Power supply integrity simulation optimization method, device, equipment and medium

Abstract

The invention discloses a power supply integrity simulation optimization method and device, equipment and a medium, and relates to the technical field of computer simulation, and the method comprises the steps: importing a to-be-simulated PCB design file into a simulation software tool; setting simulation parameters and constraint conditions in the simulation software tool, and operating the simulation software tool to perform power supply integrity simulation analysis and obtain a current simulation result; if the result is that the simulation test is not passed, inputting the current simulation result into a pre-trained power supply simulation dialogue large model to obtain a current optimization suggestion output by the model; the large power supply simulation dialogue model is obtained based on training of a training data set, and the training data set comprises historical simulation results and historical optimization adjustment measures; and adjusting the PCB design file based on the current optimization suggestion, importing the adjusted PCB design file into a simulation software tool for simulation analysis, and adding a current simulation result and a corresponding current optimization adjustment measure into a training data set after the simulation test passes.

Description

Technical Field

[0001] The present invention relates to the technical field of computer simulation, and particularly relates to a power integrity simulation optimization method, device, equipment and medium. Background Technique

[0002] With the development trend of market miniaturized products, the size of the circuit PCB (Printed Circuit Board) is getting smaller and more personalized, and the layout of each module on the board is quite compact. The power supply wiring is more irregular and the area is limited. In addition, with the development of artificial intelligence, the power consumption of chips on the PCB is getting higher and higher, and the requirements for PCB power supply design are becoming more stringent. Among them, PI (Power Integrity) simulation in PCB simulation is particularly important, which determines the reliability of the power supply design. PI simulation mainly analyzes the voltage drop and the current density of the power plane and vias, etc., and improves the power density and voltage drop according to the simulation results to meet the design requirements, such as widening the trace, changing the layout, increasing the number of vias, etc.

[0003] Traditional power integrity simulation software tools only output simulation results and do not output specific optimization solutions. Therefore, at present, R & D personnel need to give corresponding solutions according to the simulation results to optimize the PCB design, such as adjusting the trace width, increasing or decreasing the number of vias and capacitors, or changing the placement position of capacitors, etc., and then perform iterative simulation. However, when the optimization measures always fail to pass the simulation test, it is necessary to modify the same part of the circuit again or multiple times, and iterate multiple simulations to obtain a feasible design scheme, which seriously wastes the design and development time and results in low PCB design efficiency.

[0004] In summary, how to improve the simulation efficiency of power integrity, reduce the number of simulation iterations, and thus improve the design efficiency of the PCB board is an issue to be solved at present. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide a power integrity simulation optimization method, device, equipment and medium, which can improve the simulation efficiency of power integrity, reduce the number of simulation iterations, and thus improve the design efficiency of the PCB board. The specific scheme is as follows:

[0006] In the first aspect, the present application discloses a power integrity simulation optimization method, including:

[0007] Import the PCB design file to be simulated into a preset simulation software tool;

[0008] Set preset simulation parameters and preset constraint conditions in the simulation software tool, and run the simulation software tool to perform power integrity simulation analysis to obtain the current simulation result;

[0009] If the current simulation result indicates that the simulation test fails, input the current simulation result into a pre-trained power supply simulation dialogue large model to obtain the current optimization suggestions output by the power supply simulation dialogue large model; wherein, the power supply simulation dialogue large model is trained based on a training data set, and the training data set includes historical simulation results and corresponding historical optimization adjustment measures;

[0010] Adjust the PCB design file based on the current optimization suggestions, re-import the adjusted PCB design file into the simulation software tool for simulation analysis, and add the current simulation result and the current optimization adjustment measures obtained based on the current optimization suggestions to the training data set after the simulation test passes.

[0011] Optionally, the preset simulation parameters include any one or several of the conductivity of the PCB copper foil, the current-carrying capacity parameter of the via hole, the pin positions of the power supply end and the load end, the power supply voltage, the load current, and the impedance control parameter; the preset constraint conditions include any one or several of the maximum value of the voltage drop, the limit value of the via hole current density, the highest temperature value of the hot spot area, and the limit value of the ripple noise.

