Optimization scheme generation method of basic input and output system and related equipment

Abstract

The invention relates to the technical field of computers, and provides an optimization scheme generation method of a basic input and output system and related equipment, and the optimization scheme generation method of the basic input and output system comprises the following steps: obtaining system operation data, and preprocessing the system operation data; extracting time domain features and frequency domain features of the preprocessed system operation data; inputting the time domain features and the frequency domain features into the trained optimization model to obtain an optimization scheme for the basic input and output system; wherein the optimization model is obtained based on sample system operation data and corresponding optimization label training, and the optimization scheme comprises optimization of processor frequency, optimization of memory time sequence, optimization of fan speed and optimization of processor voltage. According to the method, the system operation data is scientifically analyzed by utilizing the trained optimization model, the optimization scheme is generated, static configuration or manual adjustment by an administrator according to experience is not needed, and the optimization scheme adaptive to system change can be quickly generated.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method for generating an optimization solution for a basic input and output system and related equipment. Background Art

[0002] The Basic Input Output System (BIOS) is a set of programs embedded in the ROM (Read-Only Memory) chip on the computer motherboard, which is responsible for initializing hardware components, performing hardware self-tests, and starting the operating system. The BIOS settings have a significant impact on the startup speed, stability, and overall performance of the computer system. Traditional BIOS settings are usually static and rely on users or technicians to manually configure them when the system starts. This static setting method is insufficient in the face of different application scenarios and dynamic load changes, and cannot be flexibly adjusted according to actual usage, which may result in suboptimal system performance.

[0003] In modern computing environments, especially in data centers and servers, efficient management and optimization of computing resources have become particularly important. The load and application requirements of server systems are ever-changing, and the requirements for system performance are constantly increasing. However, most existing BIOS optimization methods rely on preset static configurations or manual adjustments by administrators based on experience, which cannot adapt to system changes, are time-consuming and prone to errors. Especially in large-scale server deployment and management, manual BIOS optimization is not only inefficient, but also difficult to meet the requirements of rapid response to business needs. Summary of the invention

[0004] The present invention provides a method for generating an optimization scheme for a basic input / output system and related equipment, which are used to solve the defects of being unable to adapt to system changes and being inefficient caused by statically configuring the BIOS or manually adjusting the BIOS based on experience in the prior art, and to achieve rapid generation of an optimization scheme that adapts to system changes.

[0005] The present invention provides a method for generating an optimization scheme for a basic input / output system, comprising: Acquiring system operation data and preprocessing the system operation data; Extracting time domain features and frequency domain features of the preprocessed system operation data; The time domain features and the frequency domain features are input into the trained optimization model to obtain an optimization scheme for the basic input and output system; wherein the optimization model is trained based on sample system operation data and corresponding optimization labels, as well as sample time domain features and sample frequency domain features of the sample system operation data, the optimization labels are obtained based on sample performance data corresponding to the sample system operation data, and the optimization scheme includes optimizing processor frequency, optimizing memory timing, optimizing fan speed, and optimizing processor voltage.

[0006] According to a method for generating an optimization solution for a basic input-output system provided by the present invention, the training process of the optimization model includes: Collect several sample system operation data and sample performance data; Storing the sample system operation data and the sample performance data in a preset file in time sequence; Preprocessing the sample system operation data and the sample performance data; Based on the storage order of the sample system operation data in the preset file, grouping the sample system operation data to obtain a plurality of sample groups; Extracting sample time domain features and sample frequency domain features of the sample system operation data in each of the sample groups; Determining, based on the sample performance data, an optimization label for the sample system operation data corresponding to the sample performance data; Generate a training sample set based on the sample system operation data, the sample performance data, the sample time domain features, the sample frequency domain features and the optimization label; wherein the training sample set includes a plurality of training samples; The training samples are input into a preset basic optimization model for iterative training to obtain the trained optimization model.

[0007] According to a method for generating an optimization solution for a basic input-output system provided by the present invention, the preset basic optimization model includes a first LSTM layer, a second LSTM layer and a fully connected layer connected in sequence.

