Ex Parte Saulters et al

Patent Trials and Appeals Board

Mar 27, 2019

13754256 - (D) (P.T.A.B. Mar. 27, 2019)

UNITED STA TES p A TENT AND TRADEMARK OFFICE
APPLICATION NO. FILING DATE
13/754,256 01/30/2013
102324 7590 03/29/2019
Artegis Law Group, LLP/NVIDIA
7710 Cherry Park Drive Suite T #104
Houston, TX 77095
FIRST NAMED INVENTOR
Scott Saulters
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
ATTORNEY DOCKET NO. CONFIRMATION NO.
NVDA1TP100324TW1US 1541
EXAMINER
WANG, YUEHAN
ART UNIT PAPER NUMBER
2618
NOTIFICATION DATE DELIVERY MODE
03/29/2019 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):
kcruz@artegislaw.com
ALGdocketing@artegislaw.com
j matthews @artegislaw.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte SCOTT SAULTERS, RA TIN KUMAR, and
LIEVEN LEROY
Appeal2018-004900
Application 13/754,256
Technology Center 2600
Before JEREMY J. CURCURI, BARBARA A. BENOIT, and
ADAM J. PYONIN, Administrative Patent Judges.
CURCURI, Administrative Patent Judge.
DECISION ON APPEAL
Appellants appeal under 35 U.S.C. § 134(a) from the Examiner's
rejection of claims 1---6, 8-10, 12-17, and 19-21. Final Act. 1. We have
jurisdiction under 35 U.S.C. § 6(b ).
Claims 1-6, 8-10, 12-17, and 19-21 are rejected under pre-AIA
35 U.S.C. § 103(a) as obvious over Avkarogullari et al. (US 2013/0328890
Al, published Dec. 12, 2013) ("Avkarogullari") and Kumar (US 6,640,268
Bl, issued Oct. 28, 2003). Final Act. 2-14.
We reverse.
Appeal2018-004900
Application 13/754,256
STATEMENT OF THE CASE
Appellants' invention relates to "a method for adaptively adjusting a
framerate of a graphic processing unit (GPU), and more particularly, to a
method for changing the framerate adjustment." Spec. ,r 2. Claim 1 is
illustrative and reproduced below:
1. A method for adjusting a framerate of at least one
graphic processing unit ( GPU), comprising the steps of:
setting a polling cycle time interval for monitoring the at
least one GPU;
receiving an execution parameter associated with the at
least one GPU;
determining that a value of the execution parameter is
larger than a first reference value;
increasing a sleep time for the at least one GPU based on
the value of the execution parameter;
adjusting the polling cycle time interval based on a
relationship between the value of the execution parameter and
the first reference value;
determining that the value of the execution parameter has
become less than a second reference value, wherein the second
reference value is less than the first reference value; and
decreasing the sleep time for the at least one GPU based
on the value of the execution parameter.
PRINCIPLES OF LAW
We review the appealed rejections for error based upon the issues
identified by Appellants, and in light of the arguments and evidence
produced thereon. Ex parte Frye, 94 USPQ2d 1072, 107 5 (BP AI 2010)
(precedential).
2
Appeal2018-004900
Application 13/754,256
ANALYSIS
The Examiner finds A vkarogullari teaches all limitations of claim 1,
except for "setting a polling cycle time interval" and "adjusting the polling
cycle time interval based on a relationship between the value of the
execution parameter and the first reference value," which the Examiner finds
are taught by Kumar. See Final Act. 4--5, 8-9. 1 The Examiner reasons it
would have been obvious to combine the teachings of A vkarogullari and
Kumar "[to] permit a much more efficient allocation of bandwidth than is
available with static or fixed polling schemes in the method for adaptively
adjusting a framerate of a graphic processing unit (GPU)." Final Act. 5.
In particular, the Examiner maps Avkarogullari's decision block 80
and block 82 (Fig. 4) to claim 1 's "determining that a value of the execution
parameter is larger than a first reference value." See Final Act. 4 ( citing
Avkarogullari ,r 55). The Examiner explains as follows: "If the actual
power exceeds the target power ( decision block 80, 'yes' leg), the duty cycle
controller may decrease the duty cycle (i.e. increase the off time) (block
82)." Final Act. 4; see also Avkarogullari Fig. 3 (GPU Power Tracking
Controller 49 receives the difference between Actual GPU Power (feedback
signal) and Target GPU Power (reference signal) (see Avkarogullari ,r 54)).
In particular, the Examiner maps Avkarogullari's decision blocks 102,
106 and block 100 (Fig. 5) to claim 1 's "determining that the value of the
execution parameter has become less than a second reference value, wherein
the second reference value is less than the first reference value." See Final
1 The claims were amended after the Final Action. Among other
amendments to the claims, claim 1 was amended to incorporate the subject
matter of claim 7, and claim 7 was canceled.
3
Appeal2018-004900
Application 13/754,256
Act. 8 (citing Avkarogullari ,r,r 61, 62). The Examiner explains as follows:
"If neither of the decision blocks 102 and 106 are true ( decision blocks 102
and 106, 'no' legs), the GPU control unit 62 may be configured to transition
the GPU 24 to the lower power state (block 100) ... the second threshold is
also greater than the third threshold." Final Act. 8; see also Avkarogullari
Fig. 3 (GPU control unit 62 operates according to the flowchart of Fig. 5
(see Avkarogullari ,r 57)).
According to the Examiner, when GPU 24 (Avkarogullari, Fig. 3) is
transitioned to the lower power state (Avkarogullari, block 100, Fig. 5), this
affects the Actual GPU Power (Avkarogullari, Fig. 3), causing the duty
cycle controller to increase the duty cycle (i.e. decrease the off time)
(Avkarogullari, block 86, Fig. 4). See Final Act. 8-9 (citing Avkarogullari
,r 56) ("the duty cycle may be increased up to 100%, or the off time may be
zero"); see also Adv. Action 2; Ans. 7-11.
Appellants present the following principal argument: A vkarogullari
does not teach "determining that the value of the execution parameter has
become less than a second reference value, wherein the second reference
value is less than the first reference value; and decreasing the sleep time for
the at least one GPU based on the value of the execution parameter" as
recited in claim 1. See App. Br. 11-14.
The best simple explanation of Appellants' principal argument
appears in the Reply Brief:
Based on [the Examiner's] mappings [for the first
"determining" step ("execution parameter" reads on Actual GPU
Power, "first reference value" reads on Target GPU Power)], to
meet the second "determining" step of the independent claims set
forth above, Avkarogullari would have to disclose that (i) some
determination is made that the actual GPU power consumed falls
4
Appeal2018-004900
Application 13/754,256
below some second reference value, where that second reference
value is less than the target GPU power consumed, and (ii) the
sleep time of the GPU is then decreased based on actual GPU
power consumed. Importantly, A vkarogullari contains no such
teachings.
Reply Br. 4; see also App. Br. 12-13 (discussing inconsistent claim
mappings).
Appellants' arguments persuade us that the Examiner erred in
rejecting claim 1.
In reaching our decision, we need not speculate about the interplay
between Avkarogullari's GPU Power Tracking Controller 49 (Fig. 3, Fig. 4
(flowchart)) and Avkarogullari's GPU control unit 62 (Fig. 3, Fig. 5
(flowchart)) because the Examiner's error here arises in the way the
Examiner maps A vkarogullari to claim 1.
In finding A vkarogullari teaches claim 1 's "determining that a value
of the execution parameter is larger than a first reference value," the
Examiner maps Avkarogullari's Actual GPU Power (Fig. 3) to the recited
"execution parameter" and Avkarogullari's Target GPU Power (Fig. 3) to
the recited "first reference value." See Final Act. 4.
In finding A vkarogullari teaches claim 1 's "determining that the value
of the execution parameter has become less than a second reference value,
wherein the second reference value is less than the first reference value," the
Examiner maps A vkarogullari' s Utilization (Fig. 5) to the recited "execution
parameter" and Avkarogullari's Third Threshold (Fig. 5) to the recited
"second reference value." See Final Act. 8.
These mappings are inconsistent because Avkarogullari's Actual GPU
Power (Fig. 3) is not the same thing as Avkarogullari's Utilization (Fig. 5).
5
Appeal2018-004900
Application 13/754,256
See A vkarogullari ,r 24 ("The utilization may refer to the fraction ( or
percentage) of the execution interval [that] is actually used by the processor
to perform a task assigned to the processor for the execution interval."), ,r 40
("actual GPU power from a GPU power measurement unit 46").
Accordingly, we are persuaded of Examiner error for the reasons
given by Appellants on page 4 of the Reply Brief and quoted above.
We, therefore, do not sustain the Examiner's rejection of claim 1.
We also do not sustain the Examiner's rejection of claims 2-6, 8-10,
12, 13, and 21, which depend from claim 1.
Independent claim 14 recites the same key "determining" limitations
as claim 1. The Examiner rejects claim 14 "on similar grounds and rationale
as their corresponding limitations" in claim 1. Final Act. 12.
We, therefore, do not sustain the Examiner's rejection of claim 14.
We also do not sustain the Examiner's rejection of claims 15-1 7 and
19, which depend from claim 14.
Independent claim 20 recites the same key "determining" limitations
as claim 1. The Examiner rejects claim 20 "on similar grounds and rationale
as their corresponding limitations" in claim 1. Final Act. 13.
We, therefore, do not sustain the Examiner's rejection of claim 20.
ORDER
The Examiner's decision rejecting claims 1-6, 8-10, 12-17, and 19-
21 is reversed.
REVERSED
6