Hi Dexen,

> IIRC, the Amiga's instructions set guarantees that every function is 4-byte 
> aligned anyway, on round 4-byte boundary, simply because every instruction is 
> exactly 4 bytes (32bits). If that's the case, perhaps it could be enough to 

Strange. But why did then Jakob report that the "Unaligned Function"
error appeared at startup? Doesn't Amiga use an 68k CPU (which has
instructions of variable byte sizes)?


> @Alex
> are all functions in miniPicoLisp aligned to 2-byte boundary on IA-32? 

Yes, aligned, but to 4-byte boundaries, so that addig 2 gives the xxxx10
pattern.

> Wouldn't that hamper performance -- as opposed to the preferred 4-byte 
> alignment?

Not really, just a simple masking is involved. As this is an in-memory
operation, the penalty is not noticeable.

Cheers,
- Alex
-- 
UNSUBSCRIBE: mailto:picolisp@software-lab.de?subject=Unsubscribe

Reply via email to