.NET 11 Numerics

Presented by dotnet, this video looks at how .NET 11’s numeric types fit into generic math and how to choose between precision and range requirements.

Numeric types and APIs

The .NET 11 numerics additions covered in the video include:

• IEEE 754 decimal floating-point types
• BFloat16
• Generic complex numbers
• New generic random-number APIs

The discussion connects these additions to generic math and covers rounding, along with practical improvements to parsing and formatting.

Tensor performance and correctness

The video also covers improvements to System.Numerics.Tensors. Layout-aware tensor operations improve performance while handling shapes, overlapping memory, and special values consistently.

These topics bring both performance and correctness into focus: tensor operations need to account for how data is laid out in memory as well as the values being processed.

Looking ahead to .NET 12

The closing segment previews further tensor work planned for .NET 12, keeping those plans separate from the improvements included in .NET 11.