Appeal 2007-1573
Application 10/705,094
Narayanaswami discloses a method and system for managing a
graphical workload such that a particular pixel or display region can be
processed for a first period of time and then rendered by a second processor
for another period of time (Narayanaswami, col. 2, ll. 7-18). As shown in
Figure 1, computer 105 is coupled to graphics adapter 200 via bus 160. The
graphics adapter, in turn, is coupled to graphics output device 150 (e.g., a
display). Instructions from the computer are received by the graphics
adapter and are executed with graphics adapter processors 220. Graphics
adapter processors 220 can be a pipeline of processors in series, a set of
parallel processors, or some combination of both where each processor may
handle part of the task to be completed (Narayanaswami, col. 2, ll. 39-51;
Fig. 1).
An exemplary graphical workload allocation procedure for three
processors is shown in Figure 3C. In this approach, the breakdown and
allocation of pixels or regions of pixels to various processors can be based
on the amount of workload each processor already has. Moreover, pixels
may be reassigned to different processors over time based on “various
factors” (Narayanaswami, col. 5, ll. 4-18; Fig. 3C).
The issue, then, is whether this allocation based on processor
workload necessarily involves analyzing the time taken by the processors.
Certainly, no skilled artisan would deny that higher workloads on a single
processor will take longer to process -- a fact even admitted by Appellants.4
4 See Reply Br. 4 (“Although the Examiner’s statement [regarding the more
workload a processor has, the longer it takes to finish] may be true for a
single processor, such an assumption is clearly not necessarily true when
comparing two or more processors.”).
5
Page: Previous 1 2 3 4 5 6 7 8 9 Next
Last modified: September 9, 2013