Notebook computer performance dynamic optimization method and system based on big data analysis

Abstract

The invention discloses a notebook computer performance dynamic optimization method and system based on big data analysis, and belongs to the technical field of computer performance optimization. The method comprises the following steps: S1, generating an application feature portrait and an initial tentative sequence plan; s2, obtaining the state of the notebook computer after the initial resource test; s3, generating a negative experience index; s4, if the negative experience index deteriorates, generating a rollback instruction and generating a heuristic failure context, and if the negative experience index does not deteriorate, judging that the current heuristic is preliminarily valid; s5, obtaining an updated heuristic sequence plan and an updated notebook state; and S6, generating an updated negative experience index. According to the method, the problems of worsening user experience and ecological barriers caused by blind selection of new application cold start stage performance optimization strategies in the prior art are solved through application feature portrait construction and a dynamic optimization mechanism.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer performance optimization, and particularly relates to a notebook computer performance dynamic optimization method and system based on big data analysis. BACKGROUND

[0002] With the rapid development of mobile computing technology and the continuous improvement of hardware performance, the application scenarios that notebook computers need to handle are increasingly complex and diverse. Different applications have significant differences in demand for computing resources. How to balance performance and power consumption while ensuring user experience has become an important issue.

[0003] The existing notebook computer performance optimization system adjusts performance through a pre-set power management mode or a static strategy based on an application whitelist. For example, the power management scheme of the Windows operating system, the driver optimization tool provided by each hardware manufacturer, etc.

[0004] In the prior art, these systems mainly rely on pre-defined power consumption configuration files or user-selected working modes, and have the problems of strategy solidification and lack of adaptive ability. Specifically, when facing new applications, the computer cannot intelligently identify the resource demand characteristics of the new applications, resulting in a strategy selection dilemma during the startup phase of the new applications; due to the lack of an effective experience perception mechanism, the performance fluctuations generated during the optimization process will directly affect the user experience, causing a negative first impression; in the long run, the static optimization strategy will form a technical barrier to emerging applications, affecting the innovative development of the software ecosystem. The existence of these problems seriously restricts the development of notebook computer performance optimization technology, and an intelligent solution that can independently perceive and dynamically adjust is urgently needed. SUMMARY

[0005] The notebook computer performance dynamic optimization method and system based on big data analysis provided by the embodiments of the present application solve the problems of blind selection of performance optimization strategies when facing new applications, easy deterioration of user experience, and hindering of application ecosystem development in the prior art, realize intelligent performance optimization during the cold startup phase of new applications, guarantee the first-time use experience of users, and improve the compatibility and ecological vitality of the computing platform for diversified applications.

[0006] The notebook computer performance dynamic optimization method based on big data analysis provided by the embodiments of the present application comprises the following steps: S1: when an unknown application is detected to be started, generating an application feature portrait and an initial trial sequence plan;

[0007] S2: loading a benchmark performance strategy, performing initial resource trials according to the initial trial sequence plan, and obtaining a notebook state after the initial resource trials;

[0008] S3: generating a negative experience index in the notebook state after the initial resource trials.

[0009] S4: If the negative experience indicator deteriorates, generate rollback instructions and generate a tentative failure context, if the negative experience indicator does not deteriorate, determine that the current tentative is preliminary effective;

[0010] S5: Based on the tentative failure context, generate a secondary resource tentative, obtain an updated tentative sequence plan and an updated notebook state;

[0011] S6: Based on the updated tentative sequence plan, generate an updated negative experience indicator, if the updated negative experience indicator does not deteriorate and the performance has gain, determine that it is an effective tentative, if the updated negative experience indicator deteriorates, jump to step S4 to process the new tentative failure.

