Appeal No. 1996-3193
Application 08/160,112
of absolute values of said differences" after the addition or
substraction. As explained in the specification, the running
sum of carry outputs accounts for the overflow bits from the
partial sums (specification, page 274). "As a consequence a
register of fixed size may be used to store the running sum
regardless of the number of differences formed." (Br9.)
The Examiner applies Hill. The Examiner's position is
(Paper No. 5, page 5):
Hill discloses that carry-save addition includes storing
the carry from an addition stage and then iteratively
adding and storing succeeding carries in the carry save
register (p. 598; fig. 15.4). It would have been an
obvious modification to one of ordinary skill in the art
at the time of the invention to modify Taniguchi to use a
carry-save adder as disclosed by Hill, because a
carry-save adder is particularly well suited for a
situation in which a series of numbers is to be added
together; this would be the situation in Taniguchi if a
series of absolute values from the selection circuit,
605, were to be summed.
Appellants argue that the teachings of Hill would not
lead to the running sum of carry outputs claimed. "On the
contrary, Hill et al teaches that the carries are added during
the next following addition operation." (Br9.) We agree.
There is no suggestion in the multiplication method of Hill
"adding said carry output to a running sum of carry outputs"
as claimed and the Examiner has failed to show where this is
- 13 -
Page: Previous 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Next
Last modified: November 3, 2007