Big model-based domestic cpu architecture virtualization application compatibility method and system

Abstract

The application relates to the technical field of computers and discloses a method and system for virtualizing application compatibility of a domestic CPU architecture based on a large model, which comprises the following steps: deploying a virtual machine in a domestic CPU architecture, obtaining a compatible test function description text of the virtual machine, and establishing a test task flow chain according to the compatible test function description text; obtaining application characteristics of an application to be tested, and establishing an application requirement node according to the application characteristics; matching the application requirement node of the application to be tested with each test task flow chain, screening out a target test task flow chain, and calculating comprehensive evaluation parameters of each target test task flow chain; determining a final test task flow chain of the application to be tested according to the comprehensive evaluation parameters, performing compatibility testing on the application to be tested according to the final test task flow chain, and performing application deployment in the virtual machine according to a compatibility testing result. The application is tested for compatibility by establishing a test task flow based on a large model, the test efficiency of the application is effectively improved, and the application compatibility problem of the domestic CPU architecture is solved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and more specifically, to a method and system for application compatibility of domestic CPU architecture virtualization based on a large model. Background Technology

[0002] Currently, all drivers for industry applications, office software, and external devices on the market are developed based on the Windows system environment. This makes them incompatible with domestic CPU architectures, failing to meet users' work needs and creating significant obstacles to the promotion of domestic CPU architectures.

[0003] Existing virtualization technologies (such as QEMU's software emulation) are difficult to use for compatibility testing with domestic CPU architectures, making it impossible for applications to run directly on platforms with domestic CPU architectures. Compatibility testing is inefficient, and it is difficult to build highly compatible test task processes, resulting in poor compatibility and high deployment costs. Summary of the Invention

[0004] This invention provides a method and system for ensuring compatibility of virtualization applications based on a large-scale model with domestic CPU architecture. This addresses the problems of low testing efficiency and difficulty in constructing highly compatible test task flows for virtualization technologies based on domestic CPU architecture in existing technologies. The method includes:

[0005] A virtual machine is deployed on a domestic CPU architecture. The compatibility test function description text of the virtual machine is obtained, and a test task flow chain is established based on this text. The application characteristics of the application under test are obtained, and application requirement nodes are established based on these characteristics. The application requirement nodes of the application under test are matched with each test task flow chain to select target test task flow chains, and the comprehensive evaluation parameters of each target test task flow chain are calculated. The final test task flow chain of the application under test is determined based on the comprehensive evaluation parameters. Compatibility testing of the application under test is performed based on the final test task flow chain, and the application is deployed in the virtual machine based on the compatibility test results.

[0006] Furthermore, the step of establishing a test task flow chain based on the compatibility test function description text includes: establishing a large language model; extracting virtual machine compatibility test function keywords from the description text based on the large language model; establishing compatibility test nodes based on the virtual machine compatibility test function keywords; mining the relationships between compatibility test nodes; and establishing a test task flow chain based on the relationships between compatibility test nodes.

[0007] Furthermore, the step of mining the relationships between compatible test nodes and establishing a test task flow chain based on these relationships includes: establishing a compatible test task trajectory library; mining frequent itemsets of each compatible test function keyword in the compatible test task trajectory library based on association rule algorithms; determining the collaborative relationships of each compatible test function keyword based on the frequent itemsets; determining the relationships between compatible test nodes based on the collaborative relationships of each compatible test function keyword; connecting compatible test nodes with relationships; and establishing a test task flow chain based on the connection results of all compatible test nodes.

[0008] Furthermore, the step of establishing application requirement nodes based on application characteristics includes: acquiring historical application compatibility test data; determining historical application characteristics and corresponding application test requirements based on the historical application compatibility test data; establishing a training sample set based on the historical application characteristics and corresponding application test requirements; establishing a requirement extraction model and training the requirement extraction model based on the training sample set to obtain a trained requirement extraction model; inputting the application characteristics of the application to be tested into the trained requirement extraction model to obtain the corresponding application test requirements; and establishing application requirement nodes based on the application test requirements.

