[float] Forum Topics

I'm working on a program that requires some very low level optimization. My question is, when precision isn't an important factor, how can I automatically use the most efficient floating point type for whatever architecture the program is being run on? As far as I'm aware, floats are more efficient …