[0012] Optionally, the process of training the power supply simulation dialogue large model based on the training data set includes:

[0013] Obtain the training data set; wherein, the training data set includes historical simulation results, historical optimization adjustment measures corresponding to the historical simulation results, historical simulation parameters, and historical constraint conditions, and the historical simulation results are the simulation results of failed simulation tests exported from the simulation software tool;

[0014] Preprocess the training data set based on the data preprocessing rules to obtain the processed training data set, and perform structured processing on the processed training data set to extract key parameters and compress and generate instruction template data in a preset fixed format;

[0015] Construct a supervised fine-tuning instruction data set based on the instruction template data; wherein, each instruction data in the instruction data set includes instruction information, input parameters, and output parameters; the instruction information corresponds to the relevant problem description, the input parameters correspond to the historical simulation results, historical simulation parameters, and historical constraint conditions, and the output parameters correspond to the historical optimization adjustment measures;

[0016] Determine the initial model; wherein, the initial model is constructed based on the Qwen2.5 model;

[0017] Use the instruction data set to perform supervised fine-tuning training on the initial model to obtain the trained power supply simulation dialogue large model.

[0018] Optionally, preprocess the training data set based on data preprocessing rules to obtain a processed training data set, including:

[0019] Clean the training data set based on preset data cleaning rules to obtain a cleaned training data set; the preset data cleaning rules include data deduplication and abnormal data filtering;

[0020] Convert the data format of each training data in the cleaned training data set into a target data format recognizable by the initial model to obtain a processed training data set.

[0021] Optionally, input the current simulation result into a pre-trained power supply simulation dialogue large model to obtain the current optimization suggestions output by the power supply simulation dialogue large model, including:

[0022] Extract feature information from the current simulation result and construct a current problem description;

[0023] Input the current problem description, feature information, preset simulation parameters, and preset constraint conditions into the pre-trained power supply simulation dialogue large model to obtain the current optimization suggestions output by the power supply simulation dialogue large model.

[0024] Optionally, the current optimization suggestions include at least one of widening the power supply plane, increasing the number of decoupling capacitors, adjusting the capacitor position, increasing the number of vias, adjusting the trace width, and modifying the PCB stack-up design.

[0025] Optionally, after re-importing the adjusted PCB design file into the simulation software tool for simulation analysis, it further includes:

[0026] Obtain the simulation analysis result corresponding to the adjusted PCB design file;

[0027] If the simulation analysis result indicates that the simulation test fails, then use the simulation analysis result as the new current simulation result, and then jump back to the step of inputting the current simulation result into the pre-trained power supply simulation dialogue large model until the simulation test passes.

[0028] In a second aspect, the present application discloses a power integrity simulation optimization device, including:

[0029] A file import module for importing the PCB design file to be simulated into a preset simulation software tool;

[0030] A result acquisition module for setting preset simulation parameters and preset constraint conditions in the simulation software tool, and running the simulation software tool to perform power integrity simulation analysis to obtain the current simulation result;

[0031] An optimization suggestion acquisition module, configured to input the current simulation result into a pre-trained power supply simulation dialogue large model if the current simulation result indicates that the simulation test fails, so as to obtain the current optimization suggestion output by the power supply simulation dialogue large model; wherein, the power supply simulation dialogue large model is trained based on a training data set, and the training data set includes historical simulation results and corresponding historical optimization adjustment measures;

[0032] An optimization adjustment module, configured to adjust the PCB design file based on the current optimization suggestion, re-import the adjusted PCB design file into the simulation software tool for simulation analysis, and add the current simulation result and the current optimization adjustment measure obtained based on the current optimization suggestion to the training data set after the simulation test passes.

[0033] In a third aspect, the present application discloses an electronic device, including:

[0034] A memory, configured to store a computer program;

[0035] A processor, configured to execute the computer program to implement the steps of the power integrity simulation optimization method disclosed above.

[0036] In a fourth aspect, the present application discloses a computer-readable storage medium, configured to store a computer program; wherein, when the computer program is executed by a processor, the steps of the power integrity simulation optimization method disclosed above are implemented.

[0037] It can be seen that the present application imports the PCB design file to be simulated into a preset simulation software tool; sets preset simulation parameters and preset constraint conditions in the simulation software tool, and runs the simulation software tool to perform power integrity simulation analysis to obtain the current simulation result; if the current simulation result indicates that the simulation test fails, the current simulation result is input into a pre-trained power supply simulation dialogue large model to obtain the current optimization suggestion output by the power supply simulation dialogue large model; wherein, the power supply simulation dialogue large model is trained based on a training data set, and the training data set includes historical simulation results and corresponding historical optimization adjustment measures; adjusts the PCB design file based on the current optimization suggestion, re-imports the adjusted PCB design file into the simulation software tool for simulation analysis, and adds the current simulation result and the current optimization adjustment measure obtained based on the current optimization suggestion to the training data set after the simulation test passes.

