Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Status of Claims
This action is in reply to the application filed on 20 January 2026.
Claims 1, 3-13 and 15-22 are currently pending and have been examined.
Claims 2 and 14 have been canceled.
Claims 21-22 have been added.
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claim 1, 3-13 and 15-22 are rejected under 35 U.S.C. 103 as being unpatentable over Vijayaraghavan (US 2019/0050265 A1) in view of Dillenberger (US 8869160 B2).
Claims 1 and 13:
Vijayaraghavan discloses the following limitation:
A method comprising: executing, by a computing device, a deep learning task on a deep learning framework; (see at least figures 1, 3-5 and related text disclosing executing a deep learning task on a deep learning framework.)
Determining a combination of at least two accelerators of a plurality of accelerators to execute the deep learning task by: (see at least figures 1, 3-5 and related text disclosing executing a deep learning task on a deep learning framework. In [0050], disclosing “The example neural network processor 118 uses the neural network parameters stored in the example neural network parameter memory 260 to generate an indication of one or more accelerators to be used to execute the workload. (Block 530). The accelerator selection processor 114 then provides the workload to the selected accelerator(s) via the example accelerator interface 240. (Block 540). In some examples, multiple different accelerators may be used. In such an example, the workload may be segmented and/or divided into portions for execution among the different accelerator(s). The accelerator(s) may then execute the workload and/or portions thereof in a parallel fashion and/or in a serial fashion. The example accelerator selection processor 114 obtains the results of the workload execution via the example accelerator interface 240. (Block 550). The example accelerator selection processor 114 collects performance metrics resulting from the execution of the workload, and stores those performance metrics (and the attribute(s) of the workload) as training data in the example training data store 245. (Block 555). ”)
selecting at least two accelerators from the plurality of accelerators … wherein the plurality of accelerators comprises at least one primary accelerator and at least one secondary accelerator that is heterogeneous to the at least one primary accelerator; (see at least figures 1, 3 and 5 and related text disclosing determining at least two accelerators of a plurality of accelerators to execute the deep learning task where the accelerators are heterogeneous.)
allocating the deep learning task to the combination of at least two accelerators; and (see at least figure 5 and related text including [0050] disclosing allocating the deep learning task to multiple determined accelerators to the perform task in parallel.) (Examiner notes that Applicants specification never actually uses the word combination but does support using two accelerators to perform the task(s) in parallel (see at least Applicants specification page 21 and figure 3 (330 and 332)). Examiner will interpret the claim language “combination of at least two accelerators” to mean using at least two accelerators to perform the task(s) in parallel.)
generating, based on a result processed by the combination of at least two accelerators, result data for the deep learning task. (see at least figure 5 and related text disclosing generating result data for the deep learning task.)
Vijayaraghavan discloses the limitations as shown in the rejection above including “Determining a combination of at least two accelerators of a plurality of accelerators to execute the deep learning task by: selecting at least two accelerators from the plurality of accelerators … wherein the plurality of accelerators comprises at least one primary accelerator and at least one secondary accelerator that is heterogeneous to the at least one primary accelerator“ (see at least figures 1, 3-5 and related text). However, Vijayaraghavan does not specifically discloses “selecting at least two accelerators from the plurality of accelerators based on comparing expected response times and expected throughput values for each accelerator of the plurality of accelerators to a response time priority and a throughput requirement corresponding to the deep learning task.”
However, Dillenberger, discloses the following limitations:
selecting at least two accelerators from the plurality of accelerators based on comparing expected response times and expected throughput values for each accelerator of the plurality of accelerators to a response time priority and a throughput requirement corresponding to the deep learning task, (see Dillenberger, col 3:20-67- 4: 1-53 and col 5: 1-55 disclosing the jobs are analyzed and monitored and assigned goals and priorities including response time, queue time, execution time based on that information the accelerator assignment is configured.)
It would have been obvious to a person of ordinary skill in the art at the time the invention was made to combine the teachings of Dillenberger with Vijayaraghavan because the ability to balance accelerator resources based on workloads allows a higher overall utilization to be achieved (see Dillenberger, col 3: 1-11).
