Priya N. Vaidya et al.

Patent Trials and Appeals Board

Mar 20, 2020

14748515 - (D) (P.T.A.B. Mar. 20, 2020)

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.
14/748,515 06/24/2015 Priya N. Vaidya P83299 5725
88032 7590 03/20/2020
Jordan IP Law, LLC
12501 Prosperity Drive, Suite 401
Silver Spring, MD 20904
EXAMINER
LABUD, JONATHAN R
ART UNIT PAPER NUMBER
2196
NOTIFICATION DATE DELIVERY MODE
03/20/2020 ELECTRONIC
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.
Notice of the Office communication was sent electronically on above-indicated "Notification Date" to the
following e-mail address(es):
admin@jordaniplaw.com
info@jordaniplaw.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
________________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
________________

Ex parte PRIYA N. VAIDYA and
PREMANAND SAKARDA1
________________

Appeal 2018-004173
Application 14/748,515
Technology Center 2100
________________

Before BRADLEY W. BAUMEISTER, MIRIAM L. QUINN, and
STACEY G. WHITE, Administrative Patent Judges.

BAUMEISTER, Administrative Patent Judge.

DECISION ON APPEAL
Appellant appeals under 35 U.S.C. § 134(a) from the Examiner’s final
rejection of claims 1–24, which constitute all claims pending in this
application. Appeal Br. 5.2 We have jurisdiction under 35 U.S.C. § 6(b).
We affirm.

1 We use the word “Appellant” to refer to “applicant” as defined in
37 C.F.R. § 1.42. Appellant identifies Intel Corporation as the real party in
interest. Appeal Brief filed June 26, 2017 (“Appeal Br.”) 3.
2 Rather than repeat the Examiner’s positions and Appellant’s arguments in
their entirety, we refer to the above mentioned Appeal Brief, as well as the
following documents for their respective details: the Final Action mailed
January 27, 2017 (“Final Act.”); the Examiner’s Answer mailed January 10,
2018 (“Ans.”); and the Reply Brief filed March 9, 2018 (“Reply Br.”).
Appeal 2018-004173
Application 14/748,515

2
CLAIMED SUBJECT MATTER
Appellant describes the present invention as follows:
Systems and methods may provide for making a power
efficiency determination at runtime based on one or more
runtime usage notifications and scheduling a workload for
execution on a hardware accelerator if the power efficiency
determination indicates that execution of the workload on the
hardware accelerator will be more efficient than execution of the
workload on a host processor. Additionally, the workload may
be scheduled for execution on the host processor if the power
efficiency determination indicates that execution of the workload
on the host processor will be more efficient than execution of the
workload on the hardware accelerator. In one example, making
the power efficiency determination includes applying one or
more configurable rules to at least one of the one or more runtime
usage notifications.
Abstract.

STATEMENT OF THE REJECTIONS
Claims 7–12 stand rejected under 35 U.S.C. § 101 as being directed to
patent ineligible subject matter. Final Act. 4–5.
Claims 1, 6, 7, 12, 13, 18, 19, and 24 stand rejected under 35 U.S.C.
§ 103 as being unpatentable over Hanson et al. (US 8,301,742 B2; issued
Oct. 30, 2012) (“Hanson”) and Wang et al. (Workload Analysis and Efficient
OpenCL-based Implementation of SIFT Algorithm on a Smartphone, IEEE
Global Conference on Signal and Information Processing (GlobalSIP),
Dec. 2013) (“Wang”). Final Act. 6–9.
Claims 2, 3, 5, 8, 9, 11, 14, 15, 17, 20, 21, and 23 stand rejected under
35 U.S.C. § 103 as being unpatentable over Hanson, Wang, and Degenaro et
al. (US 8,434,087 B2; issued Apr. 30, 2013) (“Degenaro”). Final Act. 9–12.
Appeal 2018-004173
Application 14/748,515

3
Claims 4, 10, 16, and 22 stand rejected under 35 U.S.C. § 103 as
being unpatentable over Hanson, Wang, and Nguyen et al.
(US 7,870,185 B1; issued Jan. 11, 2011) (“Nguyen”). Final Act. 12–13.

STANDARD OF REVIEW
The Board conducts a limited de novo review of the appealed
rejections for error based upon the issues identified by Appellant, and in
light of the arguments and evidence produced thereon. Ex parte Frye,
94 USPQ2d 1072, 1075 (BPAI 2010) (precedential).