[0038] Beneficial effects: In this application, when importing the PCB design file to be simulated into a preset simulation software tool, setting preset simulation parameters and preset constraint conditions in the simulation software tool, and then running the simulation software tool to perform power integrity simulation analysis to obtain the current simulation result, if the current simulation result indicates that the simulation test fails, the current simulation result is input into a pre-trained power simulation dialogue large model, so as to obtain the current optimization suggestion output by the power simulation dialogue large model. It should be noted that this application pre-constructed a training dataset based on historical simulation results and corresponding historical optimization adjustment measures, and thus trained a power simulation dialogue large model based on the training dataset, enabling the power simulation dialogue large model to output corresponding optimization suggestions according to real-time simulation results. Further, this application adjusts the PCB design file based on the current optimization suggestion output by the large model, and re-imports the adjusted PCB design file into the simulation software tool for simulation analysis. That is to say, this application no longer requires researchers to manually give optimization adjustment measures according to simulation results, but uses the power simulation dialogue large model to learn the relationship between simulation results and design parameters based on historical simulation results and corresponding historical optimization adjustment measures, so as to intelligently output optimization suggestions. Compared with manual optimization, the solution of this application is more comprehensive, more accurate, and more targeted, can improve the simulation efficiency of power integrity, reduce the number of simulation iterations, and thus can further improve the design quality and design efficiency of the PCB board, and enhance the performance and reliability of the power supply. Finally, when the simulation test passes, the current simulation result and the current optimization adjustment measures obtained based on the current optimization suggestion are added to the training dataset to update the model parameters in real time. Description of the Drawings

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to the provided drawings.

[0040] Figure 1 It is a flowchart of a power integrity simulation optimization method disclosed in the present application;

[0041] Figure 2 It is a schematic diagram of an intelligent power integrity simulation optimization process disclosed in the present application;

[0042] Figure 3 It is a training flowchart of a power simulation dialogue large model disclosed in the present application;

[0043] Figure 4 It is a schematic diagram of the structure of a power integrity simulation optimization device disclosed in the present application;

[0044] Figure 5 This is a structural diagram of an electronic device disclosed in the present application. Specific implementation manners

[0045] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0046] Traditional power integrity simulation software tools only output simulation results and do not output specific optimization solutions. Therefore, at present, R & D personnel need to give corresponding solutions according to the simulation results to optimize the PCB design, such as adjusting the trace width, increasing or decreasing the number of vias and capacitors, or changing the placement position of capacitors, etc., and then perform iterative simulation. However, when the optimization measures still cannot pass the simulation test, it is necessary to modify the same part of the circuit again or multiple times, and iterate multiple simulations to obtain a feasible design solution, which seriously wastes the design and development time and results in low PCB design efficiency.

[0047] Therefore, the embodiments of the present application disclose a power integrity simulation optimization method, device, equipment and medium, which can improve the simulation efficiency of power integrity, reduce the number of simulation iterations, and thus improve the design efficiency of the PCB board.

[0048] See Figure 1 and Figure 2 As shown, the embodiments of the present application disclose a power integrity simulation optimization method, which includes:

[0049] Step S11: Import the PCB design file to be simulated into a preset simulation software tool.

[0050] In this embodiment, first, it is necessary to clarify which part of the power network needs to be simulated currently, and prepare PI simulation parameters, including information such as the pin positions of the VRM (Voltage Regulator Module, that is, the power supply end) and the SINK end (load section or power consumption end), the target voltage, the load current, etc.; find out the impedance information of devices such as series resistors or programmable electronic fuses and inductors in the power path clearly for setting simulation circuit parameters, etc.; then import the PCB design file to be simulated into a preset simulation software tool, where the simulation software tool can specifically be the cadence power DC tool.

[0051] Step S12: Set preset simulation parameters and preset constraint conditions in the simulation software tool, and run the simulation software tool to perform power integrity simulation analysis to obtain the current simulation results.

[0052] In this embodiment, after setting the preset simulation parameters and preset constraint conditions in the simulation software tool and then running the simulation software tool to perform power integrity simulation analysis to obtain the current simulation results, each test item in the simulation results includes voltage drop, plane current density, via current density, etc. Among them, the preset simulation parameters include any one or several of the conductivity of the PCB copper foil, the current-carrying capacity parameters of the vias, the pin positions of the power supply end and the load end, the power supply voltage, the load current, and the impedance control parameters; the preset constraint conditions include any one or several of the maximum value of the voltage drop, the limit value of the via current density, the highest temperature value of the hot spot area, and the limit value of the ripple noise.

