Multi-scene chip consumption control system and method
Through dynamic power consumption model and real-time monitoring algorithm, combined with adaptive scheduling and load distribution, the problem of poor chip power consumption control in the existing technology is solved, precise control of power consumption and efficiency improvement is achieved, and battery life is extended.
Patent Information
- Application Number
- CN202411488012.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-24
- Publication Date
- 2024-11-22
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing chip power consumption control technology relies on static models and simple monitoring algorithms, and cannot accurately predict and control the power consumption changes of the chip in different working scenarios, and lacks real-time response capabilities, resulting in poor power consumption efficiency and waste of resources.
Using dynamic power consumption model and real-time monitoring algorithm, voltage and frequency are adjusted through adaptive scheduling strategies, control windows and time windows are built to limit the adjustment range, and precise control and optimization of power consumption are achieved.
Improve the power consumption efficiency of the chip in different scenarios, reduce the equipment overheating problem caused by excessive power consumption, extend battery life, improve user experience, and optimize power consumption through adaptive load distribution and parallel computing, realizing Reduced overall power consumption.
Smart Images

Figure CN119005023A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of chip power consumption control technology, and more specifically, to a multi-scenario chip consumption control system and method. Background Art
[0002] With the increasing functionality of electronic devices and the diversification of application scenarios, chip power consumption control has become a key factor in improving device performance and extending battery life. Existing power consumption control technologies mainly rely on static power consumption models and simple monitoring algorithms, which often cannot accurately predict and control the power consumption changes of chips in different working scenarios. In addition, traditional power consumption control strategies usually use fixed voltage and frequency settings, lack the ability to dynamically respond to real-time power consumption data, and cannot achieve optimal power consumption efficiency in a changing usage environment.
[0003] In the process of implementing the embodiments of the present invention, the inventors found that there are at least the following problems or defects in the prior art: first, the existing power consumption control model is too simple and cannot accurately reflect the power consumption characteristics of the chip under complex working conditions; second, the lack of effective real-time monitoring and adaptive adjustment mechanism makes it impossible for power consumption control to respond to scene changes in a timely manner; finally, the existing power consumption control method often ignores the energy consumption balance in multiple scenarios, resulting in excessive power consumption in some scenarios, and may cause waste of resources in other scenarios. Summary of the invention
[0004] The present invention provides a multi-scenario chip consumption control system and method.
[0005] In a first aspect of the present invention, a multi-scenario chip consumption control system is provided, which performs the following steps:
[0006] Based on the dynamic power consumption model and real-time monitoring algorithm, the power consumption of the chip is preprocessed, and then controlled by an adaptive scheduling strategy to obtain the power consumption data in various scenarios;
[0007] Obtain the power consumption control starting point from the power consumption data of each scene, build a control window based on the power consumption control starting point, adjust the power consumption to different scenes respectively, and obtain control trajectory data; set a time window based on the control trajectory data to limit the adjustment range of the adaptive scheduling strategy;
[0008] Under the criterion of optimal power consumption efficiency, based on the principle of multi-scenario energy consumption balancing and dynamic voltage and frequency adjustment technology, the optimal voltage and frequency are searched to correct the power consumption error of the chip.
[0009] Furthermore, the dynamic power consumption model is a linear regression model; the mathematical model of the real-time monitoring algorithm is:
[0010]
[0011] in, is the power consumption, is the voltage, is the frequency, is the current, is the model parameter.
[0012] Furthermore, the preprocessing of the power consumption of the chip includes:
[0013] The chip power consumption signal with a weight adjustment of 0.1 is transmitted as a control signal, the power consumption of the current scene is estimated based on the power consumption measurement value of the previous scene, and the optimal dynamic power consumption model parameters are selected based on the power consumption and characteristics of each scene;
[0014] Obtaining a corresponding dynamic power consumption model according to the dynamic power consumption model parameters, and obtaining a power consumption response curve according to the model shape;
[0015] According to the power consumption response curve, the parameters of the control strategy are designed using the gradient descent method. The control strategy is used to adjust the power consumption signals of different scenarios respectively, and then the power consumption of the chip is adjusted using the power consumption adjustment coefficient of the dynamic power consumption model corresponding to the dynamic power consumption model parameters.
[0016] Furthermore, the optimal dynamic power consumption model parameters are dynamic power consumption model parameters corresponding to the minimum energy consumption deviation in different scenarios.