[0012] Further, the step of generating an application feature image and an initial tentative sequence plan comprises:

[0013] Performing static analysis on the executable file and associated configuration file of the unknown application, identifying and extracting the dynamically linked library dependency list explicitly declared by the unknown application, and analyzing the unknown application manifest file to obtain its declared minimum hardware resource requirement;

[0014] Integrate the extracted dynamically linked library dependency list and hardware requirement information to construct a structured data summary, i.e. an application feature image;

[0015] According to the resource dependency preference reflected in the application feature image, arrange all controllable resource items of the notebook computer in descending order according to the priority of expected impact on performance, and generate an initial tentative sequence plan.

[0016] Further, the step of performing an initial resource tentative according to the initial tentative sequence plan to obtain the notebook state after the initial resource tentative comprises:

[0017] Set all controllable hardware resource parameters of the notebook computer to the pre-defined standard running configuration, i.e. load the benchmark performance policy;

[0018] Select the resource type ranked first in the initial tentative sequence plan as the first resource item;

[0019] Adjust the first resource item by operating the performance state adjustment interface provided by the notebook computer operating system kernel;

[0020] Maintain a preset stable observation time after each adjustment, record the system performance benchmark state and hardware configuration parameters before adjustment;

[0021] After the stable observation time ends, collect the notebook computer system state data;

[0022] The collected notebook computer system state data is compared with the recorded baseline state before adjustment to form an initial resource trial notebook state description;

[0023] The first amplitude adjustment of the resource is calculated by a resource adjustment amplitude formula.

[0024] The resource adjustment amplitude formula is:

[0025] ;

[0026] In the formula, represents the resource adjustment amplitude, represents the adjustment coefficient, represents the baseline value of the resource item.

[0027] Further, the step of generating a negative experience index includes:

[0028] Obtaining a multi-dimensional performance index related to the user's subjective perception, the multi-dimensional performance index including:

[0029] Continuously capturing a rendering frame generation timestamp, calculating a sliding average value and a standard deviation of the time difference between consecutive frames to obtain rendering frame time statistical values;

[0030] Obtaining an input event timestamp, measuring the time interval from the occurrence of the input event to the completion of the graphical interface update to obtain an input response delay value;

[0031] Collecting the proportion of time that the application main thread is in a non-responsive state to obtain the main thread blocking duration;

[0032] Converting the performance index data of the three dimensions of rendering frame time statistical values, input response delay values, and main thread blocking duration into standardized scores according to a preset normalization formula;

[0033] According to the converted standardized scores, a negative experience index is calculated by a negative experience index calculation formula.

[0034] Further, if the negative experience index deteriorates, a rollback instruction is generated, and the step of generating a trial failure context includes:

[0035] When the negative experience index is not less than a preset deterioration threshold, a resource configuration recovery instruction is sent to the power management module of the notebook computer operating system;

[0036] The resource configuration recovery instruction contains a resource item identifier to be rolled back and a target recovery value, and the target recovery value is the baseline configuration parameter of the resource item recorded before the initial resource trial adjustment;

[0037] During the execution of the rollback operation, the notebook computer system state is continuously monitored until all parameters return to the pre-adjustment baseline state;

[0038] A heuristic failure context record is generated, and the initial resource heuristic failure context record is stored in a temporary buffer in a structured data format.

[0039] Further, based on the updated heuristic sequence plan, the step of generating an updated negative experience index includes:

[0040] Selecting the next resource item to be tested in order from the updated heuristic sequence plan;

[0041] Implementing configuration parameter adjustment on the resource item to be tested through the notebook computer system management driver;

[0042] Starting a new round of performance index collection period after the adjustment is completed;

[0043] During this collection period window, continuously collect the original performance data of the notebook computer;

[0044] Based on the collected original performance data, calculate and output the updated negative experience index.

[0045] Further, the negative experience index calculation formula is:

[0046] ;

[0047] In the formula, represents the negative experience index, , , respectively represent the weight coefficients of the rendering frame time, input response delay and main thread blocking time, represents the rendering frame time standardization score, represents the input response delay standardization score, represents the main thread blocking standardization score.

