On Fri, 6 Nov 1998 13:38:21 -0500 myke predko <myke@PASSPORT.CA> writes:
>Recently, I put up a 16 bit multiplication algorithm on my web page
>along
>with a 32 bit result.  Oyvind Kaurstad has pointed out to me that the
>addition routine that is on the web page does not pass the correct
>carry
>value the case:
>
> 0x0FF
>+0x001
>
>when the Carry Flag is already set.

The correct result is of course 1 with the carry set.  The classic method
for adding 8 bits with carry works in every case:

        movfw   A
        skpnc
        incfsz  A,w
        addwf   B,f

Anytime the C is set going in, the incfsz executes, replacing the value
loaded into W by the first instruction with A+1.  If A is FF, the add is
skipped, since it would clear the carry out by adding zero.  So if A =
FF, B = 1, and C = 1, the result left in B is unchanged (1) and the carry
flag is unchanged (1).

If A= 1, B = FF, and C = 1, then the incfsz moves A + 1 (which is 2) into
W and doesn't skip.  The add executes adding FF and 2 which is 1 with a
carry of 1.  Again the result is correct.

In order to add larger numbers with carry just repeat this routine for
each byte in the number starting with the LSB.




___________________________________________________________________
You don't need to buy Internet access to use free Internet e-mail.
Get completely free e-mail from Juno at http://www.juno.com/getjuno.html
or call Juno at (800) 654-JUNO [654-5866]