Ex Parte Busaba et al

Board of Patent Appeals and Interferences

Jan 28, 2008

10436219 (B.P.A.I. Jan. 28, 2008)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
____________

Ex parte FADI Y. BUSABA and TIMOTHY J. SIEGEL
____________

Appeal 2007-2287
Application 10/436,2191
Technology Center 2100
____________

Decided: January 28, 2008
____________

Before ANITA PELLMAN GROSS, JEAN R. HOMERE, and
JOHN A. JEFFERY, Administrative Patent Judges.

HOMERE, Administrative Patent Judge.

DECISION ON APPEAL
STATEMENT OF THE CASE
Appellants appeal under 35 U.S.C. § 134 from the Examiner’s final
rejection of claims 1, 3 through 10, 13, 15 through 22, and 25. Appellants
have canceled claims 2, 11, 12, 14, 23, and 24. We have jurisdiction under
35 U.S.C. § 6(b). We reverse.

1 Filed on May 12, 2003. IBM Corp. is the real party in interest.
1
Appeal 2007-2287
Application 10/436,219

The Invention
Appellants invented a method and system that compares operation
codes (opcode) to detect overlapping instruction groups, and in response, to
initiate a prescribed processor action. (Spec. 1.) As depicted in Figure 1,
the invention discloses a processor (10) that includes an instruction unit (12)
for fetching and decoding a group of instructions. The instruction unit (12)
includes an instruction register (16) for receiving and storing the group of
instructions having at least one instruction opcode. (Id. 3.) The processor
(10) further includes an execution unit (14) having a control register (22)
and a compare logic (24). The control register (22) includes a control word
having a control opcode and an action field defining an action to be taken by
the processor. (Id.) The compare logic (24) compares the instruction
opcode and the control opcode to detect a match between them. If the
compare logic (24) detects such a match, it instructs the execution unit (14)
to initiate the processor action stored in the action field. (Id.)
The compare logic (24) may operate in two modes, namely, the
sequential mode (grouped mode) and the non-sequential mode (non-grouped
mode) when the control word includes a plurality of opcodes and a field
identifying the mode. In the grouped mode, the compare logic (14) detects a
match if all the control opcodes are present in the instruction opcodes. In the
non-grouped mode, the compare logic finds a match if one of the control
opcodes is present in the instruction opcodes. (Id. 5-7.)
An understanding of the invention can be derived from exemplary
independent claim 1, which reads as follows:
2
Appeal 2007-2287
Application 10/436,219

1. A processor for detecting one or more groups of instructions and
initiating a processor action upon detecting one or more groups of
instructions, the processor comprising:

an instruction unit fetching and decoding a group of instructions;

an instruction register receiving said group of instructions having at
least one instruction opcode;

a control register including a control word including a control opcode
and an action field defining said processor action;

an execution unit having compare logic comparing said instruction
opcode and said control opcode, said execution unit initiating said processor
action upon said compare logic detecting a hit between said instruction
opcode and said control opcode;

wherein:

said control word includes multiple control opcodes and a field
indicating grouped mode or non-grouped mode;

in grouped mode, said compare logic indicates a hit if all of said
control opcodes are present in said instruction opcodes of said group of
instructions;

in non-grouped mode, said compare logic indicates a hit if one of said
control opcodes is present in said instruction opcodes of said group of
instructions.

The Examiner relies upon the following prior art to reject the claims
on appeal:
Borkenhagen US 5,790,843 Aug. 04, 1998
Slegel US 6,092,185 Jul. 18, 2000

3
Appeal 2007-2287
Application 10/436,219

The Examiner rejects the claims on appeal as follows:
A. Claims 1, 3 through 10, 13, 15 through 22, and 25 stand rejected under
35 U.S.C. § 103(a) as being unpatentable over the combination of Slegel and
Borkenhagen.

APPELLANTS’ CONTENTIONS and EXAMINER’S ASSERTIONS
Appellants contend that the combination of Slegel and Borkenhagen
does not render independent claims 1, 13, and 25 unpatentable. Particularly,
Appellants contend that the cited combination does not teach the following
limitations:
a control word includes multiple control opcodes and a field
indicating grouped mode or non-grouped mode; in grouped mode, said
compare logic indicates a hit if all of said control opcodes are present in said
instruction opcodes of said group of instructions; in non-grouped mode, said
compare logic indicates a hit if one of said control opcodes is present in said
instruction opcodes of said group of instructions, as recited in claim 1.
Particularly, Appellants state that:
Borkenhagen only teaches detecting a single instruction or an
instruction sequence that has been classified as problematic for
debugging operations . . . . There is no control field in
Borkenhagen . . . this control field is not inherent, both factually
and legally . . . Borkenhagen does not need a control field to
detect problematic instruction sequences . . . There is simply
no teaching in Borkenhagen, or in Slegel, of processing
instructions in grouped mode or non-grouped mode as recited in
claim 1.

