METHODS AND APPARATUSES FOR LOAD BALANCING BETWEEN MULTIPLE PROCESSING UNITS

• US 20090109230A1
• Filed: 10/24/2007
• Published: 04/30/2009
• Est. Priority Date: 10/24/2007
• Status: Active Grant

• Current Assignee: Apple Inc.
• Original Assignee: Apple Inc.
• Inventors: Miller, Howard, Brunner, Ralph
• Granted Patent: US 8,284,205 B2
• Time in Patent Office: Days
• US Class Current: 345/506
• CPC Class Codes: G06F 1/3203 Power management, i.e. event-based initiation of a power-saving mode
• CPC Class Codes: G06F 1/329 by task scheduling
• CPC Class Codes: G06F 15/16 Combinations of two or more digital computers each having at least an arithmetic unit, a program unit and a register, e.g. for a simultaneous processing of several programs coordinating program control therefor G06F9/52; in regulating and control system G05B
• CPC Class Codes: G06F 2209/501 Performance criteria
• CPC Class Codes: G06F 9/5044 considering hardware capabilities
• CPC Class Codes: G06F 9/5094 where the allocation takes into account power or heat criteria power management in computers in general G06F1/3203; thermal management in computers in general G06F1/206
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

Exemplary embodiments of methods and apparatuses to dynamically redistribute computational processes in a system that includes a plurality of processing units are described. The power consumption, the performance, and the power/performance value are determined for various computational processes between a plurality of subsystems where each of the subsystems is capable of performing the computational processes. The computational processes are exemplarily graphics rendering process, image processing process, signal processing process, Bayer decoding process, or video decoding process, which can be performed by a central processing unit, a graphics processing units or a digital signal processing unit. In one embodiment, the distribution of computational processes between capable subsystems is based on a power setting, a performance setting, a dynamic setting or a value setting.

Claims

1. A machine implemented method comprising:

selecting a setting of a system that has a plurality of processing units;

redistributing a processing task of the system between the processing units based on the setting.

2. A method as in claim 1 wherein the redistributing is done dynamically while the system is running and wherein the processing units comprise a central processing unit (CPU) and a special processing unit, and

wherein the processing task is created from one of a plurality of different software pipelines.

3. A method as in claim 2 wherein the special processing unit is a graphics processing unit (GPU) and wherein the GPU performs certain graphics processing operations in a high performance mode and the CPU performs those certain graphics processing operations in a power conservation mode.

4. A method as in claim 2, wherein the special processing unit is a GPU and wherein the GPU performs certain graphics processing operations in a power conservation mode and the CPU performs those certain graphics processing operations in a high performance mode.

5. A method as in claim 2 wherein the processing task is at least one of a graphics rendering process, a image processing process, a signal processing process, a Bayer filtering process, a video decoding process, or any combination thereof and wherein the redistributing a processing task of the system between the processing units is further based on one of (1) a power consumption table of the computational process for each processing unit;

(2) a performance table of the computational process for each processing unit; and

(3) a value table of the computational process for each processing unit.

6. A method as in claim 2 wherein the setting is a power setting which comprises a low power mode wherein the processing units are chosen based on achieving reduced power consumption;

a high performance mode wherein the processing units are chosen based on performance;

a dynamic mode wherein the processing units are automatically chosen to prevent power or thermal overload based on a threshold;

or a value mode wherein the processing units are chosen based on cost effectiveness between power consumption and performance.

7. A system, comprising:

a memory;

a plurality of processing units coupled to the memory through at least one bus; and

wherein the memory configures at least one of the processing units to redistribute a processing task between the processing units based on a setting.

8. The system as in claim 7 wherein the processing task is redistributed dynamically while the system is running and wherein the processing units comprise a central processing unit (CPU) and a special processing unit, and

wherein the processing task is created from one of a plurality of different software pipelines.

9. The system as in claim 8 wherein the special processing unit is a graphics processing unit (GPU) and wherein the GPU performs certain graphics processing operations in a high performance mode and the CPU performs those certain graphics processing operations in a power conservation mode.

10. The system as in claim 8, wherein the special processing unit is a GPU and wherein the GPU performs certain graphics processing operations in a power conservation mode and the CPU performs those certain graphics processing operations in a high performance mode.

11. The system as in claim 8 wherein the processing task is at least one of a graphics rendering process, a image processing process, a signal processing process, a Bayer filtering process, a video decoding process, or any combination thereof and wherein the redistributing a processing task of the system between the processing units is further based on one of (1) a power consumption table of the computational process for each processing unit;

(2) a performance table of the computational process for each processing unit; and

(3) a value table of the computational process for each processing unit.

12. A system as in claim 7 wherein the setting is a power setting which comprises a low power mode wherein the processing units are chosen based on achieving reduced power consumption;

a high performance mode wherein the processing units are chosen based on performance;

a dynamic mode wherein the processing units are automatically chosen to prevent power or thermal overload based on a threshold;

or a value mode wherein the processing units are chosen based on cost effectiveness between power consumption and performance.

13. An article of manufacturing comprising:

a machine-accessible storage medium storing data that, when accessed by a machine, cause the machine to perform operations comprising;

selecting a setting of a system that has a plurality of processing units;

redistributing a processing task of the system between the processing units based on the setting.

14. A article as in claim 13 wherein the redistributing is done dynamically while the system is running and wherein the processing units comprise a central processing unit (CPU) and a special processing unit, and

wherein the processing task is created from one of a plurality of different software pipelines.

15. The article as in claim 14 wherein the special processing unit is a graphics processing unit (GPU) and wherein the GPU performs certain graphics processing operations in a high performance mode and the CPU performs those certain graphics processing operations in a power conservation mode.

16. The article as in claim 14, wherein the special processing unit is a GPU and wherein the GPU performs certain graphics processing operations in a power conservation mode and the CPU performs those certain graphics processing operations in a high performance mode.

17. The article as in claim 14 wherein the processing task is at least one of a graphics rendering process, a image processing process, a signal processing process, a Bayer filtering process, a video decoding process, or any combination thereof and wherein the redistributing a processing task of the system between the processing units is further based on one of (1) a power consumption table of the computational process for each processing unit;

(2) a performance table of the computational process for each processing unit; and

(3) a value table of the computational process for each processing unit.

18. The article as in claim 14 wherein the setting is a power setting which comprises a low power mode wherein the processing units are chosen based on achieving reduced power consumption;

a high performance mode wherein the processing units are chosen based on performance;

a dynamic mode wherein the processing units are automatically chosen to prevent power or thermal overload based on a threshold;

or a value mode wherein the processing units are chosen based on cost effectiveness between power consumption and performance.

19. A system, comprising:

means for selecting a setting of a system that has a plurality of processing units; and

means for redistributing a computational process of the system between the processing units based on the setting.

20. A method to design a system having a plurality of subsystems, comprising:

determining at least one of the power consumption and performance for each subsystem for a processing task wherein each of the subsystems is capable of performing the processing task;

saving information about at least one of the power consumption and performance for each subsystem for use in redistributing one or more processing tasks in response to changes in settings.

21. A method as in claim 20 further comprising determining and saving information about executable codes for each subsystem for a processing task.

22. A method as in claim 20 wherein the processing task is redistributed to reduce the power consumption of the system, or to achieve a high performance but higher power consumption.

23. A method as in claim 20 wherein the determining at least one of the power consumption and performance by each of the subsystems comprises producing a matrix for the subsystems and the processing tasks.

24. A method as in claim 20 wherein the subsystems comprise a plurality of processing units.

25. A method as in claim 20 wherein the one or more processing tasks are created from a plurality of different software pipelines.