Method and system for profile learning window optimization

Patent metadata

Field	Value
Publication number	US-11029972-B2
Application number	US-201916265464-A
Country	US
Kind code	B2
Filing date	Feb 1, 2019
Priority date	Feb 1, 2019
Publication date	Jun 8, 2021
Grant date	Jun 8, 2021

Abstract

An information handling system operating a performance optimization system may comprise a processor executing computer program code instructions that interact with a plurality of computer operations and that is configured for iteratively sampling field performance data of the information handling system during learning windows having a preset duration and occurring at a preset frequency according to optimal learning window parameters, and adjusting the performance of the information handling system via adjustment of optimized system configurations based on application of a predetermined statistical model to the iteratively sampled field performance data. The optimal learning window parameters may be determined based on accuracy of previous application of the predetermined statistical model to test performance data of the information handling system.

First claim

What is claimed is:

1. An information handling system operating a performance optimization system comprising:
a data bus coupled to a processor and a non-transitory, computer-readable storage medium; and
the processor executing computer program code instructions that interact with a plurality of computer operations and that is configured for:
inputting an application workload into a predetermined statistical model to output a predicted optimized hardware configuration for the information handling system;
iteratively sampling hardware component field performance data describing performance of a hardware component of the information handling system during one or more learning windows, the learning windows having a preset duration and occurring at a preset frequency according to optimal learning window parameters to decrease computing resources consumed during determination of the predicted optimized hardware configuration by the predetermined statistical model, wherein the optimal learning window parameters are determined based on accuracy of previous application of the predetermined statistical model to previous hardware component test performance data as previously sampled during testing;
adjusting performance of the hardware component of the information handling system by configuring the hardware component according to the predicted optimized hardware configuration based on application of the predetermined statistical model to the iteratively sampled field performance data.

2. The information handling system of claim 1 , wherein the processor executing computer program code instructions is further configured for:
characterizing workloads in operation at runtime to analyze performance of the hardware component of the information handling system during usage; and
identifying a the hardware component configuration to be changed to obtain optimal performance of the hardware component based upon the characterizing of workloads.

3. The information handling system of claim 2 , wherein the adjustment of the performance of the hardware component is performed dynamically for varying workloads to account for varying workloads, the varying workloads comprising random stochastic variation in workload, or abrupt user or operating system level discontinuities in jobs.

4. The information handling system of claim 1 , wherein the optimal learning window parameters includes an optimal voting method and the optimized system configurations are determined based upon the optimal voting method.

5. The information handling system of claim 1 , wherein the accuracy of previous application of the predetermined statistical model is defined by an area under a receiver operating characteristic curve for the previous application of the predetermined statistical model.

6. The information handling system of claim 1 , wherein the iteratively sampled hardware component field performance data excludes performance data falling below a utilization threshold.

7. The information handling system of claim 1 , wherein the predetermined statistical model comprises a supervised learning operation, the supervised learning operation statistically mapping a measurement of performance of the hardware component of the information handling system to an optimal hardware component configuration.

8. A method for optimizing performance of an information handling system comprising:
sampling, via a processor, hardware component testing performance measurements of a hardware component of the information handling system over a preset test monitoring duration and partitioning the hardware component testing performance measurements within the preset test monitoring duration into one or more learning windows having a frequency and a window duration preset according to each of a plurality of candidate learning window parameter configurations;
applying a statistical model to the sampled hardware component testing performance measurements for each of the plurality of candidate learning window parameter configurations to yield modeled system settings;
determining an accuracy associated with each candidate learning window parameter configuration by determining accuracy of each modeled system setting compared with a known optimal system setting; and
identifying a most accurate window learning parameter configuration as the optimal window learning parameter configuration for use in sampling hardware component field performance measurements describing performance of a hardware component of the information handling system during field usage to determine field optimized system settings according to the statistical model using limited computing resources.

9. The method of claim 8 further comprising:
characterizing workloads in operation at runtime to analyze test performance of the hardware component of the information handling system during usage; and
identifying a hardware component configuration to be changed to obtain optimal hardware component field performance based upon the characterizing of workloads.

10. The method of claim 9 , wherein the characterizing workloads is performed dynamically for varying workloads to account for varying workloads, the varying workloads comprising random stochastic variation in workload, or abrupt user or operating system level discontinuities in jobs.

11. The method of claim 8 further comprising:
excluding from consideration as the optimal tuning parameter configuration a candidate tuning parameter configuration having a candidate accuracy falling below an accuracy threshold.

12. The method of claim 11 , wherein the accuracy threshold is a false positive rate of 20%.

13. The method of claim 8 , wherein the optimal learning window parameters include an optimal voting method for determination of the field optimized hardware component configurations.

14. The method of claim 8 , wherein the candidate accuracy for each of the plurality of candidate tuning parameter configurations is defined by an area under a receiver operating characteristic curve for the candidate laboratory optimal system setting determination.

15. An information handling system operating a performance optimization system comprising:
a data bus coupled to a processor and a non-transitory, computer-readable storage medium; and
the processor executing computer program code instructions that interact with a plurality of computer operations and that is configured for:
inputting an application workload into a predetermined statistical model to output a predicted optimized hardware configuration for the information handling system;
iteratively sampling hardware component field performance data describing performance of a hardware component of the information handling system during one or more learning windows, the learning windows having a preset duration and occurring at a preset frequency according to optimal learning window parameters to decrease computing resources consumed during determination of the predicted optimized hardware configuration by the predetermined statistical model, wherein the optimal learning window parameters are determined based on accuracy of previous application of the predetermined statistical model to previous hardware component test performance data as previously sampled during testing;
characterizing workloads in operation at runtime based on application of the predetermined statistical model to the iteratively sampled field performance data;
identifying a hardware component configuration to be changed to obtain optimal performance of the hardware component based upon the characterizing of workloads;
applying the hardware component configuration

Assignees

• Dell Products Lp

Inventors

• Vichare Nikhil M
• Khosrowpour Farzad

Classifications

• G06F11/3089

  Monitoring arrangements determined by the means or processing involved in sensing the monitored data, e.g. interfaces, connectors, sensors, probes, agents (software debugging using additional hardware using a specific debug interface G06F11/3656; performance evaluation by tracing or monitoring G06F11/3466) · CPC title

• G06F11/3447

  Performance evaluation by modeling · CPC title

• G06F9/44505 Primary

  Configuring for program initiating, e.g. using registry, configuration files · CPC title

• G06F11/3409

  for performance assessment · CPC title

• G06F11/3024

  where the computing system component is a central processing unit [CPU] · CPC title