Moreover, it would have been obvious to one of ordinary skill in the art at the time of the invention to include the selecting at least two accelerators from the plurality of accelerators based on comparing expected response times and expected throughput values for each accelerator of the plurality of accelerators to a response time priority and a throughput requirement corresponding to the deep learning task as taught by Dillenberger in Vijayaraghavan, since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art would have recognized that the results of the combination were predictable.
Claims 3 and 15:
The combination of Vijayaraghavan/ Dillenberger discloses the limitations above.
Further, Vijayaraghavan discloses the following limitation:
wherein the determining the combination of at least two accelerators comprises: determining whether the deep learning task is executable on the at least one secondary accelerator; and (see at least figures 1, 3-4 and related text disclosing determining that the deep learning task can be executed on at least one secondary accelerator.)
based on the deep learning task being executable on the at least one secondary accelerator, determining, from among the at least one secondary accelerator, an accelerator to process the deep learning task by selecting at least one of: a secondary accelerator included in a first node executing the deep learning framework; or a secondary accelerator included in a second node connected to the first node through a network. (see at least figures 1, 3-5 and related text and [0061-64] disclosing based on the deep learning task being executable on the at least one secondary accelerator, determining an accelerator to process the workload by selecting a secondary accelerator included in the second node. Also see at least [0065-90].) (Also see Dillenberger, figure 1 and related text and col 3:20-67- 4: 1-53 and col 5: 1-55.)
Claim 4:
The combination of Vijayaraghavan/ Dillenberger discloses the limitations above.
Vijayaraghavan discloses the limitations as shown in the rejection above including “Determining a combination of at least two accelerators of a plurality of accelerators to execute the deep learning task by: selecting at least two accelerators from the plurality of accelerators “ (see at least figures 1, 3-5 and related text). However, Vijayaraghavan does not specifically discloses “wherein the selecting the at least two accelerators is further based on a response time of the deep learning task.”
However, Dillenberger, discloses the following limitations:
wherein the selecting the at least two accelerators is further based on a response time of the deep learning task. (see Dillenberger, col 3:20-67- 4: 1-53 and col 5: 1-55 disclosing the jobs are analyzed and monitored and assigned goals and priorities including response time, queue time, execution time based on that information the accelerator assignment is configured.)
It would have been obvious to a person of ordinary skill in the art at the time the invention was made to combine the teachings of Dillenberger with Vijayaraghavan because the ability to balance accelerator resources based on workloads allows a higher overall utilization to be achieved (see Dillenberger, col 3: 1-11).
Moreover, it would have been obvious to one of ordinary skill in the art at the time of the invention to include the selecting the at least two accelerators is further based on a response time of the deep learning task. as taught by Dillenberger in Vijayaraghavan, since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art would have recognized that the results of the combination were predictable.
Claim 5:
The combination of Vijayaraghavan/ Dillenberger discloses the limitations above.
Further, Vijayaraghavan discloses the following limitation:
wherein the selecting the at least two accelerators comprises: based on an expected execution time of the deep learning task being shorter than a predetermined time, selecting at least one secondary accelerator included in a first node to process the deep learning task. (see at least figures 3-5 and related text.) (Also see Dillenberger, figure 1 and related text and col 3:20-67- 4: 1-53 and col 5: 1-55.)
Claim 6:
The combination of Vijayaraghavan/ Dillenberger discloses the limitations above.
Further, Vijayaraghavan discloses the following limitation:
wherein the selecting the at least two accelerators comprises: based on an expected throughput of the deep learning task being equal to or less than a predetermined throughput, selecting at least one secondary accelerator included in a first node to process the deep learning task. (see at least figures 3-5 and related text.) (Also see Dillenberger, figure 1 and related text and col 3:20-67- 4: 1-53 and col 5: 1-55.)
Claims 7 and 16:
The combination of Vijayaraghavan/ Dillenberger discloses the limitations above.