[0017] Further, obtaining the power consumption control starting point in the power consumption data of each scenario includes:
[0018] The chip power consumption signal is sequentially filtered, feature extracted, and the power consumption value of each scene is calculated to obtain the power consumption result of the scene domain;
[0019] Search for the maximum power consumption position of each scene as the power consumption control point of each scene, perform curve fitting on the power consumption control points of multiple adjacent scenes in segments, and count the relative deviation of each scene in the curve fitting result and the average deviation of the scenes in the segment;
[0020] The scene segment with the smallest average deviation is taken as the target segment, and the power consumption control point is selected according to the relative deviation threshold. The power consumption information of each selected scene power consumption control point is extracted, and the scene power consumption control point with the largest amplitude is selected as the starting point of power consumption control.
[0021] Furthermore, the coordinate range of the control window is:
[0022]
[0023] Among them, the starting point coordinates of power consumption control are ,in is the starting value of the scene sequence number, is the starting value of the time sequence number; is the search window width in the direction of scene number, is the search window height in the time sequence direction; and They respectively represent the dynamic adjustment range of the control window in the direction of scene number and time number.
[0024] Furthermore, the power consumption error correction of the chip includes: dividing the data of each chip into multiple segments along time according to the scene type, first obtaining the power consumption before and after the central voltage is adjusted according to the power consumption difference between the edge of the scene and the adjacent scene, and then obtaining the power consumption and lateral distance of other voltages before and after adjustment.
[0025] In a second aspect of the present invention, a multi-scenario chip power consumption optimization method is provided, comprising:
[0026] Acquire power consumption data obtained by the multi-scenario chip consumption control system according to any one of claims 1 to 7;
[0027] Based on the adaptive load distribution method, the power consumption data is divided into multiple data blocks with uniform energy consumption;
[0028] According to the system load, the CPU is used to load the corresponding data blocks into the memory in real time, and the power consumption blocks to be optimized are loaded into the processing unit of the optimization algorithm;
[0029] Based on the parallel computing platform, a multi-strategy power consumption optimization model is constructed in parallel using processing units to optimize power consumption.
[0030] Furthermore, the method based on adaptive load distribution divides the power consumption data into a plurality of data blocks with uniform energy consumption, including:
[0031] According to the preset energy consumption threshold, the load distribution method is adopted to distribute the power consumption data layer by layer to obtain each sub-block. According to the energy consumption in the adjacent sub-blocks, the adjacent sub-blocks are merged until the total energy consumption threshold of the sub-block is reached to obtain multiple data blocks.
[0032] Furthermore, according to the system load condition, the CPU is used to load the corresponding data block into the memory in real time, and the power consumption block to be optimized is loaded into the processing unit of the optimization algorithm, including:
[0033] Calculate the power consumption data block file address to be loaded into the memory and cache data according to the current system load; construct the memory and cache data using the queue data structure, and initialize the loading of the cache and memory power consumption data block;
[0034] When the system load changes, the type of load change is determined and the memory data is updated; the cache update threshold is calculated to determine whether it is necessary to clear useless data in the cache and whether to load external memory data into the cache, and perform the required clearing or loading operations.
[0035] The above-mentioned embodiments of the present invention have at least the following beneficial effects: the multi-scenario chip consumption control system of the present invention can accurately predict and control the power consumption of the chip in different scenarios by adopting a dynamic power consumption model and a real-time monitoring algorithm, which can effectively improve the energy efficiency ratio of the chip. The system uses an adaptive scheduling strategy to dynamically adjust the voltage and frequency according to real-time power consumption data, so as to achieve real-time optimization of power consumption. In addition, by constructing control windows and time windows, the system can limit the adjustment range of the scheduling strategy to ensure the stability and reliability of power consumption control. This method can not only reduce the problem of equipment overheating caused by excessive power consumption, but also extend battery life and improve user experience.
[0036] In addition, the multi-scenario chip power consumption optimization method of the present invention divides the power consumption data into multiple data blocks with uniform energy consumption through adaptive load distribution, which can realize efficient management and optimization of power consumption data. Utilizing the CPU real-time loading and optimization algorithm processing unit, the system can dynamically adjust the data blocks in the memory and cache according to the current system load conditions to optimize power consumption. This method can not only improve the efficiency of data processing, but also reduce the additional energy consumption caused by frequent adjustment of voltage and frequency, thereby reducing overall power consumption. Through the parallel computing platform, the system can build a multi-strategy power consumption optimization model in parallel to further improve the efficiency and effect of power consumption optimization. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the detailed description below with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, in which:
[0038] Figure 1 A schematic diagram of a flow chart of a multi-scenario chip consumption control system provided by an embodiment of the present invention;
[0039] Figure 2 A schematic flow chart of a multi-scenario chip consumption control method provided by an embodiment of the present invention;
[0040] Figure 3 The schematic diagram schematically shows the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION
[0041] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.
