A signal integrity simulation method, system, computer device and storage medium

Abstract

The application provides a signal integrity simulation method and system, computer equipment and a storage medium. The method comprises the following steps: obtaining an execution script supported by a simulation tool and a first mapping relationship between each execution script and each simulation process; obtaining text information of each execution script, inputting the text information into a preset language model, obtaining a first intention text group having a second mapping relationship with the execution script, and a second intention text group having a third mapping relationship with a calling sequence of the execution script; analyzing the input first natural language instruction through the language model to obtain a first intention instruction, wherein the first intention instruction comprises a target first intention text and a target second intention text; generating calling code of the execution script according to the first intention instruction; and controlling the simulation tool to execute the simulation process according to the calling code to obtain a simulation result. The scheme can automatically perform signal integrity simulation according to user demand, and has a lower operation threshold.

Description

Technical Field

[0001] This invention relates to the field of EDA simulation technology, and more particularly to a signal integrity simulation method, system, computer equipment, and storage medium. Background Technology

[0002] Existing EDA-based signal integrity (SI) simulation tools typically provide graphical user interfaces (GUIs) and command-line / script interfaces (such as Tcl and Python) for users to operate. Users can complete tasks such as project creation, PCB board modeling, component model import, simulation parameter configuration, and result viewing through menus, dialog boxes, parameter tables, etc.

[0003] However, while GUI interfaces are intuitive, their interaction often relies on numerous user clicks and manual operations, resulting in lengthy and repetitive task flows. A complete simulation task often requires users to repeatedly switch between multiple interfaces and set dozens of parameters, which is time-consuming and labor-intensive. While performing simulation operations by calling script APIs is more flexible than GUI interfaces, it requires users to have programming skills and be familiar with the tool's API documentation and calling conventions, making it too difficult for non-engineering professionals. Therefore, there is an urgent need for a method with a lower barrier to entry that can automatically perform signal integrity simulations based on user needs. Summary of the Invention

[0004] The purpose of this invention is to provide a signal integrity simulation method, system, computer device, and storage medium that can automatically perform signal integrity simulation according to user needs, with a lower operating threshold, greatly reducing the barrier to entry for the simulation system. Even non-professional engineers can easily issue simulation commands to perform corresponding simulations, making it more widely applicable and providing a better user experience.

[0005] The technical solution provided by this invention is as follows:

[0006] Firstly, this application provides a signal integrity simulation method, including the following steps:

[0007] Obtain the execution scripts supported by the simulation tool, and the first mapping relationship between each execution script and each simulation process in the signal integrity simulation;

[0008] The text information of each of the execution scripts is obtained and input into a preset language model for training to obtain a first intention text group that has a second mapping relationship with the execution scripts, and a second intention text group that has a third mapping relationship with the calling order of the execution scripts.

[0009] The language model is used to perform semantic parsing on the first natural language instruction input by the user to obtain a first intent instruction corresponding to the first natural language instruction. The first intent instruction includes a target first intent text and a target second intent text.

[0010] The script execution call code is generated based on the first intent instruction;

[0011] The simulation tool is controlled to execute the simulation process according to the calling code to obtain simulation results.

[0012] In some implementations, after generating the calling code for the execution script according to the first intent instruction, and before controlling the simulation tool to execute the simulation process according to the calling code, the method further includes:

[0013] The calling code is validated to determine whether each of the execution scripts in the calling code exists in the defined interface of the simulation tool, and whether the calling order of each execution script conforms to the simulation process constraints; if the validation passes, the simulation tool is controlled to execute the simulation process according to the calling code; if the validation fails, the calling code is corrected according to the reason for the validation failure.

[0014] In some implementations, inputting the text information into a preset language model for training further includes:

[0015] Obtain a third intent text group that has a fourth mapping relationship with the parameter settings of the execution script;

[0016] After obtaining the first intent instruction corresponding to the first natural language instruction, the simulation parameters of each execution script in the calling code are set according to the target third intent text included in the first intent instruction.

[0017] In some implementations, after obtaining the simulation results, the following steps are also included:

[0018] The simulation results are analyzed to obtain key performance indicators;

[0019] The simulation and analysis results will be output in the form of a structured report.

[0020] In some embodiments, the signal integrity simulation method provided in this application further includes:

[0021] The language model receives a second natural language instruction input by the user based on the structured report, parses the second natural language instruction, and combines it with the first intent instruction to generate a second intent instruction.

