Engineers targeting DSP to FPGAs have traditionally used fixed-point arithmetic, mainly because of the high cost associated with implementing floating-point arithmetic. That cost comes in the form of ...