Regulating power using a fuzzy logic control system

• US 8,700,925 B2
• Filed: 09/01/2009
• Issued: 04/15/2014
• Est. Priority Date: 09/01/2009
• Status: Active Grant

• Current Assignee: NVIDIA Corporation
• Original Assignee: NVIDIA Corporation
• Inventors: Wyatt, David
• Primary Examiner(s): Bae, Ji H
• Application Number:
  US12/552,284
• Publication Number:
  US 20110055597A1
• Time in Patent Office: 1,687 Days
• Field of Search: 713/300, 713/320, 700/50, 706/1
• US Class Current: 713/300
• CPC Class Codes: G05B 13/0275 using fuzzy logic only
• CPC Class Codes: G06F 1/206 comprising thermal management
• CPC Class Codes: G06F 1/3203 Power management, i.e. event-based initiation of a power-saving mode
• CPC Class Codes: G06F 1/3206 Monitoring of events, devices or parameters that trigger a change in power modality
• CPC Class Codes: G06F 1/3293 by switching to a less power-consuming processor, e.g. sub-CPU
• CPC Class Codes: G06F 11/3024 where the computing system component is a central processing unit [CPU]
• CPC Class Codes: G06F 11/3062 where the monitored property is the power consumption power management in a computing system G06F1/3203
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

Metrics representing a combined measure of power used by a central processing unit (CPU) and power used by a graphics processing unit (GPU) are compared to a shared supply power and thermal power budgets. Power used by the CPU and power used by the GPU are regulated in tandem using a fuzzy logic control system that can implement fuzzy rules that describe the management within thermal and supply power design constraints of the platform.

Claims

1. A system comprising:

a central processing unit (CPU); and

a graphics processing unit (GPU) coupled to the CPU;

wherein power used by the CPU and power used by the GPU are regulated in tandem using a fuzzy logic control system operable for implementing a plurality of fuzzy logic rules, wherein inputs to the fuzzy logic control system include a first value for the power used by the CPU and a second value for the power used by the GPU, wherein the sum of the first value and the second value is compared to a shared supply power and thermal power budget value, and wherein a state of the CPU and a state of the GPU are regulated to maintain the sum within the shared supply power and thermal power budget value;

wherein further the first value comprises the power used by the CPU integrated over a period of time, the second value comprises the power used by the GPU integrated over the period of time, and the shared power supply and thermal power budget shared power supply and thermal power comprises an integrated value of the budget over the period of time.

2. The system of claim 1 wherein the plurality of fuzzy logic rules comprises subsets of the fuzzy logic rules, wherein the subsets correspond to different control policies with different priorities.

3. The system of claim 2 wherein a subset of the fuzzy logic rules is implemented based on a selection made by an end user of the system.

4. The system of claim 1 wherein inputs to the fuzzy logic control system include pre-determined profile information for an executing application.

5. The system of claim 1 further comprising:

a first sensor operable for measuring the power used by the CPU; and

a second sensor operable for measuring the power used by the GPU, wherein the first and second sensors are operable for sending an event signal when a threshold is crossed and wherein, in response to the event signal, the fuzzy logic rules are evaluated.

6. The system of claim 1 wherein sets describing membership of CPU and GPU component power and temperature measurements, within natural language ranges, are embedded in firmware such that control rules can be applied independent of components in a specific platform.

7. The system of claim 6 wherein platform-specific behaviors are supported by loading the control rules from the firmware.

8. A computer-implemented method for managing a system comprising a central processing unit (CPU) and a graphics processing unit (GPU), the method comprising:

accessing a plurality of input variables having values that characterize the system in operation, wherein the values comprise a first value for the power used by the CPU and a second value for the power used by the GPU, the first value determined by integrating the power used by the CPU over a period of time, the second value determined by integrating the power used by the GPU integrated over the period of time;

evaluating a plurality of fuzzy logic rules using the values, the evaluating comprising comparing the sum of the first value and the second value to a shared power budget value, the shared power budget value determined by integrating the shared power budget over the period of time; and

regulating a state of the CPU and a state of the GPU according to a result of the evaluating, the regulating comprising maintaining the sum within the shared power budget value.

9. The method of claim 8 wherein the plurality of fuzzy logic rules comprises subsets of the fuzzy logic rules, wherein the subsets correspond to different control policies with different priorities.

10. The method of claim 8 further comprising selecting a subset of the fuzzy logic rules based on a selection made by an end user of the system.

11. The method of claim 8 wherein the values comprise pre-determined profile information for an executing application.

12. The method of claim 8 further comprising evaluating the fuzzy logic rules in response to an event signal from one of a first sensor operable for measuring the power used by the CPU and a second sensor operable for measuring the power used by the GPU, wherein the event signal is triggered when a threshold is crossed.

13. The method of claim 8 wherein the variables comprise linguistic variables having natural language values.

14. A system for managing a central processing unit (CPU) and a graphics processing unit (GPU), the system comprising:

a management unit operable for executing a fuzzy logic control system operable for implementing a plurality of fuzzy logic rules; and

a controller coupled to the management unit and operable for regulating power used by the CPU and power used by the GPU according to an output of the control system, wherein inputs to the fuzzy logic control system include a first value for the power used by the CPU and a second value for the power used by the GPU, wherein the sum of the first value and the second value is compared to a shared power budget value, and wherein a state of the CPU and a state of the GPU are regulated to maintain the sum within the shared power budget value;

wherein further the first value comprises the power used by the CPU integrated over a period of time, the second value comprises the power used by the GPU integrated over the period of time, and the shared power budget value comprises an integrated value of the shared power budget over the period of time.

15. The system of claim 14 wherein the plurality of fuzzy logic rules comprises subsets of the fuzzy logic rules, wherein the subsets correspond to different control policies with different priorities.

16. The system of claim 15 wherein a subset of the fuzzy logic rules is implemented based on a selection made by an end user of the system.

17. The system of claim 14 wherein inputs to the fuzzy logic control system comprise pre-determined profile information for an executing application.

18. The system of claim 14 wherein the fuzzy logic rules are evaluated in response to an event signal from one of a first sensor operable for measuring the power used by the CPU and a second sensor operable for measuring the power used by the GPU, wherein the event signal is triggered when a threshold is crossed.