Floating point numbers/calculations

tsfg

In searching through the Parallax website and this forum, it appears that none of the Basic stamp products can perform floating point stuff...at least, not without an add-on math co-processor. Is this true?

I have been playing around with the CUBLOC microcontroller from Comfile and it does floating point (and a whole lot more) at a lower cost (although, I gotta tell you, their basic software IDE REALLY sucks... in fact, it doesn't even work! so what's the point, eh?).

Is the co-processor the only way to go? I've had sucess with Tracy Allens method for counting beyond 65535, but doing real calculations is giving me a heaadache.

Thanks....Terry

Jon Williams

If your FP requirementss are very simple, then you can "fake" it with fixed-point values. If you need something beyond that, I think the FPU is probably the cleanest solution. If one thinks of the FPU like we treat serial LCDs or servo controllers then it's really not a big deal. What I like about the um-FPU is that it can be independently preprogrammed for specific (very complex) functions; this allows the BASIC Stamp to send it a raw (input) value, request a specific calculation on that value, then retrieve the result in the desired format.

If the CUBLOC works for you, great, but let's not start any debates about price and features here -- this forum is for supporting Parallax products, NOT for debating the relative merits of our products versus anyone else. By the same token, we never like seeing anyone criticising another vendor (in their support forums) while pointing to us as "better."

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Jon Williams
Applications Engineer, Parallax