[0008] According to a method for generating an optimization solution for a basic input-output system provided by the present invention, the preprocessing of the system operation data comprises: Calculating the median of the system operation data, and filling missing values ​​in the system operation data by using the median; Processing outliers in the system operation data by using the interquartile range method; The system operation data is normalized.

[0009] According to a method for generating an optimization solution for a basic input-output system provided by the present invention, the step of extracting the time domain features and frequency domain features of the pre-processed system operation data comprises: Calculating the time domain mean and the time domain standard deviation of the system operation data to obtain the time domain characteristics; The frequency domain feature is obtained by calculating the frequency domain mean and the frequency domain standard deviation of the system operation data through fast Fourier transform.

[0010] According to a method for generating an optimization solution for a basic input / output system provided by the present invention, the system operation data includes processor usage, memory occupancy, disk I / O rate and temperature.

[0011] According to a method for generating an optimization solution for a basic input-output system provided by the present invention, after inputting the time domain features and the frequency domain features into a trained optimization model to obtain an optimization solution for the basic input-output system, the method further includes: Determining effect parameters according to the optimization scheme; The optimization plan and the expected effect are sent to a preset display page for display.

[0012] The present invention also provides a basic input and output system optimization solution generation device, comprising: An acquisition module, configured to acquire system operation data and pre-process the system operation data; A first extraction module is configured to extract time domain features and frequency domain features of the preprocessed system operation data; An input module is configured to input the time domain features and the frequency domain features into a trained optimization model to obtain an optimization solution for a basic input-output system; wherein the optimization model is trained based on sample system operation data and corresponding optimization labels, as well as sample time domain features and sample frequency domain features of the sample system operation data, the optimization labels are obtained based on sample performance data corresponding to the sample system operation data, and the optimization solution includes optimizing processor frequency, optimizing memory timing, optimizing fan speed, and optimizing processor voltage.

[0013] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for generating an optimization solution for a basic input / output system as described above is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for generating an optimization solution for a basic input-output system as described in any one of the above is implemented.

[0015] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any one of the above-mentioned methods for generating an optimization solution for a basic input-output system.

[0016] The optimization scheme generation method and related equipment of the basic input and output system provided by the present invention pre-process the system operation data, and extract the corresponding frequency domain features and time domain features, and then input the frequency domain features and time domain features into the trained optimization model for processing to obtain the optimization scheme. The present invention uses the trained optimization model to scientifically analyze the system operation data, generates an optimization scheme based on the system operation data, and provides a reliable optimization basis without static configuration or manual adjustment by the administrator based on experience. This not only adapts to changes in the system, ensures the rationality and effectiveness of BIOS settings, saves adjustment time, but also avoids errors caused by human experience, improves the performance of the system, and can quickly respond to business needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 It is a flow chart of the method for generating an optimization scheme of a basic input-output system provided by the present invention.

[0019] Figure 2 It is a flow chart of training an optimization model in the optimization scheme generation method of a basic input-output system provided by the present invention.

[0020] Figure 3 It is a structural schematic diagram of a device for generating an optimization solution for a basic input-output system provided by the present invention.

[0021] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] Combine the following Figure 1-Figure 4 The optimization scheme generation method and related equipment of the basic input and output system of the present invention are described.

[0024] Figure 1 FIG. 1 is a flow chart of a method for generating an optimization solution for a basic input-output system according to an exemplary embodiment. Figure 1 As shown, in an exemplary embodiment, the method for generating an optimization solution for a basic input-output system includes steps 110 to 130, which are described in detail as follows.

[0025] Step 110, obtaining system operation data, and preprocessing the system operation data.

[0026] In the embodiment of the present invention, the system operation data is the hardware and software data when the system is running, and the system operation data includes processor usage, memory occupancy, disk I / O rate and temperature. The processor includes a CPU. The system operation data is collected at a certain collection period, such as the collection period is set to 1 minute, and the collected system operation data is stored in the high-performance time series database InfluxDB, and the latest system operation data is obtained from the real-time monitoring system. In the embodiment of the present invention, an example of the latest system operation data obtained is processor usage: 55%; memory occupancy: 35%; disk I / O rate: 95MB / s; temperature: 62°C.

