Hi Harold,

There are a few errors here and there in that document. They use duplicate
names for the same file registers, which made the learning curve rather
steep for me.

I did use umul0808 as the model for an 8x8 multiply that I use, but
renamed the registers to suit my needs.

I was having a heck of a time figuring out the 32 bit divide routines they
had, getting errors aplenty, and looked at the Microchip web site. They
had an updated an617 with the errors(at least in the 32 bit divide) fixed.

You might check to see how recent your an617 is.

One thing I can say, I learned quite a bit about multiple precision math
routines while trying to verify theirs!

Sympathetically,

Bob

On Wed, 17 Sep 1997, Harold Hallikainen wrote:

>         I note that the UMUL0808 macro in AN617 is quite a bit faster
> than the UMUL0808L, yet in FXM0808U, they call UMUL0808L.  I tried
> dropping UMUL0808 in there instead and got some rathr strange behavior
> (which I can't fully describe since I was in the crash and burn
> development mode).  Anyone have experience with UMUL0808?
>
> Thanks!
>
> Harold
>