On Mon, 15 Feb 2010, Isaac Marino Bavaresco wrote:

> I understand, but this method doesn't use 16-bit math at all. All the
> math is done nibble-by-nibble, no carry is propagated during the additions.
> The only part that uses carry propagation is during the shifts, that
> shifts the whole 16-bit (or 9-bit in this case) value. It works for
> values with any number of digits.
> 
> If you don't already know this method, I would suggest to study it, it
> is very enlightening. Shows a very different approach that exploits
> curious properties of binary and BCD numbers. It is being used for a
> while by the CPLD/FPGA guys, because it is easily implemented using just
> shift-registers and adders.

Thank you Isaac I'll have a look at that later (got to get the C18 
compiler and library first) :-)

I vaugely remember a method using half carrys but not well enough to do it 
off the top of my head. I also know of a method using multiply by 10 
(which itself can be broken down to shifts and add) but a method not 
propagated carry in additions? I'm not sure I've seen that one or if I 
have I've certainly forgotten all about it.

Thanks
Sergio
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist