Chip power consumption reduction method and device, computer equipment and readable storage medium

Abstract

The invention relates to a chip power consumption reduction method and device, computer equipment and a readable storage medium. The method comprises the following steps: inputting task flow information of task data into a power consumption prediction model matched with the task data to obtain data processing power consumption information of the task data; according to the data importance degree of the task data of each currently executed task and the data processing power consumption information of each task data, screening out target task data meeting a data optimization condition from each task data; according to the data type of the target task data and the task flow information of the target task data, optimizing the data volume of the target task data to obtain optimized target task data; and performing data processing on the optimized target task data and non-target task data in each task data by executing each currently executed task to obtain an execution result of each currently executed task. By adopting the method, the chip power consumption optimization effect when the chip processes multiple tasks can be improved.

Description

Technical Field

[0001] The present application relates to the field of chip technology, and in particular to a method, apparatus, computer equipment, computer-readable storage medium and computer program product for reducing chip power consumption. Background Art

[0002] With the continuous upgrading of chip hardware, the chip's anti-interference ability, computing efficiency, service life, and application scope are constantly improving. However, due to the diversity of external environments and the richness of operating conditions, the application frequency of chips is increasing, which makes the power consumption of chips continue to increase, especially when the chip processes multiple tasks of multiple modules, the power consumption increases exponentially. Therefore, how to reduce the chip power consumption when processing multiple tasks is one of the current research focuses.

[0003] The traditional method of reducing chip power consumption is to increase the hardware processing power of the chip, while increasing the bandwidth and other hardware structures to improve the chip's operating capabilities and reduce the chip's power consumption. However, this method still cannot effectively solve the problem of reducing the chip's power consumption when processing multiple tasks, resulting in poor chip power consumption optimization when processing multiple tasks. Summary of the invention

[0004] Based on this, it is necessary to provide a chip power consumption reduction method, device, computer equipment, computer readable storage medium and computer program product that can improve the chip power consumption optimization effect when processing multiple tasks in response to the above technical problems.

[0005] In a first aspect, the present application provides a method for reducing chip power consumption, comprising:

[0006] For any currently executed task among the currently executed tasks of the chip, determine the data importance of the task data according to the data format information of the task data of the any currently executed task and the task type of the any currently executed task;

[0007] Inputting the task flow information of the task data into a power consumption prediction model matching the task data to obtain data processing power consumption information of the task data;

[0008] According to the data importance of the task data of each currently executed task and the data processing power consumption information of each task data, target task data satisfying the data optimization condition is screened out from each task data;

[0009] According to the data type of the target task data and the task flow information of the target task data, the data volume of the target task data is optimized to obtain optimized target task data;

[0010] By executing each of the currently executed tasks, data processing is performed on the optimized target task data and the non-target task data in each of the task data to obtain the execution results of each of the currently executed tasks; the non-target task data is the task data in each of the task data that does not meet the data optimization conditions.

[0011] In one embodiment, determining the data importance of the task data according to the data format information of the task data of any currently executed task and the task type of any currently executed task includes:

[0012] Determining address authority information of a data source address of the task data according to data format information of the task data of any currently executed task;

[0013] According to the address authority information, query the priority authority set of the task data in the authority database;

[0014] Identifying the execution priority of the task data according to the task type of any currently executed task;

[0015] According to the execution priority, a corresponding priority weight value is determined from the priority permission set as the data importance of the task data.

[0016] In one of the embodiments, before inputting the task flow information of the task data into a power consumption prediction model matching the task data to obtain the data processing power consumption information of the task data, the method further includes:

[0017] According to the data type and data volume of the task data, query the historical task database for historical task data matching the task data; the data type and data volume of the historical task data are the same as the data type and data volume of the task data;

[0018] A power consumption prediction model matching the task data is trained based on the historical task flow information and the historical data processing power consumption information of the historical task data.

[0019] In one embodiment, the step of selecting target task data that meets data optimization conditions from each task data according to the data importance of each task data currently being executed and the data processing power consumption information of each task data includes:

[0020] Filtering out, from the task data of each currently executed task, task data whose data importance is lower than an importance threshold as initial target task data;

[0021] Sorting the initial target task data in descending order of the data processing power consumption information of the initial target task data to obtain sorted initial target task data;

[0022] In order from front to back, the initial target task data matching the preset number are screened out from the sorted initial target task data as the target task data satisfying the data optimization condition.

[0023] In one embodiment, the optimizing process of the data volume of the target task data according to the data type of the target task data and the task flow information of the target task data to obtain the optimized target task data includes:

[0024] Determining data structure information of the target task data according to the data type of the target task data;

[0025] Based on the task flow information of the target task data, identifying the data processing scope of the target task data;

[0026] Determining non-critical data content in the target task data according to the data structure information and the data processing scope;

[0027] According to a data optimization strategy that matches the data type, the data volume of the non-critical data content in the target task data is optimized to obtain optimized target task data.