[0048] The notebook computer performance dynamic optimization system based on big data analysis provided by the embodiments of the present application is used to implement the notebook computer performance dynamic optimization method based on big data analysis, which includes:

[0049] Application feature generation module, initial resource heuristic module, negative index generation module, heuristic result determination module, secondary resource heuristic module, final heuristic determination module;

[0050] The application feature generation module is used to generate an application feature portrait and an initial heuristic sequence plan when an unknown application is detected;

[0051] The initial resource trial module is configured to load a benchmark performance policy, perform initial resource trial according to an initial trial sequence plan, and obtain a notebook state after initial resource trial;

[0052] The negative indicator generation module is configured to generate a negative experience indicator in the notebook state after initial resource trial;

[0053] The trial result determination module is configured to generate a rollback instruction and a trial failure context if the negative experience indicator deteriorates, and determine that the current trial is initially effective if the negative experience indicator does not deteriorate.

[0054] The secondary resource trial module is configured to generate secondary resource trial based on the trial failure context, and obtain an updated trial sequence plan and an updated notebook state.

[0055] The final trial determination module is configured to generate an updated negative experience indicator based on the updated trial sequence plan, and determine that the trial is effective if the updated negative experience indicator does not deteriorate and performance is improved, or jump to step S4 to handle a new trial failure if the updated negative experience indicator deteriorates.

[0056] The one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0057] By applying the generation method of the feature portrait and the initial trial sequence plan, the computer can intelligently predict resource demand based on static features when encountering unknown applications, thereby avoiding the dilemma of blindly selecting conservative or aggressive strategies in the cold start stage, and effectively solving the cold start problem of new application recognition and optimal strategy selection. Further, in the cold start optimization process, the computer adjusts resources in small and reversible steps and rolls back immediately when detecting deterioration of user experience due to the use of progressive resource trial and negative experience indicator monitoring, ensuring smooth experience for users when using new applications for the first time and reducing negative first impressions caused by performance lag. Further, in the dynamic optimization process, by reconstructing the trial sequence based on the failure context and pruning high-risk options, the computer system can quickly learn and avoid invalid configurations, improving the optimization efficiency of individual applications. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 The notebook computer performance dynamic optimization method flowchart based on big data analysis provided in the embodiments of the present application;

[0059] Figure 2 The structure diagram of the notebook computer performance dynamic optimization system based on big data analysis provided in the embodiments of the present application. DETAILED DESCRIPTION

[0060] This application provides a method and system for dynamic performance optimization of laptops based on big data analysis. It solves the problems in the prior art where the selection of performance optimization strategies during the cold start phase of new applications is blind, which can easily lead to a deterioration in user experience and hinder the development of the application ecosystem. By constructing application feature profiles and implementing a dynamic optimization mechanism that includes progressive resource adjustments and monitoring of negative experience indicators, it achieves intelligent optimization of new application performance and ensures user experience, thereby improving the compatibility of the computing platform with diverse applications and the vitality of the ecosystem.

[0061] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0062] like Figure 1 The diagram shows a flowchart of a method for dynamic optimization of laptop performance based on big data analysis provided in this application embodiment. The method is applied to a system for dynamic optimization of laptop performance based on big data analysis. The method includes the following steps: S1: When an unknown application without a historical strategy is detected to start, static feature analysis is performed on the unknown application to extract information such as link libraries and hardware requirements, generate an application feature profile, and generate an initial trial sequence plan sorted by resource type based on the application feature profile.

[0063] S2: Load the baseline performance strategy and conduct initial resource testing according to the initial testing sequence plan. The initial resource testing includes: selecting the first resource item for reversible guided incremental performance improvement to obtain the laptop state after the initial resource testing.

[0064] The laptop status includes hardware resource configuration status and system performance status. Hardware resource configuration status includes CPU core frequency, CPU power limit, graphics processing unit (GPU) core frequency, memory controller timing parameters, and storage device I / O priority. System performance status includes current power consumption readings, temperature sensor values, frequency counter values, and latency measurements. For example, when adjusting the CPU power limit, the laptop status record shows that before the adjustment, the core frequency was 2.5GHz and the power limit was 15W; after the adjustment, the core frequency is 3.2GHz and the power limit is 25W. Simultaneously, the system power consumption reading changes from 20W to 30W, and the temperature sensor reading changes from 45℃ to 65℃.

