rounding up a floating point number

In the absence of ceil() can anyone give me any ideas on rounding up a large floating point number?

My first inclination is to use fixed integer math, perhaps there is a better method?