[0042] Those skilled in the art know that the embodiments of the present invention can be implemented as a system, device, apparatus, method or computer program product. Therefore, the present invention can be specifically implemented in the following forms, namely: complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.
[0043] It should be noted that any number of elements in the drawings is for illustration rather than limitation, and any naming is only for distinction and does not have any limiting meaning.
[0044] Reference below Figure 1 , Figure 1 This is a flow chart of a multi-scenario chip consumption control system provided by an embodiment of the present invention. Figure 1 As shown, a multi-scenario chip consumption control system 100 includes:
[0045] Step 101, based on a dynamic power consumption model and a real-time monitoring algorithm, pre-process the power consumption of the chip, and then use an adaptive scheduling strategy to control it, so as to obtain power consumption data in various scenarios;
[0046] Step 102, obtaining the power consumption control starting point from the power consumption data of each scene, constructing a control window according to the power consumption control starting point, adjusting the power consumption to different scenes respectively, and obtaining control trajectory data; setting a time window according to the control trajectory data to limit the adjustment range of the adaptive scheduling strategy;
[0047] Step 103, under the criterion of optimal power consumption efficiency, based on the multi-scenario energy consumption balancing principle and dynamic voltage and frequency adjustment technology, search for the optimal voltage and frequency, and correct the power consumption error of the chip.
[0048] It should be noted that the core of the multi-scenario chip consumption control system is to control the power consumption of the chip through a dynamic power consumption model and a real-time monitoring algorithm.
[0049] Specifically, the dynamic power consumption model can be a linear regression model, and the mathematical model of the real-time monitoring algorithm may include parameters such as power consumption, voltage, frequency, and current, as well as corresponding model parameters. These parameters can be adjusted according to the actual chip characteristics and usage scenarios to ensure the accuracy of the model and the effectiveness of the control strategy.
[0050] Preferably, the power consumption control starting point can be obtained by filtering the chip power consumption signal, extracting features, and calculating the power consumption value of each scene, thereby obtaining the power consumption result of the scene domain. In addition, by searching for the maximum power consumption position of each scene, it can be used as a power consumption control point, and curve fitting can be performed on the power consumption control points of multiple adjacent scenes to determine the optimal power consumption control starting point. This method can control power consumption more accurately and improve the energy efficiency of the system.
[0051] In some embodiments, the dynamic power consumption model is a linear regression model; the mathematical model of the real-time monitoring algorithm is:
[0052]
[0053] in, is the power consumption, is the voltage, is the frequency, is the current, is the model parameter.
[0054] It should be noted that the dynamic power consumption model is defined as a linear regression model, while the mathematical model of the real-time monitoring algorithm involves parameters such as power consumption, voltage, frequency and current, as well as the relationship between these parameters.
[0055] Specifically, the linear regression model is a statistical method used to establish a linear relationship between one or more independent variables (such as voltage, frequency) and a dependent variable (such as power consumption). In the real-time monitoring algorithm, the mathematical model can be expressed as ,in is the power consumption, is the voltage, is the frequency, is the current, These parameters need to be trained and optimized based on actual chip characteristics and historical power consumption data to ensure the accuracy of the model.
[0056] Preferably, for the implementation of the real-time monitoring algorithm, the voltage, frequency and current data of the chip under different working conditions can be collected, and then the linear regression model can be trained using these data. During the training process, the model parameters can be optimized by cross-validation and other methods to ensure that the model can accurately predict power consumption in different scenarios. In addition, it is also possible to consider introducing regularization technology to prevent overfitting of the model and improve the generalization ability of the model. In practical applications, the model parameters can also be adjusted as needed to adapt to different power consumption control requirements.
