On Thu, 18 May 2000, Marcelo Puhl wrote:

> Hi,
>
> Is there a way to reduce the size of this routine ?

Absolutely!

>
> Polynomial  = x0 + x12 + x17
>
> x0 is in "data" bit_0
> x12 is in "mid" bit_3
> x17 is in "old" bit_0
>
>                 rlf     mid,w        ; x12 in w bit_4
>                 movwf   temp         ; x12 in temp bit_4
>                 swapf   temp,w       ; x12 in w bit_0
>                 xorwf   old,w        ; x12 + x17 in w bit_0
>                 xorwf   data,w       ; x0 + x12 + x17 in w bit_0
>                 movwf   temp
>                 rrf     temp,f       ; x0 in Carry
>
>                 rlf     new,f        ; x0 to bit_0
>                 rlf     mid,f        ; Updates x12
>                 rlf     old,f        ; Updates x17
>
> Thanks for your help.

how 'bout

   rrf    data,w
   rlf    new,f     ;copy data.0
   rlf    mid,f     ;2nd part of 24bit shl

   swapf  mid,w     ;get mid.4 (originally mid.3)
   xorwf  old,w     ;xor'd with old.0
   rlf    old,f     ;complete the 24bit shift

   andlw  1         ;add old.0^mid.3
   xorlw new,f      ;to new.0

Saves a couple of cycles

Scott