[0022] A new calling code is generated according to the second intent instruction, and the simulation tool is controlled to execute the simulation process according to the new calling code.

[0023] In some implementations, after obtaining the first intent instruction corresponding to the first natural language instruction, the method further includes:

[0024] If the first intent text group does not include the target first intent text, and / or the second intent text group does not include the target second intent text, then a prompt message is output to prompt the user to input a third natural language command.

[0025] In some embodiments, the signal integrity simulation method provided in this application further includes: after obtaining a first intention instruction that meets the requirements based on a third natural language instruction input by the user, comparing the third natural language instruction input by the user with the first natural language instruction, and updating the first intention text group and the second intention text group according to the comparison result.

[0026] Secondly, this application provides a signal integrity simulation system, comprising:

[0027] The human-computer interaction terminal is used to receive the user's first natural language input commands;

[0028] The simulation module includes a language model module and a simulation module;

[0029] The processing end is used to obtain the execution scripts supported by the simulation tools of the simulation module, and the first mapping relationship between each execution script and each simulation process in the signal integrity simulation; the processing end is also used to obtain the text information of each execution script, and input the text information into the language model in the language model module for training, to obtain a first intention text group with a second mapping relationship with the execution script, and a second intention text group with a third mapping relationship with the calling order of the execution script;

[0030] The processing terminal performs semantic parsing on the first natural language instruction input by the user through the human-computer interaction terminal using the language model to obtain a first intent instruction corresponding to the first natural language instruction. The first intent instruction includes a target first intent text and a target second intent text. The processing terminal generates the calling code of the execution script based on the first intent instruction and controls the simulation tool to execute the simulation process according to the calling code to obtain simulation results.

[0031] Thirdly, this application provides a computer device including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the signal integrity simulation method described in the first aspect.

[0032] Fourthly, this application provides a computer storage medium storing a computer program or instructions thereon, wherein the computer program or instructions, when executed by a processor, implement the steps of the signal integrity simulation method described in the first aspect.

[0033] The signal integrity simulation method, system, computer equipment, and storage medium provided by this invention allow users to describe simulation requirements in natural language and accurately understand user semantic intent, generate script calls, and automate the entire simulation process. It unifies cumbersome simulation configuration and execution into simple natural language commands, greatly simplifying user operations and lowering the barrier to entry for the simulation system. Even non-professional engineers can easily issue simulation commands to perform corresponding simulations, making it more widely applicable and providing a better user experience. Attached Figure Description

[0034] The preferred embodiments will now be described in a clear and easy-to-understand manner, with reference to the accompanying drawings, to further explain the above-mentioned characteristics, technical features, advantages, and implementation methods of this solution.

[0035] Figure 1 This is a schematic diagram of the overall process of one embodiment of the present invention;

[0036] Figure 2 This is a flowchart illustrating an embodiment of the present invention. Detailed Implementation

[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the specific implementation methods of the present invention will be described below with reference to the accompanying drawings. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings and other implementation methods can be obtained based on these drawings without any creative effort.

[0038] To keep the drawings concise, only the parts relevant to the invention are shown schematically in each figure, and they do not represent the actual structure of the product. Furthermore, for ease of understanding, in some figures, only one of components with the same structure or function is shown schematically, or only one is labeled. In this document, "one" can mean not only "only one" but also "more than one".

[0039] Existing EDA-based signal integrity (SI) simulation tools are extremely complex and are commonly used to meet various chip and PCB design requirements. These simulation tools typically provide a graphical user interface (GUI) and a command-line / script interface (such as Tcl or Python) for users to operate. Users can complete tasks such as project creation, PCB board modeling, component model import, simulation parameter configuration, and result viewing through menus, dialog boxes, parameter tables, etc.

[0040] However, while GUI interfaces are intuitive, their interaction often relies on numerous clicks and manual operations, resulting in lengthy and repetitive task flows. A complete simulation task often requires users to repeatedly switch between multiple interfaces and set dozens of parameters, which is time-consuming and labor-intensive. To achieve automation and batch processing, users can call script interface APIs to complete project management, PCB layout data processing, parameter setting, simulation flow control, and waveform data acquisition. While using script interface APIs for simulation operations is more flexible than GUI interface operations, it requires users to have programming skills and be familiar with the tool's API documentation and calling conventions, which is too high a barrier for non-engineers. Therefore, there is an urgent need for a method with a lower operating threshold that can automatically perform signal integrity simulations according to user needs.

