Now is the time to ask, "do I really want to spend my time & effort doing this all by hand in assembler, or should I use 'C'?"

Why not get it going in 'C' - then see if there's any bits that need optimising in 'C', then see if there's still any bits that need optimising in assembler?

Browse down the Subjects list and you will see that there was extensive discussion of this a couple of weeks back!
Or try a search on "optimise" (or "optimize").