[0009] Further, the step of matching the application requirement nodes of the application to be tested with each test task process chain and selecting the target test task process chain includes: calculating the feature coverage rate of the compatible test nodes in each test task process chain to each application requirement node, and determining the compatible test nodes with feature coverage rates greater than a preset matching threshold as the functional coverage nodes of the application requirement nodes; calculating the average feature coverage rate of all functional coverage nodes in the test task process chain to obtain the total feature coverage rate, and selecting the test task process chain with a total feature coverage rate greater than a first preset threshold as the target test task process chain.

[0010] Further, the calculation of the comprehensive evaluation parameters for each target test task process chain includes: obtaining the total feature coverage of the target test task process chain for each application requirement node, and determining the total feature coverage of the target test task process chain as the first evaluation parameter; obtaining the feature coverage of each compatible test node in the target test task process chain for each application requirement node, filtering out matching compatible test nodes whose feature coverage with the application requirement node is greater than a second preset threshold, calculating the continuity of the matching compatible test nodes in the target test task process chain, and determining the second evaluation parameter based on the continuity; evaluating the test task occupancy rate of each target test task process chain, calculating the deviation between the target test task occupancy rate evaluation value and the preset allowable occupancy rate, and obtaining the third evaluation parameter; and determining the comprehensive evaluation parameters for each target test task process chain based on the first evaluation parameter, the second evaluation parameter, and the third evaluation parameter.

[0011] Further, determining the comprehensive evaluation parameters based on the first evaluation parameter, the second evaluation parameter, and the third evaluation parameter includes: establishing a center point, and drawing rays evenly outward from the center point as indicator axes; normalizing the first evaluation parameter, the second evaluation parameter, and the third evaluation parameter respectively and filling them into the indicator axes; connecting the normalized evaluation parameters on the indicator axes sequentially to obtain an evaluation parameter radar chart; calculating the area value of the evaluation parameter radar chart, and determining the comprehensive evaluation parameters based on the area value of the evaluation parameter radar chart.

[0012] Furthermore, determining the final test task process chain of the application to be tested based on the comprehensive evaluation parameters includes: selecting the target test task process chain with the highest comprehensive evaluation parameters for the application to be tested, and using the selected target test task process chain as the final test task process chain for the application to be tested.

[0013] Furthermore, the step of deploying the application in the virtual machine based on the compatibility test results includes: identifying compatibility issues of the application based on the compatibility test results, generating optimization strategies based on the compatibility issues, and deploying the application to be tested to the virtual machine for operation based on the optimization strategies.

[0014] To achieve the above objectives, the present invention also provides a domestically developed CPU architecture virtualization application compatibility system based on a large model, comprising:

[0015] The system comprises the following modules: a setup module for deploying virtual machines on a domestic CPU architecture, obtaining compatibility test function description text for the virtual machines, and establishing a test task flow chain based on the compatibility test function description text; a feature module for obtaining application features of the application under test and establishing application requirement nodes based on the application features; a matching module for matching the application requirement nodes of the application under test with each test task flow chain, filtering out target test task flow chains, and calculating the comprehensive evaluation parameters of each target test task flow chain; and a deployment module for determining the final test task flow chain of the application under test based on the comprehensive evaluation parameters, performing compatibility testing on the application under test based on the final test task flow chain, and deploying the application in the virtual machine based on the compatibility test results.

[0016] The beneficial effects of this invention are as follows:

[0017] By applying the above technical solutions, this invention establishes a test task process with high matching degree with application features based on a large model to conduct compatibility testing on applications, effectively improving the testing efficiency of applications, solving the application compatibility problem of domestic CPU architecture, and realizing the efficient and compatible operation of applications on domestic CPU platforms. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 The overall flowchart of a method for application compatibility based on a large model of domestic CPU architecture virtualization proposed in an embodiment of the present invention is shown.

