code into Math.Net Numerics before we release v1.0 of the library. As a thank you, the winner will receive a 1500$ prize! (The runner up will win 500$.) The contest rules and benchmark harness are up at <gemm.codeplex.com/>. We hope you enter our competition and share your insights on making the best .NET numerical code out there!