(App. Br. 7-8.)
In response, the Examiner asserts the following:
4
Appeal 2007-2287
Application 10/436,219

The fact that Borkenhagen differentiates from “more than problematic
instruction” and “instruction sequence” shows the presence of a
“grouped mode” . . . . The examiner concedes that Borkenhagen does
not explicitly disclose how to switch between, 1. being able to detect a
single instruction, 2. being able to detect multiple instructions, and 3.
being able to detect multiple instructions that are in a sequence.
However, one of ordinary skill in the pertinent art would have
recognized that having the multiple nodes, but having no way to
switch them, makes them potentially useless.
(Ans. 12.)
ISSUE
The pivotal issue in the appeal before us is as follows:
Have Appellants shown2 that the Examiner failed to establish that the
combination of Slegel and Borkenhagen renders the claimed invention
unpatentable under 35 U.S.C. § 103(a)? Particularly, have Appellants
shown that the Examiner erred because the Slegel-Borkenhagen disclosure

2 In the examination of a patent application, the Examiner bears the initial
burden of showing a prima facie case of unpatentability. In re Piasecki,
745 F.2d 1468, 1472 (Fed. Cir. 1984). When that burden is met, the burden
then shifts to the applicant to rebut. Id.; see also In re Harris, 409 F.3d
1339, 1343-44 (Fed. Cir. 2005) (finding rebuttal evidence unpersuasive). If
the applicant produces rebuttal evidence of adequate weight, the prima facie
case of unpatentability is dissipated. Piasecki, 745 F.2d at 1472. Thereafter,
patentability is determined in view of the entire record. Id. However,
Appellant has the burden on appeal to the Board to demonstrate error in the
Examiner’s position. See In re Kahn, 441 F.3d 977, 985-86 (Fed. Cir. 2006)
(“On appeal to the Board, an applicant can overcome a rejection [under
§ 103] by showing insufficient evidence of prima facie obviousness or by
rebutting the prima facie case with evidence of secondary indicia of
nonobviousness.”) (quoting In re Rouffet, 149 F.3d 1350, 1355 (Fed. Cir.
1998)).

5
Appeal 2007-2287
Application 10/436,219

does not teach a control word having multiple control opcodes and a field
indicating grouped mode or non-grouped mode?

FINDINGS OF FACT
The following findings of fact (FF) are supported by a preponderance
of the evidence.
Slegel
1. As depicted in Figures 1 and 2, Slegel discloses a computer system
having an instruction unit (I-Unit) for fetching and decoding instructions.
The I-Unit includes a compare logic that monitors instructions as they are
decoded to compare them with certain internal set-up by a logic design
engineer to determine if a match occurs between them. If so, it takes a
prescribed action. (Col. 3, ll. 33-47.)
2. The computer system also includes an execution unit (E-Unit) for
executing the decoded instructions. The E-Unit includes a compare logic
that matches a received instruction with another instruction that has been
programmed in the computer by the logic design engineer. It saves the
results of completed instructions in the in registers, and discards partially
completed instructions and associated opcodes. (Col. 3, ll. 33-37, col. 4, ll.
25-33.)
3. The E-Unit opcode compare logic (204) subsequently sends each
detected opcode back to the I-Unit bus. Upon detecting a new opcode, the I-
Unit compares the detected opcode with the opcode received on the bus
from the E-Unit. If they are the same, the I-Unit takes the action specified
6
Appeal 2007-2287
Application 10/436,219

by the E-Unit. If they do not match, the I-Unit executes the received
decoded instruction as it normally would. (Col. 4, ll. 33-36, ll. 43-52.)

Borkenhagen
4. As depicted in Figure 1, Borkenhagen discloses a microprocessor (10)
including an instruction matching system (12) having a plurality of
instruction match registers (IMRs) (16) to watch for known problematic
instructions. The microprocessor (10) further includes an execution control
system (14) having a plurality of execution control registers (ECRs) (18) to
alter microprocessor control for identified problematic instructions. Each
IMR is logically coupled to an associated ECR in a one-to-one
correspondence. (Col. 3, ll. 23-37.)
5. Upon receiving an instruction (20), the matching system (12)
compares the received instruction opcode with opcodes in the IMRs. It
subsequently forwards the result of the comparison to the execution system
(14). If a match results, the Execution system (14) looks to the ECR for a
control code indicating which action to take. If no match results, the
microprocessor continues with its normal execution. (Col. 3, ll. 38-57.)
6. Borkenhagen contemplates the use of multiple instruction pairs such
that more than one problematic instruction or instruction sequence can be
identified. It allows multiple IMR-ECR pairs to be combined to identify
problematic instruction sequences and to take appropriate actions. (Col. 2,
ll. 39-49.)

