Um. Excuse me being way OT here but,
is this sort of thing really worth writing in assembly or is this for personal edification. Making a bunch of OS calls can be done more easily (less typing) in C and will be easier to port to PPC that way. Why not keep all this stuff in C and do all the fancy pants number crunching and whatever really needs optimizing in assembley?
Anyway, just to drag myself back to near the topic ...
have you been
here?
Might have something useful.