[0027] The acquired system operation data is preprocessed, including missing value filling, outlier processing and standardization. The preprocessed system operation data examples are processor usage: 0.55; memory usage: 0.35; disk I / O: 0.95; temperature: 0.62.

[0028] Step 120: extracting the time domain features and frequency domain features of the preprocessed system operation data.

[0029] In the embodiment of the present invention, the system operation data is time series type data, and the time domain characteristics and frequency domain characteristics of the system operation data are extracted. The time domain characteristics characterize the changes of the system operation data on the time axis. The frequency domain characteristics characterize the components and energy distribution of the system operation data in the frequency domain.

[0030] Step 130, inputting the time domain features and the frequency domain features into the trained optimization model to obtain an optimization scheme for the basic input and output system; wherein the optimization model is trained based on sample system operation data and corresponding optimization labels, as well as sample time domain features and sample frequency domain features of the sample system operation data, the optimization labels are obtained based on sample performance data corresponding to the sample system operation data, and the optimization scheme includes optimizing processor frequency, optimizing memory timing, optimizing fan speed, and optimizing processor voltage.

[0031] In the embodiment of the present invention, an optimization model is pre-trained based on deep learning, and the obtained time domain features and frequency domain features are input into the optimization model for processing to obtain an optimization solution for the setting items of the BIOS.

[0032] In an embodiment of the present invention, the system operation data is preprocessed, and the corresponding frequency domain features and time domain features are extracted, and then the frequency domain features and time domain features are input into the trained optimization model for processing to obtain an optimization solution. The present invention uses the trained optimization model to scientifically analyze the system operation data, generates an optimization solution based on the system operation data, and provides a reliable optimization basis without static configuration or manual adjustment by the administrator based on experience. This not only adapts to changes in the system, ensures the rationality and effectiveness of BIOS settings, saves adjustment time, but also avoids errors caused by human experience, improves the performance of the system, and can quickly respond to business needs.

[0033] In the implementation of the present invention, by collecting and analyzing system operation data, using a deep learning model to predict the impact of different BIOS settings on system performance, and regularly generating optimization solutions, the BIOS settings can be adjusted according to these optimization solutions, thereby achieving dynamic optimization of system performance. The present invention provides a personalized and dynamic BIOS optimization solution in a data-driven manner. The present invention can adopt a method of regularly analyzing system operation data to continuously provide optimization solutions, so that the system can continuously optimize and improve itself and maintain a long-term high-performance operating state. This continuous optimization mechanism not only improves the long-term stability of the system, but also extends the service life of the hardware.

[0034] In an exemplary embodiment of the present invention, the training process of the optimization model includes the following steps 210 to 280, which are described in detail as follows.

[0035] Step 210, collect a number of sample system operation data and sample performance data.

[0036] In an embodiment of the present invention, hardware and software data during system operation are collected as sample system operation data for subsequent data analysis and optimization solution generation. First, the hardware monitoring tool lm_sensors is installed and configured, and information such as processor usage, memory occupancy, disk I / O rate, temperature, etc. is obtained through the hardware monitoring tool. Next, the operating system monitoring tools top, ps, and vmstat are used to monitor the system load and the performance data of the application as sample performance data. The sample performance data corresponds to the sample operation data, that is, the collection time of the two is the same. The collected sample performance data and sample system operation data are stored in the high-performance time series database InfluxDB. The high-performance time series database InfluxDB is a distributed time series, time, and indicator database that can quickly and easily store time series type system operation data.

[0037] Step 220: Store the sample system operation data and the sample performance data in a preset file in time sequence.

[0038] In an embodiment of the present invention, by installing and configuring the telegraf agent, the data collected by the hardware monitoring tool lm_sensors and the operating system monitoring tool are sent regularly (once per minute) to the high-performance time series database InfluxDB. The configured telegraf agent specifically includes setting the processor, mem, disk, system and sensors plug-ins of telegraf to ensure comprehensive and accurate monitoring of the system operation status. Finally, the integrity and accuracy of the data collection are verified through the CLI (Command-Line Interface) tool of the high-performance time series database InfluxDB. The sample system operation data stored in the high-performance time series database InfluxDB is stored in the table in sequence according to the collected time sequence.