[0053] That is, the embodiments of the present application need to set the simulation parameters and the constraint conditions of the simulation results: such as the setting of the conductivity of the PCB copper foil, which is used to calculate the current distribution and voltage drop. For example, the conductivity of pure copper is about 5.8×10 7 S / m; the setting of the current-carrying capacity of the vias, which defines the current-carrying capacity of the vias to ensure reliability in high-current scenarios, such as the via diameter, copper plating thickness, etc.; the VRM and SINK pin positions, the power supply voltage is 1.2V, the load current is set to 2A, etc.; the setting of the impedance control parameters of the series devices in the circuit, which sets the impedance target of the power plane. For example, the impedance in the low-frequency band is ≤10mΩ to ensure signal integrity. In addition, regarding the constraint requirements for voltage drop, via current density, hot spot area temperature, ripple noise, etc.; for example, the voltage drop cannot exceed 5%, the plane current density cannot exceed 0.06A / mm², the via current density cannot exceed 0.04A / mm², the highest temperature of the hot spot area cannot exceed 100°C, and the ripple amplitude of the power network cannot exceed 50mV.

[0054] Step S13: If the current simulation results indicate that the simulation test fails, input the current simulation results into a pre-trained power simulation dialogue large model to obtain the current optimization suggestions output by the power simulation dialogue large model; among them, the power simulation dialogue large model is trained based on a training data set, and the training data set includes historical simulation results and corresponding historical optimization adjustment measures.

[0055] In this embodiment, the simulation results are compared with the preset constraint conditions set in advance to determine whether the simulation test passes. Among them, if the current simulation results all meet the preset constraint conditions, it indicates that the simulation test passes, that is, it meets the power supply design requirements, and the simulation test ends; if there are test items that do not meet the preset constraint conditions, it means that the simulation test fails.

[0056] In the traditional solution, for the simulation results that fail, the R & D personnel need to optimize the layout design of this version according to the specific fail value and their own design experience, re-import the optimized layout design into the simulation software tool, repeat the above simulation work, and check whether the simulation results output by this version of the layout design meet the power supply design requirements. If they meet, the simulation is terminated immediately. If they still do not meet, the current version of the layout design needs to be further optimized, and the simulation verification is repeated iteratively until the optimized layout design meets all the constraints of the power supply simulation. Due to the relatively cramped layout of the current board and the strong irregularity of the power plane, this traditional optimization design places high requirements on the R & D personnel, has a high simulation iteration rate, a relatively long R & D time, and a high degree of uncertainty.

[0057] In this application, the current simulation results that fail the simulation test are input into a pre-trained power supply simulation dialogue large model to obtain the current optimization suggestions output by the power supply simulation dialogue large model. It should be noted that in this application, a training data set is pre-constructed based on historical simulation results and corresponding historical optimization adjustment measures, and the power supply simulation dialogue large model is trained based on the training data set, so that the power supply simulation dialogue large model can output corresponding optimization suggestions according to the real-time simulation results.

[0058] In the specific implementation manner, as Figure 3 shown, the process of training the power supply simulation dialogue large model based on the training data set includes the following steps:

[0059] Step S131: Obtain the training data set; among them, the training data set includes historical simulation results, historical optimization adjustment measures corresponding to the historical simulation results, historical simulation parameters, and historical constraint conditions. The historical simulation results are the simulation results that fail the simulation test exported from the simulation software tool.

[0060] That is, in this embodiment, it is first necessary to construct a training data set, which includes historical simulation results and historical optimization adjustment measures corresponding to the historical simulation results; among them, the historical simulation results specifically refer to the simulation results that fail the simulation test exported from the simulation software tool. In addition, the training data set can also include historical simulation parameters and historical constraint conditions corresponding to the historical simulation results.

[0061] Step S132: Preprocess the training data set based on the data preprocessing rules to obtain the processed training data set, and perform structured processing on the processed training data set to extract key parameters and compress and generate instruction template data in a preset fixed format.

[0062] In this embodiment, after obtaining the training data set, it is necessary to preprocess the training data set to obtain the processed training data set. Then, the processed training data set is structured to extract key parameters and compress and generate instruction template data in a preset fixed format. It can be understood that the data input by the user may be a long description, and through the structured processing, key parameters such as "current density = 115 A / mm², constraint = 30 A / mm²" can be automatically extracted, so as to generate instruction template data in a fixed format (i.e., Prompt).

[0063] For example, the original log is: "2023-1-1 14:00:00 WARNING: Current density value 115 A / mm²@R12 (constraint 30 A / mm²)";

[0064] Structured Prompt: "Question: Current density exceeds the standard (115 > 30); Location: R12; Current parameters: line width = 0.5 mm, copper thickness = 0.035 mm".

