Real-time dynamic test optimization method for embedded environment compiler

Abstract

The invention discloses a real-time dynamic test optimization method for an embedded environment compiler, which belongs to the technical field of embedded systems and comprises the following steps of: acquiring and processing real-time dynamic data of the embedded environment compiler, and determining real-time dynamic characteristic data of the embedded environment compiler; building a real-time dynamic test identification model of the embedded environment compiler, analyzing the real-time dynamic characteristic data of the embedded environment compiler, automatically identifying potential defects of the embedded system, and determining a real-time dynamic test identification result of the embedded environment compiler; and formulating an embedded system optimization scheme to optimize the embedded system in time. The problems that the running test state of the embedded environment compiler cannot be dynamically monitored and analyzed, potential defects cannot be found and optimized in time, and the running efficiency of an embedded system is reduced in the prior art are solved. According to the method, the running test state of the embedded environment compiler can be dynamically monitored and analyzed, potential defects can be found and optimized in time, and the running efficiency of an embedded system is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of embedded systems, in particular to a real-time dynamic test optimization method for embedded environment compiler. BACKGROUND

[0002] The embedded environment compiler refers to the compiler used in the development of embedded systems, and its main function is to convert the source code written in a high-level programming language into machine code that can be directly executed by the target embedded device. In embedded development, the compiler plays an important role because it directly affects the execution efficiency of the program and the stability of the system.

[0003] The existing technology cannot dynamically monitor and analyze the running test state of the embedded environment compiler, cannot timely discover potential defects and optimize, and reduces the running efficiency of the embedded system. SUMMARY

[0004] The present application provides a real-time dynamic test optimization method for embedded environment compiler, which can dynamically monitor and analyze the running test state of the embedded environment compiler, can timely discover potential defects and optimize, and can improve the running efficiency of the embedded system, thereby solving the problems raised in the background art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical scheme:

[0006] A real-time dynamic test optimization method for embedded environment compiler, comprising:

[0007] Collecting real-time dynamic data of the embedded environment compiler and processing it, and extracting features related to real-time dynamic test optimization of the embedded environment compiler to determine real-time dynamic feature data of the embedded environment compiler;

[0008] Building a real-time dynamic test recognition model for the embedded environment compiler, analyzing the real-time dynamic feature data of the embedded environment compiler, and automatically recognizing potential defects of the embedded system to determine the real-time dynamic test recognition result of the embedded environment compiler;

[0009] Formulating an optimization scheme for the embedded system to timely optimize the embedded system, and real-time monitoring the optimized embedded system, adjusting the optimization scheme for the embedded system according to the monitoring feedback, and forming a closed-loop test optimization for the embedded system.

[0010] Preferably, the real-time dynamic feature data of the embedded environment compiler is analyzed, and potential defects of the embedded system are automatically recognized, and the following operations are performed:

[0011] According to the real-time dynamic test optimization requirements of the embedded environment compiler, a real-time dynamic test identification model of the embedded environment compiler is constructed, and the real-time dynamic test identification model of the embedded environment compiler is deployed in an actual real-time dynamic test identification environment of the embedded environment compiler.

[0012] The real-time dynamic feature data of the embedded environment compiler is input into the real-time dynamic test identification model of the embedded environment compiler, the real-time dynamic feature data of the embedded environment compiler is analyzed according to the real-time dynamic test identification model of the embedded environment compiler, and potential defects of the embedded system are automatically identified, so as to determine the real-time dynamic test identification result of the embedded environment compiler.

[0013] Preferably, the embedded system is optimized in time by formulating an embedded system optimization scheme, and the following operations are performed:

[0014] According to the real-time dynamic test identification result of the embedded environment compiler, an embedded system optimization scheme is formulated, and the embedded system is optimized in time according to the embedded system optimization scheme;

[0015] Among them, the parameters of the embedded environment compiler are adjusted, the instructions suitable for the hardware characteristics are generated, and the hardware accelerator is used to unload the complex computing tasks, so as to reduce the load of the main processor;

[0016] The embedded system is divided into multiple modules, each module is independently optimized, the storage mode of data in the memory is adjusted, and the access delay is reduced;

[0017] The appropriate compilation optimization level is selected, the optimization options are adjusted according to the application scenario, such as enabling or disabling function inlining, deleting unused code, updating the environment compiler cache, and compiling only the changed part, which is used to improve the compilation efficiency, and dynamically adjusting the environment compiler cache, task scheduling and resource allocation to ensure the efficient operation of the system.

[0018] Preferably, the real-time dynamic data of the embedded environment compiler is collected, including:

[0019] The environmental temperature, humidity, pressure, light intensity, position, motion state and input signals of buttons, switches or touch screens during the running test process of the embedded environment compiler are monitored in real time, and the environmental perception data is collected;

[0020] The CPU load, memory usage, interrupt response time and power consumption during the running test process of the embedded environment compiler are monitored in real time, and the system running data is collected;

[0021] The function call frequency and execution time, data flow path and transmission efficiency, and code coverage during the running test process of the embedded environment compiler are monitored in real time, and the application program execution data is collected.

