Focus on ARM and ARM NEON optimizations

[recp]

trafficstars

Since macOS is switching to Apple Silicon, it is time to do more optimizations on ARM side. FWIW, cglm already applies optimizations to almost all vec4 operations and some mat4 operations.