Regulating power within a shared budget

• US 8,826,048 B2
• Filed: 09/01/2009
• Issued: 09/02/2014
• Est. Priority Date: 09/01/2009
• Status: Active Grant

• Current Assignee: NVIDIA Corporation
• Original Assignee: NVIDIA Corporation
• Inventors: Wyatt, David
• Primary Examiner(s): Spittle, Matthew D
• Application Number:
  US12/552,279
• Publication Number:
  US 20110055596A1
• Time in Patent Office: 1,827 Days
• Field of Search: 713/300, 713/310, 713/320, 713/340, 714/4.12, 714/4.21, 714/10, 714/11, 714/13
• US Class Current: 713/320
• 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 1/325 Power saving in peripheral device
• 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/or shared thermal power budget. A state of the CPU and a state of the GPU are regulated by the power management system to maintain the metrics within the shared supply power and/or thermal power budget for the purpose of managing the components to meet platform power supply, and cooling, constraints and design requirements.

Claims

1. An apparatus comprising:

a central processing unit (CPU); and

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

wherein a first thermal management system compares a metric representing a combined measure of power used by the CPU and power used by the GPU to a shared power budget, wherein a state of the CPU and a state of the GPU are regulated to maintain the metric within the shared power budget,

wherein responsibility for managing the CPU and the GPU transitions from a legacy thermal management system of the apparatus to the first thermal management system when the first thermal management system is enabled, and transitions from the first thermal management system to the legacy thermal management system when the first thermal management system is disabled, wherein the legacy thermal management system specifies the budget to the first thermal power management system in response to an event that affects the shared power budget.

2. The apparatus of claim 1 wherein the shared power budget comprises a budget selected from the group consisting of:

a shared thermal power budget; and

a shared supply power budget.

3. The apparatus of claim 1 wherein the metric further comprises a prediction of a value for the metric.

4. The apparatus of claim 1 further comprising:

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

a second sensor operable for measuring 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, a management unit coupled to the first and second sensors is operable for polling the first and second sensors to access data useful for determining the metric.

5. The apparatus of claim 4 further comprising a controller operable for receiving an alert from the management unit, wherein in response to the alert the controller initiates a change to at least one of the state of the CPU and the state of the GPU.

6. The apparatus of claim 1 wherein the state of CPU and the state of the GPU also are regulated using pre-determined profile information for an executing application.

7. The apparatus of claim 1 wherein the state of the CPU and the state of the GPU also are regulated according to preferences selected from the group consisting of:

user preferences;

operating system preferences; and

platform-specific preferences.

8. The apparatus of claim 1 further comprising a cooling fan having an adjustable fan speed, wherein the fan speed is reduced if the metric is less than the shared power budget.

9. The apparatus of claim 1 wherein the legacy thermal management system polls the first thermal management system to confirm operability of the first thermal management system.

10. The apparatus of claim 1 wherein the legacy thermal management system changes the budget in response to a change in cooling capability selected from the group consisting of:

fan speed;

internal temperature; and

external temperature.

11. The apparatus of claim 1 wherein further the power used by the CPU and the power used by the GPU are each integral values determined for sliding and overlapping windows of time.

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

a management unit operable for comparing a metric comprising an integral value of power utilized by the CPU and power utilized by the GPU over sliding and overlapping windows of time to a budget;

a controller coupled to the management unit and operable for regulating power used by the CPU and power used by the GPU to maintain the metric within the budget; and

a thermal control system coupled to the management unit, wherein the management unit performs a handshake with the thermal control system to relax thermal limits implemented by the thermal control system when the management unit is enabled, wherein the thermal limits are defaulted to when the management unit is not enabled, wherein the thermal control system determines the budget and conveys the budget to the management unit.

13. The system of claim 12 wherein the metric further comprises a prediction of power utilization by the CPU and by the GPU.

14. The system of claim 12 wherein the thermal control system comprises a first sensor operable to measure power used by the CPU and a second sensor operable to measure 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 management unit is operable for polling the first and second sensors to access data useful for determining the metric.

15. The system of claim 14 wherein the management unit is further operable for sending an alert to the controller, wherein in response to the alert the controller initiates a change in at least one of a state of the CPU and a state of the GPU.

16. The system of claim 15 wherein the state of the CPU and the state of the GPU also are regulated using pre-determined profile information for an executing application.

17. The system of claim 15 wherein the state of CPU and the state of the GPU also are regulated according to user preferences.

18. The system of claim 12 further comprising a cooling fan having an adjustable fan speed, wherein the fan speed is reduced if the metric is less than the budget.

19. A computer-implemented method of power control, the method comprising:

receiving, at a first thermal management system, a thermal power budget determined by a legacy thermal management system;

performing a handshake between the first thermal management system and the legacy thermal management system, wherein thermal limits implemented by the legacy thermal management system are relaxed in response to the handshake;

comparing a metric comprising an integral value of power utilized by a central processing unit (CPU) and power utilized by a graphics processing unit (GPU) over sliding and overlapping windows of time to the thermal power budget; and

regulating a state of the CPU and a state of the GPU to keep the metric within the thermal power budget, wherein responsibility for managing the CPU and the GPU transitions from the legacy thermal management system to the first thermal management system when the first thermal management system is enabled, and transitions from the first thermal management system to the legacy thermal management system when the first thermal management system is disabled.

20. The method of claim 19 wherein the metric further comprises a prediction of power utilized by the CPU and by the GPU.

21. The method of claim 19 further comprising:

receiving an event signal from at least one of a first sensor operable for measuring power used by the CPU and a second sensor operable for measuring power used by the GPU;

in response to the event signal, polling the first and second sensors to access data useful for determining the integral value; and

sending an alert that initiates a change in at least one of a state of the CPU and a state of the GPU.

22. The method of claim 19 wherein the regulating the state of the CPU and the state of the GPU comprises using information selected from the group consisting of:

pre-determined profile information for an executing application, and user preferences.

23. The method of claim 19 further comprising reducing a speed of a cooling fan if the metric is less than the thermal power budget.