[0057] In some embodiments, the preprocessing of the power consumption of the chip includes:
[0058] The chip power consumption signal with a weight adjustment of 0.1 is transmitted as a control signal, the power consumption of the current scene is estimated based on the power consumption measurement value of the previous scene, and the optimal dynamic power consumption model parameters are selected based on the power consumption and characteristics of each scene;
[0059] Obtaining a corresponding dynamic power consumption model according to the dynamic power consumption model parameters, and obtaining a power consumption response curve according to the model shape;
[0060] According to the power consumption response curve, the parameters of the control strategy are designed using the gradient descent method. The control strategy is used to adjust the power consumption signals of different scenarios respectively, and then the power consumption of the chip is adjusted using the power consumption adjustment coefficient of the dynamic power consumption model corresponding to the dynamic power consumption model parameters.
[0061] It should be noted that these steps include using specific weight adjustment, power consumption signal transmission, power consumption estimation, dynamic power consumption model parameter selection, power consumption response curve acquisition, control strategy parameter design and power consumption adjustment.
[0062] Specifically, the preprocessing step involves multiplying the chip power consumption signal by an applied coefficient (for example, 0.1) and transmitting it as a control signal. The purpose of this step is to adjust the power consumption signal to a suitable range for subsequent processing. Next, the system estimates the power consumption of the current scene based on the power consumption measurement of the previous scene, and selects the optimal dynamic power consumption model parameters based on the power consumption and scene characteristics. These parameters may include coefficients of voltage, frequency, and current, which can be determined through historical data and statistical analysis. Then, the system obtains the corresponding dynamic power consumption model based on these parameters, and obtains the power consumption response curve based on the shape of the model. Finally, the system uses optimization algorithms such as gradient descent to design the parameters of the control strategy and adjust the power consumption signals of different scenes.
[0063] Preferably, the preprocessing step can be further refined into the following operating points: First, a digital filter can be used to denoise the chip power consumption signal to improve the accuracy of the signal. Secondly, a machine learning algorithm can be used to estimate the power consumption of the current scene, which may involve analysis and pattern recognition of historical power consumption data. In addition, the use of genetic algorithms or other global optimization algorithms to select the optimal dynamic power consumption model parameters can be considered to improve the prediction accuracy of the model. When designing the control strategy parameters, it is possible to consider introducing an adaptive control mechanism so that the system can dynamically adjust the control parameters based on real-time feedback to adapt to changing scene conditions. Finally, the power consumption adjustment step can be achieved by introducing a feedback control loop to ensure the accuracy and stability of the power consumption adjustment.
[0064] In some embodiments, the optimal dynamic power consumption model parameters are dynamic power consumption model parameters corresponding to the minimum energy consumption deviation in different scenarios.
[0065] Specifically, the optimal dynamic power consumption model parameters refer to the set of parameters that can minimize the power consumption deviation in a specific scenario. These parameters may include coefficients of voltage, frequency, and current, which are optimized through historical power consumption data and real-time monitoring data. In actual operation, the power consumption deviation under different parameter combinations can be compared to determine which set of parameters can provide the best power consumption control effect. This usually involves a lot of data collection and analysis, as well as complex mathematical calculations.
[0066] Preferably, the process of determining the optimal dynamic power consumption model parameters can be further refined into the following steps: First, data mining technology can be used to analyze historical power consumption data to identify the patterns of power consumption changes in different scenarios. Secondly, machine learning algorithms, such as support vector machines (SVM) or neural networks, can be used to predict power consumption deviations under different parameter combinations. In addition, multi-objective optimization algorithms, such as genetic algorithms or particle swarm optimization (PSO), can be introduced to simultaneously consider multiple objectives such as power consumption deviation and system response time to find the optimal parameter combination. During the implementation process, real-time adjustment of parameters can also be considered to adapt to the impact of scene changes and ensure the dynamic adaptability and control effect of the system.
[0067] In some embodiments, obtaining the power consumption control starting point in the power consumption data of each scenario includes:
[0068] The chip power consumption signal is sequentially filtered, feature extracted, and the power consumption value of each scene is calculated to obtain the power consumption result of the scene domain;
[0069] Search for the maximum power consumption position of each scene as the power consumption control point of each scene, perform curve fitting on the power consumption control points of multiple adjacent scenes in segments, and count the relative deviation of each scene in the curve fitting result and the average deviation of the scenes in the segment;
[0070] The scene segment with the smallest average deviation is taken as the target segment, and the power consumption control point is selected according to the relative deviation threshold. The power consumption information of each selected scene power consumption control point is extracted, and the scene power consumption control point with the largest amplitude is selected as the starting point of power consumption control.
[0071] It should be noted that in a multi-scenario chip consumption control system, the specific method of obtaining the starting point of power consumption control is the key to the system's ability to accurately control power consumption. It involves the processing of chip power consumption signals and the analysis and processing of power consumption data in different scenarios.
