Ex Parte Eichenberger et al

Patent Trial and Appeal Board

Sep 23, 2014

11339591 (P.T.A.B. Sep. 23, 2014)

UNITED STATES PATENT AND TRADEMARK OFFICE
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O. Box 1450
Alexandria, Virginia 22313-1450
www.uspto.gov
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO.
11/339,591 01/25/2006 Alexandre E. Eichenberger AUS920050823US1 9145
50170 7590 09/24/2014
IBM CORP. (WIP)
c/o WALDER INTELLECTUAL PROPERTY LAW, P.C.
17304 PRESTON ROAD
SUITE 200
DALLAS, TX 75252
EXAMINER
BULLOCK JR, LEWIS ALEXANDER
ART UNIT PAPER NUMBER
2199
MAIL DATE DELIVERY MODE
09/24/2014 PAPER
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
PTOL-90A (Rev. 04/07)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________

Ex parte ALEXANDRE E. EICHENBERGER
and JOHN KEVIN O’BRIEN
____________

Appeal 2012-003050
Application 11/339,591
Technology Center 2100
____________

Before ALLEN R. MacDONALD, CARLA M. KRIVAK, and
JEFFREY S. SMITH, Administrative Patent Judges.

SMITH, Administrative Patent Judge.

DECISION ON APPEAL
Appeal 2012-003050
Application 11/339,591

2
STATEMENT OF THE CASE
This is an appeal under 35 U.S.C. § 134(a) from the Examiner’s final
rejection of claims 1 and 21–34. We have jurisdiction under 35 U.S.C.
§ 6(b).
We affirm.

Representative Claim
1. A method, in a data processing system comprising a
processor, for performing a scalar operation, comprising:
identifying, by the processor, the scalar operation in
source code;
generating, by the processor, an expression tree for the
scalar operation, wherein the expression tree has leaf nodes
corresponding to scalar operands of the scalar operation;
determining, by the processor, an alignment of the scalar
operands in respective registers that store the scalar operands;
propagating, by the processor, the alignment of the scalar
operands up the expression tree;
annotating, by the processor, the expression tree based on
the propagation of the alignment of the scalar operands up the
expression tree, wherein annotating the expression tree based
on the propagation of the alignment comprises annotating
selected nodes of the expression tree to identify shift operations
to be performed with respect to the selected nodes, wherein the
shift operations are determined with reference to all other shift
operations within a given region of the source code by means of
an optimization process which attempts to find a minimum
number of the shift operations within the given region of the
source code, wherein the optimization process determines
common sub-expressions of scalar operations, and wherein the
shift operations are selected such that for a number of instances
of the common sub-expressions a same shift operation is
utilized; and
Appeal 2012-003050
Application 11/339,591

3
generating, by the processor, executable code for the
scalar operation based on the expression tree and the alignment
of the scalar operands, wherein generating the executable code
for the scalar operation includes inserting, by the processor,
shift instructions into the executable code based on the
alignment of the scalar operands.

Prior Art
Alexander E. Eichenberger et al., Vectorization for SIMD
Architectures with Alignment Constraints, ACM SIGPLAN
Notices, Vol. 39, No. 6, 82–93 (May 2004) (“Eichenberger”).

Examiner’s Rejections
Claims 1 and 21–34 stand rejected under 35 U.S.C. § 102(b) as being
anticipated by Eichenberger.

ANALYSIS
We adopt the findings of fact from the Final Rejection and
Examiner’s Answer as our own. We agree with the Examiner that claims 1
and 21–34 are anticipated by Eichenberger for the reasons given in the
Examiner’s Answer. We highlight the following for emphasis.
Appellants contend Eichenberger does not describe “wherein the shift
operations are determined with reference to all other shift operations within
a given region of the source code by means of an optimization process which
attempts to find a minimum number of the shift operations within the given
region of the source code” as recited in claim 1. App. Br. 7. According to
Appellants, Eichenberger does not disclose determining shift operations with
reference to all other shift operations within a given region of the source
code. App. Br. 8–13. Appellants further contend Eichenberger does not
Appeal 2012-003050
Application 11/339,591

4
disclose an optimization process which attempts to find a minimum number
of shift operations within a given region of source code. Reply Br. 2–6.
The Examiner finds the lazy-shift and dominant-shift policies of
Eichenberger describe the disputed limitation of claim 1. Ans. 6.
Eichenberger describes a lazy-shift policy as improving an eager-shift policy
by delaying stream shifts as long as two constraints are satisfied. In this
case, rather than performing two or three shift operations, only the result of
the add needs to be shifted. Eichenberger, 87. A dominant-shift policy
further reduces the number of shift operations by shifting register streams to
the most dominant stream offset in the graph. Id.
Appellants find support for the disputed limitation of claim 1 on page
17 of Appellants’ Specification (App. Br. 3), which states:
There may be more than one way to choose an alignment
and operand to shift. This choice can be made arbitrarily, in a
simple compiler, or it may be made with reference to all other
shifts within some given region of code, such as a loop, by
means of an optimization process which attempts to find a
minimum number of shifts within the region. For example,
common sub-expressions of scalar operations may be
determined and shift operations may be selected such that for
some number of instances of the common sub-expression the
same shift may be utilized.
Spec. 17:7–16. The scope of the claimed “shift operations are determined
with reference to all other shift operations within a given region of the
source code by means of an optimization process which attempts to find a
minimum number of the shift operations within the given region of the
source code,” when read in light of page 17 of Appellants’ Specification,
encompasses determining common sub-expressions of scalar operations and
Appeal 2012-003050
Application 11/339,591

5
selecting shift operations such that for some number of instances of the
common sub-expression, the same shift may be utilized.
Appellants have not provided persuasive evidence or argument to
distinguish the lazy-shift and dominant-shift policies of Eichenberger, which
utilize the same shift for common sub-expressions, from “shift operations
are determined with reference to all other shift operations within a given
region of the source code by means of an optimization process which
attempts to find a minimum number of the shift operations within the given
region of the source code” as recited in claim 1.
We sustain the rejection of claim 1 under 35 U.S.C. § 102. Appellants
do not provide arguments for separate patentability of claims 21–34, which
fall with claim 1.

DECISION
The rejection of claims 1 and 21–34 under 35 U.S.C. § 102(b) as
being anticipated by Eichenberger is affirmed.
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a). See 37 C.F.R.
§ 41.50(f).

AFFIRMED

msc