[0065] S3: In the laptop state after initial resource exploration, obtain multi-dimensional performance indicators that are strongly correlated with the user's subjective perception and generate negative experience indicators;

[0066] S4: If the negative experience index worsens, generate rollback instructions and generate a trial failure context combining the initial resource trial notebook state, if the negative experience index does not worsen, determine that the current trial is initially effective, and continue to execute further gradual performance improvement along the initial resource trial direction;

[0067] S5: Based on the trial failure context, generate a secondary resource trial, obtain an updated trial sequence plan and an updated notebook state;

[0068] The steps of the secondary resource trial include:

[0069] According to the trial failure context, the root cause of performance deterioration is determined using a decision tree analysis method, and the decision tree nodes include resource type, adjustment amplitude, temperature change rate and power consumption change rate.

[0070] Using association rule mining technology, the association relationship between resource adjustment and performance deterioration in historical trial failure records is analyzed, and a resource adjustment taboo rule table is generated.

[0071] According to the taboo rule table, all resource adjustment options that violate the taboo rules are removed from the initial trial sequence plan.

[0072] Through the remaining adjustable resource items, the system performance is reordered according to the sensitivity of the resource type, and the sensitivity is calculated by the ratio of performance improvement amplitude to resource adjustment amplitude in historical successful trial records.

[0073] S6: Based on the updated trial sequence plan, generate an updated negative experience index, if the updated negative experience index does not worsen and the performance has gain, determine that it is an effective trial and jump to step S2 to continue execution along the secondary resource trial, if the updated negative experience index worsens, jump to step S4, input the updated negative experience index and the updated notebook state, and handle the new trial failure, that is, generate a trial failure context according to the updated negative experience index and the updated notebook state;

[0074] The basis for judging the deterioration of the negative experience index is to obtain the deterioration index of the current negative experience index through the deterioration degree calculation formula:

[0075] ;

[0076] In the formula, The deterioration index is a percentage value, The negative experience index before adjustment is obtained from the negative experience index history record generated in step S3, The negative experience index after adjustment is obtained from the negative experience index generated in real time in step S3.

[0077] When the derived deterioration index is not less than a preset degree threshold value obtained through a historical database, it is determined that the negative experience index deteriorates.

[0078] When the guided progressive exploration cycle is normally ended due to the completion of all trial sequences, all resource items determined to be valid trials and their final configuration parameters are summarized to form a final optimization strategy corresponding to the application feature image;

[0079] The mapping relationship between the application feature image, the trial failure context, the valid trial, and the finally formed optimization strategy is stored in real time in a lightweight local knowledge base for rapid strategy loading of the application or other similar feature image applications.

[0080] Further, the steps of generating the application feature image and the initial trial sequence plan include:

[0081] Performing static analysis on the executable file and associated configuration file of the unknown application, identifying and extracting the dynamically linked library dependency list explicitly declared by the unknown application through binary scanning technology, and analyzing the unknown application manifest file to obtain its declared minimum hardware resource requirements, including processor instruction set, memory capacity lower limit, and graphics processing unit capability requirement;

[0082] Integrating the extracted dynamically linked library dependency list and hardware requirement information to construct a structured data summary, i.e., an application feature image;

[0083] The step of integrating the extracted dynamically linked library dependency list and hardware requirement information includes:

[0084] Using feature vector normalization technology to construct a structured data summary:

[0085] Converting the dynamically linked library dependency list to a binary feature vector through one-hot encoding.

[0086] Mapping the hardware requirement information to the [0, 1] interval through the min-max normalization method.

[0087] Using principal component analysis method to reduce the dimension of the feature vector, and retaining the principal components with a variance contribution rate of more than 95%.