[0072] Specifically, the process of obtaining the starting point of power consumption control includes filtering the chip power consumption signal, feature extraction, and calculating the power consumption value of each scene. Filtering is to remove noise from the signal, feature extraction is to identify the key factors affecting power consumption, and calculating the power consumption value of each scene is to determine the power consumption characteristics of each scene. Next, the system will search for the maximum power consumption position of each scene, which will be used as power consumption control points for subsequent power consumption adjustments. The system will also perform curve fitting on the power consumption control points of multiple adjacent scenes to statistically calculate the relative deviation of each scene in the curve fitting results and the average deviation of the scenes within the segment.
[0073] Preferably, the operation steps for obtaining the starting point of power consumption control can be further refined or alternative solutions can be provided. For example, the filtering process can use different types of filters, such as low-pass filters, high-pass filters or band-pass filters, and select the appropriate filter type according to the characteristics of the signal. Feature extraction can use methods such as principal component analysis (PCA) or independent component analysis (ICA) to extract the features that best represent the power consumption characteristics. When calculating the power consumption value, techniques such as time series analysis or Fourier transform can be used to more accurately capture the trend of power consumption changes. When searching for the maximum power consumption position, a peak detection algorithm such as gradient descent or Newton's method can be used to quickly and accurately find the maximum value. In addition, curve fitting can use polynomial fitting, exponential fitting or other nonlinear fitting methods to obtain a more accurate power consumption control point. When determining the starting point of power consumption control, an optimization algorithm such as a genetic algorithm or a simulated annealing algorithm can also be introduced to take into account other performance indicators of the system while satisfying the minimization of power consumption deviation.
[0074] In some embodiments, the coordinate range of the control window is:
[0075]
[0076] Among them, the starting point coordinates of power consumption control are ,in is the starting value of the scene sequence number, is the starting value of the time sequence number; is the search window width in the direction of scene number, is the search window height in the time sequence direction; and They respectively represent the dynamic adjustment range of the control window in the direction of scene number and time number.
[0077] It should be noted that the method for determining the coordinate range of the control window is a key technical feature in the multi-scenario chip consumption control system. The control window is a virtual space used to limit the adjustment range of the adaptive scheduling strategy. The setting of its coordinate range is crucial to ensure the accuracy and effectiveness of power consumption control.
[0078] Specifically, the coordinate range of the control window can be obtained by the formula To determine, Represents the starting value of the scene sequence number. Represents the starting value of the time sequence. and They represent the dynamic adjustment range of the control window in the scene number and time number directions respectively. The setting of these parameters needs to be based on a deep understanding of the power consumption behavior of the chip in different scenarios and an accurate grasp of the system performance requirements. For example, and The value of can be determined by experimental measurement, simulation analysis or statistical methods based on historical data.
[0079] Preferably, the determination of the coordinate range of the control window can be further refined or an alternative solution can be provided. For example, an adaptive algorithm can be used to dynamically adjust and The value of can be adjusted to better adapt to the real-time changing scene conditions. In addition, intelligent control technologies such as fuzzy logic control or neural networks can be introduced to improve the flexibility and accuracy of control window adjustment. In actual operation, it is also possible to consider introducing a safety margin to ensure that the system can maintain stable power consumption control performance even in extreme or unknown scenarios. In addition, the parameters of the control window can be adjusted manually or automatically through user input or system feedback to meet specific application requirements or optimize the power consumption control effect.
[0080] In some embodiments, the power consumption error correction of the chip includes: dividing the data of each chip into multiple segments along time according to the scene type, first obtaining the power consumption before and after the central voltage is adjusted according to the power consumption difference between the edge of the scene and the adjacent scene, and then obtaining the power consumption and lateral distance of other voltages before and after adjustment.
[0081] Specifically, the process of correcting the power consumption error of the chip includes the following key steps: First, the data of each chip is divided into multiple segments along time according to the scene type. This means that the system needs to be able to identify different working scenarios and divide the data according to the characteristics of these scenarios. Next, the system needs to obtain the power consumption before and after the central voltage adjustment from the power consumption difference between the edge of the scene and the adjacent scene. This involves comparing and analyzing the power consumption data to determine the impact of voltage adjustment on power consumption. Finally, the system also needs to obtain the power consumption before and after other voltage adjustments and lateral distances, which may involve further processing and analysis of the power consumption data.