[0028] In one embodiment, before performing data processing on the optimized target task data and non-target task data in each task data by executing each currently executed task to obtain the execution result of each currently executed task, the method further includes:

[0029] According to the data importance and data volume of each of the optimized target task data and each of the non-target task data, each of the optimized target task data and each of the non-target task data is sorted to obtain a task data sequence;

[0030] For the task data at any sequence position in the task data sequence, according to the data amount of the task data at any sequence position, the task data at any sequence position is divided into a plurality of data groups to obtain a plurality of data groups corresponding to the any sequence position; the extreme difference of the data amount of each task data contained in the data group is lower than the difference threshold;

[0031] The step of performing data processing on the optimized target task data and the non-target task data in each of the task data by executing each of the currently executed tasks to obtain the execution results of each of the currently executed tasks includes:

[0032] By executing each of the currently executed tasks, data processing is performed on a plurality of data groups corresponding to each of the sequence positions in the task data sequence in turn, so as to obtain the execution results of each of the currently executed tasks.

[0033] In a second aspect, the present application also provides a chip power consumption reduction device, comprising:

[0034] A determination module, configured to determine, for any currently executed task among various currently executed tasks of the chip, the data importance of the task data according to the data format information of the task data of any currently executed task and the task type of any currently executed task;

[0035] An input module, used for inputting the task flow information of the task data into a power consumption prediction model matching the task data to obtain data processing power consumption information of the task data;

[0036] A screening module, used for screening target task data satisfying data optimization conditions from each task data according to the data importance of each task data currently being executed and the data processing power consumption information of each task data;

[0037] An optimization module, used for optimizing the data volume of the target task data according to the data type of the target task data and the task flow information of the target task data, to obtain optimized target task data;

[0038] A processing module is used to perform data processing on the optimized target task data and the non-target task data in each of the task data by executing each of the currently executed tasks, so as to obtain the execution results of each of the currently executed tasks; the non-target task data is the task data in each of the task data that does not meet the data optimization conditions.

[0039] In a third aspect, the present application further provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0040] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above method when executed by a processor.

[0041] In a fifth aspect, the present application also provides a computer program product, including a computer program, which implements the steps of the above method when executed by a processor.

[0042] The above-mentioned chip power consumption reduction method, device, computer equipment, computer-readable storage medium and computer program product determine the data importance of the task data for any currently executed task among the currently executed tasks of the chip according to the data format information of the task data of any currently executed task and the task type of any currently executed task; input the task flow information of the task data into a power consumption prediction model matching the task data to obtain the data processing power consumption information of the task data; according to the data importance of the task data of each currently executed task and the data processing power consumption information of each task data, select the target task data that meets the data optimization conditions from each task data; according to the data type of the target task data and the task flow information of the target task data, optimize the data volume of the target task data to obtain the optimized target task data; by executing each currently executed task, perform data processing on the optimized target task data and the non-target task data in each task data to obtain the execution result of each currently executed task, and the non-target task data is the task data in each task data that does not meet the data optimization conditions. By analyzing the data importance of task data of each currently executed task during multi-data processing and predicting the data processing power consumption information of each task data, the target task data that can be used for data optimization is screened, and then the data content of the target task data is optimized according to the data type and task flow information of the target task data. Then, each currently executed task is executed to process the optimized target task data and the task data that does not need to be optimized, thereby reducing the data processing amount of the chip when processing the target task data, improving the data processing power consumption optimization effect of the chip, avoiding the high cost problem of hardware structure improvement, and thus improving the chip power consumption optimization effect when processing multiple tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0044] FIG1 is a diagram showing an application environment of a method for reducing chip power consumption in one embodiment;

[0045] FIG2 is a schematic flow chart of a method for reducing chip power consumption in one embodiment;

[0046] FIG3 is a schematic diagram of another process of reducing chip power consumption in one embodiment;

[0047] FIG4 is a structural block diagram of a device for reducing chip power consumption in one embodiment;

[0048] FIG. 5 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0050] The chip power consumption reduction method provided in the embodiment of the present application can be applied in the application environment shown in FIG1. ​​The terminal 102 communicates with the server 104 through a network. The data storage system can store data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The terminal 102 determines the data importance of the task data for any currently executed task among the currently executed tasks of the chip according to the data format information of the task data of any currently executed task and the task type of any currently executed task; the terminal 102 inputs the task flow information of the task data into the power consumption prediction model matching the task data to obtain the data processing power consumption information of the task data; the terminal 102 selects the target task data that meets the data optimization conditions from each task data according to the data importance of the task data of each currently executed task and the data processing power consumption information of each task data; the terminal 102 optimizes the data volume of the target task data according to the data type of the target task data and the task flow information of the target task data to obtain the optimized target task data; the terminal 102 performs data processing on the optimized target task data and the non-target task data in each task data by executing each currently executed task to obtain the execution result of each currently executed task; the non-target task data is the task data in each task data that does not meet the data optimization conditions. The terminal 102 may be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, IoT devices, and portable wearable devices. The IoT devices may be smart speakers, smart TVs, smart air conditioners, smart car devices, projection devices, etc. Portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. Head-mounted devices may be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The server 104 may be an independent physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server that provides cloud computing services.