[0065] In the specific implementation manner, the training data set is preprocessed based on the data preprocessing rules to obtain the processed training data set, including: cleaning the training data set based on the preset data cleaning rules to obtain the cleaned training data set; the preset data cleaning rules include data deduplication processing and abnormal data filtering processing; converting the data formats of the training data in the cleaned training data set into the target data formats recognizable by the initial model to obtain the processed training data set. That is to say, the preprocessing process specifically includes data cleaning and format conversion. First, data deduplication processing and abnormal data filtering processing are performed on the training data set, that is, invalid, duplicate, or incorrect data records are eliminated to complete data cleaning and obtain the cleaned training data set. Then, the data formats of the training data in the cleaned training data set are converted into the target data formats recognizable by the initial model to obtain the processed training data set. In addition, the data preprocessing rules may also include feature engineering, standardization / normalization, data augmentation, structured processing, etc.

[0066] Step S133: Construct a supervised fine-tuning instruction data set based on the instruction template data; wherein, each instruction data in the instruction data set includes instruction information, input parameters, and output parameters; the instruction information corresponds to the relevant problem description, the input parameters correspond to the historical simulation results, historical simulation parameters, and historical constraint conditions, and the output parameters correspond to the historical optimization adjustment measures.

[0067] In this embodiment, it is necessary to convert the instruction template data into "question-answer" instruction pairs that can be understood by the model to adapt to the requirements of supervised fine-tuning (SFT). Among them, each piece of instruction data in the instruction dataset includes instruction information, input parameters, and output parameters; the instruction information corresponds to the relevant problem description, the input parameters correspond to the historical simulation results, historical simulation parameters, and historical constraint conditions, and the output parameters correspond to the historical optimization adjustment measures.

[0068] That is, each piece of data contains "Instruction" + "Input" + "Output", for example:

[0069] "instruction": "Generate optimization suggestions based on the current density simulation results";

[0070] "input": "Current density = 114.29 A / mm² (constraint value 30 A / mm²), line width = 0.5mm, copper thickness = 0.035mm";

[0071] "output": "Increase the line width to 2mm, or keep the line width at 1mm but increase the copper thickness to 0.07mm (2 ounces)".

[0072] The above process is also the annotation process of the training data. In addition, in addition to the above annotation information, information such as the product type, application scenario, and design difficulty level corresponding to each data sample can also be added. In this way, during model training, the model can better understand the relationships and applicable ranges between different data, improving the accuracy of the optimization suggestions output by the model. And, power simulation data of different types of electronic products (such as consumer electronics, industrial control equipment, automotive electronics, etc.) can be further collected, because the power design requirements and problems faced by different types of products vary greatly, and adding this data can enable the model to learn more extensive application scenarios.

[0073] Step S134: Determine the initial model; among them, the initial model is constructed based on the Qwen2.5 model.

[0074] In this embodiment, the initial model is constructed based on the Qwen2.5 (Tongyi Qianwen open-source base model) model.

[0075] Step S135: Use the instruction dataset to perform supervised fine-tuning training on the initial model to obtain the trained power simulation dialogue large model.

[0076] In this embodiment, the initial model is supervised fine-tuned using an instruction dataset, with the aim of enabling the model to learn to map from simulation results (input) to optimization suggestions (output), thereby obtaining the trained power simulation dialogue large model. Additionally, during the model training process, the learning rate needs to be set to avoid overfitting, and the batch size needs to be set to balance the video memory and training efficiency.

[0077] In the specific implementation, the current simulation result is input into the pre-trained power simulation dialogue large model to obtain the current optimization suggestion output by the power simulation dialogue large model, including: extracting feature information from the current simulation result and constructing the current problem description; inputting the current problem description, feature information, preset simulation parameters, and preset constraint conditions into the pre-trained power simulation dialogue large model to obtain the current optimization suggestion output by the power simulation dialogue large model. That is, this application needs to extract feature information from the current simulation result to construct the current problem description, and then use the current problem description, feature information, preset simulation parameters, and preset constraint conditions as instructions and input parameters and input them into the pre-trained power simulation dialogue large model, so as to obtain the current optimization suggestion output by the power simulation dialogue large model. This optimization suggestion is the optimal solution selected through internal comparison, and the simulation passing rate is high, such as widening a certain part of the power plane to a certain width, adding decoupling capacitors, adding via designs, etc.

[0078] Among them, the current optimization suggestions include but are not limited to widening the power plane, increasing the number of decoupling capacitors, adjusting the capacitor position, increasing the number of vias, adjusting the trace width, modifying the PCB stack-up design, and so on.

[0079] Step S14: Adjust the PCB design file based on the current optimization suggestion, re-import the adjusted PCB design file into the simulation software tool for simulation analysis, and add the current simulation result and the current optimization adjustment measures obtained based on the current optimization suggestion to the training dataset after the simulation test passes.