[0082] Preferably, the process of power consumption error correction can be further refined or alternative solutions can be provided. For example, advanced data analysis techniques, such as time series analysis or machine learning algorithms, can be used to more accurately identify and divide different working scenarios. In addition, an adaptive control algorithm can be introduced to dynamically adjust the voltage to achieve more precise power consumption control. In actual operation, it is also possible to consider using a multivariable control strategy to adjust multiple parameters, such as voltage and frequency, at the same time to optimize power consumption. In addition, the power consumption control strategy can be continuously adjusted and optimized through real-time monitoring and feedback mechanisms to adapt to changing working conditions and scenario requirements. These methods can improve the flexibility and adaptability of the system and ensure that the best power consumption control effect can be achieved in various scenarios.
[0083] The above-mentioned embodiments of the present invention have the following beneficial effects: the multi-scenario chip consumption control system of the present invention can realize precise control of chip power consumption according to the power consumption requirements in different usage scenarios through dynamic power consumption models and real-time monitoring algorithms. The system can dynamically adjust the voltage and frequency according to the real-time power consumption data through an adaptive scheduling strategy, thereby optimizing power consumption. In addition, by constructing control windows and time windows, the system can limit the adjustment range of the scheduling strategy to ensure the stability and reliability of power consumption control. This method can not only reduce the problem of equipment overheating caused by excessive power consumption, but also extend battery life and improve user experience.
[0084] like Figure 2 As shown, a multi-scenario chip consumption control method 200 of some embodiments includes:
[0085] Step 201, obtaining power consumption data obtained by the multi-scenario chip consumption control system according to any one of claims 1-7;
[0086] Step 202, based on the adaptive load distribution method, the power consumption data is divided into a plurality of data blocks with uniform energy consumption;
[0087] Step 203, according to the system load condition, use the CPU to load the corresponding data block into the memory in real time, and load the power consumption block to be optimized into the processing unit of the optimization algorithm;
[0088] Step 204 , based on the parallel computing platform, using the processing units to construct a multi-strategy power consumption optimization model in parallel to optimize the power consumption.
[0089] It can be understood that the modules recorded in the multi-scenario chip power consumption optimization method 200 are similar to the reference Figure 1Therefore, the operations, features and beneficial effects described above for the multi-scenario chip power consumption optimization system are also applicable to the multi-scenario chip power consumption optimization method 200 and the modules contained therein, and will not be repeated here.
[0090] In some embodiments, the method for dividing the power consumption data into a plurality of data blocks with uniform energy consumption based on the adaptive load distribution method includes:
[0091] According to the preset energy consumption threshold, the load distribution method is adopted to distribute the power consumption data layer by layer to obtain each sub-block. According to the energy consumption in the adjacent sub-blocks, the adjacent sub-blocks are merged until the total energy consumption threshold of the sub-block is reached to obtain multiple data blocks.
[0092] Specifically, the adaptive load distribution method involves distributing the power consumption data layer by layer using a load distribution method according to a preset energy consumption threshold. This includes decomposing the power consumption data into multiple sub-blocks and merging adjacent sub-blocks according to the energy consumption in the adjacent sub-blocks until the total energy consumption threshold of the sub-block is reached. This process requires precise setting of the energy consumption threshold to ensure uniform energy consumption of the data block. The energy consumption threshold can be determined based on historical data, statistical analysis or simulation testing. In addition, the merging process of sub-blocks also needs to consider the relevance and processing efficiency of the data to ensure the accuracy and real-time performance of power consumption optimization.
[0093] Preferably, the implementation of the adaptive load distribution method can be further refined or provide alternatives. For example, a method of dynamically adjusting the energy consumption threshold can be adopted to automatically adjust the threshold according to the real-time power consumption data and system load changes to adapt to different working conditions. In addition, intelligent algorithms, such as genetic algorithms or particle swarm optimization algorithms, can be introduced to optimize the merging process of sub-blocks to improve the efficiency of data processing and the effect of power consumption optimization. In actual operation, it is also possible to consider using parallel processing technology to distribute power consumption data to multiple processing units for parallel processing to speed up data processing and improve the overall performance of the system. These methods can improve the flexibility and adaptability of the system and ensure that the best power consumption optimization effect can be achieved in various scenarios.
