Appeal 2007-1959
Application 10/039,789
instead of the claimed using a run-time library or a library to perform a
reduction operation (Br. 14). We disagree. As discussed above and argued
by the Examiner (Answer 20-21), the rejection is based on using the
reduction operation or the algebraic operation of Sundaresan as the operation
to be parallel processed in Poulsen (FF 9).
Claims 22, 24, and 25
Appellants argue that Hardwick does not teach performing a plurality
of “vector operations,” as recited in claim 24 (Br. 14) and may not be
combined with Poulsen and Sundaresan since Hardwick is directed to
reducing inter-processor communications during parallel processing, instead
of being concerned with program translation (Br. 15). The Examiner
responds that there is no need for multiple vector operations to be disclosed
by all the references since Hardwick’s vector operation may be used in each
of the plurality of reduction operations of Sundaresan (Answer 22).
We agree with the Examiner and find that Hardwick does indeed
describe a vector operation to be carried out over a group of processors
operating in parallel (FF 12 & 13) such that each of the reduction operations
of Sundaresan may benefit from implementing such vector operation. Not
only does Hardwick discuss the benefit of portability, but also one of
ordinary skill in the art would have performed program translations to
implement the vector operation as Hardwick does provide for translation of
the vector operation for parallel processing (FF 13).
10
Page: Previous 1 2 3 4 5 6 7 8 9 10 11 Next
Last modified: September 9, 2013