[0041] This solution proposes a natural language-based SI simulation method. This method allows users to describe simulation requirements in natural language. The system pre-inputs supported script APIs or interface documents as prompts into a Large Language Model (LLM), which then performs semantic parsing of the input information. Since hardware description languages ​​and simulation scripts are essentially text, this aligns perfectly with the text processing capabilities of LLMs. This solution leverages this characteristic, automatically generating corresponding Python API call sequences through LLM to execute the simulation process, thus achieving end-to-end automation from natural language to simulation. Specifically, the system generates Python scripts for the SI simulation tool based on the parsed task intent: sequentially calling the tool's modeling, parameter configuration, simulation execution, and result acquisition APIs to automate the entire simulation process. After simulation, the system can also automatically analyze the output waveforms and data, generating structured reports (HTML or PDF format) for users to view and share the simulation results. The following is a detailed description of this solution with reference to the accompanying figures:

[0042] In one embodiment, refer to the appendix to the specification. Figure 1 and attached Figure 2 This application provides a signal integrity simulation method, including the following steps:

[0043] S100. Obtain the execution scripts supported by the simulation tool, and the first mapping relationship between each execution script and each simulation process in the signal integrity simulation.

[0044] This step is a preliminary step before the user actually performs the simulation operation. Due to the differences between various simulation tools, it is first necessary to obtain the execution scripts supported by the simulation tool. Execution scripts are usually in Python format, but other script formats can also be used; this application does not impose any restrictions. Each execution script corresponds to a different simulation process in the signal integrity simulation process. For example, execution script A corresponds to the circuit or PCB model building operation in the signal integrity simulation process, and execution script B corresponds to the netlist generation operation in the signal integrity simulation process. Therefore, this solution needs to first determine the initial mapping relationship between each execution script and each simulation process in the signal integrity simulation. Based on this initial mapping relationship, the simulation tool can be controlled to execute the corresponding simulation operation when the execution script is determined.

[0045] S200. Obtain the text information of each execution script, and input the text information into a preset language model for training to obtain a first intention text group with a second mapping relationship with the execution script, and a second intention text group with a third mapping relationship with the calling order of the execution script.

[0046] This step is also a preparatory step before the user actually performs the simulation operation. After determining the execution scripts that the simulation tool can support, the text information of each execution script is obtained, such as API documentation or interface descriptions. This text information is input into a preset language model for training, which yields a first intention text group with a second mapping relationship to the execution scripts, and a second intention text group with a third mapping relationship to the order in which the execution scripts are called. Both the first and second intention text groups include multiple keyword information, which corresponds to the user's intent. Based on these keyword information, the execution scripts participating in the simulation and their order can be determined. Both the first and second intention text groups are updatable text groups; as the simulation progresses, both the first and second intention text groups can be continuously updated to become increasingly accurate.

[0047] The language model used in this solution is a Large Language Model (LLM). Other embodiments may also use OpenAI's GPT series, locally deployed LLaMA models, or other proprietary models; this application does not impose any restrictions. To enable customized access to different language models, this application provides an abstraction layer for model call interfaces, tailored to different scenarios and performance requirements. Through this interface, users or developers can flexibly access various LLM services. The interface layer is responsible for passing user input to the selected model and receiving the output text generated by the model for use by subsequent modules. This design ensures the system's scalability and customizability, allowing for seamless replacement or upgrades of the large model in the future without affecting the overall architecture.

[0048] S300. Semantically analyze the first natural language instruction input by the user through a language model to obtain the first intention instruction corresponding to the first natural language instruction. The first intention instruction includes the target first intention text and the target second intention text.

[0049] This application uses user-input natural language commands for simulation control. Users' actual simulation needs are typically expressed in natural language, such as "extract S-parameters from the DDR4 channel and generate an eye diagram report." However, existing simulation tools cannot directly understand this natural language. In existing technologies, users must manually convert it into GUI operation steps or script statements, which carries cognitive burden and error risks. This solution pre-acquires the text information of each execution script and inputs it into a preset language model for training. This yields a first intention text group with a second mapping relationship to the execution scripts, and a second intention text group with a third mapping relationship to the calling order of the execution scripts. This allows, after obtaining the user-input natural language command, semantic parsing to obtain the corresponding intention command. The intention command includes a target first intention text and a target second intention text. Through the target first intention text and the target second intention text, the execution scripts participating in the simulation operation corresponding to the natural language command, as well as the order of execution of each script, can be determined.