[0051] In an exemplary embodiment, as shown in FIG. 2 , a method for reducing chip power consumption is provided, which is described by taking the method applied to the terminal 102 in FIG. 1 as an example, and includes:

[0052] Step S202 , for any currently executed task among the currently executed tasks of the chip, determine the data importance of the task data according to the data format information of the task data of any currently executed task and the task type of any currently executed task.

[0053] In a specific implementation, the terminal can collect each currently executed task of the chip and the task data of each currently executed task through a data detection instrument set on the chip. The data detection instrument includes a data detection unit and a data transmission unit. When the data detection unit detects that the chip is performing a data processing task, the currently executed task being processed by the chip and the task data of the currently executed task are transmitted to the terminal.

[0054] Among them, the currently executed tasks include but are not limited to image optimization tasks, image recognition tasks, data statistics tasks, data calculation tasks, data analysis tasks and other task data, and different currently executed tasks correspond to different task data.

[0055] The task data is the data information that needs to be processed when executing the current task. The data information may include data of various data types, such as text data type, picture data type, video data type, table data type, audio data type, and other types of data.

[0056] In a specific implementation, the terminal can identify the data format information of each task data. The data format information can be the data profile information of the task data, and the data profile information of each task data is the information contained when each task data is transmitted from the data source to the chip. The data format information may include but is not limited to the data volume, data source, and data flow of the task data.

[0057] In a specific implementation, the terminal can identify the task type of each currently executed task, and the task type may include but is not limited to text data recognition tasks, image data recognition tasks, video data recognition tasks, table data recognition tasks, audio data recognition tasks, etc.

[0058] The terminal can analyze the data importance of the task data according to the data format information of the task data of the currently executed task and the task type of the currently executed task. The data importance can be represented as a percentage value. As an example, the terminal can construct a mapping table for different data formats and task types, and associate each combination with an importance weight, that is, assign different importance weights to each combination of data format information and task type. For example, image data is more important in computer vision tasks, but less important in text analysis tasks; then, according to the data format information and task type and their corresponding weights, a weighted calculation is performed to obtain the data importance of the task data.

[0059] Step S204 , inputting the task flow information of the task data into a power consumption prediction model matching the task data to obtain data processing power consumption information of the task data.

[0060] The task flow information includes a detailed description of the task data during execution, such as task objectives, task steps, time schedule, input and output, decision points, etc. The task flows of currently executed tasks of the same task type can be the same or different. For example, the task flows of a text data recognition task and an image data recognition task, both of which are data recognition types, are different.

[0061] Specifically, the task type and task flow information of the currently executed task can be stored in the task database. The terminal extracts the task identification information of each currently executed task, and then searches the task database for the task type and task flow information containing the task identification information. The task identification information of the currently executed task can be, but is not limited to, symbol identification, color identification, text identification, letter identification and other identification information. The task database includes the corresponding relationship between each task type and task identification information, and the corresponding relationship between each task flow information and task identification information.

[0062] The terminal can select a power consumption prediction model that matches the task data. For example, the terminal can determine the matching power consumption prediction model based on the task type of the task data. For example, task data of an image processing task type and task data of an audio analysis task type can match different power consumption prediction models.

[0063] Optionally, the power consumption prediction model may be a reinforcement learning neural network, which is used to predict the data processing power consumption information of each task data. The power consumption prediction model may be trained based on historical task flow information and historical data processing power consumption information of historical task data.

[0064] The data processing power consumption information may include the power consumed by the task data during execution and processing. The data processing power consumption information may be used to effectively screen out the task data that needs to be optimized. The data processing power consumption information may include the power consumed by the chip during the calculation, storage, and transmission of the task data when executing the task.

[0065] Step S206 , based on the data importance of each task data of the currently executed task and the data processing power consumption information of each task data, target task data that meets the data optimization condition is screened out from each task data.

[0066] The terminal selects target task data that meets the data optimization condition from various task data according to the data importance of each task data currently being executed and the data processing power consumption information of each task data.

[0067] As an example, the data optimization condition may be that the data importance is lower than a certain value and the power consumption value indicated by the data processing power consumption information is higher than a certain value, that is, the target task data may be the task data with lower data importance and higher power consumption value indicated by the data processing power consumption information among all the task data.