[0022] According to the environmental perception data, system operation data and application program execution data, the embedded environment compiler real-time dynamic data is determined.

[0023] Preferably, the embedded environment compiler real-time dynamic data is processed to perform the following operations:

[0024] The embedded environment compiler real-time dynamic data is cleaned to remove noise in the embedded environment compiler real-time dynamic data that is not valuable for embedded environment compiler real-time dynamic test optimization, thereby reducing the interference of noise on embedded environment compiler real-time dynamic test optimization.

[0025] The embedded environment compiler real-time dynamic data is normalized to convert the embedded environment compiler real-time dynamic data into a unified data format, remove dimensional differences in the embedded environment compiler real-time dynamic data, and form standardized embedded environment compiler real-time dynamic data.

[0026] Preferably, the embedded environment compiler real-time dynamic data is processed to perform the following operations:

[0027] The embedded environment compiler real-time dynamic data is integrated into a unified data view, and the integrated embedded environment compiler real-time dynamic data is verified for integrity. After the data integrity verification is passed, the integrated embedded environment compiler real-time dynamic data is securely stored.

[0028] The embedded environment compiler real-time dynamic data is feature extracted to extract features related to embedded environment compiler real-time dynamic test optimization from the embedded environment compiler real-time dynamic data, and determine embedded environment compiler real-time dynamic feature data.

[0029] Preferably, the embedded environment compiler real-time dynamic test recognition model is constructed to perform the following operations:

[0030] The embedded environment compiler real-time dynamic test historical data is collected and divided to determine the training set and the test set.

[0031] Based on deep learning technology, the training set is used to train the deep learning model, so that the deep learning model learns the embedded environment compiler real-time dynamic test analysis behavior from the training set and automatically identifies potential defects of the embedded system, and determines the embedded environment compiler real-time dynamic test recognition model based on deep learning.

[0032] The performance of the deep learning-based embedded environment compiler real-time dynamic test identification model is tested based on the test set, so as to evaluate whether the deep learning-based embedded environment compiler real-time dynamic test identification model can achieve the expected effect of automatically identifying potential defects of the embedded system.

[0033] When the deep learning-based embedded environment compiler real-time dynamic test identification model cannot achieve the expected effect of automatically identifying potential defects of the embedded system, the parameters of the deep learning-based embedded environment compiler real-time dynamic test identification model are adjusted and iteratively optimized until the deep learning-based embedded environment compiler real-time dynamic test identification model can achieve the expected effect of automatically identifying potential defects of the embedded system, so as to determine the optimal embedded environment compiler real-time dynamic test identification model.

[0034] Preferably, the dynamically adjusted environment compiler cache, task scheduling and resource allocation are monitored in real time, and the embedded system optimization scheme is adjusted according to the real-time monitoring feedback to form a closed-loop test optimization of the embedded system.

[0035] Preferably, the embedded environment compiler real-time dynamic test identification model is constructed to analyze the embedded environment compiler real-time dynamic characteristic data and automatically identify potential defects of the embedded system, and the following operations are further performed:

[0036] The analysis and identification history of the embedded environment compiler real-time dynamic test identification model is obtained; wherein the analysis and identification history at least includes: analysis results of analyzing the embedded environment compiler real-time dynamic characteristic data in the history, and results of automatically identifying potential defects of the embedded system in the history;

[0037] The feedback guidance confidence of each target data in the analysis and identification history is calculated by the following formula:

[0038]

[0039] Wherein, Z t is the feedback guidance confidence of the tth target data in the analysis and identification history, L t,p is the index value of the tth target data in the analysis and identification history under the pth feedback guidance confidence index, Q p is the preset weight corresponding to the pth feedback guidance confidence index, M is the total number of feedback guidance confidence indexes, and W is the preset error coefficient;

[0040] The top N target data of the feedback guidance confidence is selected as the feedback guidance basis; wherein N is a preset positive integer;

[0041] Based on the pre-trained big data compensation model, the feedback guidance basis is compensated;

[0042] decide the feedback guidance strategy based on the compensated feedback guidance basis;

[0043] based on the feedback guidance strategy, feedback guidance is provided for the subsequent process of analyzing the real-time dynamic characteristic data of the embedded environment compiler by using the embedded environment compiler real-time dynamic test identification model and automatically identifying potential defects of the embedded system.

[0044] Preferably, the feedback adjustment of the embedded system optimization scheme according to the monitoring situation is performed as follows:

[0045] The monitoring situation is analyzed, and the malady degree curve under each adverse index is updated in real time;

[0046] When the total number of wave peaks exceeding the wave peak threshold in the malady degree curve under each adverse index exceeds the quantity threshold, the optimization importance anchor point is determined based on the data related to the adverse index corresponding to the malady degree curve in the monitoring situation;

[0047] Based on the pre-trained system optimization scheme decision model, the system optimization scheme is decided according to the optimization importance anchor point;

[0048] Based on the system optimization scheme, feedback adjustment is made to the embedded system;

[0049] The multi-dimensional effect data after the feedback adjustment of the embedded system is tracked;

[0050] Based on the effect evaluation template corresponding to the optimization importance anchor point, the effect feature distribution is evaluated according to the multi-dimensional effect data;

[0051] Based on the effect feature distribution, the supplementary anchor point of the optimization importance anchor point is determined;

[0052] Based on the supplementary anchor point, the system optimization supplementary scheme is decided;

[0053] Based on the system optimization supplementary scheme, supplementary feedback adjustment is made to the embedded system.