[0088] Concatenating the reduced dimension feature vector and the hardware requirement normalized value to form a fixed dimension application feature image.

[0089] According to the resource dependency preference reflected in the application feature image, such as high processor load or high memory bandwidth occupancy, all controllable resource items of the notebook computer, such as processor core frequency, memory timing, and input-output scheduler mode, are arranged in descending order according to the priority of expected performance impact to generate an initial trial sequence plan.

[0090] The priority of the expected impact on performance is determined by performance counter sampling analysis.

[0091] Further, the step of performing an initial resource probe according to the initial probe sequence plan to obtain an initial resource-probed notebook state comprises:

[0092] Setting all controllable hardware resource parameters of the notebook computer to a predefined standard running configuration, i.e., loading a benchmark performance policy;

[0093] Selecting the resource type ranked first in the initial probe sequence plan as a first resource item;

[0094] Implementing a first magnitude adjustment on the resource item through a performance state adjustment interface provided by the notebook computer operating system kernel, and the adjustment mode comprises increasing a central processing unit power consumption limit value, adjusting a graphics processing unit core frequency offset, modifying a memory timing parameter, or improving a storage device input / output priority;

[0095] Maintaining a preset stable observation duration after each adjustment, which is sufficient to complete at least three complete performance indicator sampling periods, and recording the system performance benchmark state and hardware configuration parameters before adjustment;

[0096] After the stable observation duration ends, collecting notebook computer system state data including current power consumption readings (reading current sensor sampling values through the I2C interface of the motherboard power management chip), temperature sensor values (reading digital temperature sensor register values through the embedded controller), frequency count values (directly reading through the processor performance state register), and delay measurement values (obtained by measuring interrupt response time through a high-precision event timer);

[0097] Comparing and analyzing the collected notebook computer system state data with the recorded benchmark state before adjustment to form an initial resource-probed notebook state description containing resource item type, adjustment direction, adjustment magnitude, and system response data;

[0098] The magnitude of the first magnitude adjustment of the resource is calculated through a resource adjustment magnitude formula;

[0099] The resource adjustment magnitude formula is:

[0100] ;

[0101] In the formula, represents the resource adjustment magnitude, which is a dimensionless proportion value, represents an adjustment coefficient, which is a preset constant, and the value range is 0.01 to 0.1, The reference value representing the resource item is derived from the current resource configuration parameters obtained from the notebook computer operating system kernel performance state adjustment interface, which ensures that the adjustment range is small and reversible, and is based on the performance state transition mechanism implemented in the existing power management specifications, such as the ACPI standard.

[0102] The adjustment coefficient is calculated based on the thermal design power margin of the notebook computer:

[0103] ;

[0104] In the formula, where represents the current thermal design power, represents the reference thermal design power, which is obtained by reading the processor model-specific register.

[0105] Further, the step of generating a negative experience index includes:

[0106] Obtaining a multi-dimensional performance index strongly related to the user's subjective perception, the multi-dimensional performance index including:

[0107] Continuously capturing rendering frame generation timestamps through a graphical application interface, calculating the sliding average and standard deviation of the time difference between consecutive frames to obtain rendering frame time statistics; the calculation formula is:

[0108] ;

[0109] ;

[0110] where, represents the timestamp of the frame, represents the sampling window size (typical value 60);

[0111] Obtaining input event timestamps through a human-computer interface device driver, measuring the time interval from the occurrence of an input event to the completion of graphical interface update to obtain an input response delay value; the calculation formula is:

[0112] ;

[0113] where, represents the input event timestamp, represents the interface update completion timestamp;

[0114] Obtaining the main thread blocking duration by collecting the proportion of time that the application main thread is in a non-responsive state through a notebook computer system performance monitor; the calculation formula is:

[0115] ;

[0116] where represents the timestamp of the Duration of the second blockage. Indicates the total observation time. Indicates the number of times the device was blocked.

[0117] The performance metrics data of three dimensions—rendering frame time statistics, input response latency, and main thread blocking duration—are converted into standardized scores according to a preset normalization formula.

