Automatic Floating to Fixed Point Translation and its Application to Post-Rendering 3D Warping

The automatic conversion of oating point software implementations of algorithms to a equivalent xed point implementation which can be eeciently implemented in an FCCM remains an obstacle in the rapid systems prototyping design ow. Floating point to xed point conversion is tedious, error prone and requires a good knowledge of xed point computer arithmetic. This paper describes a software system called fp designed to automate the process. It consists of a xed point C++ class; a prooler which is used to determine the number of bits of precision required for each signal in the hardware implementation; an optimiser which nds the minimal number of bits required for a speciied degree of accuracy in the implementation and nally and a compiler which takes the information collected by the system and outputs synthesisable VHDL code. A post{rendering 3D image warping application designed using this system is used as an example.