초록 ▼

An apparatus that generates a power model for an electronic device, an apparatus that operates in accordance with a generated power model, and methods for generating a power for an electronic device are disclosed. In a particular embodiment, a method of generating a power model for the electronic de

An apparatus that generates a power model for an electronic device, an apparatus that operates in accordance with a generated power model, and methods for generating a power for an electronic device are disclosed. In a particular embodiment, a method of generating a power model for the electronic device includes reducing training data to identify a subset of operating parameters of an electronic device that contribute most to power consumption of the electronic device and generating the power model for the electronic device based on the reduced training data. The power model is operative to predict a power consumption value corresponding to the electronic device responsive to a set of operating parameter values corresponding to operation of the electronic device.

대표청구항 ▼

1. A system comprising: a processor configured to: identify a subset of operating parameters of an electronic device that contributes most to power consumption of the electronic device by reducing training data; andgenerate a power model for the electronic device based on the reduced training data,

1. A system comprising: a processor configured to: identify a subset of operating parameters of an electronic device that contributes most to power consumption of the electronic device by reducing training data; andgenerate a power model for the electronic device based on the reduced training data, wherein the power model is operable to predict, responsive to a set of operating parameter values corresponding to operation of the electronic device, a power consumption value corresponding to the electronic device. 2. The system of claim 1, wherein the processor is further configured to reduce the training data by: generating an initial model based on the training data, the initial model comprising a plurality of basis functions, each of the plurality of basis functions corresponding to an operating parameter of the electronic device that contributes to power consumption of the electronic device; andreducing the initial model by iteratively removing one or more of the plurality of basis functions. 3. The system of claim 2, wherein the processor reduces the initial model within a predetermined number of iterations, with each iteration comprising: identifying a basis function of the plurality of basis functions of the initial model contributing least to an overall goodness-of-fit; andremoving the identified basis function from the initial model. 4. The system of claim 3, wherein the predetermined number of iterations is determined based on a defined value corresponding to a maximum number of operating parameters to identify as the operating parameters contributing most to the power consumption of the electronic device. 5. The system of claim 2, wherein the processor generates the initial model in a number of iterations corresponding to a predetermined number of basis functions, with each iteration comprising: selecting a basis function to add to the initial model to maximize a goodness-of-fit value of the initial model with respect to the training data; andadding the selected basis function to the initial model. 6. The system of claim 5, wherein the predetermined number of basis functions is a user defined value. 7. The system of claim 1, wherein the training data includes a plurality of operating parameter values and corresponding power consumption values of the electronic device. 8. The system of claim 1, integrated in at least one semiconductor die. 9. The system of claim 1, further comprising at least one of: a communications device, a fixed location data unit, or a computer, into which the processor is integrated. 10. The system of claim 1, wherein the electronic device comprises a power management circuit that is responsive to the power model. 11. The system of claim 10, wherein the power management circuit is configured to set at least one operating parameter value in accordance with the power model to dynamically manage power consumption of the electronic device in real-time. 12. The system of claim 10, further comprising a processor core, wherein the power model excludes processor on-chip memory accesses and excludes processor instruction branching performance. 13. The system of claim 10, wherein the electronic device comprises an electrical interface. 14. The system of claim 13, wherein the processor reduces the training data by removing: a parameter indicating a number of masters communicating via the electrical interface;a parameter indicating a number of slaves communicating via the electrical interface; anda parameter indicating burst length of data communications via the electrical interface. 15. The system of claim 13, wherein the electrical interface is compliant with an advanced extensible interface (AXI) protocol. 16. The system of claim 10, integrated in at least one semiconductor die. 17. The system of claim 10, further comprising at least one of: a set top box, a music player, a video player, an entertainment unit, a navigation device, a communications device, a personal digital assistant (PDA), a fixed location data unit, or a computer, into which the power management circuit is integrated. 18. The system of claim 1, wherein the electronic device includes an electrical interface comprising one or more operating parameter values that are set in accordance with the power model, wherein the power model is operable to predict power consumption of the electrical interface based on values of a plurality of operating parameters excluding a parameter indicating a number of masters communicating via the electrical interface, a parameter indicating a number of slaves communicating via the electrical interface, and a parameter indicating burst length of data communications via the electrical interface. 