[0039] Step 230: pre-process the sample system operation data and the sample performance data.

[0040] In the embodiment of the present invention, the sample system operation data and sample performance data are preprocessed including missing value filling, outlier processing and standardization. The collected sample system operation data is cleaned and standardized to ensure the accuracy and consistency of the data for subsequent feature extraction and model training. Specifically, the Pandas library function can be used to complete these three steps respectively.

[0041] Step 240: grouping the sample system operation data based on the storage order of the sample system operation data in the preset file to obtain a plurality of sample groups.

[0042] In the embodiment of the present invention, based on the storage order of the sample system operation data in the preset file, the sample system operation data is grouped to obtain multiple sample groups. Specifically, one column in the table can be used as a sample group.

[0043] Step 250: extracting sample time domain features and sample frequency domain features of the sample system operation data in each of the sample groups.

[0044] In the embodiment of the present invention, the sample time domain features and sample frequency domain features of the sample system operation data in each sample group are extracted, and the sample system operation data in the same sample group have the same sample time domain features and sample frequency domain features. The sample time domain features include the time domain mean and the time domain standard deviation, and the sample frequency domain features include the frequency domain mean (FFT mean) and the frequency domain standard deviation (FFT standard deviation).

[0045] Specifically, the mean and standard deviation of each column of sample system operation data are calculated. Then the frequency domain features are extracted using Fast Fourier Transformation (FFT), the frequency domain mean and standard deviation of each column are calculated, and finally the extracted feature data is saved in a new CSV (Comma Separated Values) file for subsequent model construction and training.

[0046] Step 260: Determine an optimization tag of the sample system operation data corresponding to the sample performance data based on the sample performance data.

[0047] In an embodiment of the present invention, an optimization tag of the sample system operation data corresponding to the sample performance data is determined based on the sample performance data. The optimization tag is the value of the BIOS setting item corresponding to each set of characteristic data when the performance is excellent. The optimization tag includes optimizing processor frequency, optimizing memory timing, optimizing fan speed, and optimizing processor voltage. An example of the optimization tag in an embodiment of the present invention is as follows: Optimize processor frequency: 3.5GHz, 3.4GHz, 3.6GHz, ...; Optimized memory timings: 16-18-18-36, 16-18-18-36, 16-19-19-38, ...; Optimize fan speed: 1200RPM, 1150RPM, 1300RPM, ...; Optimize processor voltage: 1.25V, 1.24V, 1.26V, ...

[0048] Step 270: Generate a training sample set based on the sample system operation data, the sample performance data, the sample time domain features, the sample frequency domain features and the optimization label; wherein the training sample set includes a plurality of training samples.

[0049] In an embodiment of the present invention, a data set is generated based on sample system operation data, sample performance data, sample time domain characteristics, sample frequency domain characteristics and optimized labels, and then the data set is divided according to a ratio of 80% training sample set and 20% verification sample set. When dividing, a random division method is adopted to ensure uniform distribution of data, and finally complete the construction of the training sample set.

[0050] An example of a sample of the data set is shown in Listing 1 below: Table 1

[0051] Step 280: input the training samples into a preset basic optimization model for iterative training to obtain the trained optimization model.

[0052] In an embodiment of the present invention, the training samples in the training sample set are input into a preset basic optimization model for iterative training to obtain a trained optimization model, and then the trained optimization model is verified by the sample data in the verification sample set. When the verification fails, the model is retrained, and when the verification passes, the optimization model can be obtained.

[0053] In the embodiment of the present invention, by regularly collecting and analyzing system operation data and using a deep learning model to generate an optimization solution, the BIOS settings can dynamically adapt to the actual operating status of the system. Compared with the traditional static setting method, the present invention can flexibly adjust the BIOS settings according to changes in system load and application requirements, significantly improving system performance and stability.

