Ex Parte Feind

Patent Trials and Appeals BoardApr 18, 2019

13900757 - (D) (P.T.A.B. Apr. 18, 2019)

UNITED STA TES p A TENT AND TRADEMARK OFFICE APPLICATION NO. FILING DATE FIRST NAMED INVENTOR 13/900,757 05/23/2013 Karl Allan Feind 56436 7590 04/22/2019 Hewlett Packard Enterprise 3404 E. Harmony Road Mail Stop 79 Fort Collins, CO 80528 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. 90404856 1000 EXAMINER ASHLEY, BRUCE S ART UNIT PAPER NUMBER 2199 NOTIFICATION DATE DELIVERY MODE 04/22/2019 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): hpe.ip.mail@hpe.com chris.mania@hpe.com PTOL-90A (Rev. 04/07) UNITED STATES PATENT AND TRADEMARK OFFICE BEFORE THE PATENT TRIAL AND APPEAL BOARD Ex parte KARL ALLAN FEIND Appeal2017-006387 Application 13/900,757 1 Technology Center 2100 Before MICHAEL J. STRAUSS, IRVINE. BRANCH, and JASON J. CHUNG, Administrative Patent Judges. CHUNG, Administrative Patent Judge. DECISION ON APPEAL This is a decision on appeal under 35 U.S.C. Â§ 134(a) of the Non- Final Rejection of claims 1---6, 8-17, 19-28, and 30-36. We have jurisdiction under 35 U.S.C. Â§ 6(b). We affirm. INVENTION The invention is directed to allocating accelerators in a high performance computing system among threads. Spec. 1: 13-15. Claim 1 is illustrative of the invention and is reproduced below: 1. A method of distributing threads among accelerators in a high performance computing system, the method comprising: 1 According to Appellant, Silicon Graphics International Corp. is the real party in interest. App. Br. 3. Appeal2017-006387 Application 13/900,757 receiving, by an allocator, a request to assign an accelerator in the computing system to a thread, the request including a mode indicative of ( 1) a location of the accelerator and (2) exclusivity of the accelerator for use by the thread, corresponding to whether the accelerator will be (a) used by the thread exclusively or (b) used by the thread concurrently with other threads; selecting, by the allocator, the accelerator according to a processor assigned to the thread; and assigning, by the allocator, the accelerator to the thread with the exclusivity specified in the request. REJECTION AT ISSUE Claims 1---6, 8-17, 19-28, and 30-36 stand rejected under 35 U.S.C. Â§ 103 as being unpatentable over the combination of Aasheim (US 2012/0317568 Al; published Dec. 13, 2012), Koenen (US 2004/0019891 Al; published Jan. 29, 2004), and Garney (US 8,146,089 B2; published Mar. 27, 2012). Non-Final Act. 3-19. ANALYSIS The Examiner finds Aasheim teaches the limitation "receiving, by an allocator, a request to assign an accelerator in the computing system to a thread, the request including a mode" recited in claim 1. Non-Final Act. 3- 4. In addition, the Examiner finds Aasheim teaches a system determines a particular instruction set used by the scheduled thread such as whether one or more instructions set coprocessors are being requested, which the Examiner maps to the limitation "selecting, by the allocator, the accelerator according to a processor assigned to the thread; and assigning, by the allocator, the accelerator to the thread" recited in claim 1. Ans. 9-10 ( citing 2 Appeal2017-006387 Application 13/900,757 Aasheim ,r 46). The Examiner finds Garney teaches making a core exclusive or shared between partitions (i.e., the limitation "exclusivity of the accelerator for use by the thread, corresponding to whether the accelerator will be (a) used by the thread exclusively or (b) used by the thread concurrently with other threads" recited in claim 1). Non-Final Act. 5-7; Ans. 5---6 ( citing Aasheim ,r,r 30, 39; Garney 4:44--48). Appellant contends one skilled in the art would have understood from the description of the four modes "Local," "Local_Shared," "Near," and "Near Shared" in the Specification that assignment of an accelerator to a thread is either such that the thread obtains exclusive or shared use of the accelerator. App. Br. 8. Thus, Appellant argues the limitation "exclusivity of the accelerator for use by the thread" should be interpreted to mean whether the accelerator is assigned to a thread to execute the thread (i) exclusively or (ii) concurrently with other threads. Id. Based on the proffered claim interpretation, Appellant argues Garney merely teaches making a core exclusive or shared between partitions rather than obviating making a core exclusive or shared between threads. App. Br. 8-9; Reply Br. 5-7. And Appellant argues Aasheim and Koenen fail to teach "assigning ... the accelerator to the thread," wherein the accelerator has been selected based on a processor that has already been assigned to the thread because Aasheim and Koenen merely teach assigning an initial CPU to a thread. App. Br. 9-10. We disagree with Appellant. As an initial matter, we note