Appeal 2007-1573
Application 10/705,094
But whether a higher workload for one processor as compared to another
necessarily results in longer processing times is questionable at best.
First, processing times for processors with different workloads can
vary widely depending on various factors as Appellants indicate. In fact, in
certain instances, a processor with fewer tasks (i.e., a smaller workload)
could actually take longer to process than a faster processor with more
tasks.5
To be sure, identical processors processing different amounts of
workload in operating conditions such that all other variables are the same
for both processors would result in different processing times for each
processor. Under these conditions, a higher workload for a given processor
will naturally result in more processing time as compared to another
identical processor with less workload.
This, however, is the crux of the Examiner’s position -- a position that
assumes the individual processors of the pipeline of processors in
Narayanaswami are the same type. The Examiner, however, points to
nothing in the reference that supports this theory -- an essential requirement
for anticipation. In fact, the Examiner readily admits that Narayanaswami is
completely silent regarding whether the processors are identical.6 Merely
5 For example, assuming that individual tasks within a given workload (i.e.,
Tasks 1, 2, and 3) consume the same amount of time to process, a slower
processor with fewer tasks could have a longer processing time than a faster
processor with more tasks. That is, a 100 MHz processor with two tasks
(Task 1 and Task 2) would have a longer processing time than a 200 MHz
processor with three tasks (Task 1, Task 2, and Task 3).
6 See Answer, at 5 (admitting that “Narayanaswami doesn’t mention the
multiprocessors are homogeneous type processors”); see also Answer, at 6
6
Page: Previous 1 2 3 4 5 6 7 8 9 Next
Last modified: September 9, 2013