Question on int vs. floating point math

[Fur]

So, in trying to optimzie the performance of some of my scripts, I was going to look at converting some floating pt calcs into integer ones. But in reading the help file it looks like all math is going to be done in floating point anways. And doing bitwise shifts to multiply and divde look like they would convert to int, shift and convert back to double, just increasing overhead? Are these all correct assumptions?

Looks like I'll just need to move my stuff into a DLL and use DLLCall() to do the grunt work.