Modernize PorterDuffFunctions with operators and add tests

[JimBobSquarePants]

Prerequisites

• I have written a descriptive pull-request title
• I have verified that there are no overlapping pull-requests open
• I have verified that I am following the existing coding patterns and practice as demonstrated in the repository. These follow strict Stylecop rules 👮.
• I have provided test coverage for my change (where applicable)

Description

This pull request introduces AVX-512 (Vector512) support for image blending operations in ImageSharp, enabling faster processing for large spans of pixel data on hardware that supports AVX-512 instructions. It also refactors and extends several vectorized helper methods to support the new vector width, and moves some utility methods to more appropriate files. The main focus is on improving performance and maintainability by leveraging the latest SIMD capabilities.

AVX-512 SIMD support for pixel blending:

• Added AVX-512 (Vector512) code paths to the generated pixel blender implementations in DefaultPixelBlenders.Generated.tt, enabling batch processing of 16 floats at a time for various blend functions. This includes handling for both constant and variable opacity, as well as proper remainder handling for non-multiple-of-four lengths. [1] [2] [3] [4]

• Implemented AVX-512 (Vector512) versions of the main Porter-Duff blend functions (Src, SrcAtop, SrcOver, SrcIn) in PorterDuffFunctions.Generated.tt, using AVX-512-specific blend and mask intrinsics. [1] [2] [3] [4]

Vectorized math utilities and refactoring:

• Added new UnPremultiply and MultiplyAddNegated methods for Vector512<float> in Numerics.cs and Vector512Utilities.cs to support AVX-512 math operations. [1] [2]

• Moved and refactored MultiplyAddNegated for Vector256<float> from SimdUtils.HwIntrinsics.cs to Vector256Utilities.cs for better organization and maintainability. [1] [2]

Benchmarks

NormalSrcOver shows significant improvement.

BenchmarkDotNet v0.15.8, Windows 11 (10.0.26200.8117/25H2/2025Update/HudsonValley2)
AMD RYZEN AI MAX+ 395 w/ Radeon 8060S 3.00GHz, 1 CPU, 32 logical and 16 physical cores
.NET SDK 10.0.201
  [Host]     : .NET 8.0.25 (8.0.25, 8.0.2526.11203), X64 RyuJIT x86-64-v4
  DefaultJob : .NET 8.0.25 (8.0.25, 8.0.2526.11203), X64 RyuJIT x86-64-v4

Method	Mean	Error	StdDev	Ratio
Scalar	384.5 ns	0.44 ns	0.35 ns	1.00
Avx2	181.1 ns	0.21 ns	0.20 ns	0.47
Avx512	102.5 ns	0.08 ns	0.08 ns	0.27