Feedback Normalization

Feedback normalization is the term used to describe the conversion of feedbacks to a standard polarity. The standard most often used is the polarity observed on the device output pin. Since feedbacks (both registered and combinational) can originate from a wide variety of macrocell locations, it is important to normalize feedbacks so that the same design description can be mapped to a wide variety of devices.

Register feedback is normalized to one of two locations in a device. When a feedback is specified as .FB, or with no dot extension, the signal fed back to the array is normalized to the output pin, regardless of the actual source of the feedback. When register feedback is specified with .Q, the signal fed back to the array is normalized to the Q output of the associated flip-flop — again regardless of the actual source of the feedback.