19. The system of claim 18, integrated in at least one semiconductor die. 20. The system of claim 18, wherein the electronic device is at least one of: a set top box, a music player, a video player, an entertainment unit, a navigation device, a communications device, a personal digital assistant (PDA), a fixed location data unit, or a computer. 21. An apparatus comprising: means for identifying a subset of operating parameters of an electronic device that contributes most to power consumption of the electronic device by reducing training data; andmeans for generating a power model for the electronic device based on the reduced training data, wherein the power model is operable to predict, responsive to a set of operating parameter values corresponding to operation of the electronic device, a power consumption value corresponding to the electronic device. 22. The apparatus of claim 21, integrated in at least one semiconductor die. 23. The apparatus of claim 21, further comprising at least one of: a communications device, a fixed location data unit, or a computer, into which the means for reducing the training data and the means for generating the power model are integrated. 24. A method comprising: identifying a subset of operating parameters of an electronic device that contributes most to power consumption of the electronic device by reducing training data; andgenerating a power model for the electronic device based on the reduced training data, wherein the power model is operable to predict, responsive to a set of operating parameter values corresponding to operation of the electronic device, a power consumption value corresponding to the electronic device. 25. The method of claim 24, wherein generating the power model for the electronic device further comprises performing a multivariable adaptive regression splines operation. 26. The method of claim 24, further comprising verifying the generated power model. 27. The method of claim 26, wherein verifying the generated power model further comprises: performing a factor analysis of the training data to identify a plurality of influencers prior to generating the power model; andcomparing basis functions of the generated power model to the identified plurality of influencers. 28. The method of claim 24, further comprising establishing a design of experiments to generate the power model, the design of experiments specifying a method of collecting the training data. 29. The method of claim 24, further comprising establishing a design of experiments to generate a second power model for a second electronic device, the design of experiments specifying a method of collecting training data for the second electronic device, the training data for the second electronic device comprising a plurality of operating parameter values and corresponding power consumption values for the second electronic device. 30. The method of claim 24, wherein reducing the training data and generating the power model for the electronic device are performed at a processor integrated into a second electronic device. 31. The method of claim 24, wherein the electronic device includes a processor, and further comprising setting one or more operating parameters of the processor in accordance with the power model, wherein the power model is operable to predict power consumption of the processor based on values of a plurality of operating parameters excluding a processor on-chip memory access parameter and excluding a processor instruction branching performance parameter. 32. The method of claim 31, wherein the processor is integrated in at least one semiconductor die. 33. The method of claim 31, wherein the electronic device is at least one of: a set top box, a music player, a video player, an entertainment unit, a navigation device, a communications device, a personal digital assistant (PDA), a fixed location data unit, or a computer. 34. A method comprising: a first step for identifying a subset of operating parameters of an electronic device that contributes most to power consumption of the electronic device by reducing training data; anda second step for generating a power model for the electronic device based on the reduced training data, wherein the power model is operable to predict, responsive to a set of operating parameter values corresponding to operation of the electronic device, a power consumption value corresponding to the electronic device. 35. The method of claim 34, wherein the first step and the second step are performed by a processor integrated into a second electronic device. 36. A non-transitory computer readable medium storing instructions executable by a computer, the instructions comprising: instructions that are executable by the computer to identify a subset of operating parameters of an electronic device that contributes most to power consumption of the electronic device by generating an initial model based on training data and by reducing the training data in the initial model, wherein the initial model comprises a plurality of basis functions and each of the plurality of basis functions corresponds to an operating parameter of the electronic device; andinstructions that are executable by the computer to generate a power model for the electronic device based on the reduced training data, wherein the power model is operable to predict, responsive to a set of operating parameter values corresponding to operation of the electronic device, a power consumption value corresponding to the electronic device. 37. The non-transitory computer readable tangible medium of claim 36, wherein the instructions are executable by at least one of: a communications device, a fixed location data unit, or a second computer.