[0068] As another example, the terminal can perform weighted calculation based on data importance and data processing power consumption information to obtain a weighted calculation result, and filter out target task data that meets data optimization conditions from various task data based on the weighted calculation result. The data optimization condition may mean that the weighted calculation result is within a preset numerical range.

[0069] Step S208, optimizing the data volume of the target task data according to the data type of the target task data and the task flow information of the target task data to obtain optimized target task data.

[0070] Among them, the task process information can be used to indicate the data processing method and data processing range of the target task data. For example, the task process information shows that a rough analysis is required within the first data processing range in the task data, and a precise analysis is required within the second data processing range in the task data. Then, the data volume of the data within the first data processing range in the target task data can be optimized.

[0071] The data type may include but is not limited to an image data type, a text data type, and a video data type.

[0072] The terminal can determine the optimization strategy to be adopted according to the data type of the target task data, and optimize the data volume of the target task data according to the optimization strategy. For example, for the target task data of the image data type, the method of optimizing the data volume of the target task data may include image resolution reduction processing, blanking processing, image dimension reduction processing, etc.; for the target task data of the text data type, the method of optimizing the data volume of the target task data may include text deletion processing, text semantic extraction processing, text format simplification processing, etc.

[0073] Step S210 , by executing each currently executed task, data processing is performed on the optimized target task data and the non-target task data in each task data to obtain the execution result of each currently executed task.

[0074] The non-target task data is task data that does not meet the data optimization condition among various task data. The non-target task data may be task data with higher data importance among the task data, or task data with lower power consumption value indicated by data processing power consumption information.

[0075] After optimizing the target task data, the terminal can start data processing for each currently executed task on the chip, that is, processing the optimized target task data and non-target task data that does not need to be optimized. Since the target task data has been optimized in each task data, the amount of task data is reduced, and the computing resources and power consumption of the chip in processing these task data will be reduced.

[0076] The terminal performs data processing on the optimized target task data and non-target task data in each task data, which may be to perform specific calculation, conversion or analysis operations on the task data to complete the current execution task to which the task data belongs, and obtain the execution results of each current execution task. These execution results may include recognized images, analyzed text content, processed video clips, etc., which are related to the task type of the current execution task.

[0077] In the above-mentioned chip power consumption reduction method, for any currently executed task of the chip, the data importance of the task data is determined according to the data format information of the task data of any currently executed task and the task type of any currently executed task; the task flow information of the task data is input into the power consumption prediction model matching the task data to obtain the data processing power consumption information of the task data; according to the data importance of the task data of each currently executed task and the data processing power consumption information of each task data, the target task data that meets the data optimization conditions is screened out from each task data; according to the data type of the target task data and the task flow information of the target task data, the data volume of the target task data is optimized to obtain the optimized target task data; by executing each currently executed task, the optimized target task data and the non-target task data in each task data are processed to obtain the execution result of each currently executed task, and the non-target task data is the task data in each task data that does not meet the data optimization conditions. By analyzing the data importance of task data of each currently executed task during multi-data processing and predicting the data processing power consumption information of each task data, the target task data that can be used for data optimization is screened, and then the data content of the target task data is optimized according to the data type and task flow information of the target task data. Then, each currently executed task is executed to process the optimized target task data and the task data that does not need to be optimized, thereby reducing the data processing amount of the chip when processing the target task data, improving the data processing power consumption optimization effect of the chip, avoiding the high cost problem of hardware structure improvement, and thus improving the chip power consumption optimization effect when processing multiple tasks.

[0078] In another embodiment, the data importance of the task data is determined based on the data format information of the task data of any currently executed task and the task type of any currently executed task, including: determining the address permission information of the data source address of the task data based on the data format information of the task data of any currently executed task; querying the priority permission set of the task data in a permission database based on the address permission information; identifying the execution priority of the task data based on the task type of any currently executed task; and determining the corresponding priority weight value from the priority permission set based on the execution priority as the data importance of the task data.

[0079] The data source information may record the data source of the task data and the address permission information of the data source address. The terminal may determine the address permission information of the data source address based on the data format information of the task data. Optionally, the data source address may be an Internet Protocol address (IP address), and the address permission information may be a permission level associated with the data source address, which is used to indicate the credibility, sensitivity or access restriction of the task data. For example, some key data may come from a high-privilege address, and ordinary data may come from a low-privilege address.

[0080] The terminal can identify the execution priority of task data according to the task type of the currently executed task. Different task types affect the execution priority of task data. For example, task data under the video analysis task type has high real-time requirements and high importance, so it has a higher execution priority.

[0081] Among them, the permission database can store address permission information and the corresponding priority permission set, and the priority permission set includes the priority weight value corresponding to each execution priority. The higher the execution priority, the greater the priority weight value. The priority weight value can be a numerical value that quantifies the execution priority, which can be directly used as the data importance of the task data, or it can be obtained by normalizing the priority weight value to obtain the data importance.