[0054] The deep learning model used in the present invention can accurately predict the impact of different BIOS settings on system performance by analyzing a variety of system operation data, such as processor usage, memory occupancy, disk I / O rate and temperature. By integrating multiple data sources in this way, the accuracy and reliability of model prediction are improved, ensuring the scientificity and effectiveness of the optimization solution.

[0055] In an exemplary embodiment of the present invention, the preset basic optimization model includes a first LSTM layer, a second LSTM layer and a fully connected layer connected in sequence.

[0056] In the embodiment of the present invention, since the system operation data is time series type data, the embodiment of the present invention selects LSTM (Long Short-Term Memory) as the preset basic optimization model. LSTM is good at processing time series data and can remember long-term dependencies. When constructing the preset basic optimization model, first create a Keras Sequential model, add the first LSTM layer and the second LSTM layer in sequence, and finally add the fully connected layer as the structure of the entire model. In other embodiments of the present invention, other models that are good at processing time series data can also be selected as the preset basic optimization model.

[0057] In the embodiment of the present invention, when training the preset basic optimization model, the learning rate is set to 0.001, the batch size is set to 32, the number of training rounds is set to 50, the Adam optimizer is used, and the loss function is set to mean square error. Then the aforementioned training sample set is used for model training, and finally the trained model is saved as an HDF (Hierarchical Data Format, h5) file for subsequent use of the optimization model.

[0058] In an exemplary embodiment of the present invention, the preprocessing of the system operation data includes the following steps, which are described in detail as follows.

[0059] The median of the system operation data is calculated, and missing values ​​in the system operation data are filled by the median.

[0060] Outliers in the system operation data are processed by the interquartile range method.

[0061] The system operation data is normalized.

[0062] In the embodiment of the present invention, the collected system operation data is cleaned and standardized to ensure the accuracy and consistency of the system operation data for subsequent feature extraction. The preprocessing is specifically divided into three steps, and the Pandas library function is used to complete the processing of these three steps respectively.

[0063] Specifically, for the processing of missing values: first export the system operation data from the high-performance time series database InfluxDB, save it as a CSV file, calculate the median of the system operation data, and fill the missing values ​​in the system operation data with the median to ensure the integrity of the data. The median filling method can effectively reduce the impact of outliers on data analysis.

[0064] Processing of outliers: The outliers in the system operation data are determined by the interquartile range method (IQR), which includes extreme values, and then the determined outliers are removed to remove the extreme values ​​in the system operation data. The interquartile range method calculates the first quartile (Q1) and the third quartile (Q3) of the system data, and uses 1.5 times the interquartile range to define the outlier range. The system operation data in the outlier range can be determined as outliers.

[0065] The system operation data is normalized to the interval [0, 1] to eliminate the differences between different feature dimensions and facilitate subsequent model processing. Finally, the cleaned and standardized system operation data is saved in a new CSV file for the next step of feature extraction.

[0066] In an exemplary embodiment of the present invention, the extracting of the time domain features and frequency domain features of the preprocessed system operation data includes the following steps, which are described in detail as follows.

[0067] The time domain mean and the time domain standard deviation of the system operation data are calculated to obtain the time domain characteristics.

[0068] The frequency domain feature is obtained by calculating the frequency domain mean and the frequency domain standard deviation of the system operation data through fast Fourier transform.

[0069] In the embodiment of the present invention, the time domain mean and the time domain standard deviation of the system operation data are calculated, and then the frequency domain mean and the frequency domain standard deviation of the system operation data are calculated by fast Fourier transform.

[0070] In an exemplary embodiment of the present invention, after inputting the time domain features and the frequency domain features into the trained optimization model to obtain an optimization solution for the basic input-output system, the method further includes the following steps, which are described in detail as follows.

[0071] The effect parameters are determined according to the optimization scheme.

[0072] The optimization plan and the expected effect are sent to a preset display page for display.

[0073] In the embodiment of the present invention, a trained optimization model is used to generate a preferred solution for BIOS setting items based on the system operation data collected in real time. Then, the effect parameter is determined based on the optimization solution. The effect parameter is the average value of the system operation data obtained through multiple tests when the optimization solution is obtained. Then, the optimization solution and the expected effect are displayed through the user interface-web front end to facilitate the understanding of relevant personnel.

