Yeah indeed, but I definitely do that a the end, after the asm lib/axiom is complete
Also I like how I have speed up triangle filling code : asm version is 4x faster than axe version XD, and is +-300 bytes smaller... that put the theorical triangle output rate at almost 192 Tri/s, almost 2x more than the axe version ... and I haven't finished to optimize filling and many other things.
Also I like how I have speed up triangle filling code : asm version is 4x faster than axe version XD, and is +-300 bytes smaller... that put the theorical triangle output rate at almost 192 Tri/s, almost 2x more than the axe version ... and I haven't finished to optimize filling and many other things.