[0082] The technical solution of this embodiment identifies the data importance of task data through data source and task type, thereby improving the accuracy of identifying the data importance of task data.

[0083] In another embodiment, before inputting the task flow information of the task data into the power consumption prediction model that matches the task data to obtain the data processing power consumption information of the task data, it also includes: based on the data type and data volume of the task data, querying the historical task data that matches the task data in the historical task database; the data type and data volume of the historical task data are the same as the data type and data volume of the task data; and training the power consumption prediction model that matches the task data based on the historical task flow information and the historical data processing power consumption information of the historical task data.

[0084] In a specific implementation, for each task data, based on the data type and data volume of the task data, the terminal queries the historical task database for each historical task data with the same data type and the same data volume as the task data, as well as the historical data processing power consumption information of each historical task data when executing each historical task flow and the corresponding historical task flow information. Then, the terminal trains a power consumption prediction model based on the historical task flow information and the historical data processing power consumption information.

[0085] The technical solution of this embodiment, by screening historical task data of the same data type and the same data volume, respectively trains the power consumption prediction models, obtains the power consumption prediction network corresponding to each task data, and then predicts the data processing power consumption information of the task data, thereby improving the comprehensiveness and accuracy of predicting the data processing power consumption information of each task data.

[0086] In another embodiment, based on the data importance of the task data of each currently executed task and the data processing power consumption information of each task data, target task data that meets the data optimization conditions are screened out from each task data, including: from the task data of each currently executed task, task data whose data importance is lower than an importance threshold is screened out as initial target task data; according to the order of the data processing power consumption information of each initial target task data from large to small, each initial target task data is sorted to obtain each sorted initial target task data; according to the order from front to back, initial target task data that matches a preset number are screened out from each sorted initial target task data as target task data that meets the data optimization conditions.

[0087] In the specific implementation, the terminal screens the task data below the importance threshold as the initial target task data, and sorts each initial target task data in descending order according to the data processing power consumption information of each initial target task data, that is, sorts each initial target task data in descending order according to the power consumption value indicated in the data processing power consumption information of each initial target task data to obtain the sorted initial target task data; then, according to the order from front to back of the sorted initial target task data, a preset number of initial target task data are screened as the target task data.

[0088] Among them, the preset number is a preset screening number, which is a number obtained through a large number of creative experiments and expert experience. By limiting the screening number, it is possible to avoid the problem of data loss and incomplete data caused by the optimization processing of all task data. At the same time, it can also effectively reduce the amount of task data that the chip needs to process and improve the degree of power consumption optimization.

[0089] For example, if the preset number is 5, the first 5 initial target task data corresponding to the largest data processing power consumption information are selected from the sorted initial target task data as the target task data.

[0090] The technical solution of this embodiment selects task data with high data processing power consumption information and low data importance as target task data, thereby avoiding affecting all task data when optimizing task data, and effectively improving the optimization effect of chip power consumption.

[0091] In another embodiment, the data volume of the target task data is optimized according to the data type of the target task data and the task flow information of the target task data to obtain the optimized target task data, including: determining the data structure information of the target task data according to the data type of the target task data; identifying the data processing scope of the target task data based on the task flow information of the target task data; determining the non-critical data content in the target task data according to the data structure information and the data processing scope; and optimizing the data volume of the non-critical data content in the target task data according to a data optimization strategy that matches the data type to obtain the optimized target task data.

[0092] In a specific implementation, the data structure information may include the data components of the target task data. For example, the data structure information of the image data type includes the image data structure in which each pixel is arranged in order according to the position information; the data structure information of the text data type includes the text data structure in which each text content is arranged in order, etc.

[0093] The data processing scope can refer to the specific content and areas involved in data processing, such as the areas in an image that need to be focused on and the paragraphs in a text that need to be analyzed.

[0094] Among them, non-critical data content can be non-critical and unimportant data content in the target task data. For example, in image processing, the background area of ​​the image may belong to non-critical data content; in text processing, additional comments may belong to non-critical data content.

[0095] Among them, the data optimization strategy matches the data type. For example, the data optimization strategy of the image data type includes image resolution reduction processing, non-critical processing data content blanking processing, and image dimension reduction processing; the data optimization strategy of the text data type includes text deletion processing, text semantic extraction processing, text format simplification processing, etc. Among them, the terminal can identify the data optimization strategy according to the correspondence between the preset data type and the data optimization strategy. The data optimization strategy of each data type is different. Each data optimization strategy is used to optimize the data volume of each task data, thereby reducing the data volume of the task data while ensuring the accuracy of data processing, thereby reducing the data processing power consumption information of the chip processing the task data.