[0054] Compared with the prior art, the beneficial effects of the present application are:

[0055] The application collects and processes real-time dynamic data of the embedded environment compiler, extracts features related to real-time dynamic test optimization of the embedded environment compiler, determines real-time dynamic feature data of the embedded environment compiler, analyzes the real-time dynamic feature data of the embedded environment compiler by constructing a real-time dynamic test identification model of the embedded environment compiler, automatically identifies potential defects of the embedded system, determines real-time dynamic test identification results of the embedded environment compiler, timely optimizes the embedded system by formulating an embedded system optimization scheme, and monitors the optimized embedded system in real time, adjusts the embedded system optimization scheme according to the monitoring feedback, forms a closed-loop test optimization of the embedded system, dynamically monitors and analyzes the running test state of the embedded environment compiler, can timely find potential defects and optimize, and can improve the running efficiency of the embedded system.

[0056] By constructing a real-time dynamic test identification model of the embedded environment compiler, combining historical analysis identification results and automatically identified potential defects, intelligent feedback guidance of the embedded system is realized. By calculating the feedback guidance confidence and compensating the feedback basis by big data, missing data and related information can be effectively supplemented, so as to optimize the feedback guidance strategy. Specifically, the process includes two preprocessing steps: first, calculate the confidence of the target data based on multiple feedback guidance confidence indicators, and second, compensate the feedback basis by using a big data compensation model to ensure that the compensated data is more comprehensive and accurate. Finally, the compensated feedback basis provides strong support for subsequent defect analysis of the embedded system, can perform in-depth analysis and identification on key problems, and thus improves the identification effect of the embedded compiler real-time dynamic test and the overall performance of the system.

[0057] By monitoring multiple adverse indicators in real time, abnormal fluctuations in the system are accurately identified and performance bottlenecks are located. In addition, through an intelligent decision-making model, a targeted optimization scheme is formulated according to the optimization importance anchor point, ensuring that key problems are quickly solved. In addition, multi-dimensional effect data is continuously tracked to comprehensively evaluate the optimization results, ensuring that each improvement brings real and quantifiable progress. At the same time, effect feature distribution analysis helps to find potential optimization opportunities, and the supplement of anchor points and supplement of schemes further improves the depth and breadth of system performance. Overall, the system stability, reliability and maintenance cost are significantly improved, and efficient and intelligent operation and maintenance of the embedded system are realized. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 The flowchart of the embedded environment compiler real-time dynamic test optimization method of the application. DETAILED DESCRIPTION

[0059] With reference to the drawings and embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part 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 skilled in the art without creative work fall within the scope of the present application.

[0060] In order to solve the problem that the existing embedded environment compiler cannot dynamically monitor and analyze the running test state, cannot discover potential defects in time and optimize, and reduces the running efficiency of the embedded system, please refer to Figure 1 The embodiment provides the following technical solutions.

[0061] An embedded environment compiler real-time dynamic test optimization method comprises the following steps.

[0062] Real-time dynamic data of the embedded environment compiler is collected and processed, and features related to real-time dynamic test optimization of the embedded environment compiler are extracted to determine real-time dynamic feature data of the embedded environment compiler.

[0063] In the embodiment, the real-time dynamic data of the embedded environment compiler is collected, comprising the following steps.

[0064] The environment temperature, humidity, pressure, light intensity, position, motion state and input signals of buttons, switches or touch screens in the running test process of the embedded environment compiler are monitored in real time to collect environment perception data.

[0065] It should be noted that the temperature, humidity, pressure, light intensity and other data are collected by the environment sensor to monitor the running environment of the device, and in the embedded system related to navigation and positioning, the GPS sensor is used to obtain the position and motion state in real time, and the input signals of the buttons, switches or touch screens are used for user interaction or system control.

[0066] The CPU load, memory usage, interrupt response time and power consumption in the running test process of the embedded environment compiler are monitored in real time to collect system running data.

[0067] It should be noted that the CPU load and memory usage reflect the resource occupation of the embedded system, which is used for dynamically adjusting task scheduling and memory management, the interrupt response time records the response time of the embedded system to the interrupt event, which is used for evaluating the real-time performance of the embedded system, and the power consumption is used for optimizing power management.

[0068] The function call frequency and execution time, data flow path and transmission efficiency, and code coverage in the running test process of the embedded environment compiler are monitored in real time to collect application execution data.

[0069] It should be noted that the function call frequency and execution time are used to analyze the code performance bottleneck, the dynamic optimization strategy of the optimization compiler, the data flow path and transmission efficiency are used to optimize the data access mode, and the code coverage is used to analyze the coverage of the test case to ensure that the key code path is fully tested.

