Display card operation condition identification method and system and storage medium

Abstract

The invention provides a video card operation condition identification method and system and a storage medium. The method comprises the following steps: acquiring a time sequence of a plurality of working parameters when a video card operates; selecting a corresponding classification model for each working parameter, inputting the time sequence of each working parameter into the corresponding classification model, and obtaining a pre-judgment working condition result corresponding to each working parameter; querying a credibility matrix corresponding to each classification model in a database based on the classification model corresponding to each working parameter; according to the credibility matrix, constructing a BPA function of each classification model, and calculating a fused BPA function value of each working condition; and according to the fused BPA function value corresponding to the pre-judgment working condition result of each working parameter, determining the operation working condition of the display card. According to the method, multi-source and multi-mode data are adopted, the pre-judgment result is used as the final judgment evidence, and video card operation condition recognition and fault recognition with high data type compatibility, low calculation power requirement and high robustness are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of graphics card monitoring, and more particularly to a graphics card running condition recognition method and system and a storage medium. BACKGROUND

[0002] During the operation of a graphics card, the working condition of the graphics card may change with changes in external environment (such as temperature, humidity, etc.) and display stress, and even malfunction. It is very important to monitor the working condition of the graphics card in real time and identify faults in a timely manner.

[0003] During the operation of a GPU, the working condition of the graphics card may change suddenly or even malfunction with changes in various application stresses and external environment. Therefore, it is very important to monitor the working condition and faults of the graphics card in real time during operation. In the existing similar solutions, there are certain deficiencies, such as single data source, high demand for hardware resources, etc. SUMMARY

[0004] The present application provides a graphics card running condition recognition method, system and storage medium to overcome the problem of high computing resource requirement of the existing solution, which significantly increases the overall cost of the graphics card.

[0005] According to a first aspect of the present application, a graphics card running condition monitoring method is provided, comprising:

[0006] collecting time series of multiple working parameters during the operation of the graphics card;

[0007] selecting a corresponding classification model for each working parameter, inputting the time series of each working parameter into the corresponding classification model, and obtaining a pre-judgment working condition result corresponding to each working parameter;

[0008] querying a credibility matrix corresponding to each classification model in the database based on the classification model corresponding to each working parameter;

[0009] constructing a BPA function of each classification model according to the credibility matrix corresponding to each classification model;

[0010] calculating a fusion BPA function value of each working condition based on the BPA function of each classification model;

[0011] determining the running condition of the graphics card according to the fusion BPA function value corresponding to the pre-judgment working condition result of each working parameter.

[0012] According to a second aspect of the present application, a graphics card running condition monitoring system is provided, comprising:

[0013] a collection module for collecting time series of multiple working parameters during the operation of the graphics card;

[0014] The acquisition module is configured to select a corresponding classification model for each working parameter, input a time sequence of each working parameter into the corresponding classification model, and acquire a pre-judgment working condition result corresponding to each working parameter;

[0015] The query module is configured to query a confidence matrix corresponding to each classification model in a database based on the classification model corresponding to each working parameter;

[0016] The construction module is configured to construct a BPA function of each classification model according to the confidence matrix corresponding to each classification model;

[0017] The calculation module is configured to calculate a fusion BPA function value of each working condition based on the BPA function of each classification model;

[0018] The determination module is configured to determine a running working condition of the graphics card according to the fusion BPA function value corresponding to the pre-judgment working condition result of each working parameter.

[0019] According to a third aspect of the present application, an electronic device is provided, comprising a memory and a processor, wherein the processor is configured to implement the steps of the graphics card running working condition monitoring method when executing a computer management program stored in the memory.

[0020] According to a fourth aspect of the present application, a computer readable storage medium is provided, which stores a computer management program, wherein the computer management program is configured to implement the steps of the graphics card running working condition monitoring method when executed by a processor.

