Bringing some of the new GBA renderer optimizations to uxngba :uxn:
I'm specially happy with the new way I'm doing blending modes, removing per-nibble for loops in favor of some bit wrangling wizardry.
2.9x speedup for 2bpp sprites (chr) and 2.55x for 1bpp (icn).
Still have some unmerged changes with an updated uxn core, since I have that on a separate branch where I was doing oquonie work, but will be there soon.