THE 101 REJECTION
Illustrative Claim 7
Independent claim 7, reproduced below, illustrates the subject matter
of claims 7–12, which stand rejected under 35 U.S.C. § 101:
7. An apparatus comprising:
logic, implemented at least partly in one or more of
configurable logic or fixed functionality logic hardware, to:
make a power consumption determination at
runtime based on one or more runtime usage notifications;
schedule a workload for execution on a hardware
accelerator if the power consumption determination
indicates that execution of the workload on the hardware
accelerator will be more efficient than execution of the
workload on a host processor; and
schedule the workload for execution on the host
processor if the power consumption determination
indicates that execution of the workload on the host
processor will be more efficient than execution of the
workload on the hardware accelerator.
Appeal 2018-004173
Application 14/748,515

4
Determinations and Contentions
The Examiner determines that the language of claim 7 is broad
enough to read on software, per se, and, therefore, claim 7 is broad enough
to read on non-statutory subject matter. Final Act. 4. According to the
Examiner, software constitutes a description or expression of a computer
program, which is not a physical thing. Id.
The Examiner cites to both the language of claim 7 and Appellant’s
Specification to support the determination that the claim reads on software
per se. Ans. 3–4. First, the Examiner notes that the express language of
claim 7 states that the logic may be implemented either in “configurable
logic or fixed functionality logic hardware.” Id. at 4. The Examiner also
notes that Appellant’s Specification discloses that the “configurable logic
may be, for example, programmable logic arrays (PLAs), field
programmable gate arrays (FPGAs), complex programmable logic devices
(CPLDs).” Id. (citing Spec. ¶ 13) (emphasis added). The Examiner reasons
that the Specification’s use of the permissive term “may be” indicates that
that the Specification does not limit the claimed “configurable logic” to the
hardware examples listed, and, therefore, does not exclude software
embodiments. Id.
Appellant argues that “there is no legal basis for the contention that
software is per se non-statutory subject matter.” Appeal Br. 11. According
to Appellant, claims that are directed to software, as opposed to hardware,
are not inherently abstract. Id. (citing Enfish, LLC v. Microsoft Corp., 822
F.3d 1327, 1335 (Fed. Cir. 2016)).
Appellant further argues “claims directed to a specific improvement to
the way computers operate are patentable, even in the context of software.”
Appeal 2018-004173
Application 14/748,515

5
Id. at 11. According to Appellant, claims recite patent-eligible subject
matter when they are “directed to a specific improvement–a particular
database technique–in how computers could carry out their functions of
storage and retrieval of data.” Id. at 11–12 (citing Enfish, LLC, 822 F.3d at
1336). Appellant also contends that claims are patent eligible when they are
“directed to a software-based invention that improves the performance of the
computer system itself.” Id. at 12 (citing also BASCOM Global Internet
Servs., Inc., v. AT&T Mobility LLC, 827 F.3d 1341 (Fed. Cir. 2016)).
Analysis
Appellant’s arguments are unpersuasive. The question posed is not
whether a claim is per se patent-ineligible merely because some of the
limitations may recite software, as well as hardware. Rather, the relevant
question is whether a claim reciting only software per se is patent ineligible.
Enfish and BASCOM, cited by Appellant, are not on point because the
claims at issue in those cases recited physical elements, as well as
improvements to software.
More specifically, claim 17 of Enfish recited “A data storage and
retrieval system for a computer memory, comprising means for configuring
said tables. . .” Enfish, 822, F.3d at 1336. Claim 1 of BASCOM recited
“A content filtering system for filtering content retrieved from an Internet
computer network. . . , said filtering system comprising: a local client
computer; . . . [and] a remote [Internet Service Provider] server coupled to
said client computer and said Internet computer network. BASCOM, 827
F.3d at 1345.
Appellant points to no persuasive authority for the proposition that
software per se is patent eligible. Instead, it is well settled that “[a]bstract
Appeal 2018-004173
Application 14/748,515

6
software code is an idea without physical embodiment . . . .” Microsoft
Corp. v. AT&T Corp., 550 U.S. 437, 449 (2007). Moreover, a patent claim
that is broad enough to cover both statutory subject matter and nonstatutory
subject matter is subject to rejection under 35 U.S.C. § 101. Amgen, Inc. v.
Hoechst Marion Roussel, Inc., 314 F.3d at 1329.
Because a claim to computer software per se constitutes non-statutory
subject matter, we do not perform an analysis pursuant to the 2019 Revised
Patent Subject Matter Eligibility Guidance (84 Fed. Reg. 50). Only after a
claim is determined to fall within a statutory category of subject matter, e.g.,
a useful process or machine (35 U.S.C. § 101), is this patent-eligibility
analysis performed. Alice Corp. v. CLS Bank Int’l, 573 U.S. 208, 216–17
(2014). Such a patent-eligibility analysis is undertaken in order to address
the separate question of whether a claim reciting a statutory category of
invention, nonetheless, is directed to an exception to patent-eligible subject
matter without reciting significantly more. Id.
Furthermore, Appellant merely argues that software per se is not
necessarily patent ineligible. Appellant does not set forth any persuasive
substantive arguments as to why the Examiner is incorrect in determining
that claim 7 does, in fact, only recite software per se. Appeal Br. 11–14;
Reply Br. 5–7.
Accordingly, we sustain the rejection of claim 7 under § 101 as being
directed to non-statutory subject matter. We, likewise, sustain the rejection
of dependent claims 8–12, which Appellant does not argue separately. See
37 C.F.R. § 41.37(c)(1)(iv).

