Smoothscale assembly routines should be converted to intrinsics (3586)

[GalacticEmperor1]

Issue №3586 opened by Starbuck5 at 2022-11-24 06:19:47

This is the only handwritten assembly in Pygame, and it is stuck in the past. Modern assembly we use "intrinsics" which can be written as lines of C and translate into assembly. This is more maintainable and portable, even allows header-file-translation to other SIMD standards (like we have an SSE2Neon header to convert x86 SSE routines to ARM Neon routines).

See scale_mmx.c and related files VS something like simd_blitters_sse2.c

---

Comments

[Starbuck5]

I've been working this up for a bit, I'll have an initial PR out soonTM.

[MyreMylar]

I think we should drop the mmx backend in the SDL 3.0 transition.