And with modern 8052's operating at anywhere from 12 MIPS up to close to 100 MIPS it's not even clear whether floating point is necessarily a problem even with speed-sensitive applications. I think things have changed a lot since the era when the "floating point shouldn't be used on the 8052" rule was originally formulated. That may have been true when typical 8052 designs operated at 1 MIPS but it isn't as strong an argument these days with MCUs that are one or two orders of magnitude faster.

I do agree with Andy, of course, that non-FP approaches should be explored first and I also believe that in most cases floating point isn't necessary. But I think it should be avoided in the name of efficiency and not necessarily in the name of lack of capability on the part of the processor.

Just my two cents.

Regards,
Craig Steiner