[0080] In this embodiment, the PCB design file is adjusted based on the current optimization suggestions output by the large model, such as modifying the power network layout, increasing the number of vias, adjusting the capacitor position, etc., and the adjusted PCB design file is re-imported into the simulation software tool for simulation analysis. That is, this application no longer requires researchers to manually give optimization adjustment measures according to the simulation results, but uses the power simulation dialogue large model to learn the relationship between the simulation results and design parameters based on the historical simulation results and the corresponding historical optimization adjustment measures, so as to intelligently output optimization suggestions. Compared with manual optimization, the solution of this application is more comprehensive, more accurate, and more targeted, which can improve the simulation efficiency of power integrity, reduce the number of simulation iterations, and thus can further improve the design quality and design efficiency of the PCB board, and enhance the performance and reliability of the power supply. Finally, when the simulation test passes, the current simulation results and the current optimization adjustment measures obtained based on the current optimization suggestions are added to the training data set to update the model parameters in real time.

[0081] It should be noted that after re-importing the adjusted PCB design file into the simulation software tool for simulation analysis, it further includes: obtaining the simulation analysis results corresponding to the adjusted PCB design file; if the simulation analysis results indicate that the simulation test fails, then taking the simulation analysis results as the new current simulation results, and then re-jumping to the step of inputting the current simulation results into the pre-trained power simulation dialogue large model until the simulation test passes. That is, when the adjusted PCB design file is imported into the simulation software tool, if the simulation test still fails, then inputting the new current simulation results into the pre-trained power simulation dialogue large model to obtain new optimization suggestions until the simulation test passes.

[0082] It can be seen that in this application, when the PCB design file to be simulated is imported into a preset simulation software tool, preset simulation parameters and preset constraint conditions are set in the simulation software tool, and then the simulation software tool is run to perform power integrity simulation analysis to obtain the current simulation result. If the current simulation result indicates that the simulation test fails, the current simulation result is input into a pre-trained power simulation dialogue large model, so as to obtain the current optimization suggestion output by the power simulation dialogue large model. It should be noted that in this application, a training data set is pre-constructed based on historical simulation results and corresponding historical optimization adjustment measures, and the power simulation dialogue large model is trained based on the training data set, so that the power simulation dialogue large model can output corresponding optimization suggestions according to real-time simulation results. Further, this application adjusts the PCB design file based on the current optimization suggestion output by the large model, and re-imports the adjusted PCB design file into the simulation software tool for simulation analysis. That is, this application does not need to manually give optimization adjustment measures by researchers according to the simulation results, but uses the power simulation dialogue large model to learn the relationship between simulation results and design parameters based on historical simulation results and corresponding historical optimization adjustment measures, so as to intelligently output optimization suggestions. Compared with manual optimization, the solution of this application is more comprehensive, more accurate and more targeted, can improve the simulation efficiency of power integrity, reduce the number of simulation iterations, and thus can further improve the design quality and design efficiency of the PCB board, and enhance the performance and reliability of the power supply. Finally, when the simulation test passes, the current simulation result and the current optimization adjustment measure obtained based on the current optimization suggestion are added to the training data set to update the model parameters in real time.

[0083] The following takes optimizing the current density of the PCB power line as an example to illustrate the above content in detail:

[0084] In PCB design, reasonable current density is crucial for ensuring the performance and reliability of the circuit board;

[0085] Current density is the magnitude of the current passing through a unit area, and the calculation formula is J = I / A;

[0086] Among them, J is the current density, and the unit is amperes per square millimeter (A / mm²);

[0087] I is the current passing through the conductor, and the unit is amperes (A);

[0088] A is the cross-sectional area of the conductor, and the unit is square millimeters (mm²);

[0089] Among them, the cross-sectional area A can be calculated by the wire width W and the copper foil thickness T: A = W × T; where W is the wire width in millimeters (mm), and T is the PCB copper foil thickness in millimeters (mm). Usually, the thickness of 1 ounce of copper foil is about 0.035 mm;

[0090] In PI simulation, the general power density constraint value is 30 A / mm². Once this value is exceeded, the PCB needs to be optimized, which can effectively avoid problems such as overheating and electromigration, and improve the reliability and service life of the PCB.