Further, Vijayaraghavan discloses the following limitation:
wherein the determining the combination of the at least two accelerators of the plurality of accelerators to execute the deep learning task comprises: dividing the deep learning task into a plurality of partial deep learning tasks; and assigning, to each of the at least two accelerators, a different task of the plurality of partial deep learning tasks. (see at least figures 3-5 and related text.) (Also see Dillenberger, figure 1 and related text and col 3:20-67- 4: 1-53 and col 5: 1-55.)
Claims 8 and 17:
The combination of Vijayaraghavan/ Dillenberger discloses the limitations above.
Further, Vijayaraghavan discloses the following limitation:
wherein the dividing the deep learning task into a plurality of partial deep learning tasks comprises at least one of: based on the throughput requirement of the deep learning task being equal to or greater than a predetermined throughput, dividing the deep learning task into the plurality of partial deep learning tasks; or based on the response time priority of the deep learning task being equal to or longer than a predetermined time, dividing the deep learning task into the plurality of partial deep learning tasks. (see at least figures 3-5 and related text and [0042] disclosing “In some examples, a portion of the workload is provided to the accelerator. Providing a portion of the workload ensures that, for workloads that would otherwise take long amounts of time to complete, the workload can be completed in a shorter amount of time. As a result, the amount of time required to determine which accelerator should be selected is reduced.” and [0044] disclosing that “the accelerator selection processor 114 may select an accelerator that resulted in the shortest execution time.” Further, [0050] disclosing that “the workload may be segmented and/or divided into portions for execution among the different accelerator(s). The accelerator(s) may then execute the workload and/or portions thereof in a parallel fashion and/or in a serial fashion. ”) (Also see Dillenberger, figure 1 and related text and col 3:20-67- 4: 1-53 and col 5: 1-55.)
Claims 9 and 18:
The combination of Vijayaraghavan/ Dillenberger discloses the limitations above.
Further, Vijayaraghavan discloses the following limitation:
wherein the allocating the deep learning task comprises: providing the plurality of partial deep learning tasks to a scheduler that manages a plurality of secondary accelerators included in a second node; selecting, by the scheduler, one or more executable secondary accelerators from among the plurality of secondary accelerators included in the second node; and allocating, by the scheduler, the plurality of partial deep learning tasks to the selected one or more executable secondary accelerators. (see at least figures 3-5 and related text and [0042] disclosing “In some examples, a portion of the workload is provided to the accelerator. Providing a portion of the workload ensures that, for workloads that would otherwise take long amounts of time to complete, the workload can be completed in a shorter amount of time. As a result, the amount of time required to determine which accelerator should be selected is reduced.” and [0044] disclosing that “the accelerator selection processor 114 may select an accelerator that resulted in the shortest execution time.” Further, [0050] disclosing that “the workload may be segmented and/or divided into portions for execution among the different accelerator(s). The accelerator(s) may then execute the workload and/or portions thereof in a parallel fashion and/or in a serial fashion. ”) (Also see Dillenberger, figure 1 and related text and col 3:20-67- 4: 1-53 and col 5: 1-55.)
Claims 10 and 19:
The combination of Vijayaraghavan/ Dillenberger discloses the limitations above.
Further, Vijayaraghavan discloses the following limitation:
wherein the dividing the deep learning task into a plurality of partial deep learning tasks comprises dividing input data of the deep learning task into a plurality of partial input data sets, and wherein the allocating the plurality of partial deep learning tasks to the selected one or more executable secondary accelerators comprises allocating a function of the deep learning task and each of the plurality of partial input data sets to the selected one or more executable secondary accelerators. (see at least figures 3-5 and related text.) (Also see Dillenberger, figure 1 and related text and col 3:20-67- 4: 1-53 and col 5: 1-55.)
Claim 11:
The combination of Vijayaraghavan/ Dillenberger discloses the limitations above.
