Browse Prior Art Database

Floating Point Integer Power Instruction for Phong Light Model and Animation in Graphics and Multimedia Applications

IP.com Disclosure Number: IPCOM000105276D
Original Publication Date: 1993-Jul-01
Included in the Prior Art Database: 2005-Mar-19
Document File: 2 page(s) / 68K

Publishing Venue

IBM

Related People

Karim, F: AUTHOR [+3]

Abstract

An integer power of a floating point number needs to be evaluated rapidly in high-speed graphics and multimedia applications for calculating the specular highlight component using Phong light model. The specular highlight component for Phong lighting is evaluated in graphics as follows:

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Floating Point Integer Power Instruction for Phong Light Model and Animation in Graphics and Multimedia Applications

      An integer power of a floating point number needs to be
evaluated rapidly in high-speed graphics and multimedia applications
for calculating the specular highlight component using Phong light
model.  The specular highlight component for Phong lighting is
evaluated in graphics as follows:

                      COS**n X
where COS X is the dot product of two vectors and is a floating point
number between -1.0 and 1.0.  n is called the material's specular
reflec tion coefficient or shininess and is an integer between 0 and
128.  If COS X is less than or equal to 0.0 then no specular
highlights will ensue.

      The state-of-the-art for implementing the Phong light model in
graphics is by a table-look-up.  A table is generated for X as an
inde pendent variable and the function COS**n X as a dependent
variable for a particular n. The slope and intercept of the function
are stored for specific values of X and can be retrieved by using X
as address.  If the value of X lies in between the points stored, the
function is interpo lated.

      In high-speed graphics and multimedia applications, animation
of an object involves rapidly calculating the specular highlight
component.  A table-look-up method makes the process serial and
places severe con straints on how fast the object moves.  In short,
the table-look-up method does not support parallelism of multiple
execution units.  Superscalar RISC processors and specialized
graphics processors support multiple execution units, executing more
than one instruction per cycle.  Using a table-look-up, will
serialize the floating point instruction stream that are evaluating
the specular highligh...