[0091] There is a certain section of PCB circuit that needs to carry a current of 2 A. The circuit width is 0.5 mm, and the copper foil thickness is 35 μm (i.e., 0.035 mm). Then the circuit cross-sectional area is: A = 0.5 × 0.035 = 0.0175 mm². The current density is: J = 2 / 0.0175 ≈ 114.29 A / mm². This value has obviously exceeded the constraint value. Traditionally, when R & D personnel obtain the current density value and fail flag here in the simulation software, they will manually calculate or optimize the trace width according to the empirical value. After the optimized PCB is given to the simulation software for re-simulation, if there is still a deviation, the simulation will be iterated until the simulation result passes the test. However, through the design of the present invention, the number of iterations will be greatly reduced. The simulation results and circuit parameters will be input into the trained power simulation dialogue large model, and then targeted PCB optimization measures can be output, such as adjusting the trace width to 2 mm, or adjusting the trace width to 1 mm and increasing the copper thickness to 2 ounces, etc. This optimization measure is fast and accurate for R & D personnel to refer to.

[0092] In a more complex environment such as irregular power traces, this large model can still give accurate and targeted optimization solutions, which are more efficient and accurate than manually checking and calculating to obtain optimization solutions. For complex environments, such as the same power plane from the VRM power output end to the SINK power consumption end, the trace width is not the same width. In fact, the line width will be adjusted multiple times according to the actual board wiring situation, while manual calculation can only calculate an approximate value, which is not accurate. However, the large model will give very accurate optimization measures according to the board situation, greatly improving the R & D efficiency and quality.

[0093] See Figure 4 As shown in the figure, an embodiment of the present application discloses a power integrity simulation optimization device, which includes:

[0094] A file import module 11, configured to import the PCB design file to be simulated into a preset simulation software tool;

[0095] A result acquisition module 12, configured to set preset simulation parameters and preset constraint conditions in the simulation software tool, and run the simulation software tool to perform power integrity simulation analysis to obtain the current simulation result;

[0096] An optimization suggestion acquisition module 13, configured to, if the current simulation result indicates that the simulation test fails, input the current simulation result into a pre-trained power supply simulation dialogue large model to obtain the current optimization suggestion output by the power supply simulation dialogue large model; wherein, the power supply simulation dialogue large model is trained based on a training data set, and the training data set includes historical simulation results and corresponding historical optimization adjustment measures;

[0097] An optimization adjustment module 14, configured to adjust the PCB design file based on the current optimization suggestion, re-import the adjusted PCB design file into the simulation software tool for simulation analysis, and add the current simulation result and the current optimization adjustment measure obtained based on the current optimization suggestion to the training data set after the simulation test passes.

[0098] Since the embodiments of the device part correspond to the above-mentioned embodiments, the embodiments of the device part are described with reference to the embodiments of the above-mentioned method part and will not be elaborated here.

[0099] It can be seen that in this application, after importing the PCB design file to be simulated into a preset simulation software tool, setting preset simulation parameters and preset constraint conditions in the simulation software tool, and then running the simulation software tool to perform power integrity simulation analysis to obtain the current simulation result, if the current simulation result indicates that the simulation test fails, the current simulation result is input into a pre-trained power supply simulation dialogue large model, so as to obtain the current optimization suggestion output by the power supply simulation dialogue large model. It should be noted that in this application, a training data set is pre-constructed based on historical simulation results and corresponding historical optimization adjustment measures, and the power supply simulation dialogue large model is trained based on the training data set, so that the power supply simulation dialogue large model can output corresponding optimization suggestions according to the real-time simulation result. Further, in this application, the PCB design file is adjusted based on the current optimization suggestion output by the large model, and the adjusted PCB design file is re-imported into the simulation software tool for simulation analysis, that is, in this application, it is no longer necessary for researchers to manually give optimization adjustment measures according to the simulation result, but the power supply simulation dialogue large model is used to learn the relationship between the simulation result and the design parameters according to the historical simulation result and the corresponding historical optimization adjustment measure, so as to intelligently output optimization suggestions. Compared with manual optimization, the solution of this application is more comprehensive, more accurate, and more targeted, can improve the simulation efficiency of power integrity, reduce the number of simulation iterations, and thus can further improve the design quality and design efficiency of the PCB board, and improve the performance and reliability of the power supply. Finally, when the simulation test passes, the current simulation result and the current optimization adjustment measure obtained based on the current optimization suggestion are added to the training data set to update the model parameters in real time.

[0100] Figure 5A schematic structural diagram of an electronic device provided by an embodiment of the present application. Specifically, it may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. Among them, the memory 22 is used to store a computer program, and the computer program is loaded and executed by the processor 21 to implement the relevant steps in the power integrity simulation optimization method executed by the electronic device disclosed in any of the foregoing embodiments.

[0101] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows is any communication protocol applicable to the technical solution of the present application, and no specific limitation is imposed on it here; the input / output interface 25 is used to obtain external input data or output data to the outside, and its specific interface type can be selected according to specific application needs, and no specific limitation is made here.

