Implement [ Variable ] Precision Floating Point [ Arithmetic ] in CAS Mode

[Wes Loewer]

I don't know if this would be of any help to anybody, but some 20+ years ago, I wrote the arbitrary precision floating point library that was used in Fractint. The copyright on the library was exceedingly loose (give credit, etc...). I had to google to find it myself, but it can be found at http://www.nahee.com/spanky/www/fractint/arb-prec.html

Originally, it was written mostly in 80x86 assembly and some C (in Fractint speed was everything), but I later rewrote the assembly into C for general use and ports to xwindows. The lib had arbitrary length binary floating point, decimal floating point, and binary fixed point formats. (Fixed point was especially useful for generating those deep zoom fractals.)

If I needed a library today, I'd use one of the professional quality GPL/LGPL libraries that are available, but if anybody needs something without those licenses, feel free to use mine. Just keep in mind that it was written for a specific purpose (Fractint) and as such does not have the IEEE rounding, nan, error checking and such. (Who cares if a pixel is off by 1?)

The documentation has the details of use, but to those I'll add:

Quote: *** No publicly making fun of my amateurish code. *** :-)

Also keep in mind that I have not looked at the code in many years. I have no idea what you might find. ;-)

-wes