Here comes another necro. This thread is the top result on Google for "game maker studio int to float", and the information in here is wrong, so I thought I'd save some fellow Googlers Of The Future a bit of annoyance. Maybe things were different in 2018, but in 2024, the answer is to use the real() function.
I was having a problem where I was dividing to get a percentage, like this:
a is always smaller than b, and x was always 0. I tracked the problem down to GMS treating a and b as ints. Per the manual, "If A and B are integers (either int32 or int64) then the division will be done as integers (and divide by 0 will be checked and error'd). Otherwise it will be done as a floating point division (with no divide by 0 check)."
I changed it to this:
x is now a proper percentage because real() forces GMS to treat a and b as floats (or "reals", as they're called in GMS) and thus performs non-integer division. Interestingly, x doesn't need to be forced into being a real, or declared with a decimal point. GMS is smart enough to recognize that the act of performing non-integer division and storing it in x means x must be a non-integer number. Nifty!