[0096] The technical solution of this embodiment screens non-critical data content by identifying data structure information and data processing range, and then optimizes the non-critical data content of the target task data through a data optimization strategy that matches the data type to obtain the optimized target task data. While ensuring the accuracy of task execution of the optimized target task data, the data volume of the target task data is greatly reduced.

[0097] In another embodiment, before performing data processing on the optimized target task data and the non-target task data in each task data by executing each currently executed task to obtain the execution result of each currently executed task, the method further includes:

[0098] According to the data importance and data volume of each optimized target task data and each non-target task data, each optimized target task data and each non-target task data are sorted to obtain a task data sequence;

[0099] By executing each currently executed task, data processing is performed on the optimized target task data and the non-target task data in each task data to obtain the execution results of each currently executed task, including:

[0100] By executing each currently executed task, data processing is performed on the optimized target task data and the non-target task data in each task data according to the task data sequence, so as to obtain the execution result of each currently executed task.

[0101] In a specific implementation, task data with high importance and large data volume may be placed in front of a task data sequence, and task data with high importance and large data volume may be processed preferentially through the task data sequence.

[0102] The technical solution of this embodiment sorts the data through the optimized task data, which ensures that the chip can give priority to high-priority data information while reducing the parallel data processing capacity of the chip, thereby greatly optimizing the chip's power consumption per unit time and avoiding the high cost problem of hardware structure improvement, thereby comprehensively improving the chip power consumption optimization effect when processing multiple tasks.

[0103] In another embodiment, before the execution of each currently executed task, data processing is performed on the optimized target task data and the non-target task data in each task data to obtain the execution result of each currently executed task, it also includes: sorting the optimized target task data and the non-target task data according to the data importance and data volume of each optimized target task data and each non-target task data to obtain a task data sequence; for the task data at any sequence position in the task data sequence, according to the data volume of the task data at any sequence position, dividing the task data at any sequence position into multiple data groups to obtain multiple data groups corresponding to any sequence position; the extreme difference of the data volume of each task data contained in the data group is lower than the difference threshold;

[0104] The method comprises: performing data processing on multiple data groups corresponding to each sequence position in the task data sequence in turn by executing each currently executing task, and obtaining the execution results of each currently executing task.

[0105] In the specific implementation, the terminal divides the task data at any sequence position in the task data sequence into multiple data groups, so that the data volume in the data group is relatively balanced to avoid extremely large data differences. The extreme difference of the data volume refers to the difference between the maximum and minimum values ​​of the data volume, which needs to be lower than the difference threshold. In other words, each task data is divided into each data group according to the size of the data volume, thereby reducing the power consumption change of the chip processing each data group, so as to reduce the power consumption of the chip caused by the change of power consumption.

[0106] The terminal processes the multiple data groups corresponding to the respective sequence positions in the task data sequence in turn by executing each currently executed task, and obtains the execution results of each currently executed task. For example, the task data at the respective sequence positions in the task data sequence are A, B, and C in turn, and the task data A is divided into multiple data groups a1, a2, and a3, the task data B is divided into multiple data groups b1 and b2, and the task data C is divided into multiple data groups c1 and c2. The multiple data groups corresponding to the respective sequence positions in the task data sequence are processed in turn, that is, the order of processing in turn is: a1, a2, a3, b1, b2, c1, and c2.

[0107] The technical solution of this embodiment groups the task data at each sequence position according to the data volume, thereby reducing the power consumption variation of the chip in processing each data group, thereby reducing the power consumption information generated by the chip due to the change in power consumption, and improving the optimization effect of chip power consumption.

[0108] In another embodiment, as shown in FIG. 3 , a method for reducing chip power consumption is provided, which is described by taking the method applied to the terminal 102 in FIG. 1 as an example, and includes the following steps:

[0109] S302, based on the data format information of the task data of any currently executed task, determine the address permission information of the data source address of the task data; based on the address permission information, query the priority permission set of the task data in the permission database; based on the task type of any currently executed task, identify the execution priority of the task data; based on the execution priority, determine the corresponding priority weight value from the priority permission set as the data importance of the task data.

[0110] S304, based on the data type and data volume of the task data, query the historical task data that matches the task data in the historical task database; wherein the data type and data volume of the historical task data are the same as the data type and data volume of the task data; and based on the historical task process information of the historical task data and the historical data processing power consumption information, train a power consumption prediction model that matches the task data.

[0111] S306 , inputting the task flow information of the task data into a power consumption prediction model matching the task data to obtain data processing power consumption information of the task data.

[0112] S308, from the task data of each currently executed task, filter out the task data whose data importance is lower than the importance threshold as the initial target task data; sort the initial target task data in order from large to small according to the data processing power consumption information of each initial target task data to obtain the sorted initial target task data; in order from front to back, filter out the initial target task data that matches the preset number from the sorted initial target task data as the target task data that meets the data optimization conditions.