that Appellant's argument that Aasheim and Koenen fail to teach "assigning ... the accelerator to the thread," (App. Br. 9-10) is unpersuasive because it fails to address the Examiner's finding that Aasheim's teaching of a system that determines a particular instruction 3 Appeal2017-006387 Application 13/900,757 set used by the scheduled thread ( such as whether one or more instructions set coprocessors are being requested) teaches or suggests the disputed limitation "selecting, by the allocator, the accelerator according to a processor assigned to the thread; and assigning, by the allocator, the accelerator to the thread" recited in claim 1. Ans. 9--10 ( citing Aasheim if 46). Regarding Appellant's argument (App. Br. 8 (citing Spec. 23:19-- 24:9)) that a person having ordinary skill in the art would understand the limitation "exclusivity ... for use" of the accelerator to mean four modes, we disagree because this is not recited in the claims. 2 Although we interpret claim limitations consistent with the specification, we do not import limitations from the specification into the claims. In re Van Geuns, 988 F.2d 1181, 1184 (Fed. Cir. 1993) (citing In re Zietz, 893 F.2d 319,321 (Fed. Cir. 1989)). In particular, when construing claim terminology during prosecution before the Office, claims are to be given their broadest reasonable interpretation consistent with the specification, reading claim language in light of the specification as it would have been interpreted by one of ordinary skill in the art. In re Am. Acad. of Sci. Tech Ctr., 367 F.3d 1359, 1364 (Fed. Cir. 2004). "Construing claims broadly during prosecution is not unfair to the applicant ... because the applicant has the opportunity to 2 Claim 1, a method claim, recites "corresponding to whether the accelerator will be (a) used by the thread exclusively or (b) used by the thread concurrently with other threads" ( emphasis added). In the event of further prosecution, the Examiner may want to consider if the italicized language above is a conditional limitation and, if so, whether the conditional language should be given patentable weight. This footnote applies only to the method claims. 4 Appeal2017-006387 Application 13/900,757 amend the claims to obtain more precise claim coverage." Id. Moreover, the Specification characterizes these "four modes" as an "embodiment" meaning that this is not a limiting definition. Spec. 23: 19--20. And significantly, Appellant admits that these are examples. App. Br. 8. As for Appellant's argument that Garney is deficient because it merely teaches making a core exclusive or shared between partitions rather than obviating making a core exclusive or shared between threads (App. Br. 8-9; Reply Br. 5-7), we disagree with Appellant because one cannot show nonobviousness "by attacking references individually" where the rejections are based on combinations of references. In re Merck & Co., Inc., 800 F.2d 1091, 1097 (Fed. Cir. 1986) (citing In re Keller, 642 F.2d 413,425 (CCPA 1981)). In this case, the Examiner finds Aasheim teaches scheduling threads and Garney teaches making a core exclusive or shared between partitions. Ans. 5---6 (citing Aasheim ,r,r 30, 39; Garney 4:44--48). Additionally, the Examiner finds a person having ordinary skill in the art would have combined Garney with Aasheim and Koenen because doing so provides another degree of control to the allocation method. Non-Final Act. 6-7 ( citing Garney, 2: 5---6). We, therefore, determine the Examiner has set forth sufficient "articulated reasoning with some rational underpinning to support the legal conclusion of obviousness." KSR Int 'l Co. v. Teleflex Inc., 550 U.S. 398,418 (2007) (quoting Kahn, 441 F.3d at 988). For the reasons discussed above, Appellant's arguments are unpersuasive of Examiner error. Appellant does not argue separately claims 2-6, 8-17, 19--28, and 30-36 with particularity, but asserts the rejections of those claims should be withdrawn for at least the same reasons as argued in 5 Appeal2017-006387 Application 13/900,757 independent claim 1. App. Br. 7-10; Reply Br. 5-8. Accordingly, we sustain the Examiner's rejection of: (1) independent claims 1, 12, 23, and 34; and (2) dependent claims 2---6, 8-11, 13-17, 19-22, 24--28, 30-33, 35, and 36 under 35 U.S.C. Â§ 103. We have only considered those arguments that Appellant actually raised in the Briefs. Arguments Appellant could have made, but chose not to make, in the Briefs have not been considered and are deemed to be waived. See 37 C.F.R. Â§ 4I.37(c)(l)(iv). DECISION We affirm the Examiner's decision rejecting claims 1-6, 8-17, 19-28, and 30-36 under 35 U.S.C. Â§ 103. No time period for taking any subsequent action in connection with this appeal may be extended under 37 C.F.R. Â§ 1.136(a)(l)(iv). AFFIRMED 6