[0021] The graphics card running working condition monitoring method, system and storage medium provided by the present application collect time sequences of multiple working parameters during graphics card running, select a corresponding classification model for each working parameter, input a time sequence of each working parameter into the corresponding classification model, acquire a pre-judgment working condition result corresponding to each working parameter, query a confidence matrix corresponding to each classification model in a database based on the classification model corresponding to each working parameter, construct a BPA function of each classification model according to the confidence matrix, calculate a fusion BPA function value of each working condition, and determine a running working condition of the graphics card according to the fusion BPA function value corresponding to the pre-judgment working condition result of each working parameter. The present application uses multi-source and multi-modal data, uses a pre-judgment result as a final judgment evidence, and realizes high data type compatibility, low computing power requirement, high robustness of graphics card running working condition recognition and fault recognition. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 A graphics card running working condition monitoring method flowchart is provided for an embodiment of the present application;

[0023] Figure 2 A principle block diagram of a graphics card running condition recognition method according to an embodiment of the present application is shown in the figure.

[0024] Figure 3 A structure block diagram of a graphics card running condition recognition system according to an embodiment of the present application is shown in the figure.

[0025] Figure 4 A hardware structure schematic diagram of a possible electronic device according to the present application is shown in the figure.

[0026] Figure 5 A hardware structure schematic diagram of a possible computer readable storage medium according to the present application is shown in the figure. DETAILED DESCRIPTION

[0027] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application. In addition, the technical features in each embodiment or in a single embodiment provided by the present application can be combined with each other arbitrarily to form a feasible technical solution, and this combination is not restricted by the order of steps and / or structure mode, but should be based on the fact that it can be realized by those of ordinary skill in the art. When the combination of technical solutions appears contradictory or unfeasible, it should be considered that the combination of technical solutions does not exist, and is not within the protection scope of the present application.

[0028] The embodiments of the present application provide a method for graphics card running condition and fault state recognition. The temperature, power consumption, power supply voltage, working frequency and other data can be monitored by the built-in sensor of GPU, so as to realize automatic monitoring and recognition of the graphics card condition and fault condition.

[0029] Figure 1 A flow chart of a graphics card running condition recognition method according to an embodiment of the present application is shown in the figure. Figure 1 and Figure 2 The method comprises the following steps.

[0030] Step 1: collecting time series of multiple working parameters during graphics card running.

[0031] In an embodiment of the present application, the multiple working parameters at least include temperature, power consumption, power supply voltage and working frequency. The collecting of time series of multiple working parameters during graphics card running comprises the following steps.

[0032] The temperature time sequence, the power consumption time sequence, the power supply voltage time sequence and the working frequency time sequence of the graphics card are respectively collected in a preset time segment.

[0033] It can be understood that during the operation of the graphics card, a plurality of working parameters of the graphics card are collected, and in the embodiment of the present application, a plurality of sensors built in the graphics card are used to collect the parameters such as temperature, power consumption, power supply voltage and working frequency of the graphics card. Specifically, the running conditions of the graphics card are periodically monitored and identified, so that each working parameter in a period of time can be periodically collected, and the time sequence of each working parameter can be obtained.

[0034] After the time sequence of each working parameter is collected, the time sequence of each working parameter is preprocessed, and the preprocessing includes filtering and normalization.

[0035] First, filtering is performed: considering the characteristics of sensor data, physical quantities such as temperature and power consumption will not appear high-frequency alternating conditions, so low-pass filtering is performed on the time sequence of each working parameter: , , wherein, is a low-pass filtering parameter, is an output value of this filtering, is an output value of last filtering, is a measurement value of this time, is a sampling period, is a cut-off frequency. After filtering, data normalization is further performed: according to , the data is mapped to interval, which is convenient for subsequent model to process data, is the minimum value in the parameter sequence, is the maximum value in the parameter sequence, is the normalized parameter value, is the parameter value before normalization.

[0036] Step 2: select a corresponding classification model for each working parameter, input the time sequence of each working parameter into the corresponding classification model, and obtain the pre-judgment working condition result corresponding to each working parameter.