[0113] S310, determine the data structure information of the target task data according to the data type of the target task data; identify the data processing scope of the target task data based on the task flow information of the target task data; determine the non-critical data content in the target task data according to the data structure information and the data processing scope; optimize the data volume of the non-critical data content in the target task data according to the data optimization strategy matching the data type to obtain the optimized target task data.

[0114] S312, sorting each optimized target task data and each non-target task data according to their data importance and data volume to obtain a task data sequence; for the task data at any sequence position in the task data sequence, dividing the task data at any sequence position into multiple data groups according to the data volume of the task data at any sequence position to obtain multiple data groups corresponding to any sequence position.

[0115] Among them, the extreme difference of the data volume of each task data contained in the data group is lower than the difference threshold.

[0116] S314, by executing each currently executed task, data processing is performed on multiple data groups corresponding to each sequence position in the task data sequence in turn to obtain the execution result of each currently executed task.

[0117] It should be noted that the specific definition of the above steps can refer to the specific definition of a chip power consumption reduction method mentioned above.

[0118] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0119] Based on the same inventive concept, the embodiment of the present application also provides a chip power consumption reduction device for implementing the chip power consumption reduction method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more chip power consumption reduction device embodiments provided below can refer to the limitations of the chip power consumption reduction method above, and will not be repeated here.

[0120] In an exemplary embodiment, as shown in FIG4 , a device for reducing chip power consumption is provided, comprising:

[0121] The determination module 410 is used to determine the data importance of the task data of any currently executed task among the currently executed tasks of the chip according to the data format information of the task data of any currently executed task and the task type of any currently executed task;

[0122] An input module 420, configured to input the task flow information of the task data into a power consumption prediction model matching the task data, to obtain data processing power consumption information of the task data;

[0123] A screening module 430, configured to screen target task data satisfying a data optimization condition from each of the task data according to the data importance of each of the task data currently being executed and the data processing power consumption information of each of the task data;

[0124] An optimization module 440 is used to optimize the data volume of the target task data according to the data type of the target task data and the task flow information of the target task data to obtain optimized target task data;

[0125] The processing module 450 is used to perform data processing on the optimized target task data and the non-target task data in each of the task data by executing each of the currently executed tasks, so as to obtain the execution results of each of the currently executed tasks; the non-target task data is the task data in each of the task data that does not meet the data optimization conditions.

[0126] In one embodiment, the determination module 410 is specifically used to determine the address permission information of the data source address of the task data based on the data format information of the task data of any currently executed task; query the priority permission set of the task data in the permission database based on the address permission information; identify the execution priority of the task data based on the task type of any currently executed task; and determine the corresponding priority weight value from the priority permission set based on the execution priority as the data importance of the task data.

[0127] In one embodiment, the input module 420 is specifically used to query historical task data matching the task data in a historical task database based on the data type and data volume of the task data; the data type and data volume of the historical task data are the same as the data type and data volume of the task data; and based on the historical task flow information of the historical task data and the historical data processing power consumption information, a power consumption prediction model matching the task data is trained.

[0128] In one embodiment, the filtering module 430 is specifically used to filter out the task data whose data importance is lower than the importance threshold from the task data of each currently executed task as the initial target task data; sort the initial target task data in order from large to small according to the data processing power consumption information of each initial target task data to obtain the sorted initial target task data; and filter out the initial target task data that matches a preset number from the sorted initial target task data in order from front to back as the target task data that meets the data optimization condition.

[0129] In one embodiment, the optimization module 440 is specifically used to determine the data structure information of the target task data according to the data type of the target task data; identify the data processing scope of the target task data based on the task flow information of the target task data; determine the non-critical data content in the target task data according to the data structure information and the data processing scope; and optimize the data volume of the non-critical data content in the target task data according to a data optimization strategy matching the data type to obtain optimized target task data.

[0130] In one embodiment, the processing module 450 is specifically used to sort each of the optimized target task data and each of the non-target task data according to their data importance and data volume to obtain a task data sequence; for the task data at any sequence position in the task data sequence, the task data at any sequence position is divided into multiple data groups according to the data volume of the task data at any sequence position to obtain multiple data groups corresponding to the any sequence position; the extreme difference of the data volume of each task data contained in the data group is lower than the difference threshold; by executing each of the currently executed tasks, the multiple data groups corresponding to each of the sequence positions in the task data sequence are processed in turn to obtain the execution results of each of the currently executed tasks.