7
Appeal 2007-2287
Application 10/436,219

PRINCIPLES OF LAW
OBVIOUSNESS (Prima Facie)
The Supreme Court in Graham v. John Deere Co., 383 U.S. 1, 17-18,
(1966), stated that the following factual inquiries underpin any
determination of obviousness:
Under § 103, [1] the scope and content of the prior art are to be
determined; [2] differences between the prior art and the claims
at issue are to be ascertained; and [3] the level of ordinary skill
in the pertinent art resolved. Against this background, the
obviousness or nonobviousness of the subject matter is
determined. Such [4] secondary considerations as commercial
success, long felt but unsolved needs, failure of others, etc.,
might be utilized to give light to the circumstances surrounding
the origin of the subject matter sought to be patented. As
indicia of obviousness or nonobviousness, these inquiries may
have relevancy.
Where the claimed subject matter involves more than the simple
substitution of one known element for another or the mere application of a
known technique to a piece of prior art ready for the improvement, a holding
of obviousness must be based on “an apparent reason to combine the known
elements in the fashion claimed.” KSR Int’l v. Teleflex, Inc., 127 S. Ct.
1727, 1740-41 (2007). That is, “there must be some articulated reasoning
with some rational underpinning to support the legal conclusion of
obviousness.” Id. at 1741 (quoting In re Kahn, 441 F.3d 977, 988 (Fed. Cir.
2006)). Such reasoning can be based on interrelated teachings of multiple
patents, the effects of demands known to the design community or present in
the marketplace, and the background knowledge possessed by a person
having ordinary skill in the art. KSR, 127 S. Ct. at 1740-41.

8
Appeal 2007-2287
Application 10/436,219

ANALYSIS
Independent claims 1, 13, and 25 require, in pertinent part, a control
word that includes multiple control opcodes and a field indicating grouped
mode or non-grouped mode. In the grouped mode, the compare logic
indicates a hit if all of the control opcodes are present in the instruction
opcodes of the group of instructions. In the non-grouped mode, however,
the compare logic indicates a hit if one of the control opcodes is present in
the instruction opcodes of the group of instructions. (App. Br. 11, Claims
Appendix.) As indicated above, Appellants argue that the Slegel-
Borkenhagen combination does not teach those limitations. The Examiner
responds that Borkenhagen’s disclosure of combining a plurality of IMR-
ECR pairs to identify problematic instructions or instruction sequences
teaches the cited limitations. We do not agree with the Examiner that the
Slegel-Borkenhagen combination reasonably teaches or even suggests this
limitation.
Slegel teaches an I-Unit and an E-Unit, each having a compare logic
that compares an incoming instruction opcode with a design instruction
opcode to detect design deficiencies in a processor. (FF 1-2.) Slegel further
teaches that I-Unit opcodes are compared with E-Unit opcodes to determine
which prescribed action to take upon finding a match between them. (FF 3.)
Borkenhagen teaches comparing received instruction opcodes with opcodes
in IMR and ECR to detect a problematic instruction or instruction sequence.
(FF 4-5). Borkenhagen further teaches combining a plurality of IMR-ECR
pairs to detect multiple instruction sequences or problematic instructions.
One of ordinary skill in the art would have readily recognized that the
9
Appeal 2007-2287
Application 10/436,219

combined disclosures of Slegel and Borkenhagen, at best, teach a control
word having multiple control opcodes being used to detect multiple
problematic instructions or instruction sequences in a microprocessor. The
ordinarily skilled artisan would have recognized, however, that Slegel and
Borkenhagen’s teachings do not extend to a control word having a field
indicating a grouped mode or a non-grouped mode. Even if we were to
loosely construe the disclosure of detecting problematic instructions or
instruction sequences as separate groups, the Slegel-Borkenhagen
combination would still fall short of suggesting that such groups are fields
indicating a grouped mode and a non-grouped mode. To somehow construe
Borkenhagen’s detection of problematic instructions or instruction
sequences as including fields that indicate that the compare logic is
operating in a grouped mode or a non-grouped mode would require us to
resort to speculation and strain the reference’s teachings beyond reasonable
limits. Further, the Slegel-Borkenhagen is totally silent on the specific
manner the compare logic detects a hit while operating in either the grouped
mode or the non-grouped mode.
We have considered the Examiner’s rationale for the rejection
including the hypothetical example provided in the Answer. (Ans. 12-15.)
We find the Examiner’s conclusion of obviousness to be unreasonable. We
also find the hypothetical example to be purely speculative, and unsupported
by the record before us. It follows that the Examiner erred in rejecting
independent claims 1, 13, and 25 as being unpatentable over the combination
of Slegel and Borkenhagen. We conclude for these same reasons that the
10
Appeal 2007-2287
Application 10/436,219

cited combination does not render dependent claims 3 through 10 and 15
through 22 unpatentable.
CONCLUSION OF LAW
Appellants have shown that the Examiner failed to establish that
claims 1, 3 through 10, 13, 15 through 22, and 25 are unpatentable over the
combination of Slegel and Borkenhagen under 35 U.S.C. § 103(a).

DECISION
We reverse the Examiner’s decision rejecting claims 1, 3 through 10,
13, 15 through 22, and 25.

REVERSED

rwk

Philmore H. Colburn II
CANTOR COLBURN LLP
55 Griffin Road South
Bloomfield, CT 06002

11