[0074] Specifically, the generated optimization solution example is optimized processor frequency: 3.6 GHz; optimized memory timing: 16-18-18-36; optimized fan speed: 1250 RPM; optimized processor voltage: 1.25V.

[0075] In an exemplary implementation of the present invention, after the optimization solution and expected effects are displayed on a preset display page, a confirmation page pops up on the preset display page. After the user confirms on the confirmation page, the user can directly modify the relevant setting items of the BIOS according to the optimization solution.

[0076] In an exemplary embodiment of the present invention, an optimization scheme is generated in real time, and the BIOS setting items are adjusted in real time. The system performance can also be detected in real time. When the system performance reaches the condition that the BIOS setting items need to be adjusted, the BIOS setting items are adjusted according to the latest generated optimization scheme.

[0077] In an exemplary embodiment of the present invention, a method for generating an optimization solution for a basic input / output system using traditional rules and algorithms is also provided. Specifically, it can be implemented based on experience and statistical analysis methods, and the optimization settings of BIOS setting items are performed through fixed logical rules. The implementation is simple and does not require complex model training, but its dynamic adaptability and optimization effect may not be as good as the deep learning method.

[0078] In an exemplary embodiment of the present invention, a method for generating an optimization solution for a basic input / output system based on simulation and emulation is also provided. Specifically, by establishing a simulation and emulation model of a computer system, the influence of different BIOS settings on system performance is simulated, thereby generating an optimization solution. Although this method can provide an accurate optimization solution, it requires complex modeling and a large amount of computing resources, and the actual effect depends on the accuracy of the simulation model.

[0079] Most of the existing methods for optimizing BIOS rely on static settings and fixed configurations, and cannot be dynamically adjusted according to the actual operating status of the system. Moreover, manual adjustments by professionals are usually required, which is cumbersome and time-consuming, especially in large-scale server deployment and management. The efficiency is lower, and different application scenarios and load modes have different requirements for BIOS setting items. Traditional fixed configuration methods are difficult to meet diverse needs. The present invention regularly collects and analyzes system operation data to generate personalized optimization solutions, so that BIOS setting items can dynamically adapt to system changes and improve overall performance and stability. At the same time, the present invention also greatly reduces the workload of manual adjustment and improves optimization efficiency through automated data analysis and optimization solution generation, making BIOS optimization in large-scale application scenarios more convenient and efficient. In addition, the present invention can generate optimization solutions for different application scenarios and system states through the training and data analysis of deep learning models to meet various needs and enhance the system's adaptability in different scenarios.

[0080] The following is a description of the optimization solution generation device for the basic input-output system provided by the present invention. The optimization solution generation device for the basic input-output system described below and the optimization solution generation method for the basic input-output system described above can be referred to each other. It should be noted that the device provided in the following embodiment and the method provided in the above embodiment belong to the same concept, and the specific manner in which each module and unit performs the operation has been described in detail in the method embodiment, which will not be repeated here.

[0081] In an exemplary embodiment of the present invention, see Figure 3 , Figure 3 A device for generating an optimization solution for a basic input-output system according to an exemplary embodiment includes the following modules.

[0082] The acquisition module 310 is configured to acquire system operation data and pre-process the system operation data.

[0083] The first extraction module 320 is configured to extract the time domain features and frequency domain features of the pre-processed system operation data.

[0084] The input module 330 is configured to input the time domain features and the frequency domain features into the trained optimization model to obtain an optimization solution for the basic input and output system; wherein the optimization model is trained based on sample system operation data and corresponding optimization labels, as well as sample time domain features and sample frequency domain features of the sample system operation data, the optimization labels are obtained based on sample performance data corresponding to the sample system operation data, and the optimization solution includes optimizing processor frequency, optimizing memory timing, optimizing fan speed, and optimizing processor voltage.

[0085] In an exemplary embodiment of the present invention, the training process of the optimization model includes the following modules.

[0086] The collection module is configured to collect a number of sample system operation data and sample performance data.