Appeal 2018-004173
Application 14/748,515

7
THE 103 REJECTIONS
Illustrative claim 1
Independent claim 1, reproduced below, illustrates the appealed
claims:
1. A system comprising:
a hardware accelerator;
a host processor; and
logic, implemented at least partly in one or more of
configurable logic or fixed functionality logic hardware, to:
make a power consumption determination at
runtime based on one or more runtime usage notifications,
schedule a workload for execution on the hardware
accelerator if the power consumption determination
indicates that execution of the workload on the hardware
accelerator will be more efficient than execution of the
workload on the host processor, and
schedule the workload for execution on the host
processor if the power consumption determination
indicates that execution of the workload on the host
processor will be more efficient than execution of the
workload on the hardware accelerator.
Determinations and Contentions
The Examiner finds that Hanson discloses a host processor that
includes logic to make power consumption determinations at runtime based
on runtime usage notifications. Final Act. 6 (citing Hanson col. 2, l. 53–
col. 3, l. 8, col 7, ll. 6–43). According to the Examiner,
Hanson does not explicitly teach a hardware accelerator;
scheduling a workload for execution on the hardware accelerator
if the power consumption determination indicates that execution
of the workload on the hardware accelerator will be more
efficient than execution of the workload on the host processor;
Appeal 2018-004173
Application 14/748,515

8
and scheduling the workload for execution on the host processor
if the power consumption determination indicates that execution
of the workload on the host processor will be more efficient than
execution of the workload on the hardware accelerator.
Final Act. 6.
The Examiner finds that Wang teaches this limitation and that
motivation existed to combine the teachings of Hanson and Wang. Id. at 7–
8. More specifically, the Examiner finds that Wang teaches a graphics
processing unit (GPU) and that this GPU discloses a hardware accelerator,
as claimed. Id. at 7. The Examiner further finds that Wang teaches
measuring processing performance and processing times of workloads that
are performed on a central processing unit (CPU) and comparing those
parameters to when the workloads are performed on the GPU. Id. (citing
Wang p. 3 § D; p. 4, § V). According to the Examiner, Wang teaches
workloads either being kept on the CPU or alternatively offloaded to the
GPU dependent upon which device provides better performance for the
workload. Id. The Examiner also determines that Wang’s disclosure of
comparing workload processing times is synonymous with disclosing
comparing power consumption. Id.
Appellant acknowledges in the Appeal Brief that “Wang teaches to
offload functions to the GPU if the processing time of the GPU is shorter
than the CPU processing time.” Appeal Br. 13. Appellant argues, though,
that Wang does not teach scheduling the workload for execution on the host
processor or hardware accelerator based upon a determination of which
execution will produce the more efficient power consumption. Id. at 13–14.
Appellant also argues that Hanson does not compensate for this purported
failure of Wang’s teaching. Id. at 14.
Appeal 2018-004173
Application 14/748,515

9
In response, the Examiner further explains that the following passage
of Wang provides the basis for reaching the Examiner’s conclusion that was
set forth in the Final Action:
To measure energy efficiency, we compute [the scale-
invariant feature transform (SIFT)] on the same dataset for
several minutes, measure the average system power, and then
subtract the idle system power. We measured l490mW and
l429mW power consumptions for the CPU-only
implementation and the heterogeneous implementation,
respectively. The average energy consumption per image is
413.0mJ for the CPU-only implementation, and 242.3mJ for the
heterogeneous one. Therefore, a 41% reduction in energy
consumption is achieved.
Ans. 5 (citing Wang, p. 5 § V. Experimental Results) (emphasis added in the
Examiner’s Answer); see also id. (citing Wang p. 2. ¶ 2, titled “IV.
ALGORITHM PROFILING AND WORKLOAD PARTITIONING”).
The Examiner also cites to Wang’s more general discussion of
using Open Computing Language (OpenCL) to achieve an efficient,
high-performance implementation of GPUs on mobile processors:
the clock frequency of mobile GPUs is typically between 200
and 400 MHz, which is much slower than their desktop
counterparts, which typically reach l GHz clock frequency. For
these reasons, it is critical that targeted algorithms are carefully
analyzed to find a good mapping to mobile GPUs. On the other
hand, the low clock frequency of mobile GPUs suggests the
possibility of efficient low power designs, in which offloading
Appeal 2018-004173
Application 14/748,515

