Methods and apparatuses for load balancing between multiple processing units

Patent metadata

Field	Value
Publication number	US-9311152-B2
Application number	US-201213616856-A
Country	US
Kind code	B2
Filing date	Sep 14, 2012
Priority date	Oct 24, 2007
Publication date	Apr 12, 2016
Grant date	Apr 12, 2016

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.

First claim

What is claimed is:

1. A machine implemented method comprising:
distributing a processing task for processing by a first processor having a first processor type, in a system comprising a plurality of processors each having a processor type;
determining a performance optimization setting that optimizes a plurality of performance criteria including processing throughput, temperature of the system, or power consumption of the system;
monitoring system performance including: a processing load of each of the plurality of processors, a temperature of the system, and a power availability of the system;
looking up a performance profile for the processing task, wherein the performance profile comprises a matrix that includes a performance value for each of the plurality of performance criteria for each of the plurality of processor types;
automatically and dynamically redistributing, by a redistribution process switching the processing task for processing by a second processor having a second processor type, different from the first processor type, wherein the processing task is executable on both the first and second processor types, the switching based at least in part upon:
the performance optimization setting for the system;
the monitored system performance; and
the matrix of performance criteria values of the processing task for the first and second processor types in the performance profile for the processing task.

2. The method of claim 1 , wherein the processor type comprises one of a central processing unit (CPU), a graphics processing unit (GPU), and a digital signal processing unit (DSP) and wherein determining a performance setting for the system comprises a user input determining the performance setting.

3. The method of claim 1 , wherein redistributing the processing task comprises, at least one of:
compiling the processing task to be executable on both the first and second processor types;
emulating the first processor type on the second processor, or;
emulating the processing task on the second processor type.

4. The method of claim 1 , wherein the performance optimization setting comprises setting the system for high performance and the performance criterion comprises determining that a processing load of the first processor exceeds a threshold value.

5. The method of claim 1 , wherein the performance optimization setting comprises setting the system to control temperature due to processing and the performance criterion comprises determining that a temperature of the first processor exceeds a threshold value.

6. The method of claim 1 , further wherein the performance optimization setting comprises setting the system to reduce power consumption and the performance criterion comprises determining that power consumption will be reduced by processing the processing task on the second processor.

7. The method of claim 1 , wherein the performance optimization setting comprises setting the system for high performance and the performance criterion comprises determining that the processing task will be performed faster on the second processor than the first processor.

8. The method of claim 1 , wherein distributing the processing task for processing by a first processor comprises distributing the processing task to a default processor.

9. The method of claim 1 , wherein the performance optimization setting comprises at least one of optimizing power consumption, optimizing performance throughput, optimizing a ratio of performance to power consumed, or optimizing a dynamic state of the system.

10. The method of claim 9 , wherein a dynamic state of the system comprises at least one of: the temperature of one or more subsystems, the charge level of a battery in the system, or the processing load on one or more of the plurality of processors.

11. The method of claim 1 , further comprising:
while running the system, determining that the optimization setting has changed; and
automatically redistributing the first processing task in accordance with the changed optimization setting.

12. The method of clam 1 , wherein determining a performance setting for the system comprises one of:
in response to determining that the system temperature is above a predetermined threshold, selecting, by the system, a temperature performance setting wherein the processing task is redistributed to a processor having a processor type for which the performance profile for the processing task has performance values for performance criteria that minimize a temperature performance criteria and maximize a throughput criteria;
in response to determining that the system temperature is below a predetermined threshold, selecting, by the system, a throughput performance setting wherein the processing task is redistributed to a processor having a processor type for which the performance profile for the processing task has performance values for performance criteria that maximize processing throughput for the processing task;
in response to determining that the system battery level is below a predetermined threshold, selecting, by the system, a power availability performance setting wherein the processing task is redistributed to a processor having a processor type for which the performance profile for the processing task has performance values for performance criteria that minimize power consumption and maximize throughput.

13. An article comprising a computer-readable non-transitory storage medium having stored thereon executable instructions that, when executed, cause one or more processors to:
distribute a processing task for processing by a first processor having a first processor type, in a system comprising a plurality of processors each having a processor type;
determine a performance optimization setting that optimizes a plurality of performance criteria including processing throughput, temperature of the system, and power consumption of the system;
monitor system performance including: a processing load of each of the plurality of processors, a temperature of the system, and a power availability of the system;
look up a performance profile for the processing task, wherein the performance profile comprises a matrix that includes a performance value for each of the plurality of performance criteria for each of the plurality of processor types;
automatically and dynamically redistribute, by a redistribution process switching the processing task for processing by a second processor having a second processor type, different from the first processor type, wherein the processing task is executable on both the first and second processor types, the switching based at least in part upon:
the performance setting for the system;
the monitored system performance; and
the matrix of performance criteria values of the processing task for the first and second processor types in the performance profile for the processing task.

14. The article of claim 13 , wherein the processor type comprises one of a central processing unit (CPU), a graphics processing unit (GPU), and a digital signal processing unit (DSP) and wherein determining a performance setting for the system comprises a user input determining the performance setting.

15. The article of claim 13 , wherein redistributing the processing task comprises, at least one of:
compile the processing task to be executable on both the first and second processor types,
emulating the first processor type on the second processor, or;
emulating the processing task on the second processor type.

16. The article of claim 13 , wherein the performance optimization setting comprise

Assignees

• Miller Howard
• Brunner Ralph
• Apple Inc

Inventors

• Miller Howard
• Brunner Ralph

Classifications

• G06F1/3203 Primary

  Power management, i.e. event-based initiation of a power-saving mode · CPC title

• G06F2209/501

  Performance criteria · CPC title

• Y02B60/142

  Cross-Sectional Technologies · mapped topic

• Y02B60/144

  Cross-Sectional Technologies · mapped topic

• G06F9/5044 Primary

  considering hardware capabilities · CPC title