I have been writing programs with a 32k code size limitation license, I think. I got the message:
C5300: Limitation Code Size 1969 > 1024 bytes
I have compiled much larger code than that in the past. I still can too. The difference this time is that I am compiling a test program with 32 bit floating point.
Does floating point have a more severe licensing limitation?
it would be good to know which architecture/compiler (S12? S08?) you are using, and which tools. But I know that C++ code had a different 'free' limitation of 1K of code. So could it be that you are using somehow C++ code or libraries?
I am programming a 9s12xdt512 demo board with CodeWarrior. I have compiled the code as c++ and as c with both using floating point math. Both produce the same error. It looks like the floating point compiles up a lot of code fast. The library alone is compiling up 21509 bytes, startup 62 bytes, main.c is 11 bytes, datapage.c is 546, an my code causes an error somewhere after 1025 bytes. So it appears that an error occures at just over 23153 bytes total used.
Arduino compiles no problem. I would like to do a speed comparison, but I can't.