Alberto

> When I read the title, I assumed this was a SIMD implementation of `alphablit_colorkey`. As far as I can understand, that's not it at all. What you've made is...

I believe that with pygame having "game" in its name and many softwares made with it being pixel art games, this change is really important and worth implementing it directly...

A little test program for reproducibility (without this PR i get around 200/250 fps and with this PR 470/520): Images:     ```Py import pygame from random import...

I don't get if this is a strong/weak no or a strong/weak yes @Starbuck5.

May I remind you of these? - https://github.com/pygame-community/pygame-ce/pull/1839 - https://github.com/pygame-community/pygame-ce/pull/1841 - https://github.com/pygame-community/pygame-ce/pull/1834 - https://github.com/pygame-community/pygame-ce/pull/1835

And this: https://github.com/pygame-community/pygame-ce/issues/2208

I don't know if you want to move this forward or what but i have done a lot of testing and have some findings to share. Basically the LOOP_UNROLLED isn't...

Anyway this implementation could still use some simplifications like we don't actually need two shuffle masks for the 16 bit shuffle macro since we can directly use `_mm256_unpacklo_epi8` / `_mm256_unpackhi_epi8`...

But I'd really encourage you to revive this PR if possible or maybe consider working on this together? Your choice.

I've checked on this and I believe this one is on SDL.