[0020] Figure 2 The diagram shows a structural schematic of a domestic CPU architecture virtualization application compatibility system based on a large model, as proposed in an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] This application provides a method for virtualization application compatibility based on a large model of domestic CPU architecture, such as... Figure 1 As shown, it includes:

[0023] S101: Deploy virtual machines in a domestic CPU architecture, obtain the compatibility test function description text of the virtual machines, and establish a test task flow chain based on the compatibility test function description text;

[0024] In some embodiments of this application, the step of establishing a test task flow chain based on the compatibility test function description text includes: establishing a large language model; extracting virtual machine compatibility test function keywords from the description text based on the large language model; establishing compatibility test nodes based on the virtual machine compatibility test function keywords; mining the correlation between compatibility test nodes; and establishing a test task flow chain based on the correlation between compatibility test nodes.

[0025] In this embodiment, a virtual machine is deployed in a domestic CPU architecture using virtualization software. The compatibility test function description text of the virtual machine is obtained through technical manuals, design documents, etc. The description text is then input into a large language model to parse out the functional description keywords of the functional units involved in the compatibility test task of the virtual machine. The compatibility test function description text consists of key test items or functional points related to the compatibility function, performance, and compatibility of the virtual machine.

[0026] In some embodiments of this application, the step of mining the association relationships between compatible test nodes and establishing a test task flow chain based on the association relationships between compatible test nodes includes: establishing a compatible test task trajectory library; mining frequent itemsets of each compatible test function keyword in the compatible test task trajectory library based on association rule algorithms; determining the collaboration relationships of each compatible test function keyword based on the frequent itemsets; determining the association relationships between compatible test nodes based on the collaboration relationships of each compatible test function keyword; connecting compatible test nodes with association relationships; and establishing a test task flow chain based on the connection results of all compatible test nodes.

[0027] In this embodiment, the sequence of compatible test nodes encountered during the compatibility test process is collected through historical compatibility test data. A compatibility test task trajectory library is established based on the sequence of compatible test nodes in the entire compatibility test process. Frequently occurring compatibility test node sequences in the compatibility test task trajectory library are mined based on the Apriori association rule algorithm to obtain frequent itemsets. The compatibility test node sequences in the frequent itemsets are determined as the cooperation relationship of each compatibility test function keyword. For example, if the sequence [A, B, C] is a frequent itemset, then A and B, and B and C have a cooperation relationship. B is often executed after A, and C is often executed after B. The association relationship between compatibility test nodes is determined through the cooperation relationship of all compatibility test function keywords. Thus, multiple test task flow chains are established based on the compatibility test nodes with existing relationships.

[0028] S102, Obtain the application characteristics of the application to be tested, and establish application requirement nodes based on the application characteristics;

[0029] In some embodiments of this application, the step of establishing application requirement nodes based on application characteristics includes: acquiring historical application compatibility test data; determining historical application characteristics and corresponding application test requirements based on the historical application compatibility test data; establishing a training sample set based on the historical application characteristics and corresponding application test requirements; establishing a requirement extraction model and training the requirement extraction model based on the training sample set to obtain a trained requirement extraction model; inputting the application characteristics of the application to be tested into the trained requirement extraction model to obtain the corresponding application test requirements; and establishing application requirement nodes based on the application test requirements.

[0030] In this embodiment, historical application features and application testing requirements are extracted from historical application compatibility test data through static analysis and questionnaire configuration. A requirement extraction model is established based on a deep learning neural network model and trained on it using historical application compatibility test data to obtain multiple application testing requirements corresponding to the application to be tested. Each application testing requirement is then abstracted into an application requirement node.

[0031] S103, match the application requirement nodes of the application to be tested with each test task process chain, filter out the target test task process chains, and calculate the comprehensive evaluation parameters of each target test task process chain.

[0032] In some embodiments of this application, the step of matching the application requirement nodes of the application to be tested with each test task process chain and filtering out the target test task process chain includes: calculating the feature coverage rate of the compatible test nodes in each test task process chain to each application requirement node, determining the compatible test nodes with feature coverage rates greater than a preset matching threshold as the functional coverage nodes of the application requirement nodes; calculating the average feature coverage rate of all functional coverage nodes in the test task process chain to obtain the total feature coverage rate, and filtering out the test task process chain with a total feature coverage rate greater than a first preset threshold as the target test task process chain.