[0118] The negative experience index is calculated based on the standardized score of the conversion using the formula for calculating the negative experience index.

[0119] Furthermore, if negative experience metrics worsen, the steps to generate a rollback instruction and a trial failure context include:

[0120] When the negative experience index is not less than the preset deterioration threshold obtained from the historical database, a resource configuration recovery command is sent to the power management module of the laptop operating system.

[0121] The resource configuration recovery instruction contains the identifier of the resource item to be rolled back and the target recovery value. The target recovery value is the baseline configuration parameter of the resource item recorded before the initial resource trial adjustment.

[0122] During the rollback operation, the laptop system status is continuously monitored until all parameters (hardware configuration parameters, power management parameters, temperature control parameters, performance status parameters) are restored to the baseline state before the adjustment;

[0123] Generate a resource probing failure context record, which includes the type of resource item that caused the deterioration, a description of the specific adjustment operation, the adjustment magnitude, the baseline value of the negative experience index before the deterioration, the peak value after the deterioration, the duration of the deterioration, the system temperature reading change curve, and the power consumption reading change curve. Store the initial resource probing failure context record in a temporary buffer in a structured data format.

[0124] Furthermore, based on the updated trial sequence plan, the steps for generating updated negative experience metrics include:

[0125] Select the next resource item to be tested from the updated test sequence plan in sequence;

[0126] The order refers to the descending order of the sensitivity of resource types to system performance. The sensitivity is determined by calculating the average performance improvement rate brought about by adjustments to each resource using historical data. The calculation formula is as follows:

[0127] ;

[0128] in Representing resources Sensitivity Indicates the first This adjustment of resources Performance improvement rate Indicates the first This adjustment of resources The adjustment range.

[0129] The system management driver on the laptop computer adjusts the configuration parameters of the resource item being tested, and the adjustment range is controlled within a preset safe change range.

[0130] After the adjustment is completed, a new round of performance indicator collection cycle will be started. The duration of this collection cycle shall not be less than the shortest time window required for the stability of the laptop system.

[0131] The method for determining the shortest time window required for the stability of a laptop system is through step response analysis.

[0132] Within this acquisition period window, raw performance data of the laptop is continuously acquired through performance counters. The raw performance data includes raw rendering frame time data, input response latency data, and main thread blocking duration data.

[0133] Based on the collected raw performance data, an updated negative experience index is calculated and output. This value reflects the level of performance experience quality that users may perceive in the current updated laptop state.

[0134] Furthermore, the formula for calculating the negative experience index is:

[0135] ;

[0136] In the formula, This represents a negative experience indicator, which is a dimensionless comprehensive score. , , These represent the weighting coefficients for rendering frame time, input response latency, and main thread blocking duration, respectively; all are preset constants. , This represents the normalized score for rendering frame time, calculated and derived from the moving average of rendering frame time and a preset frame time threshold. This represents the standardized score for input response latency, calculated and derived from the input event latency measurement and a preset latency threshold. This represents the standardized score for main thread blocking, calculated and derived from the proportion of main thread non-responsive time and a preset blocking threshold.

[0137] The formula for calculating the standardized score is as follows:

[0138] ;

[0139] In the formula, represents an ideal frame time, represents a maximum acceptable frame time.

[0140] The formula for calculating the input response latency normalized score is:

[0141] ;

[0142] wherein represents a maximum acceptable latency.

[0143] The formula for calculating the main thread blocking normalized score is:

[0144] ;

[0145] wherein represents a maximum acceptable blocking ratio.

[0146] As shown in Figure 2 FIG. 1 is a structural schematic diagram of a notebook computer performance dynamic optimization system based on big data analysis provided by an embodiment of the present application. The notebook computer performance dynamic optimization system based on big data analysis provided by the embodiment of the present application comprises an application feature generation module, an initial resource trial module, a negative index generation module, a trial result determination module, a secondary resource trial module, and a final trial determination module.