[0070] According to the environment perception data, system running data and application program execution data, the real-time dynamic data of the embedded environment compiler is determined.

[0071] It should be noted that the environment perception data, system running data and application program execution data monitored in real time can determine the real-time dynamic data of the embedded environment compiler, facilitate subsequent dynamic monitoring and analysis of the running test state of the embedded environment compiler, so that potential defects can be found in time and optimized.

[0072] In this embodiment, the real-time dynamic data of the embedded environment compiler is processed, and the following operations are performed:

[0073] The real-time dynamic data of the embedded environment compiler is cleaned to remove noise in the real-time dynamic data of the embedded environment compiler that is worthless for real-time dynamic test optimization of the embedded environment compiler, so as to reduce the interference of noise on the real-time dynamic test optimization of the embedded environment compiler;

[0074] The real-time dynamic data of the embedded environment compiler is normalized to convert the real-time dynamic data of the embedded environment compiler into a unified data format, remove the dimensional differences in the real-time dynamic data of the embedded environment compiler, and form standardized real-time dynamic data of the embedded environment compiler;

[0075] The real-time dynamic data of the embedded environment compiler is integrated into a unified data view, and the integrated real-time dynamic data of the embedded environment compiler is subjected to integrity verification. After the data integrity verification is qualified, the integrated real-time dynamic data of the embedded environment compiler is stored safely;

[0076] The real-time dynamic data of the embedded environment compiler is subjected to feature extraction to extract features related to real-time dynamic test optimization of the embedded environment compiler from the real-time dynamic data of the embedded environment compiler, and determine real-time dynamic feature data of the embedded environment compiler.

[0077] It should be noted that by cleaning, normalizing, integrating and extracting features from the real-time dynamic data of the embedded environment compiler, features related to real-time dynamic test optimization of the embedded environment compiler can be extracted, and real-time dynamic feature data of the embedded environment compiler can be determined, so that potential defects of the embedded system can be better identified in the future.

[0078] The embedded environment compiler real-time dynamic test recognition model is constructed, the embedded environment compiler real-time dynamic characteristic data is analyzed, and the potential defects of the embedded system are automatically identified to determine the embedded environment compiler real-time dynamic test recognition result.

[0079] In the embodiment, the embedded environment compiler real-time dynamic characteristic data is analyzed, and the potential defects of the embedded system are automatically identified to perform the following operations:

[0080] According to the embedded environment compiler real-time dynamic test optimization requirement, the embedded environment compiler real-time dynamic test recognition model is constructed;

[0081] The embedded environment compiler real-time dynamic test historical data is collected, and the collected embedded environment compiler real-time dynamic test historical data is divided to determine the training set and the test set;

[0082] Based on the deep learning technology, the training set is used to train the deep learning model, so that the deep learning model learns the embedded environment compiler real-time dynamic test analysis behavior from the training set, and automatically identifies the potential defects of the embedded system to determine the embedded environment compiler real-time dynamic test recognition model based on deep learning;

[0083] The embedded environment compiler real-time dynamic test recognition model based on deep learning is tested based on the test set to evaluate whether the embedded environment compiler real-time dynamic test recognition model based on deep learning can achieve the expected effect of automatically identifying the potential defects of the embedded system;

[0084] When the embedded environment compiler real-time dynamic test recognition model based on deep learning cannot achieve the expected effect of automatically identifying the potential defects of the embedded system, the parameters of the embedded environment compiler real-time dynamic test recognition model based on deep learning are adjusted and iteratively optimized until the embedded environment compiler real-time dynamic test recognition model based on deep learning can achieve the expected effect of automatically identifying the potential defects of the embedded system, so as to determine the optimal embedded environment compiler real-time dynamic test recognition model;

[0085] The optimal embedded environment compiler real-time dynamic test recognition model is deployed, and the embedded environment compiler real-time dynamic test recognition model is deployed in the actual embedded environment compiler real-time dynamic test recognition environment;

[0086] The embedded environment compiler real-time dynamic feature data is input into the embedded environment compiler real-time dynamic test identification model, the embedded environment compiler real-time dynamic feature data is analyzed according to the embedded environment compiler real-time dynamic test identification model, and potential defects of the embedded system are automatically identified, so as to determine the embedded environment compiler real-time dynamic test identification result, potential defects can be found in time and optimized, and the running efficiency of the embedded system can be improved.

[0087] The embedded system optimization scheme is formulated to optimize the embedded system in time, the optimized embedded system is monitored in real time, the embedded system optimization scheme is adjusted according to the monitoring feedback, and a closed-loop test optimization of the embedded system is formed.