[0087] The storage module is configured to store the sample system operation data and the sample performance data in a preset file in a time sequence.

[0088] The preprocessing module is configured to preprocess the sample system operation data and the sample performance data.

[0089] The group processing module is configured to perform group processing on the sample system operation data based on the storage order of the sample system operation data in the preset file to obtain a plurality of sample groups.

[0090] The second extraction module is configured to extract the sample time domain features and the sample frequency domain features of the sample system operation data in each of the sample groups.

[0091] The first determination module is configured to determine, based on the sample performance data, an optimization label of the sample system operation data corresponding to the sample performance data.

[0092] A generation module is configured to generate a training sample set based on the sample system operation data, the sample performance data, the sample time domain features, the sample frequency domain features and the optimization label; wherein the training sample set includes multiple training samples.

[0093] The iterative training module is configured to input the training samples into a preset basic optimization model for iterative training to obtain the trained optimization model.

[0094] In an exemplary embodiment of the present invention, the preset basic optimization model includes a first LSTM layer, a second LSTM layer and a fully connected layer connected in sequence.

[0095] In an exemplary embodiment of the present invention, the acquisition module 310 includes the following sub-modules.

[0096] The first calculation submodule is configured to calculate the median of the system operation data and fill in the missing values ​​in the system operation data by using the median.

[0097] The processing submodule is configured to process abnormal values ​​in the system operation data by using the interquartile range method.

[0098] The normalization processing submodule is configured to perform normalization processing on the system operation data.

[0099] In an exemplary embodiment of the present invention, the extraction module 320 includes the following sub-modules.

[0100] The second calculation submodule is configured to calculate the time domain mean and the time domain standard deviation of the system operation data to obtain the time domain feature.

[0101] The third calculation submodule is configured to calculate the frequency domain mean and frequency domain standard deviation of the system operation data through fast Fourier transform to obtain the frequency domain feature.

[0102] In an exemplary embodiment of the present invention, the system operation data includes processor usage, memory occupancy, disk I / O rate and temperature.

[0103] In an exemplary embodiment of the present invention, the device for generating an optimization solution for a basic input-output system further includes the following modules.

[0104] The second determination module is configured to determine effect parameters according to the optimization scheme.

[0105] The sending module is configured to send the optimization plan and the expected effect to a preset display page for display.

[0106] Figure 4 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 4 As shown, the electronic device may include: a processor 410, a communications interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communications interface 420 and the memory 430 communicate with each other via the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute the optimization solution generation method of the basic input and output system, the method comprising: obtaining system operation data, and preprocessing the system operation data; Extracting time domain features and frequency domain features of the preprocessed system operation data; The time domain features and the frequency domain features are input into the trained optimization model to obtain an optimization scheme for the basic input and output system; wherein the optimization model is trained based on sample system operation data and corresponding optimization labels, as well as sample time domain features and sample frequency domain features of the sample system operation data, the optimization labels are obtained based on sample performance data corresponding to the sample system operation data, and the optimization scheme includes optimizing processor frequency, optimizing memory timing, optimizing fan speed, and optimizing processor voltage.

[0107] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0108] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored in a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute the basic input and output system optimization solution generation method provided by the above methods, the method includes: obtaining system operation data, and preprocessing the system operation data; Extracting time domain features and frequency domain features of the preprocessed system operation data; The time domain features and the frequency domain features are input into the trained optimization model to obtain an optimization scheme for the basic input and output system; wherein the optimization model is trained based on sample system operation data and corresponding optimization labels, as well as sample time domain features and sample frequency domain features of the sample system operation data, the optimization labels are obtained based on sample performance data corresponding to the sample system operation data, and the optimization scheme includes optimizing processor frequency, optimizing memory timing, optimizing fan speed, and optimizing processor voltage.

