Egads!  As I was typing up that message, there was a fresh
entry from Dmitriy -- very slick solution, that now takes the
top spot with the fewest # of steps.  I haven't actually tried it,
but a mental walk-thru tells me it should work.  And it actually
uses a neat XOR swap technique.  My hat's off to you sir.

Cheers,
-Neil.



----- Original Message -----
From: "Pic Dude" <picdude@AVN-TECH.COM>
To: <PICLIST@MITVMA.MIT.EDU>
Sent: Sunday, March 10, 2002 12:01 AM
Subject: [PIC}: Re: Bit operation quiz...


> Ack! forgot the [PIC]: prefix.  Anyway, been looking thru the
> code snippets that were posted, and my resulting thought is that
> I need to stop making life so complicated on myself.  In my
> past life, when I coded assembly on the TRS-80, I remember
> something about conditionals being very expensive, so I tend
> to avoid them.  Gotta get into the mindset that things are not
> the same anymore.  As Bob Ammerman demonstrated, the
> solution with the fewest steps, was the most brute force method
> possible -- to test each of the 4 lowest bits and set the
> corresponding bit in the result.  Only 11 steps!
>
> That said, I must say that I like the way Tom Messenger thinks
> -- really slick dataflow.  Sequentially squeeze the 4 lowest bits
> out the right and push it into the result from the right (to the left).
> Tom, as for the other solutions, I am actually running out of
> space on the F872.  I'm sure however, that I can do a lot of
> cleanup on my code -- this is still my first PIC code ever.  As
> for wiring one port into the other -- nice thought, but I started
> this bit-reversal thing to clean up wires, and the other ports are
> in use.
>
> Jinx, your solution had the top 4 bits swapped.  With some
> code adjustments I managed to get the result in 14 steps.
>
> Nigel unfortunately I'm not using C, but your code flow is
> pretty much the same as Bob's -- check each bit and set the
> output bits accordingly.
>
> This was fun!  Must do this again sometime! :-)
>
> Thanks much guys!
> -Neil.
>
>
>
> ----- Original Message -----
> From: "Pic Dude" <picdude@AVN-TECH.COM>
> To: <PICLIST@MITVMA.MIT.EDU>
> Sent: Friday, March 08, 2002 9:29 PM
> Subject: Bit operation quiz...
>
>
> Anyone up for a pop quiz?...
>
> I have all Port C outputs going to 8 LED's.  To save PCB space, I can
> re-route
> the 4 lowest lines to the LED's, but the last 4 bits get reversed.  What's
> the fastest
> way, in code, to reverse just those 4 lowest bits, and leave the 4 highest
> bits
> untouched?  Fastest = smallest # of steps.
>
> Assume input = ABCDEFGH, then output should be ABCDHGFE.
>
> So far I've come up with this (untested), but it's seems like a very
> "sledgehammer" way to do it.  There's gotta be something slicker.  I
> remember learning about the XOR swap trick many years ago, and
> thought it was the coolest thing ever (at the time).  Any thing like
> that exist for doing this partial reverse?
>
>
> movf    DATAIN,TEMP
> rrf    TEMP,F
> movf    TEMP,W
> andlw    H'02'
> movwf    DATAOUT    ; Bit 1 done
> rrf    TEMP,F
> rrf    TEMP,W
> andlw    H'01'
> andwf    DATAOUT,F    ; Bit 0 done
> movf    DATAIN,TEMP
> rlf    TEMP,F
> movf    TEMP,W
> andlw    H'04'
> andwf    DATAOUT,F    ; Bit 2 done
> rlf    TEMP,F
> rlf    TEMP,W
> andlw    H'08'
> andwf    DATAOUT,F    ; Bit 3 done
> movf    DATAIN,W
> andlw    H'F0'        ; Strip out low nibble
> andwf    DATAOUT,F    ; Merge with result
>
>
> Cheers,
> -Neil.
>
> --
> http://www.piclist.com hint: PICList Posts must start with ONE topic:
> [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
>
> --
> http://www.piclist.com hint: The PICList is archived three different
> ways.  See http://www.piclist.com/#archives for details.
>
>
>

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