Re: [f-cpu] about the scheduler...

[Michael Riepe <michael@stud.uni-hannover.de>]

On Fri, Aug 02, 2002 at 02:04:15AM -0700, jaap stolk wrote:
[...]
> ah, i think i get your point now, as far as i
> can see, its ok to write to a register that is
> "being computed" as long as the second write will
> be performed _after_ the first one.
[...]

Or at the same time. In that case, the second result should overwrite
the first; that is, the second write takes precedence. This is a rare
case but it may happen with 2r2w instructions if only the `secondary'
result is used, e.g.:

    addc r1, r2, r4 // we only want the carry bits from r5
    and r6, r7, r4  // re-use of r4 overwrites previous result

`addc' and `and' may write at the same time, with the `and' instruction
taking precedence.

-- 
 Michael "Tired" Riepe <Michael.Riepe@stud.uni-hannover.de>
 "All I wanna do is have a little fun before I die"
*************************************************************
To unsubscribe, send an e-mail to majordomo@seul.org with
unsubscribe f-cpu       in the body. http://f-cpu.seul.org/