[0037] Understandably, a classification model is used to determine the operating condition using the preprocessed time series of each operating parameter as input, resulting in multiple pre-judged operating condition results. Specifically, a corresponding classification model is selected for each operating parameter; different operating parameters can use the same classification model or different models. There are no restrictions on the type of classification model used; models such as CNN, decision tree, support vector machine, and Naive Bayes can be selected based on computing resources. Since multiple sets of data (time series of multiple operating parameters) are used, the time series of each operating parameter is input into the corresponding selected classification model to obtain the pre-judged operating condition results output by the classification model. Multiple pre-judged operating condition results can be obtained as "evidence" for subsequent steps.

[0038] Step 3: Based on the classification model corresponding to each working parameter, query the database for the confidence matrix corresponding to each classification model.

[0039] Understandably, based on the classification model corresponding to each working parameter, the confidence matrix corresponding to the classification model is queried from the database. The confidence matrix represents the accuracy of the classification for each working condition. The confidence matrix is ​​represented as follows:

[0040]

[0041] in, The elements in the matrix represent the classification model. This indicates the actual working conditions are as follows: The samples were classified by the classification model as The probability of a certain type of working condition.

[0042] Step 4: Construct the BPA function for each classification model based on the confidence matrix corresponding to each classification model.

[0043] Understandably, the basic probability assignment function (BPA) is used to construct the set of working conditions for each working parameter based on the confidence matrix obtained from the lookup table.

[0044] The working condition identification framework is defined as follows: ,in, This represents the set of operating conditions for the nth operating parameter. Includes all possible working conditions. It represents uncertainty.

[0045] Define the BPA function: from the power set arrive mapping ,but Can represent the The degree of accurate trust.

[0046] calculating local credibility based on the credibility matrix corresponding to each classification model and the first global credibility , the local credibility represents the local credibility of a working condition result in the classification model, and the global credibility represents the global credibility of the classification model;

[0047] calculating the second global credibility based on the local credibility and the first global credibility , the second global credibility represents the global credibility of a working condition result in the classification model.

[0048] constructing the BPA function of each classification model according to the second global credibility

[0049] ;

[0050] wherein, respectively represent the probability of .

[0051] Step 5, calculating the fusion BPA function value of each working condition based on the BPA function of each classification model.

[0052] It can be understood that the fusion BPA function value of each single working condition is calculated according to the constructed BPA function of each classification model.

[0053] Taking n working parameters as an example, the BPA function value of single working condition a is:

[0054] wherein, , represents the BPA function value of single working condition a, respectively represent the BPA function of the classification model corresponding to the n working parameters, represents that the single working condition a is included in the working condition set of the kth classification model .

[0055] The fusion BPA function value of each single working condition is calculated.

[0056] Step 6, determining the running working condition of the graphics card according to the fusion BPA function value corresponding to the pre-judgment working condition result of each working parameter.

[0057] It can be understood that the fusion BPA function value of each single working condition calculated in the above step 5 finds the fusion BPA function value corresponding to the pre-judgment working condition result of each working parameter, and finally the running working condition of the graphics card is identified according to the fusion BPA function value corresponding to the pre-judgment working condition result of each working parameter.

[0058] Specifically, according to the fusion BPA function value corresponding to the pre-judgment working condition result of each working parameter, the running working condition of the graphics card is determined, including:

[0059] Suppose that the pre-judgment working condition results corresponding to a plurality of working parameters are , , , and the fusion BPA function values of , , are obtained respectively;

[0060] The maximum fusion BPA function value is obtained, and the pre-judgment working condition result corresponding to the maximum fusion BPA function value is represented as

[0061] The BPA function value representing uncertainty is obtained, and the uncertainty is represented as

[0062] When and , the running working condition of the graphics card is determined as ; otherwise, the pre-judgment result of the running working condition of the graphics card is invalid, and the running working condition recognition in the next cycle is waited.