[0033] In this embodiment, the matching degree between the compatible test nodes and the application requirement nodes in the test task flow chain is calculated to obtain the feature coverage of the two nodes. In this embodiment, the semantic similarity between the compatible test nodes and the application requirement nodes is determined as the matching degree. Compatible test nodes with a matching degree greater than a preset matching threshold are determined as functional coverage nodes of the application requirement nodes. The total feature coverage is determined by the average feature coverage of all functional coverage nodes in a test task flow chain. The target test task flow chain is obtained by using the total feature coverage.

[0034] In some embodiments of this application, the calculation of the comprehensive evaluation parameters of each target test task process chain includes: obtaining the total feature coverage of the target test task process chain for each application requirement node, and determining the total feature coverage of the target test task process chain as a first evaluation parameter; obtaining the feature coverage of each compatible test node in the target test task process chain for each application requirement node, filtering out matching compatible test nodes whose feature coverage with the application requirement node is greater than a second preset threshold, calculating the continuity of the matching compatible test nodes in the target test task process chain, and determining the second evaluation parameter based on the continuity; evaluating the test task occupancy rate of each target test task process chain, calculating the deviation between the target test task occupancy rate evaluation value and the preset allowable occupancy rate, and obtaining a third evaluation parameter; and determining the comprehensive evaluation parameters of each target test task process chain based on the first evaluation parameter, the second evaluation parameter, and the third evaluation parameter.

[0035] In this embodiment, the target test task flow chain and the first evaluation parameter are obtained by the total feature coverage of all application requirement nodes and compatible test nodes in the test task flow chain. Matching compatible test nodes are filtered based on the feature coverage. The connectivity of the matching compatible test nodes is calculated based on their positions in the target test task flow chain. Continuity is determined based on the connectivity to obtain the second evaluation parameter. The task utilization rate is evaluated using the CPU utilization rate of the target test task flow chain in historical application compatibility test data. The third evaluation parameter is obtained by the deviation between the target test task utilization rate evaluation value and the preset allowable utilization rate.

[0036] In some embodiments of this application, determining the comprehensive evaluation parameter based on the first evaluation parameter, the second evaluation parameter, and the third evaluation parameter includes: establishing a center point, and drawing rays evenly outward from the center point as indicator axes; normalizing the first evaluation parameter, the second evaluation parameter, and the third evaluation parameter respectively and filling them into the indicator axes; connecting the normalized evaluation parameters on the indicator axes sequentially to obtain an evaluation parameter radar chart; calculating the area value of the evaluation parameter radar chart, and determining the comprehensive evaluation parameter based on the area value of the evaluation parameter radar chart.

[0037] In this embodiment, an evaluation parameter radar chart is established to comprehensively evaluate the task matching of each target test task process chain based on the first evaluation parameter, the second evaluation parameter, and the third evaluation parameter. The area value of the evaluation parameter radar chart is then determined as the comprehensive evaluation parameter.

[0038] S104. Determine the final test task flow chain of the application to be tested based on the comprehensive evaluation parameters, perform compatibility testing on the application to be tested based on the final test task flow chain, and deploy the application in a virtual machine based on the compatibility test results.

[0039] In some embodiments of this application, determining the final test task flow chain of the application to be tested based on comprehensive evaluation parameters includes: selecting the target test task flow chain with the highest comprehensive evaluation parameters of the application to be tested, and using the selected target test task flow chain as the final test task flow chain of the application to be tested.

[0040] In some embodiments of this application, the step of deploying the application in a virtual machine based on the compatibility test results includes: identifying compatibility issues of the application based on the compatibility test results, generating optimization strategies based on the compatibility issues, and deploying the application to be tested to the virtual machine for operation based on the optimization strategies.

