Sorry, I live in the 14 bit PIC world... I guess I should do a little
research before putting in my 2 bits!!!
:)

----------
> From: Jennifer Wilson <jenniferx@JUNO.COM>
> To: PICLIST@MITVMA.MIT.EDU
> Subject: Re: Absolute Difference Algorithm
> Date: Wednesday, January 14, 1998 12:24 PM
>
> On Wed, 14 Jan 1998 12:02:07 GMT Lynn Richardson
> <REMOVE_TO_REPLY_jnlrich@qni.com> writes:
> >Not elegant but seven instructions and free:
>
> >    movf    X,W             ; get first number.
> >    subwf   Y,W             ; sub second number.
> >    bc      AA              ; result positive? - Yes, go to AA.
> >    xorlw   h'FF'           ; convert from two's compliment
> >    addlw   1               ;   negative to positive.
> >AA  subwf   DIFF,w          ; sub difference from DIFF.
> >    bc      LESS            ; result + or 0? - Yes, goto LESS.
> >
>
>     I think that will assemble to 9 instructions when the psuedo-ops are
> expanded, equal to the original...
>
> >->
> >->          MOVF       X,W        ;Try it this way
> >->          SUBWF   Y,W
> >->
> >->          SKPNC                  ; Positive?
> >->          GOTO      POS      ; Yes
> >->
> >->          MOVF       Y,W       ; No, then do it the other way
> >->          SUBWF   X,W
> >->
> >->POS  SUBWF  DIFF,W  ; Compare against the limit
> >->
> >->          SKPNC                  ; The carry knows...
> >->          GOTO     LESS
>
>
> On Wed, 14 Jan 1998 10:22:48 -0700 Justin Crooks <jcrooks@WYOMING.COM>
> writes:
> >That's pretty close to perfect.  You could also try this...
> >
> >        movf    X,W
> >        subwf   Y,W
> >
> >        skpc
> >        sublw   0x00    ;If it is negative, make it positive.
> >
> >        subwf   DIFF,W
> >        skpnc
> >        goto    LESS...
> >
> >        Not much of an improvement, but maybe a little cleaner :)
> >
>
>     Correct me if wrong but, I don't believe that "SUBLW" is part of the
> 12 bit PIC world...
>
>     Andy Warren, who reputation exceeds himself in this venue, sent me a
> version using the same logic employing a file register either
> conveniently or purposefully cleared to 0.
>
>                 ********************************
>
>    Perhaps, I should expand a bit on the conditions...
>
>    If it helps, DIFF can be a constant instead of a file register.
>
>    X & Y are in the range of 0 to 32
>
>    DIFF ranges between 4 & 9
>
>    It would be nice to optimize this, as there are several similar
> operations in the application.
>
> Still more thanks in advance!
>    Jen :-)