[0063] Referring to Figure 3 , a graphics card running working condition recognition system of an embodiment of the present application is provided, and the system comprises:

[0064] A collection module 301 is configured to collect time series of a plurality of working parameters during running of a graphics card;

[0065] An acquisition module 302 is configured to select a corresponding classification model for each working parameter, input the time series of each working parameter into the corresponding classification model, and acquire a pre-judgment working condition result corresponding to each working parameter;

[0066] A query module 303 is configured to query a credibility matrix corresponding to each classification model in a database based on the classification model corresponding to each working parameter;

[0067] A construction module 304 is configured to construct a BPA function of each classification model according to the credibility matrix corresponding to each classification model;

[0068] A calculation module 305 is configured to calculate a fusion BPA function value of each working condition based on the BPA function of each classification model;

[0069] The determination module 306 is used to determine the operating condition of the graphics card based on the fusion BPA function value corresponding to the pre-judgment result of each working parameter.

[0070] It is understood that the graphics card operating condition identification system provided by the present invention corresponds to the graphics card operating condition identification method provided in the foregoing embodiments. The relevant technical features of the graphics card operating condition identification system can be referred to the relevant technical features of the graphics card operating condition identification method, and will not be repeated here.

[0071] Please see Figure 4 , Figure 4 This is a schematic diagram illustrating an embodiment of the electronic device provided in this invention. For example... Figure 4 As shown, an embodiment of the present invention provides an electronic device 400, including a storage 410, a processor 420, and a computer program 411 stored in the storage 410 and executable on the processor 420. When the processor 420 executes the computer program 411, it implements the steps of the graphics card operating condition identification method.

[0072] Please see Figure 5 , Figure 5 This is a schematic diagram illustrating an embodiment of a computer-readable storage medium provided by the present invention. (See diagram below.) Figure 5 As shown, this embodiment provides a computer-readable storage medium 500, on which a computer program 511 is stored. When the computer program 511 is executed by a processor, it implements the steps of a graphics card operating condition identification method:

[0073] The present invention provides a method, system, and storage medium for identifying the operating conditions of a graphics card, which has the following beneficial effects:

[0074] (1) Using multi-source, multi-modal data can comprehensively reflect the working condition of the graphics card. This patent does not have specific requirements on the types or number of data, only that the data used is related to the working condition of the graphics card.

[0075] (2) During the process of judging the working condition of the graphics card, there is no need to manually determine the parameters, thus avoiding the problem of inaccurate judgment results caused by different operators;

[0076] (2) The results of multi-source prediction are used as the final judgment basis, which does not rely on the prediction accuracy of a single model and has strong robustness.

[0077] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0078] Those skilled in the art will appreciate that embodiments of the present application can be devised for a variety of applications. It is intended that the present application be limited only by the scope of the appended claims, and it is intended that various modifications and alterations made by those skilled in the art be considered as within the scope of the present application. The embodiments of the present application will be described with reference to the attached drawings, wherein:

[0079] The present application is described in reference to the drawings using a flowchart and / or a block diagram of the method, apparatus (system) and computer program product according to embodiments of the application. It will be understood that each block of the flowchart and / or block diagram, and combinations of blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks.

[0080] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks.

[0081] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks.

[0082] While the preferred embodiments of the application have been described, additional variations and modifications can be made to the embodiments by those skilled in the art once they learn of the basic inventive concepts. Therefore, the appended claims are intended to cover all such modifications and variations as fall within the scope of the present application.

[0083] Obviously, many modifications and variations of the present application are possible in light of the above teachings. It is, therefore, to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A method for identifying a graphics card operating condition, characterized in that, The method comprises the following steps: collecting time series of a plurality of working parameters of a graphics card in operation; selecting a corresponding classification model for each working parameter, inputting the time series of each working parameter into the corresponding classification model, and obtaining a pre-judgment working condition result corresponding to each working parameter; querying a credibility matrix corresponding to each classification model in a database based on the classification model corresponding to each working parameter; constructing a BPA function of each classification model according to the credibility matrix corresponding to each classification model; calculating a fusion BPA function value of each working condition based on the BPA function of each classification model; determining the running working condition of the graphics card according to the fusion BPA function value corresponding to the pre-judgment working condition result of each working parameter.