[0041] Based on the same technological concept, such as Figure 2 As shown, this invention also provides a domestically developed CPU architecture virtualization application compatibility system based on a large model, comprising:

[0042] The system comprises the following modules: a setup module for deploying virtual machines on a domestic CPU architecture, obtaining compatibility test function description text for the virtual machines, and establishing a test task flow chain based on the compatibility test function description text; a feature module for obtaining application features of the application under test and establishing application requirement nodes based on the application features; a matching module for matching the application requirement nodes of the application under test with each test task flow chain, filtering out target test task flow chains, and calculating the comprehensive evaluation parameters of each target test task flow chain; and a deployment module for determining the final test task flow chain of the application under test based on the comprehensive evaluation parameters, performing compatibility testing on the application under test based on the final test task flow chain, and deploying the application in the virtual machine based on the compatibility test results.

[0043] By applying the above technical solutions, this invention deploys a virtual machine in a domestic CPU architecture, obtains the virtual machine's compatibility test function description text, and establishes a test task flow chain based on the compatibility test function description text; it obtains the application characteristics of the application to be tested and establishes application requirement nodes based on the application characteristics; it matches the application requirement nodes of the application to be tested with each test task flow chain, filters out target test task flow chains, and calculates the comprehensive evaluation parameters of each target test task flow chain; it determines the final test task flow chain of the application to be tested based on the comprehensive evaluation parameters, performs compatibility testing on the application to be tested based on the final test task flow chain, and deploys the application in the virtual machine based on the compatibility test results. By establishing a test task flow based on a large model to perform compatibility testing on the application, the testing efficiency of the application is effectively improved, and the application compatibility problem of domestic CPU architecture is solved.

[0044] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for virtualizing application compatibility based on a large model of a domestic CPU architecture, characterized in that, The application comprises the following steps: Deploying a virtual machine in a domestic CPU architecture, obtaining a compatible test function description text of the virtual machine, and establishing a test task flow chain according to the compatible test function description text; Obtaining application characteristics of the application to be tested, and establishing an application requirement node according to the application characteristics; Matching the application requirement node of the application to be tested with each test task flow chain, screening out a target test task flow chain, and calculating a comprehensive evaluation parameter of each target test task flow chain; Determining a final test task flow chain of the application to be tested according to the comprehensive evaluation parameter, performing compatibility testing on the application to be tested according to the final test task flow chain, and deploying the application in the virtual machine according to the compatibility testing result; The step of establishing a test task flow chain according to the compatible test function description text comprises the following steps: Establishing a large language model, extracting compatible test function keywords of the virtual machine based on the large language model according to the description text; Establishing a compatible test node according to the compatible test function keywords of the virtual machine, mining the association relationship between the compatible test nodes, and establishing a test task flow chain according to the association relationship between the compatible test nodes; The step of establishing an application requirement node according to the application characteristics comprises the following steps: Obtaining historical application compatibility testing data, determining historical application characteristics and corresponding application testing requirements according to the historical application compatibility testing data; Establishing a training sample set according to the historical application characteristics and corresponding application testing requirements, training a requirement extraction model according to the training sample set, and obtaining a trained requirement extraction model; Inputting the application characteristics of the application to be tested into the trained requirement extraction model to obtain corresponding application testing requirements, and establishing an application requirement node according to the application testing requirements.

2. The method of claim 1, wherein the method comprises: The step of mining the association relationship between the compatible test nodes and establishing a test task flow chain according to the association relationship between the compatible test nodes comprises the following steps: Establishing a compatible test task trajectory library, and mining frequent item sets of the compatible test function keywords in the compatible test task trajectory library based on an association rule algorithm; Determining the collaboration relationship of each compatible test function keyword according to the frequent item sets, and determining the association relationship between the compatible test nodes according to the collaboration relationship of each compatible test function keyword; Connecting the compatible test nodes with the association relationship, and establishing a test task flow chain according to the connection result of all the compatible test nodes.