[0131] Each module in the above chip power consumption reduction device can be implemented in whole or in part by software, hardware and their combination. Each module can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0132] In an exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be shown in FIG5. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit, and an input device. The processor, the memory, and the input / output interface are connected via a system bus, and the communication interface, the display unit, and the input device are connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be implemented through WIFI, a mobile cellular network, near field communication (NFC) or other technologies. When the computer program is executed by the processor, a method for reducing chip power consumption is implemented. The display unit of the computer device is used to form a visually visible picture, which may be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0133] Those skilled in the art will understand that the structure shown in FIG. 5 is merely a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0134] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0135] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0136] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0137] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0138] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0139] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0140] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for reducing chip power consumption, characterized in that: The method comprises: For any currently executed task among the currently executed tasks of the chip, determine the data importance of the task data according to the data format information of the task data of the any currently executed task and the task type of the any currently executed task; Inputting the task flow information of the task data into a power consumption prediction model matching the task data to obtain data processing power consumption information of the task data; According to the data importance of the task data of each currently executed task and the data processing power consumption information of each task data, target task data satisfying the data optimization condition is screened out from each task data; According to the data type of the target task data and the task flow information of the target task data, the data volume of the target task data is optimized to obtain optimized target task data; By executing each of the currently executed tasks, data processing is performed on the optimized target task data and the non-target task data in each of the task data to obtain the execution results of each of the currently executed tasks; the non-target task data is the task data in each of the task data that does not meet the data optimization conditions.

2. The method according to claim 1, characterized in that: The determining the data importance of the task data according to the data format information of the task data of any currently executed task and the task type of any currently executed task includes: Determining address authority information of a data source address of the task data according to data format information of the task data of any currently executed task; According to the address authority information, query the priority authority set of the task data in the authority database; Identifying the execution priority of the task data according to the task type of any currently executed task; According to the execution priority, a corresponding priority weight value is determined from the priority permission set as the data importance of the task data.

3. The method according to claim 1, characterized in that Before inputting the task flow information of the task data into a power consumption prediction model matching the task data to obtain the data processing power consumption information of the task data, the method further includes: According to the data type and data volume of the task data, query the historical task database for historical task data matching the task data; the data type and data volume of the historical task data are the same as the data type and data volume of the task data; A power consumption prediction model matching the task data is trained based on the historical task flow information and the historical data processing power consumption information of the historical task data.

4. The method according to claim 1, characterized in that The step of selecting target task data satisfying data optimization conditions from each task data according to the data importance of each task data currently being executed and the data processing power consumption information of each task data comprises: Filtering out, from the task data of each currently executed task, task data whose data importance is lower than an importance threshold as initial target task data; Sorting the initial target task data in descending order of the data processing power consumption information of the initial target task data to obtain sorted initial target task data; In order from front to back, the initial target task data matching the preset number are screened out from the sorted initial target task data as the target task data satisfying the data optimization condition.

5. The method according to claim 1, characterized in that: The step of optimizing the data volume of the target task data according to the data type of the target task data and the task flow information of the target task data to obtain the optimized target task data includes: Determining data structure information of the target task data according to the data type of the target task data; Based on the task flow information of the target task data, identifying the data processing scope of the target task data; Determining non-critical data content in the target task data according to the data structure information and the data processing scope; According to a data optimization strategy that matches the data type, the data volume of the non-critical data content in the target task data is optimized to obtain optimized target task data.

6. The method according to claim 1, characterized in that Before performing data processing on the optimized target task data and the non-target task data in each of the task data by executing each of the currently executed tasks to obtain the execution results of each of the currently executed tasks, the method further includes: According to the data importance and data volume of each of the optimized target task data and each of the non-target task data, each of the optimized target task data and each of the non-target task data is sorted to obtain a task data sequence; For the task data at any sequence position in the task data sequence, according to the data amount of the task data at any sequence position, the task data at any sequence position is divided into a plurality of data groups to obtain a plurality of data groups corresponding to the any sequence position; the range of the data amount of each task data included in the data group is lower than the difference threshold; The step of performing data processing on the optimized target task data and the non-target task data in each of the task data by executing each of the currently executed tasks to obtain the execution results of each of the currently executed tasks includes: By executing each of the currently executed tasks, data processing is performed on a plurality of data groups corresponding to each of the sequence positions in the task data sequence in turn, so as to obtain the execution results of each of the currently executed tasks.

7. A chip power consumption reduction device, characterized in that: The device comprises: A determination module, configured to determine, for any currently executed task among various currently executed tasks of the chip, the data importance of the task data according to the data format information of the task data of any currently executed task and the task type of any currently executed task; An input module, used for inputting the task flow information of the task data into a power consumption prediction model matching the task data to obtain data processing power consumption information of the task data; A screening module, used for screening target task data satisfying data optimization conditions from each task data according to the data importance of each task data currently being executed and the data processing power consumption information of each task data; An optimization module, used for optimizing the data volume of the target task data according to the data type of the target task data and the task flow information of the target task data, to obtain optimized target task data; A processing module is used to perform data processing on the optimized target task data and the non-target task data in each of the task data by executing each of the currently executed tasks, so as to obtain the execution results of each of the currently executed tasks; the non-target task data is the task data in each of the task data that does not meet the data optimization conditions.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images