[0050] In practice, users submit simulation task instructions through a human-computer interaction interface (such as a text input box or a dialog-based interactive window). For example, a user can input "build a 2cm × 2cm rectangular microstrip antenna and simulate its S-parameters," or make a similar request in a dialog-like format. After receiving the natural language instruction, the system sends it to the backend parsing module for semantic analysis to determine the user's intent.

[0051] S400: Generate the script execution call code based on the first intent instruction.

[0052] After parsing the user's input natural language commands to obtain the corresponding execution scripts for the simulation operations and the order in which these scripts are executed, the calling code for the execution scripts can be generated. Specifically, since the system pre-inputs the supported Python script API documentation or interface specifications as prompts into the LLM, it possesses the contextual knowledge for calling simulation tools. When the user inputs a natural language command, the language model performs semantic parsing of the command and, combined with the existing API specifications, can directly generate the corresponding Python calling code. The generated code includes the function calls and parameter settings required for the simulation process, such as project creation, PCB layout operations, simulation parameter configuration, simulation execution, and result extraction. The simulation module can directly run this calling code to enable the simulation tool to automatically execute the simulation process according to the user's requirements.

[0053] The S500 control simulation tool executes the simulation process according to the calling code and obtains the simulation results.

[0054] The system executes the simulation process sequentially according to the generated API scripts. During this process, the Python API of the simulation tool can be used to complete the following operations in sequence: building circuit or PCB models, generating netlists, meshing, configuring simulation parameters (such as frequency points, excitation signals, etc.), starting the simulation engine to run calculations, and acquiring simulation outputs (such as electromagnetic field waveforms, voltage / current time-domain responses, etc.). The entire simulation process runs automatically without user intervention.

[0055] Unlike existing manual configuration or script-driven methods, this solution uses natural language as the human-computer interaction interface, greatly simplifying user operations. It leverages LLM to accurately understand semantic intent and automatically generate script calls, unifying cumbersome simulation configuration and execution into simple natural language commands. Based on this, the invention specifically develops an interpretation and execution scheme for SI simulation tasks, achieving a closed loop from natural language to simulation results. Furthermore, to ensure flexibility, this solution designs a model call interface abstraction framework, allowing access to various large language models. Users can customize and use public or private large model services as needed.

[0056] The natural language control method of this invention significantly improves the usability and automation level of simulation tools. Through natural language interaction, users do not need to memorize complex commands or programming syntax. The natural language interface "reduces the learning curve" and "requires no technical knowledge," greatly lowering the barrier to entry. Even non-technical engineers can easily issue simulation commands. Simultaneously, the entire simulation setup and execution process is automated, transforming traditional manual operations into a single command. This not only significantly shortens project cycles but also improves execution efficiency. Verification shows that EDA systems with natural language interfaces can increase production efficiency by more than 10 times. This invention also significantly accelerates simulation configuration and execution, reducing human error. Furthermore, because natural language aligns with human expression habits, this solution is more user-friendly for non-technical users, promoting the application and promotion of simulation tools in a wider range of fields.

[0057] In one embodiment, based on the foregoing embodiments, after generating the calling code for the execution script according to the first intent instruction, and before controlling the simulation tool to execute the simulation process according to the calling code, the method further includes:

[0058] The calling code is validated to determine whether each execution script exists in the defined interface of the simulation tool and whether the calling order of each execution script conforms to the simulation process constraints. If the validation passes, the simulation tool is controlled to execute the simulation process according to the calling code. If the validation fails, the calling code is corrected according to the reason for the validation failure.

[0059] Specifically, to ensure the reliability of the generated results, the system needs to perform a validity check on the Python code output by the language model before execution. This includes checking whether API calls exist in the defined interfaces and whether the call order conforms to the simulation process constraints. After the verification is passed, the system submits the Python code to the simulation tool for execution, achieving an end-to-end mapping from natural language to simulation execution. If errors are found, the system can automatically provide feedback to the language model for secondary correction until the correct API call sequence is generated.

[0060] In one embodiment, based on the foregoing embodiments, the method of inputting text information into a preset language model for training further includes:

[0061] Obtain a third intent text group that has a fourth mapping relationship with the parameter settings of the executed script;

[0062] After obtaining the first intent instruction corresponding to the first natural language instruction, the simulation parameters of each execution script in the calling code are set according to the target third intent text included in the first intent instruction.