[0088] In this embodiment, the embedded system optimization scheme is formulated to optimize the embedded system in time, and the following operations are performed:

[0089] According to the embedded environment compiler real-time dynamic test identification result, the embedded system optimization scheme is formulated, and the embedded system is optimized in time according to the embedded system optimization scheme;

[0090] Among them, the parameters of the embedded environment compiler are adjusted, instructions suitable for hardware characteristics are generated, and complex computing tasks are unloaded using hardware accelerators to reduce the load of the main processor;

[0091] The embedded system is divided into multiple modules, each module is optimized independently, the storage mode of data in the memory is adjusted to reduce access delay;

[0092] The appropriate compilation optimization level is selected, the optimization options are adjusted according to the application scenario, such as enabling or disabling function inlining, deleting unused code, updating the environment compiler cache, and compiling only the changed part to improve the compilation efficiency, and dynamically adjusting the environment compiler cache, task scheduling and resource allocation to ensure efficient operation of the system.

[0093] Among them, the environment compiler cache, task scheduling and resource allocation after dynamic adjustment are monitored in real time, the embedded system optimization scheme is adjusted according to the real-time monitoring feedback, and a closed-loop test optimization of the embedded system is formed, which can improve the running efficiency of the embedded system.

[0094] In this embodiment, the embedded environment compiler real-time dynamic test identification model is constructed, the embedded environment compiler real-time dynamic feature data is analyzed, and potential defects of the embedded system are automatically identified, and the following operations are also performed:

[0095] The analysis and identification history of the embedded environment compiler real-time dynamic test identification model is acquired, and the analysis and identification history at least includes: analysis results of the embedded environment compiler real-time dynamic characteristic data in the past, and results of automatically identifying potential defects of the embedded system in the past; the analysis results of the embedded environment compiler real-time dynamic characteristic data in the past at least include: memory management optimization analysis results, performance bottleneck analysis results, etc.; and the results of automatically identifying potential defects of the embedded system in the past at least include: memory leakage and overflow, concurrent errors, timing problems, etc.

[0096] The feedback guidance confidence degree of each target data in the analysis and identification history is calculated by the following formula:

[0097]

[0098] Wherein, Z t is the feedback guidance confidence degree of the tth target data in the analysis and identification history, L t,p is the index value of the tth target data in the analysis and identification history under the pth feedback guidance confidence index, Q p is the preset weight corresponding to the pth feedback guidance confidence index, M is the total number of feedback guidance confidence indexes, and W is a preset error coefficient.

[0099] The target data with the top N feedback guidance confidence degrees is selected as the feedback guidance basis; wherein, N is a preset positive integer.

[0100] The feedback guidance basis is compensated based on the pre-trained big data compensation model; wherein, the analysis and identification history can reflect which problems can be analyzed and identified in depth by the embedded environment compiler real-time dynamic test identification model in the subsequent process, and the feedback guidance is realized, but due to the complexity of the analysis and identification history, not all of them can be subjected to the feedback guidance process, and the embodiment of the application is subjected to two-step preprocessing: first, the feedback guidance confidence of each target data in the analysis and identification history is calculated; the feedback guidance confidence represents the confidence degree of the target data used for feedback guidance, a plurality of feedback guidance confidence indicators (the indicators at least include: the positive effect degree of the feedback guidance given by the same type of target data in the history, the reuse rate of the feedback guidance given by the same type of target data in the history, etc.) are pre-set, the index values of the target data under different feedback guidance confidence indicators are determined (the index values can be directly determined based on the indicators), and the weights of the influence degree of the index values under different feedback guidance confidence indicators on the final calculation of the feedback guidance confidence are pre-set, the index values of the target data under different feedback guidance confidence indicators are weighted and calculated, and then the average value is taken, in addition, an error coefficient is pre-set, the average value is error corrected, so that the final calculation result is a value representing the confidence degree of the target data used for feedback guidance; second, the feedback guidance basis is compensated; when the big data compensation model is pre-trained, a large number of feedback guidance bases labeled with compensation rules (labels of rules such as related data supplement and missing data supplement) are used as training samples for machine learning training, and the model trained to convergence is used as the big data compensation model, and the big data compensation model can perform data compensation measures such as related data supplement and missing data supplement on the feedback guidance basis, so that the feedback guidance basis after compensation is more comprehensive.

[0101] Based on the compensated feedback guidance basis, the feedback guidance strategy is decided;

[0102] Based on the feedback guidance strategy, the subsequent process of analyzing and automatically identifying potential defects of the embedded system by the embedded environment compiler real-time dynamic test identification model on the real-time dynamic characteristic data of the embedded environment compiler is guided. Wherein, the feedback guidance basis after compensation can be used to decide the feedback guidance strategy based on which key problems in the feedback guidance basis after compensation are analyzed, and the model in the subsequent process is guided to analyze and identify these key problems in depth.

[0103] The embodiment of the application realizes intelligent feedback guidance of the embedded system by constructing an embedded environment compiler real-time dynamic test recognition model, combining historical analysis recognition results with potential defects of automatic recognition. By calculating the feedback guidance confidence and compensating the feedback basis by big data, missing data and related information can be effectively supplemented, so that the feedback guidance strategy is optimized. Specifically, the process includes two preprocessing steps: one is to calculate the confidence of the target data based on multiple feedback guidance confidence indicators, and the other is to compensate the feedback basis by using a big data compensation model to ensure that the compensated data is more comprehensive and accurate. Finally, the compensated feedback basis provides strong support for subsequent defect analysis of the embedded system, which can perform in-depth analysis and recognition on key problems, thereby improving the recognition effect of the embedded compiler real-time dynamic test and the overall performance of the system.