2. The method of claim 1, wherein, The plurality of working parameters at least include temperature, power consumption, power supply voltage and working frequency, and the collecting of the time series of the plurality of working parameters of the graphics card in operation comprises: collecting temperature time series, power consumption time series, power supply voltage time series and working frequency time series of the graphics card respectively within a preset time segment.

3. The method of claim 1, wherein, The collecting of the time series of the plurality of working parameters of the graphics card in operation is followed by the preprocessing of the time series of each working parameter: low-pass filtering the time series of each working parameter. ， ； wherein, is a low-pass filter coefficient, is a current filter output value, is a previous filter output value, is a current measurement value, is a sampling period, is a cutoff frequency; After low-pass filtering, the time series of each working parameter is normalized according to mapping the data to the interval [0, 1], for the normalized data, for the minimum value in the time series, for the maximum value in the time series, is the measurement value after this filtering.

4. The method of claim 1, wherein, The classification model comprises a convolutional neural network (CNN), a decision tree, a support vector machine and a naive Bayes model, and the selecting of the corresponding classification model for each working parameter comprises: selecting one classification model corresponding to each working parameter according to computing resource.

5. The method of claim 1, wherein, The credibility matrix represents the accuracy of the classification model in classifying each working condition, and is expressed as: ； wherein, represents a classification model, an element in the matrix indicates the probability that a sample with a true working condition is classified by the classification model as a working condition class.

6. The method of claim 5, wherein, The constructing of the BPA function of each classification model according to the credibility matrix corresponding to each classification model comprises: The working condition recognition framework is defined as: wherein, represents the nth working condition set, contains all possible working conditions, represents the uncertainty; Definition of BPA function: a mapping from power set to Then may represent the exact trustworthiness of ;​ calculating a local credibility based on the credibility matrix corresponding to each classification model and a first global credibility , the local credibility representing a local credibility of a working condition result in the classification model, and the global credibility representing a global credibility of the classification model; According to the local credibility and the first global credibility, a second global credibility is calculated The second global credibility represents a global credibility of a working condition result of a classification model. constructing the BPA function of each classification model according to the second global credibility: ； wherein respectively represent the probability.

7. The method of claim 6, wherein, The calculating of the fusion BPA function value of each working condition based on the BPA function of each classification model comprises: the BPA function value of a single working condition a is: wherein, , represents the BPA function value of the single working condition a, respectively represent the BPA functions of the n working parameters respectively corresponding classification models, represents the working condition set of the kth classification model including the single working condition a; iteratively calculating the fusion BPA function value of each single working condition.

8. The method of claim 1, wherein, The determining of the running working condition of the graphics card according to the fusion BPA function value corresponding to the pre-judgment working condition result of each working parameter comprises: A plurality of working parameters correspond to the pre-judgment working condition results 、 、....、 , and the fusion BPA function values of 、 ,...., are obtained respectively; acquiring the maximum fusion BPA function value , representing a pre-judgment working condition result corresponding to the maximum fusion BPA function value Obtaining a BPA function value representative of uncertainty , representative of uncertainty; When and , it is determined that the running condition of the graphics card is ; otherwise, the pre-judgment result of the running condition of the graphics card this time is invalid.

9. A graphics card operating condition identification system, characterized in that, The method comprises the following steps: a collecting module for collecting time series of a plurality of working parameters of a graphics card in operation; an obtaining module for selecting a corresponding classification model for each working parameter, inputting the time series of each working parameter into the corresponding classification model, and obtaining a pre-judgment working condition result corresponding to each working parameter; a querying module for querying a credibility matrix corresponding to each classification model in a database based on the classification model corresponding to each working parameter; a constructing module for constructing a BPA function of each classification model according to the credibility matrix corresponding to each classification model; a calculating module for calculating a fusion BPA function value of each working condition based on the BPA function of each classification model; a determining module for determining the running working condition of the graphics card according to the fusion BPA function value corresponding to the pre-judgment working condition result of each working parameter.

10. A computer-readable storage medium, characterized in that, A computer management program is stored on the computer readable medium, and the computer management program is executed by the processor to implement the steps of the graphic card running condition identification method according to any one of claims 1-8.

Images