[0063] Because the simulation parameters for each simulation step or operation need to be determined during actual simulation execution, the final simulation results will differ even for the same simulation steps if the simulation parameters are different. Therefore, this application also needs to determine the simulation parameters for each execution script when generating the calling code.

[0064] Specifically, this solution, when inputting text information into a preset language model for training, also includes obtaining a third intent text group that has a fourth mapping relationship with the parameter settings of the execution scripts. After obtaining the intent instructions corresponding to the user's natural language commands, the simulation parameters of each execution script can be determined based on the intent instructions, and the simulation parameters of each execution script can be set so that after generating the calling code, the simulation tool can execute the simulation operation according to the set simulation parameters. Similarly, the third intent text group can also be updated to improve the accuracy of parameter settings.

[0065] In one embodiment, after obtaining the simulation results based on the foregoing embodiments, the method further includes:

[0066] Analyze the simulation results to obtain key performance indicators;

[0067] The simulation and analysis results will be output in the form of a structured report.

[0068] After simulation, the system automatically collects the output results and performs preliminary analysis. The analysis module calculates key performance indicators (such as reflection loss, amplitude and phase response, and time-domain delay) and generates charts and data summaries. Finally, the system formats the simulation configuration, simulation result images, and analysis conclusions into a structured report, which can be output in HTML or PDF format. Users can view this report to evaluate the design effectiveness, or it can be archived or shared as a document. For example, the report includes circuit diagrams, S-parameter curves, tables of key indicators, and textual explanations of the simulation results.

[0069] Preferably, the signal integrity simulation method provided in this application further includes:

[0070] The system receives second natural language instructions from the user based on a structured report through a language model, parses the second natural language instructions, and combines them with the first intention instruction to generate a second intention instruction.

[0071] The system generates new calling code based on the second intent instruction and controls the simulation tool to execute the simulation process according to the new calling code.

[0072] To enhance usability, this system supports context-based multi-turn conversational interaction. If a user submits a new request after viewing the report (e.g., "Please change the antenna size to 3cm and resimulate"), the system can continue parsing the instruction while maintaining the existing context, updating the intermediate representation, and regenerating the corresponding API calls, automatically executing the new simulation process. This multi-turn conversational mode allows users to progressively refine their requirements, enabling the simulation tool to iteratively complete tasks like an intelligent assistant.

[0073] In one embodiment, after obtaining the first intent instruction corresponding to the first natural language instruction, based on the foregoing embodiments, the method further includes:

[0074] If the first intent text group does not include the target first intent text, and / or the second intent text group does not include the target second intent text (i.e., the execution scripts required for this simulation and / or the order of execution scripts cannot be determined based on the natural language instructions entered by the user), then a prompt message is output to prompt the user to enter a third natural language instruction (i.e., a new natural language instruction).

[0075] Preferably, the signal integrity simulation method provided in this application further includes: after obtaining a first intention instruction that meets the requirements based on the third natural language instruction input by the user, comparing the third natural language instruction input by the user with the first natural language instruction, and updating the first intention text group and the second intention text group according to the comparison result, so that the language model can become more and more accurate.

[0076] In one embodiment, based on the foregoing embodiments, this application provides a signal integrity simulation system, including a human-computer interaction terminal, a simulation terminal, and a processing terminal. The human-computer interaction terminal is used to receive a first natural language command input by a user. The simulation terminal includes a language model module and a simulation module.

[0077] The processing end is used to obtain the execution scripts supported by the simulation tools of the simulation module, as well as the first mapping relationship between each execution script and each simulation process in the signal integrity simulation; the processing end is also used to obtain the text information of each execution script, and input the text information into the language model in the language model module for training, to obtain the first intention text group with the second mapping relationship with the execution script, and the second intention text group with the third mapping relationship with the calling order of the execution script.

[0078] The processing unit performs semantic parsing on the first natural language instruction input by the user through the human-computer interaction terminal using a language model to obtain the first intention instruction corresponding to the first natural language instruction. The first intention instruction includes the target first intention text and the target second intention text. Based on the first intention instruction, the processing unit generates the calling code for the execution script and controls the simulation tool to execute the simulation process according to the calling code to obtain the simulation results.

[0079] The technical concept of the signal integrity simulation system provided in this embodiment is the same as that of the signal integrity simulation method in the previous embodiment, so it will not be described again.

[0080] In one embodiment, based on the foregoing embodiments, this application provides a computer device including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the signal integrity simulation method of the foregoing embodiments.