[0104] In the embodiment, the feedback adjustment of the embedded system optimization scheme according to the monitoring situation is performed as follows:

[0105] The monitoring situation is analyzed, and a malady degree curve under a plurality of adverse indicators is updated in real time. The adverse indicators at least include delay, performance bottleneck, memory consumption, etc. The malady degree curve is drawn based on the real-time change value of the monitoring situation under different adverse indicators. The horizontal axis of the malady degree curve is time, and the vertical axis is the real-time change value of the monitoring situation under the adverse indicators. The malady degree curve is updated in real time to reflect the current change of the malady of the system;

[0106] When the total number of wave peaks exceeding the wave peak threshold value in the malady degree curve under each adverse indicator exceeds the number threshold value, the optimization importance anchor point is determined based on the related data of the adverse indicator corresponding to the malady degree curve in the monitoring situation. The wave peak threshold value is a threshold value representing a larger wave peak value, and the number threshold value is a threshold value representing a larger number of wave peaks. In the malady degree curve, the wave peak represents an abnormal or malady peak value. If a plurality of wave peaks exceeding the preset wave peak threshold value appear in the curve, and the total number exceeds the number threshold value, it indicates that the plurality of adverse indicators of the system have abnormalities. The related data of these wave peaks is used to determine the optimization importance anchor point. These anchor points refer to key problems or key performance indicators, which need to be optimized to improve the system performance;

[0107] Based on the pre-trained system optimization scheme decision model, the system optimization scheme is decided according to the optimization importance anchor point. The system optimization scheme decision model has been trained by a large amount of historical data (a large number of optimization importance anchor points labeled with different system optimization scheme labels), and can automatically generate optimization schemes according to different optimization targets (i.e. optimization importance anchor points). Through the analysis of these anchor points, the decision model determines which optimization scheme is most suitable for solving the current system problem;

[0108] Based on the system optimization scheme, feedback adjustment is performed on the embedded system; wherein, after determining the optimization scheme, the system will perform real-time adjustment according to the scheme, which includes modifying system configuration, reallocating resources, optimizing memory management or adjusting processing logic, etc., the purpose is to improve the performance and stability of the system;

[0109] Tracking the multi-dimensional effect data of the embedded system after feedback adjustment; wherein, after adjustment, the system will collect and track multi-dimensional effect data, which includes at least system response time, processing speed, resource utilization and other indicators, which helps to evaluate the effect of system optimization;

[0110] Based on the effect evaluation template corresponding to the optimization importance anchor point, the effect feature distribution is evaluated according to the multi-dimensional effect data; wherein, the optimization importance anchor point corresponds to the preset effect evaluation template, and the effect evaluation template can be used to quantitatively analyze the optimization effect according to the collected multi-dimensional effect data, and the evaluation result of quantitative analysis forms the effect feature distribution, which reflects the performance and distribution of each performance index after optimization;

[0111] Based on the effect feature distribution, the supplementary anchor point of the optimization importance anchor point is determined; wherein, through effect feature distribution analysis, it can be found that some optimization measures still have problems or do not achieve the expected effect, at this time, according to the distribution of effect features, the system will determine the supplementary anchor point, that is, the supplementary optimization target;

[0112] Based on the supplementary anchor point, the system optimization supplementary scheme is decided; wherein, based on the supplementary anchor point, the system optimization scheme decision model will generate a system optimization supplementary scheme according to the supplementary anchor point;

[0113] Based on the system optimization supplementary scheme, supplementary feedback adjustment is performed on the embedded system. Wherein, finally, based on the system optimization supplementary scheme, supplementary feedback adjustment is performed on the embedded system.

[0114] The embodiment of the application accurately identifies abnormal fluctuations and locates performance bottlenecks in the system by monitoring a plurality of adverse indicators in real time. It also formulates targeted optimization schemes according to optimization importance anchor points through an intelligent decision model, ensuring that key problems are quickly solved. In addition, multi-dimensional effect data is continuously tracked to comprehensively evaluate optimization results, ensuring that each improvement brings real and quantifiable progress. Meanwhile, effect feature distribution analysis helps to discover potential optimization opportunities, and supplementary anchor points and supplementary schemes further improve the depth and breadth of system performance. Overall, the system stability, reliability and maintenance cost are significantly improved, realizing efficient and intelligent operation and maintenance of embedded systems.

[0115] It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting; it is not intended to exclude myriad other embodiments of the present application that other inventors can develop based on the same general inventive concepts embodied by the described embodiments. That is, although the present application is described in terms of particular embodiments and illustrative figures, it should be apparent that the scope of the present application is not limited to these specific embodiments.

[0116] While the embodiments of the application have been shown and described herein, it will be understood by those skilled in the art that many changes, modifications, substitutions and alterations to these embodiments can be made without departing from the principles and spirits of the application, and it is intended that the scope of the application be limited solely by the scope of the appended claims and the equivalents thereof.