[0147] The application feature generation module is configured to generate an application feature portrait and an initial trial sequence plan when detecting that an unknown application is started.

[0148] The initial resource trial module is configured to load a benchmark performance strategy, perform initial resource trial according to the initial trial sequence plan, and obtain a notebook state after initial resource trial.

[0149] The negative index generation module is configured to generate a negative experience index in the notebook state after initial resource trial.

[0150] The trial result determination module is configured to generate a rollback instruction and a trial failure context if the negative experience index deteriorates, and determine that the current trial is initially effective if the negative experience index does not deteriorate.

[0151] The secondary resource trial module is configured to generate secondary resource trial based on the trial failure context, obtain an updated trial sequence plan and an updated notebook state.

[0152] The final exploration decision module is configured to generate an updated negative experience index based on the updated exploration sequence plan, determine that the exploration is effective if the updated negative experience index is not deteriorated and performance is improved, and jump to step S4 if the updated negative experience index is deteriorated.

[0153] The above formulas are dimensionless values calculated, and the formulas are obtained by collecting a large amount of data to simulate a formula of the most recent real situation, and preset parameters in the formula are set by a person skilled in the art according to actual conditions.

[0154] The above embodiments can be realized wholly or partially by software, hardware, firmware or any other combination. When realized by software, the above embodiments can be realized wholly or partially in the form of a computer program product.

[0155] Those skilled in the art can realize that the modules and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0156] In addition, each functional module in each embodiment of the present application can be integrated in one processing module, or each module can exist physically independently, or two or more modules can be integrated in one module.

[0157] The above is merely a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0158] Finally, the above is only the preferred embodiment of the present application, and is not used to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for dynamic optimization of notebook computer performance based on big data analysis, characterized in that, The method comprises the following steps: S1: generating an application feature image and an initial trial sequence plan when detecting that an unknown application is started; S2: loading a benchmark performance strategy, performing initial resource trials according to the initial trial sequence plan, and obtaining a notebook state after the initial resource trials; S3: generating a negative experience index in the notebook state after the initial resource trials; S4: if the negative experience index deteriorates, generating a rollback instruction and a trial failure context, and if the negative experience index does not deteriorate, determining that the current trial is initially effective; S5: generating secondary resource trials based on the trial failure context, obtaining an updated trial sequence plan and an updated notebook state; S6: generating an updated negative experience index based on the updated trial sequence plan, and if the updated negative experience index does not deteriorate and performance has increased, determining that the trial is effective, and if the updated negative experience index deteriorates, jumping to step S4 to handle the new trial failure.

2. The notebook computer performance dynamic optimization method based on big data analysis of claim 1, wherein, The step of generating an application feature image and an initial trial sequence plan comprises: performing static analysis on the executable file and the associated configuration file of the unknown application, identifying and extracting the dynamically linked library dependency list explicitly declared by the unknown application, and analyzing the unknown application manifest file to obtain the minimum hardware resource requirement declared by the unknown application; integrating the extracted dynamically linked library dependency list and the hardware requirement information to construct a structured data summary, i.e., an application feature image; arranging all controllable resource items of the notebook computer in descending order according to the priority of expected performance impact based on the resource dependency preference reflected in the application feature image, and generating an initial trial sequence plan.

3. The notebook computer performance dynamic optimization method based on big data analysis of claim 1, wherein, The step of performing initial resource trials according to the initial trial sequence plan to obtain a notebook state after the initial resource trials comprises: setting all controllable hardware resource parameters of the notebook computer to a predefined standard running configuration, i.e., loading a benchmark performance strategy; selecting the resource type ranked first in the initial trial sequence plan as the first resource item; implementing a first amplitude adjustment on the resource item through a performance state adjustment interface provided by the notebook computer operating system kernel; maintaining a preset stable observation time after each adjustment, recording the system performance benchmark state and hardware configuration parameters before adjustment; collecting notebook computer system state data after the stable observation time ends; comparing and analyzing the collected notebook computer system state data with the recorded benchmark state before adjustment to form a notebook state description after the initial resource trials; wherein the amplitude of the first amplitude adjustment is calculated by a resource adjustment amplitude formula; The resource adjustment amplitude formula is: ； In the formula, represents the resource adjustment range, represents the adjustment coefficient, represents the reference value of the resource item.