[0094] In some embodiments, according to the system load condition, using the CPU to load the corresponding data block into the memory in real time, and loading the power consumption block to be optimized into the processing unit of the optimization algorithm, includes:
[0095] Calculate the power consumption data block file address to be loaded into the memory and cache data according to the current system load; construct the memory and cache data using the queue data structure, and initialize the loading of the cache and memory power consumption data block;
[0096] When the system load changes, the type of load change is determined and the memory data is updated; the cache update threshold is calculated to determine whether it is necessary to clear useless data in the cache and whether to load external memory data into the cache, and perform the required clearing or loading operations.
[0097] It should be noted that in the multi-scenario chip power consumption optimization method, according to the system load, the CPU is used to load the corresponding data blocks into the memory in real time, and the power consumption blocks to be optimized are loaded into the processing unit of the optimization algorithm. This step is to ensure that the system can effectively manage and optimize power consumption according to the current workload.
[0098] Specifically, this process includes the following key steps: First, calculate the power consumption data block file address that needs to be loaded into the memory and cache data according to the current system load. This involves real-time monitoring and analysis of the current state of the system to determine which data blocks need to be processed first. Next, use the queue data structure to build memory and cache data, and initialize the loading of cache and memory power consumption data blocks. The queue data structure allows the system to process data blocks in a specific order, improving the efficiency of data processing. Finally, when the system load changes, the system needs to determine the type of load change and update the memory data. This may involve reordering or adjusting the data blocks in memory to adapt to the new load conditions.
[0099] Preferably, the operation steps can be further refined. For example, a prediction algorithm can be used to predict the changing trend of the system load, so as to adjust the data blocks in the memory and cache in advance to reduce the impact of load changes on system performance. In addition, an adaptive algorithm can be introduced to dynamically adjust the loading and clearing strategies of data blocks to optimize the use efficiency of memory and cache. In actual operation, it is also possible to consider using a multi-level cache system to improve the system's response speed and power consumption optimization effect by intelligently allocating data blocks between different storage levels. These methods can improve the flexibility and adaptability of the system and ensure that the best power consumption optimization effect can be achieved in various scenarios.
[0100] Reference below Figure 3 , which shows a schematic diagram of a structure 300 of an electronic device suitable for implementing some embodiments of the present invention. The electronic devices in some embodiments of the present invention may include, but are not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.
[0101] like Figure 3 As shown, the electronic device 300 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 to a random access memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the electronic device 300 are also stored. The processing device 301, the ROM 302, and the RAM 303 are connected to each other via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.
[0102] Typically, the following devices may be connected to the I / O interface 305: an input device 306 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 308 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 309. The communication device 309 may allow the electronic device 300 to communicate with other devices wirelessly or by wire to exchange data. Figure 3 The electronic device 300 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead. Figure 3 Each block shown in the figure may represent one device, or may represent multiple devices as required.
[0103] Furthermore, the storage medium of the embodiment of the present application stores program instructions that can implement all the above methods, wherein the program instructions can be stored in the above storage medium in the form of a software product, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, a server, a mobile phone, and a tablet.
[0104] The above descriptions are only some preferred embodiments of the present invention and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.
Claims
1. A multi-scenario chip consumption control system, characterized in that: The system performs the following steps: Based on the dynamic power consumption model and real-time monitoring algorithm, the power consumption of the chip is preprocessed, and then controlled by an adaptive scheduling strategy to obtain the power consumption data in various scenarios; Obtain the power consumption control starting point from the power consumption data of each scene, build a control window based on the power consumption control starting point, adjust the power consumption to different scenes respectively, and obtain control trajectory data; Set the time window according to the control trajectory data to limit the adjustment range of the adaptive scheduling strategy; Under the criterion of optimal power consumption efficiency, based on the principle of multi-scenario energy consumption balancing and dynamic voltage and frequency adjustment technology, the optimal voltage and frequency are searched to correct the power consumption error of the chip.
2. The multi-scenario chip consumption control system according to claim 1, characterized in that: The dynamic power consumption model is a linear regression model; the mathematical model of the real-time monitoring algorithm is: ; in, is the power consumption, is the voltage, is the frequency, is the current, is the model parameter.
3. The multi-scenario chip consumption control system according to claim 2, characterized in that: The preprocessing of the power consumption of the chip includes: The chip power consumption signal with a weight adjustment of 0.1 is transmitted as a control signal, the power consumption of the current scene is estimated based on the power consumption measurement value of the previous scene, and the optimal dynamic power consumption model parameters are selected based on the power consumption and characteristics of each scene; Obtaining a corresponding dynamic power consumption model according to the dynamic power consumption model parameters, and obtaining a power consumption response curve according to the model shape; According to the power consumption response curve, the parameters of the control strategy are designed using the gradient descent method. The control strategy is used to adjust the power consumption signals of different scenarios respectively, and then the power consumption of the chip is adjusted using the power consumption adjustment coefficient of the dynamic power consumption model corresponding to the dynamic power consumption model parameters.