[0081] In one embodiment, based on the foregoing embodiments, this application provides a computer storage medium storing a computer program or instructions thereon, wherein the computer program or instructions, when executed by a processor, implement the steps of the signal integrity simulation method of the foregoing embodiments.

[0082] It should be noted that the above embodiments can be freely combined as needed. The above description is only a preferred embodiment of the present invention. It should be pointed out that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A signal integrity simulation method, characterized in that, Including the following steps: Obtain the execution scripts supported by the simulation tool, and the first mapping relationship between each execution script and each simulation process in the signal integrity simulation; The text information of each execution script is obtained and input into a preset language model for training to obtain a first intention text group with a second mapping relationship to the execution script, a second intention text group with a third mapping relationship to the calling order of the execution script, and a third intention text group with a fourth mapping relationship to the parameter settings of the execution script. The language model is used to perform semantic parsing on the first natural language instruction input by the user to obtain a first intent instruction corresponding to the first natural language instruction. The first intent instruction includes a target first intent text and a target second intent text. The execution script is generated according to the first intent instruction, and the simulation parameters of each execution script in the calling code are set according to the target third intent text included in the first intent instruction. The simulation tool is controlled to execute the simulation process according to the calling code to obtain simulation results; The simulation results are analyzed to obtain key performance indicators; The simulation and analysis results will be output in the form of a structured report; The language model receives a second natural language instruction input by the user based on the structured report, parses the second natural language instruction, and combines it with the first intent instruction to generate a second intent instruction. A new calling code is generated according to the second intent instruction, and the simulation tool is controlled to execute the simulation process according to the new calling code.

2. The signal integrity simulation method according to claim 1, characterized in that, After generating the calling code for the execution script according to the first intent instruction, and before controlling the simulation tool to execute the simulation process according to the calling code, the method further includes: The calling code is validated to determine whether each of the execution scripts in the calling code exists in the defined interface of the simulation tool, and whether the calling order of each execution script conforms to the simulation process constraints; if the validation passes, the simulation tool is controlled to execute the simulation process according to the calling code; if the validation fails, the calling code is corrected according to the reason for the validation failure.

3. The signal integrity simulation method according to claim 1, characterized in that, After obtaining the first intent instruction corresponding to the first natural language instruction, the method further includes: If the first intent text group does not include the target first intent text, and / or the second intent text group does not include the target second intent text, then a prompt message is output to prompt the user to input a third natural language command.

4. The signal integrity simulation method according to claim 3, characterized in that, Also includes: After obtaining the first intent instruction that meets the requirements based on the third natural language instruction input by the user, the third natural language instruction input by the user is compared with the first natural language instruction, and the first intent text group and the second intent text group are updated according to the comparison result.

5. A signal integrity simulation system, characterized in that, include: The human-computer interaction terminal is used to receive the user's first natural language input commands; The simulation module includes a language model module and a simulation module; The processing end is used to obtain the execution scripts supported by the simulation tools of the simulation module, and the first mapping relationship between each execution script and each simulation process in the signal integrity simulation. The processing terminal is also used to acquire the text information of each of the execution scripts, and input the text information into the language model in the language model module for training, to obtain a first intention text group with a second mapping relationship with the execution script, a second intention text group with a third mapping relationship with the calling order of the execution script, and a third intention text group with a fourth mapping relationship with the parameter settings of the execution script. The processing terminal performs semantic parsing on the first natural language instruction input by the user through the human-computer interaction terminal using the language model to obtain a first intent instruction corresponding to the first natural language instruction. The first intent instruction includes a target first intent text and a target second intent text. It then generates calling code for the execution script based on the first intent instruction and sets simulation parameters for each execution script in the calling code based on the target third intent text included in the first intent instruction. Finally, it controls the simulation tool to execute the simulation process according to the calling code to obtain simulation results. The simulation results are analyzed to obtain key performance indicators; The simulation and analysis results will be output in the form of a structured report; The language model receives a second natural language instruction input by the user based on the structured report, parses the second natural language instruction, and generates a second intent instruction in combination with the first intent instruction; it then generates a new calling code based on the second intent instruction and controls the simulation tool to execute the simulation process according to the new calling code.

6. A computer device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the signal integrity simulation method according to any one of claims 1-4.

7. A computer storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instructions are executed by the processor, they implement the steps of the signal integrity simulation method according to any one of claims 1-4.

Images