In your common FFT program, you want to make sure that the frame you're working with has a size that is a power of 2. This tells you in just a few operations. Granted, you won't be using this algorithm inside a loop, so the savings aren't that great, but every little hack helps ;)

// Quit if size isn't a power of 2

if ((-size ^ size) & size) return;

// If size is an unsigned int, the above might not compile.

// You'd want to use this instead:

if (((~size + 1) ^ size) & size) return;