4. The notebook computer performance dynamic optimization method based on big data analysis of claim 1, wherein, The step of generating a negative experience index comprises: obtaining multi-dimensional performance indexes strongly related to user subjective perception, which include: continuously capturing rendering frame generation timestamps, calculating the sliding average and standard deviation of the time difference between consecutive frames to obtain rendering frame time statistics; obtaining input event timestamps, measuring the time interval from the occurrence of an input event to the completion of graphical interface update to obtain input response delay values; collecting the proportion of time that the application program main thread is in a non-responsive state to obtain the main thread blocking time; The performance indicator data of the three dimensions of the rendering frame time statistics value, the input response delay value and the main thread blocking duration are converted into standardized scores according to a preset normalization formula; According to the converted standardized scores, a negative experience index is calculated by a negative experience index calculation formula.

5. The notebook computer performance dynamic optimization method based on big data analysis of claim 1, wherein, If the negative experience index deteriorates, a rollback instruction is generated, and the step of generating a trial failure context includes: When the negative experience index is not less than a preset deterioration threshold, a resource configuration recovery instruction is sent to a power management module of the notebook computer operating system; The resource configuration recovery instruction contains a resource item identifier to be rolled back and a target recovery value, which is a baseline configuration parameter recorded before the initial resource trial adjustment; During the execution of the rollback operation, the notebook computer system state is continuously monitored until all parameters are restored to the baseline state before adjustment; An initial resource trial failure context record is stored in a temporary buffer in a structured data format.

6. The notebook computer performance dynamic optimization method based on big data analysis of claim 1, wherein, Based on the updated trial sequence plan, the step of generating an updated negative experience index includes: The next resource item to be tried is selected in order from the updated trial sequence plan; The notebook computer system management driver adjusts the configuration parameters of the resource item to be tried; After the adjustment is completed, a new round of performance indicator collection period is started; During this collection period window, the original performance data of the notebook computer is continuously collected; Based on the collected original performance data, an updated negative experience index is calculated and output.

7. The notebook computer performance dynamic optimization method based on big data analysis of claim 4, wherein, The negative experience index calculation formula is: ； In the formula, represents a negative experience index, , , respectively represent the weight coefficients of the rendering frame time, the input response delay and the main thread blocking time length, represents the rendering frame time standardization score, represents the input response delay standardization score, represents the main thread blocking standardization score.

8. A notebook computer performance dynamic optimization system based on big data analysis, for implementing the notebook computer performance dynamic optimization method based on big data analysis of any one of claims 1-7, characterized in that, It includes: An application feature generation module, an initial resource trial module, a negative index generation module, a trial result determination module, a secondary resource trial module, and a final trial determination module; The application feature generation module is configured to generate an application feature portrait and an initial trial sequence plan when an unknown application is detected; The initial resource trial module is configured to load a baseline performance strategy, perform an initial resource trial according to the initial trial sequence plan, and obtain a notebook state after the initial resource trial; The negative index generation module is configured to generate a negative experience index in the notebook state after the initial resource trial; The trial result determination module is configured to generate a rollback instruction and a trial failure context if the negative experience index deteriorates, and determine that the current trial is preliminary effective if the negative experience index does not deteriorate; The secondary resource trial module is configured to generate a secondary resource trial based on the trial failure context, obtain an updated trial sequence plan and an updated notebook state; The final trial determination module is configured to generate an updated negative experience index based on the updated trial sequence plan, determine that the trial is effective if the updated negative experience index does not deteriorate and the performance is improved, and jump to step S4 to handle a new trial failure if the updated negative experience index deteriorates.

Images