Using a C compiler for the specific MCU you use will enable you to have variables of float types and supportes also math (multiplication / division / root, even some trigonometry). Also when having variables and a C language enviroment is no problme 'store results' in memory.... However all these ends up with a lot of code size increase, since float math nedds a lot of code to be implemented, but this should be no major problem if you have decent FLASH memory (eg 64 K or more...). Also, MCUs support integer math internally and up to multiplication / division, so doing a sqrt or a sine is also time consuming compared to sipler 8 / 16 bits math.
What's the need for doing such math ?