I've found the math library. It seems to be a bit large though. I hoped to minimize code size as much as possible

Floating point processing is very complicated, and as such tahes a lot of code space. I recommend you try to avoid it. A simple technique is to keep the decimal point in the same place throughout and just insert it in the output.
say you have 123.45 and 234.56
treat as 12345 and 23456 till you display the value, then just pop a dot in before the last 2 digits.

Have fun,

Erik