Method and system for performing perspective divide operations on

Abstract

A method and system for performing perspective divide operations on three-dimensional graphics data within a computer system is disclosed. In accordance with the method and system of the present invention, a 3-D graphical image is represented by three-dimensional graphical object data via a number of vertices. In a first iteration, the following steps are performed concurrently: computing a set of clip coordinates for a vertex N, wherein the set of clip coordinates is a four-component vector �x', y', z', w'!; refining an estimate value of 1/w' for a vertex N-1, wherein vertex N-1 is a vertex calculated one iteration previous to vertex N; and, generating a clip code for the vertex N-1 . Then, the following steps are performed in a next iteration: performing a clip test on the set of clip coordinates for the vertex N; generating an estimate value of 1/w' for the vertex N; and computing a set of normalized device coordinates for the vertex N-1 utilizing a refined estimate value of 1/w' for the vertex N-1 and a set of clip coordinates for the vertex N-1. By so doing, perspective divide operations can be performed more efficiently.