[0109] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to execute the method for generating an optimization solution for a basic input / output system provided by the above methods, the method comprising: obtaining system operation data, and preprocessing the system operation data; Extracting time domain features and frequency domain features of the preprocessed system operation data; The time domain features and the frequency domain features are input into the trained optimization model to obtain an optimization scheme for the basic input and output system; wherein the optimization model is trained based on sample system operation data and corresponding optimization labels, as well as sample time domain features and sample frequency domain features of the sample system operation data, the optimization labels are obtained based on sample performance data corresponding to the sample system operation data, and the optimization scheme includes optimizing processor frequency, optimizing memory timing, optimizing fan speed, and optimizing processor voltage.

[0110] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0111] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0112] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for generating an optimization solution for a basic input / output system, characterized in that: include: Acquiring system operation data and preprocessing the system operation data; Extracting time domain features and frequency domain features of the preprocessed system operation data; The time domain features and the frequency domain features are input into the trained optimization model to obtain an optimization scheme for the basic input and output system; wherein the optimization model is trained based on sample system operation data and corresponding optimization labels, as well as sample time domain features and sample frequency domain features of the sample system operation data, the optimization labels are obtained based on sample performance data corresponding to the sample system operation data, and the optimization scheme includes optimizing processor frequency, optimizing memory timing, optimizing fan speed, and optimizing processor voltage.

2. The method for generating an optimization solution for a basic input / output system according to claim 1, characterized in that: The training process of the optimization model includes: Collect several sample system operation data and sample performance data; Storing the sample system operation data and the sample performance data in a preset file in time sequence; Preprocessing the sample system operation data and the sample performance data; Based on the storage order of the sample system operation data in the preset file, grouping the sample system operation data to obtain a plurality of sample groups; Extracting sample time domain features and sample frequency domain features of the sample system operation data in each of the sample groups; Determining, based on the sample performance data, an optimization label for the sample system operation data corresponding to the sample performance data; Generate a training sample set based on the sample system operation data, the sample performance data, the sample time domain features, the sample frequency domain features and the optimization label; wherein the training sample set includes a plurality of training samples; The training samples are input into a preset basic optimization model for iterative training to obtain the trained optimization model.

3. The method for generating an optimization solution for a basic input / output system according to claim 2, characterized in that: The preset basic optimization model includes a first LSTM layer, a second LSTM layer and a fully connected layer connected in sequence.

4. The method for generating an optimization solution for a basic input / output system according to claim 1, characterized in that: The preprocessing of the system operation data includes: Calculating the median of the system operation data, and filling missing values ​​in the system operation data by using the median; Processing outliers in the system operation data by using the interquartile range method; The system operation data is normalized.

5. The method for generating an optimization solution for a basic input / output system according to claim 1, characterized in that: The extracting of the time domain features and frequency domain features of the preprocessed system operation data includes: Calculating the time domain mean and the time domain standard deviation of the system operation data to obtain the time domain characteristics; The frequency domain feature is obtained by calculating the frequency domain mean and the frequency domain standard deviation of the system operation data through fast Fourier transform.

6. The method for generating an optimization solution for a basic input / output system according to claim 1, characterized in that: The system operation data includes processor usage, memory occupancy, disk I / O rate and temperature.

7. The method for generating an optimization solution for a basic input-output system according to any one of claims 1 to 6, characterized in that: After inputting the time domain features and the frequency domain features into the trained optimization model to obtain an optimization solution for the basic input and output system, the method further includes: Determining effect parameters according to the optimization scheme; The optimization plan and the expected effect are sent to a preset display page for display.

8. A device for generating an optimization plan for a basic input / output system, characterized in that: include: An acquisition module, configured to acquire system operation data and pre-process the system operation data; A first extraction module is configured to extract time domain features and frequency domain features of the preprocessed system operation data; An input module is configured to input the time domain features and the frequency domain features into a trained optimization model to obtain an optimization solution for a basic input-output system; wherein the optimization model is trained based on sample system operation data and corresponding optimization labels, as well as sample time domain features and sample frequency domain features of the sample system operation data, the optimization labels are obtained based on sample performance data corresponding to the sample system operation data, and the optimization solution includes optimizing processor frequency, optimizing memory timing, optimizing fan speed, and optimizing processor voltage.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the optimization solution generation method of the basic input and output system according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for generating an optimization solution for a basic input / output system as claimed in any one of claims 1 to 7 is implemented.

Images