3. The method of claim 1, wherein the method comprises: The step of matching the application requirement node of the application to be tested with each test task flow chain and screening out a target test task flow chain comprises the following steps: Calculating the feature coverage rate of the compatible test nodes in each test task flow chain to each application requirement node, and determining the compatible test nodes with a feature coverage rate greater than a preset matching threshold as function coverage nodes of the application requirement node; Calculating the average feature coverage rate of all the function coverage nodes in the test task flow chain to obtain a total feature coverage rate, and screening out a test task flow chain with a total feature coverage rate greater than a first preset threshold as a target test task flow chain.

4. The method of claim 3, wherein the method further comprises: The step of calculating a comprehensive evaluation parameter of each target test task flow chain comprises the following steps: Obtaining the total feature coverage rate of the target test task flow chain to each application requirement node, and determining the total feature coverage rate of the target test task flow chain as a first evaluation parameter; Obtaining feature coverage of each application demand node by each compatible test node in the target test task flow chain, screening out a matching compatible test node with feature coverage of the application demand node greater than a second preset threshold, calculating continuity of the matching compatible test node in the target test task flow chain, and determining a second evaluation parameter according to the continuity; Performing test task occupancy rate evaluation on each target test task flow chain, calculating a deviation value of the target test task occupancy rate evaluation value and a preset allowable occupancy rate, and obtaining a third evaluation parameter; Determine the comprehensive evaluation parameter according to the first evaluation parameter, the second evaluation parameter and the third evaluation parameter.

5. The method of claim 4, wherein the method further comprises: The determination of the comprehensive evaluation parameter according to the first evaluation parameter, the second evaluation parameter and the third evaluation parameter comprises: Establish a center, and evenly draw rays from the center as index axes; Respectively normalize the first evaluation parameter, the second evaluation parameter and the third evaluation parameter and fill them into the index axes, connect the normalized evaluation parameters on the index axes in turn to obtain an evaluation parameter radar chart; Calculate the area value of the evaluation parameter radar chart, and determine the comprehensive evaluation parameter according to the area value of the evaluation parameter radar chart.

6. The method of claim 1, wherein the method further comprises: The determination of the final test task flow chain of the application to be tested according to the comprehensive evaluation parameter comprises: Screening out a target test task flow chain with the highest comprehensive evaluation parameter of the application to be tested, and taking the screened target test task flow chain as the final test task flow chain of the application to be tested.

7. The method of claim 1, wherein the method further comprises: The application deployment in the virtual machine according to the compatibility test result comprises: Identifying compatibility problems of the application according to the compatibility test result, generating an optimization strategy according to the compatibility problems, and deploying the application to be tested to the virtual machine for running according to the optimization strategy.

8. A large model-based domestic CPU architecture virtualization application compatibility system, characterized in that, Comprise: The establishment module is used for deploying a virtual machine in a domestic CPU architecture, obtaining a compatible test function description text of the virtual machine, and establishing a test task flow chain according to the compatible test function description text; The feature module is used for obtaining application features of the application to be tested, and establishing application demand nodes according to the application features; The matching module is used for matching the application demand nodes of the application to be tested with each test task flow chain, screening out target test task flow chains, and calculating comprehensive evaluation parameters of each target test task flow chain; The deployment module is used for determining a final test task flow chain of the application to be tested according to the comprehensive evaluation parameters, performing compatibility test on the application to be tested according to the final test task flow chain, and deploying the application in the virtual machine according to the compatibility test result; The establishment module comprises: Establishing a large language model, and extracting compatible test function keywords of the virtual machine based on the large language model according to the description text; Establishing compatible test nodes according to the compatible test function keywords of the virtual machine, mining association relationships between the compatible test nodes, and establishing a test task flow chain according to the association relationships between the compatible test nodes; The feature module comprises: Obtaining historical application compatibility test data, determining historical application features and corresponding application test demands according to the historical application compatibility test data; A training sample set is established according to historical application features and corresponding application test requirements, a requirement extraction model is established, and the requirement extraction model is trained according to the training sample set, so as to obtain the trained requirement extraction model; The application features of the application to be tested are input into the trained requirement extraction model, and the corresponding application test requirements are obtained, and the application requirement node is established according to the application test requirements.

Images