Byron A Jeff wrote:

>Algorithm 2 is definitely along the right lines. It's better because it
>takes a constant amount of time to execute no matter which bit is being =
set.
>In my bytecode interpreter I used the
>following fragment which is analogous to Algorithm 2

BCF STATUS, C
RLF DATAIN, W
ANDLW 0x0F
ADDWF PCL, F
BSF INDF, 0

I have well simulated your code moving DATAOUT to FSR
and it takes 12 cycles, cool, but looking at Scott proposal,
I see:
1.No PCLATH issues,
2.It don't cares what the mask is used for: BFS or BCF (IORWF or XORWF)
3.And it takes 10 cycles max.

But I'll steal your idea to re-arrange some parts of the code.

To Scott, many thanks for this "piece of art".
Is very simple and very tricky and very helpful.
I saw it first on your pwm256, but until now,
I didn't "really *see* it" at all.
After a few seconds I found the original code
from Mike Keitz on the PicList web site and the file is mask.htm
(techref/microchip/math/bit/mask.htm?key=3DMike+Keitz&from=3D)
So I have to thanks and recognize the
great job all the admins do for us.


_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com


--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.