Browse Prior Art Database

Faster Geometric Rendering by Referencing Shared Vertices in Polyhedra

IP.com Disclosure Number: IPCOM000122047D
Original Publication Date: 1991-Oct-01
Included in the Prior Art Database: 2005-Apr-04
Document File: 1 page(s) / 39K

Publishing Venue

IBM

Related People

Vanover, MT: AUTHOR [+3]

Abstract

Complicated shapes can be displayed on computer graphics screens by using hardware assisted rendering. These shapes are composed of various polygons which make up the surface. For example, a pyramid is composed of 4 triangles and a square for its base. The hardware renders these polygons by performing calculations for each vertex in each separate polygon. In the case of the pyramid, and other polyhedra, the vertex calculations performed for one polygon, such as lighting calculations, are redundantly recalculated for the same vertex in other polygons of the same polyhedra. When viewed as separate polygons, there are 16 vertex calculations to perform for 4 triangles and a square, but there are actually only 5 vertices in a pyramid.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 100% of the total text.

Faster Geometric Rendering by Referencing Shared Vertices in Polyhedra

      Complicated shapes can be displayed on computer graphics
screens by using hardware assisted rendering.  These shapes are
composed of various polygons which make up the surface. For example,
a pyramid is composed of 4 triangles and a square for its base.  The
hardware renders these polygons by performing calculations for each
vertex in each separate polygon.  In the case of the pyramid, and
other polyhedra, the vertex calculations performed for one polygon,
such as lighting calculations, are redundantly recalculated for the
same vertex in other polygons of the same polyhedra.  When viewed as
separate polygons, there are 16 vertex calculations to perform for 4
triangles and a square, but there are actually only 5 vertices in a
pyramid. Instead of recalculating the transformations and lighting
calculations for vertices which are shared among the sides of a
polyhedra, hardware could store a list of "completed" vertices which
may then be referenced by index in subsequent vertices of the same
polyhedra.  This idea requires a modification of the current hardware
interface to allow vertex information to be indirectly referenced.
This change could provide dramatic speed up for cases where vertex
data is reused, as is the case in closed objects, such as polyhedra.

      Anonymous.