10
some parallelizable computationally[]intensive algorithms to
GPUs may reduce power consumption.
Ans. 5 (citing Wang, p. 1, col. 2, § II. OpenCL on Mobile
Processors)3 (emphasis added in the Examiner’s Answer).
The Examiner concludes that because Wang recognized that
processing time is reasonably correlated to power consumption,
Wang’s disclosure of making decisions to offload workloads from a
CPU to a GPU reasonably can be understood to be a teaching that
decisions to offload workloads to a hardware accelerator were known
to be based on the processor and accelerator’s relative power
consumptions. Id. at 6.
In the Reply Brief, Appellant responds to the Examiner’s reliance on
Section V of Wang, as set forth in the Examiner’s Answer:
While section V of Wang does discuss experimental results of
using the SIFT algorithm, this section is comparing and
contrasting the results of implementing the SIFT algorithm using
only a CPU of a mobile device versus implementing the
algorithm in a heterogen[e]ous manner[] (i.e., using both the
CPU and a GPU of a mobile device). That is, the energy
efficiency that is mentioned in section V is only mentioned with
respect to: 1) only using the mobile CPU to implement the SIFT
algorithm, and, separately, 2) implementing the SIFT algorithm
using a GPU+CPU of a mobile device.
Reply Br. 9–10.
Appellant reiterates the argument that “Wang only discusses the
energy efficiency result of experiments based on processing times of the
mobile device’s GPU and CPU, and there is absolutely no discussion of how

3 The Examiner’s Answer incorrectly refers to page 1 of Wang as page 2.
This clerical error constitutes harmless error.
Appeal 2018-004173
Application 14/748,515

11
workloads are partitioned onto a mobile device CPU and GPU based on
energy consumption.” Id. at 10.
Analysis
The Examiner’s position reasonably can be understood to be that
Wang teaches that measuring processing time is synonymous with
measuring processing power. Furthermore, the Examiner sets forth a factual
basis for why one of ordinary skill who is in possession of Wang would have
understood the teaching of the former additionally to be a teaching of the
latter. See, e.g., Ans. 4–6.
Moreover, and contrary to Appellant’s assertion (Reply Br. 10), Wang
does discuss how workloads are determined to be partitioned. In Section D.
Profiling Results and Workload Partitioning, Wang states that “[b]ased on
the profiling results, we partition the SIFT algorithm to a CPU-GPU
heterogeneous implementation to minimize the total processing and memory
transfer time, as shown in in Fig. 2.” Wang. p. 3, col. 2, cited in Final
Act. 7.
Accordingly, Appellant does not persuade us of error in the
Examiner’s obviousness rejection of claim 1 based upon the combination of
Hanson and Wang. We, therefore, sustain the rejection of that claim and of
claims 6, 7, 12, 13, 18, 19, and 24, which Appellant does not argue
separately. Appeal Br. 12–14; see 37 C.F.R. § 41.37(c)(1)(iv).
We, likewise, sustain the Examiner’s obviousness rejections of
claims 2, 3, 5, 8, 9, 11, 14, 15, 17, 20, 21, and 23 over the combination of
Hanson, Wang, and Degenaro, as well as the claims 4, 10, 16, and 22 over
the combination of Hanson, Wang, and Nguyen. Appellant does not
particularly pointed out errors in the Examiner’s reasoning regarding the
Appeal 2018-004173
Application 14/748,515

12
additional teachings of Degenaro or Nguyen, but merely argues that neither
of the references cures the purported deficiencies of Hanson and Wang that
Appellant alleges in relation to independent claim 1. Appeal Br. 14–15.

DECISION SUMMARY
In summary:

TIME PERIOD FOR RESPONSE
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(1). See 37 C.F.R.
§ 1.136(a)(1)(iv).

AFFIRMED

Claims
Rejected
35 U.S.C. § Basis Affirmed Reversed
7–12 101 Eligibility 7–12
1, 6, 7, 12, 13,
18, 19, and 24
103 Hanson, Wang, 1, 6, 7, 12,
13, 18, 19,
and 24

2, 3, 5, 8, 9, 11,
14, 15, 17, 20,
21, and 23
103 Hanson, Wang,
Degenaro
2, 3, 5, 8, 9,
11, 14, 15,
17, 20, 21,
and 23

4, 10, 16, and 22 103 Hanson, Wang,
Nguyen
4, 10, 16,
and 22

Overall
Outcome
1–24