[Discussion] SIMD

[Tryibion]

Any thoughts on using the .NET SIMD as underlying types? I assume there would be a performance increase. There was quite a bit of work done on SIMD for .NET 10.

I can take a stab at implementing it if you believe it is worth the time.

[PaperPrototype]

Hmm, this would be cool. cc @michaelsakharov it I think if this does get implemented maybe do it as a separate project within this one (eg. Prowl.Vector.SIMD) with static SIMD methods for different math ops using [MethodImpl(MethodImplOptions.AggressiveInlining)] for each, and then progressively update the math in each of the Prowl.Vector to use them over time.

[PaperPrototype]

Also note Prowl.Vector 2.0 branch is the latest branch and is already in use in most of the Prowl libraries (README is very out of date in both branches), make sure to check it out: https://github.com/ProwlEngine/Prowl.Vector/tree/2.0

Example usage of Prowl.Vector 2.0 is here: https://github.com/ProwlEngine/Prowl.Vector/blob/2.0/Samples/SoftwareRasterizer/RaylibDemo.cs

[michaelsakharov]

SIMD support would be nice to have, I honestly dont know much about SIMD myself, so if you wanna take a stab at it i would greatly appreciate it. Though A bit part of Prowl Libraries is to follow KISS, so im a bit on edge of SIMD's performance gains outweigh the complexity and if its borderline crossing outside the KISS guideline.

So open to discussing what the implementation would look like exactly to get a better idea of if its complexity isn't worth the gains.
Especially since i think Prowl.Vector being aimed at game development, it may not see many cases where SIMD shines.