Further, Vijayaraghavan discloses the following limitation:
wherein the dividing the deep learning task into a plurality of partial deep learning tasks comprises dividing parameter data of the deep learning task into a plurality of partial parameter data sets, and wherein the allocating the plurality of partial deep learning tasks to the selected one or more executable secondary accelerators comprises allocating a function of the deep learning task and each of the plurality of partial parameter data sets to the selected one or more executable secondary accelerators. (see at least figures 3-5 and related text and [0042] disclosing “In some examples, a portion of the workload is provided to the accelerator. Providing a portion of the workload ensures that, for workloads that would otherwise take long amounts of time to complete, the workload can be completed in a shorter amount of time. As a result, the amount of time required to determine which accelerator should be selected is reduced.” and [0044] disclosing that “the accelerator selection processor 114 may select an accelerator that resulted in the shortest execution time.” Further, [0050] disclosing that “The example accelerator selection processor 114 provides the attribute(s) of the workload to the neural network processor 118 for selection of an accelerator to be used to execute the workload. The example neural network processor 118 uses the neural network parameters stored in the example neural network parameter memory 260 to generate an indication of one or more accelerators to be used to execute the workload. (Block 530). The accelerator selection processor 114 then provides the workload to the selected accelerator(s) via the example accelerator interface 240. (Block 540). In some examples, multiple different accelerators may be used. In such an example, the workload may be segmented and/or divided into portions for execution among the different accelerator(s). The accelerator(s) may then execute the workload and/or portions thereof in a parallel fashion and/or in a serial fashion. The example accelerator selection processor 114 obtains the results of the workload execution via the example accelerator interface 240. (Block 550). The example accelerator selection processor 114 collects performance metrics resulting from the execution of the workload, and stores those performance metrics (and the attribute(s) of the workload) as training data in the example training data store 245.”) (Also see Dillenberger, figure 1 and related text and col 3:20-67- 4: 1-53 and col 5: 1-55.)
Claims 12 and 20:
The combination of Vijayaraghavan/ Dillenberger discloses the limitations above.
Further, Vijayaraghavan discloses the following limitation:
further comprising: prior to the determining the combination of the at least two accelerators, determining that an operation of the deep learning task requires a plurality of accelerators; and based on the operation of the deep learning task requiring a plurality of accelerators and based on a quantity of accelerators allocated to a first node executing the deep learning framework being less than a quantity of the required plurality of accelerators, scheduling at least a portion of the deep learning task for execution on at least one of the accelerators allocated to the first node, wherein the deep learning task includes scheduling information for the deep learning task scheduled to be executed on the at least one of the accelerators allocated to the first node. (see at least figures 3-5 and related text.) (Also see Dillenberger, figure 1 and related text and col 3:20-67- 4: 1-53 and col 5: 1-55.)
Claims 21 and 22:
The combination of Vijayaraghavan/ Dillenberger discloses the limitations above.
Further, Vijayaraghavan discloses the following limitation:
wherein the combination of the at least two accelerators comprises at least two secondary accelerators. (see at least figures 1, 3-5 and related text disclosing executing a deep learning task on a deep learning framework. In [0050], disclosing “The accelerator selection processor 114 then provides the workload to the selected accelerator(s) via the example accelerator interface 240. (Block 540). In some examples, multiple different accelerators may be used. In such an example, the workload may be segmented and/or divided into portions for execution among the different accelerator(s). The accelerator(s) may then execute the workload and/or portions thereof in a parallel fashion...”) (Also see Dillenberger, figure 1 and related text and col 3:20-67- 4: 1-53 and col 5: 1-55.)
Response to Arguments
Applicant’s arguments with respect newly added limitation of claims 1, 3-13 and 15-22 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Sengupta (US 2020/0004595 A1) fault-tolerant accelerator based inference service.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to PATRICIA H MUNSON whose telephone number is (571)270-5396. The examiner can normally be reached M-F 7:30am-4:30pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Tariq Hafiz can be reached at 571-272-5350. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/PATRICIA H MUNSON/Supervisory Patent Examiner, Art Unit 3624
Prosecution Insights