Claims

1. A real-time dynamic testing and optimization method for an embedded environment compiler, characterized in that, include: Collect and process real-time dynamic data of the embedded environment compiler, and extract features related to real-time dynamic testing and optimization of the embedded environment compiler to determine the real-time dynamic feature data of the embedded environment compiler. A real-time dynamic test identification model for embedded environment compilers is constructed, the real-time dynamic feature data of embedded environment compilers is analyzed, and potential defects in embedded systems are automatically identified to determine the real-time dynamic test identification results of embedded environment compilers. Among them, based on the feedback guidance strategy, the embedded environment compiler real-time dynamic test identification model is used to analyze the real-time dynamic feature data of the embedded environment compiler and automatically identify potential defects in the embedded system to provide feedback guidance in the subsequent process. Develop an embedded system optimization plan to optimize the embedded system in a timely manner, monitor the optimized embedded system in real time, and adjust the embedded system optimization plan based on the monitoring feedback to form a closed-loop test optimization of the embedded system.

2. The embedded environment compiler real-time dynamic testing and optimization method as described in claim 1, characterized in that, The embedded system optimization scheme is adjusted based on monitoring feedback, and the following operations are performed: Analyze the monitoring data and update the defect rate curves under multiple defect indicators in real time; among them, defect indicators include at least: latency, performance bottleneck, and memory consumption; When the total number of peaks exceeding the peak threshold of the defect degree curve under each defect indicator exceeds the quantity threshold, the optimization importance anchor point is determined based on the relevant data of the defect indicator corresponding to the defect degree curve in the monitoring situation. Based on a pre-trained system optimization decision model, the system optimization scheme is determined according to the optimization importance anchor point; Based on the system optimization scheme, feedback adjustments are made to the embedded system; Track multi-dimensional effect data after feedback adjustments to the embedded system; Based on the effect evaluation template corresponding to the optimized importance anchor point, the effect feature distribution is evaluated according to the multi-dimensional effect data. Based on the distribution of effect characteristics, supplementary anchor points for optimizing importance anchor points are determined; Based on the supplementary anchor points, the decision-making system optimizes the supplementary plan; Based on the system optimization and supplementary scheme, supplementary feedback adjustments are made to the embedded system.

3. The embedded environment compiler real-time dynamic testing and optimization method as described in claim 2, characterized in that, Construct a real-time dynamic test identification model for embedded environment compilers, analyze the real-time dynamic feature data of embedded environment compilers, and automatically identify potential defects in embedded systems, performing the following operations: Based on the optimization requirements of real-time dynamic testing of embedded environment compilers, a real-time dynamic testing identification model for embedded environment compilers is constructed, and the model is deployed in the actual real-time dynamic testing identification environment for embedded environment compilers. The real-time dynamic feature data of the embedded environment compiler is input into the real-time dynamic test identification model of the embedded environment compiler. The real-time dynamic feature data of the embedded environment compiler is analyzed according to the real-time dynamic test identification model of the embedded environment compiler, and potential defects of the embedded system are automatically identified, thereby determining the real-time dynamic test identification result of the embedded environment compiler.

4. The embedded environment compiler real-time dynamic testing and optimization method as described in claim 3, characterized in that, A real-time dynamic test identification model for embedded environment compilers is constructed, which analyzes the real-time dynamic feature data of embedded environment compilers, automatically identifies potential defects in embedded systems, and performs the following operations: Obtain the analysis and identification history of the real-time dynamic test identification model for the embedded environment compiler; wherein, the analysis and identification history includes at least: the analysis results of the real-time dynamic feature data of the embedded environment compiler in the past, and the results of automatically identifying potential defects in the embedded system in the past; wherein, the analysis results of the real-time dynamic feature data of the embedded environment compiler in the past include at least: memory management optimization analysis results and performance bottleneck analysis results; the results of automatically identifying potential defects in the embedded system in the past include at least: memory leaks and overflows, concurrency errors, and timing problems; The confidence level of feedback guidance for identifying each target data point in the historical data is calculated using the following formula: Among them, Z t To analyze and identify the confidence level of the feedback guidance for the t-th target data in history, where the confidence level represents the degree of confidence in using the target data as feedback guidance, L t,p To analyze and identify the index value of the t-th target data in history under the p-th feedback guidance confidence index, Q p The preset weight is the weight corresponding to the p-th feedback guidance confidence index, M is the total number of feedback guidance confidence indices, and W is the preset error coefficient; the feedback guidance confidence indices include at least: the positive effect of historical feedback guidance provided by similar target data, and the reuse rate of historical feedback guidance provided by similar target data; The top N target data points with the highest confidence levels for feedback guidance are selected as the basis for feedback guidance; where N is a preset positive integer. Based on a pre-trained big data compensation model, the feedback guidance is compensated. Based on the feedback guidance after compensation, decision-making feedback guidance strategies; Based on the feedback guidance strategy, feedback guidance is provided for the subsequent process of analyzing the real-time dynamic feature data of the embedded environment compiler and automatically identifying potential defects in the embedded system using the real-time dynamic test identification model of the embedded environment compiler.