4. The multi-scenario chip consumption control system according to claim 3, characterized in that: The optimal dynamic power consumption model parameters are the dynamic power consumption model parameters corresponding to the minimum energy consumption deviation in different scenarios.
5. The multi-scenario chip consumption control system according to claim 1, characterized in that: The obtaining of the power consumption control starting point from the power consumption data of each scenario includes: The chip power consumption signal is sequentially filtered, feature extracted, and the power consumption value of each scene is calculated to obtain the power consumption result of the scene domain; Search for the maximum power consumption position of each scene as the power consumption control point of each scene, perform curve fitting on the power consumption control points of multiple adjacent scenes in segments, and count the relative deviation of each scene in the curve fitting result and the average deviation of the scenes in the segment; The scene segment with the smallest average deviation is taken as the target segment, and the power consumption control point is selected according to the relative deviation threshold. The power consumption information of each selected scene power consumption control point is extracted, and the scene power consumption control point with the largest amplitude is selected as the starting point of power consumption control.
6. The multi-scenario chip consumption control system according to claim 1, characterized in that: The coordinate range of the control window is: ; Among them, the starting point coordinates of power consumption control are ,in is the starting value of the scene sequence number, is the starting value of the time sequence number; is the search window width in the direction of scene number, is the search window height in the time sequence direction; and They respectively represent the dynamic adjustment range of the control window in the direction of scene number and time number.
7. The multi-scenario chip consumption control system according to claim 1, characterized in that: The power consumption error correction of the chip includes: dividing the data of each chip into multiple segments along time according to the scene type, first obtaining the power consumption before and after the central voltage is adjusted according to the power consumption difference between the edge of the scene and the adjacent scene, and then obtaining the power consumption and lateral distance of other voltages before and after adjustment.
8. A multi-scenario chip power consumption optimization method, characterized in that: The steps include: Obtain power consumption data obtained by the multi-scenario chip consumption control system according to any one of claims 1 to 7; Based on the adaptive load distribution method, the power consumption data is divided into multiple data blocks with uniform energy consumption; According to the system load, the CPU is used to load the corresponding data blocks into the memory in real time, and the power consumption blocks to be optimized are loaded into the processing unit of the optimization algorithm; Based on the parallel computing platform, a multi-strategy power consumption optimization model is constructed in parallel using processing units to optimize power consumption.
9. The multi-scenario chip power consumption optimization method according to claim 8, characterized in that: The method based on adaptive load distribution divides the power consumption data into multiple data blocks with uniform energy consumption, including: According to the preset energy consumption threshold, the load distribution method is adopted to distribute the power consumption data layer by layer to obtain each sub-block. According to the energy consumption in the adjacent sub-blocks, the adjacent sub-blocks are merged until the total energy consumption threshold of the sub-block is reached to obtain multiple data blocks.
10. The multi-scenario chip power consumption optimization method according to claim 9, characterized in that: According to the system load condition, the CPU is used to load the corresponding data block into the memory in real time, and the power consumption block to be optimized is loaded into the processing unit of the optimization algorithm, including: Calculate the power consumption data block file address to be loaded into the memory and cache data according to the current system load; construct the memory and cache data using the queue data structure, and initialize the loading of the cache and memory power consumption data block; When the system load changes, the type of load change is determined and the memory data is updated; the cache update threshold is calculated to determine whether it is necessary to clear useless data in the cache and whether to load external memory data into the cache, and perform the required clearing or loading operations.
Citation Information
Patent Citations
Server power consumption management method oriented to scene prediction
CN103902016A
Power consumption control method and device of electronic equipment, storage medium and terminal
CN112162625A
Power consumption scheduling strategy sending method and device, electronic equipment and storage medium
CN115379541A
Low-power-consumption control method and system for multi-core heterogeneous chip
CN117289782A
Power consumption control method for intelligent terminal
CN118655972A
Cited By
Satellite communication SoC chip power consumption control method and device
CN120980653A
Resource utilization rate determination method, chip, device, equipment and storage medium
CN121433884A
Optimization method and device for dynamically adjusting power consumption and performance of chip
CN122261363A