5. The embedded environment compiler real-time dynamic testing and optimization method as described in claim 4, characterized in that, Develop an embedded system optimization plan to optimize the embedded system in a timely manner, and perform the following operations: Based on the real-time dynamic test results of the embedded environment compiler, an embedded system optimization plan is formulated, and the embedded system is optimized in a timely manner according to the embedded system optimization plan; Among these measures, the embedded environment compiler parameters are adjusted to generate instructions suitable for hardware characteristics, and hardware accelerators are used to offload complex computing tasks and reduce the load on the main processor. The embedded system is divided into multiple modules, and each module is optimized independently. By adjusting the way data is stored in memory, access latency is reduced. Choose the appropriate compilation optimization level and adjust optimization options according to the application scenario, such as enabling or disabling function inlining, deleting unused code, updating the environment compiler cache, and compiling only the changed parts to improve compilation efficiency. Dynamically adjust the environment compiler cache, task scheduling, and resource allocation to ensure efficient system operation.

6. The embedded environment compiler real-time dynamic testing and optimization method as described in claim 5, characterized in that, Collect real-time dynamic data from the embedded environment compiler, including: Real-time dynamic monitoring of ambient temperature, humidity, pressure, light intensity, position, motion status, and input signals from buttons, switches, or touchscreens during the embedded environment compiler runtime test is performed to collect environmental perception data. Real-time dynamic monitoring of CPU load, memory usage, interrupt response time, and power consumption during the embedded environment compiler runtime test, and collection of system operation data; Real-time dynamic monitoring of function call frequency and execution time, data flow path and transmission efficiency, and code coverage during the embedded environment compiler runtime test, and collection of application execution data; Based on environmental perception data, system operation data, and application execution data, the real-time dynamic data of the embedded environment compiler is determined.

7. The embedded environment compiler real-time dynamic testing and optimization method as described in claim 6, characterized in that, To process real-time dynamic data from the embedded environment compiler, perform the following operations: Clean the real-time dynamic data of the embedded environment compiler to remove noise that is not valuable for real-time dynamic testing and optimization of the embedded environment compiler, thereby reducing the interference of noise on real-time dynamic testing and optimization of the embedded environment compiler. The real-time dynamic data of the embedded environment compiler is normalized to convert it into a unified data format, remove the dimensional differences in the real-time dynamic data of the embedded environment compiler, and form standardized real-time dynamic data of the embedded environment compiler.

8. The embedded environment compiler real-time dynamic testing and optimization method as described in claim 7, characterized in that, The embedded environment compiler processes real-time dynamic data and also performs the following operations: The real-time dynamic data of the embedded environment compiler is integrated, and the real-time dynamic data of the embedded environment compiler from different sources is integrated into a unified data view. The integrity of the integrated real-time dynamic data of the embedded environment compiler is verified. After the data integrity verification is qualified, the integrated real-time dynamic data of the embedded environment compiler is securely stored. Feature extraction is performed on the real-time dynamic data of the embedded environment compiler. Features related to the real-time dynamic testing and optimization of the embedded environment compiler are extracted from the real-time dynamic data of the embedded environment compiler, and the real-time dynamic feature data of the embedded environment compiler is determined.

9. The embedded environment compiler real-time dynamic testing and optimization method as described in claim 8, characterized in that, To build a real-time dynamic test recognition model for an embedded environment compiler, perform the following operations: Collect historical data of real-time dynamic testing of embedded environment compilers, and divide the collected historical data of real-time dynamic testing of embedded environment compilers to determine the training set and test set; Based on deep learning technology, a training set is used to train the deep learning model, enabling the deep learning model to autonomously learn the real-time dynamic test analysis behavior of the embedded environment compiler from the training set, and automatically identify potential defects in the embedded system, thus determining the deep learning-based real-time dynamic test identification model for the embedded environment compiler. The performance of the deep learning-based real-time dynamic test identification model for embedded environment compilers is tested based on the test set to evaluate whether the deep learning-based real-time dynamic test identification model for embedded environment compilers can achieve the expected effect of automatically identifying potential defects in embedded systems. When the deep learning-based real-time dynamic test identification model for embedded environment compilers fails to achieve the expected effect of automatically identifying potential defects in embedded systems, the parameters of the deep learning-based real-time dynamic test identification model for embedded environment compilers are adjusted and iteratively optimized until the deep learning-based real-time dynamic test identification model for embedded environment compilers can achieve the expected effect of automatically identifying potential defects in embedded systems, thereby determining the optimal real-time dynamic test identification model for embedded environment compilers.

10. The embedded environment compiler real-time dynamic testing and optimization method as described in claim 9, characterized in that, The system monitors the dynamically adjusted environment compiler cache, task scheduling, and resource allocation in real time, and adjusts the embedded system optimization scheme based on the feedback from the real-time monitoring, forming a closed-loop test optimization of the embedded system.

Images