[0102] Among them, the processor 21 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 21 can be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 21 may also include a main processor and a coprocessor. The main processor is a processor used to process data in the wake state, also known as the CPU (Central Processing Unit); the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 21 may be integrated with a GPU (Graphics Processing Unit), and the GPU is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 21 may also include an AI (Artificial Intelligence) processor, and the AI processor is used to process computing operations related to machine learning.

[0103] In addition, as a carrier for resource storage, the memory 22 may be a read-only memory, a random access memory, a disk, or an optical disc, etc. The resources stored thereon include an operating system 221, a computer program 222, data 223, etc., and the storage method may be temporary storage or permanent storage.

[0104] Among them, the operating system 221 is used to manage and control each hardware device and computer program 222 on the electronic device 20, so as to implement the operation and processing of the massive data 223 in the memory 22 by the processor 21. It can be Windows, Unix, Linux, etc. In addition to the computer program that can be used to complete the power integrity simulation optimization method executed by the electronic device 20 disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs that can be used to complete other specific tasks. The data 223 may include not only the data transmitted by external devices received by the electronic device, but also the data collected by its own input / output interface 25, etc.

[0105] Furthermore, an embodiment of the present application also discloses a computer-readable storage medium. When the computer program stored in the storage medium is loaded and executed by a processor, the steps of the power integrity simulation optimization method disclosed in any of the foregoing embodiments are implemented.

[0106] An embodiment of the present invention also discloses a computer program product, including a computer program / instructions. When the computer program / instructions are executed by a processor, the steps of the power integrity simulation optimization method disclosed in any of the foregoing embodiments are implemented.

[0107] In this specification, each embodiment is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple. For the relevant parts, refer to the description of the method part.

[0108] Those skilled in the art can further realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0109] The steps of the methods or algorithms described in connection with the embodiments disclosed herein may be implemented directly in hardware, in software modules executed by a processor, or in a combination thereof. The software modules may be located in a random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, a hard disk, a removable disk, a compact disc read-only memory (CD-ROM), or any other form of storage medium well known in the art.

[0110] Finally, it should also be noted that in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0111] The above has introduced in detail a power integrity simulation optimization method, apparatus, device and storage medium provided by the present invention. Specific examples are used herein to illustrate the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A power integrity simulation optimization method, characterized in that Including: Import the PCB design file to be simulated into a preset simulation software tool; Set preset simulation parameters and preset constraint conditions in the simulation software tool, and run the simulation software tool to perform power integrity simulation analysis to obtain the current simulation result; If the current simulation result indicates that the simulation test fails, input the current simulation result into a pre-trained power simulation dialogue large model to obtain the current optimization suggestions output by the power simulation dialogue large model; wherein, the power simulation dialogue large model is trained based on a training data set, and the training data set includes historical simulation results and corresponding historical optimization adjustment measures; Adjust the PCB design file based on the current optimization suggestions, re-import the adjusted PCB design file into the simulation software tool for simulation analysis, and add the current simulation result and the current optimization adjustment measures obtained based on the current optimization suggestions to the training data set after the simulation test passes.

2. The power integrity simulation optimization method according to claim 1, wherein The preset simulation parameters include any one or several of the conductivity of the PCB copper foil, the current-carrying capacity parameter of the via, the pin positions of the power supply end and the load end, the power supply voltage, the load current, and the impedance control parameter; the preset constraint conditions include any one or several of the maximum value of the voltage drop, the limit value of the via current density, the highest temperature value of the hot spot area, and the limit value of the ripple noise.

3. The power integrity simulation optimization method according to claim 1, wherein The process of training the power simulation dialogue large model based on the training data set includes: Obtain a training data set; wherein, the training data set includes historical simulation results, historical optimization adjustment measures corresponding to the historical simulation results, historical simulation parameters, and historical constraint conditions, and the historical simulation results are simulation results of failed simulation tests exported from the simulation software tool; Preprocess the training data set based on data preprocessing rules to obtain a processed training data set, and perform structured processing on the processed training data set to extract key parameters and compress and generate instruction template data in a preset fixed format; Construct a supervised fine-tuning instruction data set based on the instruction template data; wherein, each instruction data in the instruction data set includes instruction information, input parameters, and output parameters; the instruction information corresponds to the relevant problem description, the input parameters correspond to the historical simulation results, the historical simulation parameters, and the historical constraint conditions, and the output parameters correspond to the historical optimization adjustment measures; Determine an initial model; wherein, the initial model is constructed based on the Qwen2.5 model; Use the instruction data set to perform supervised fine-tuning training on the initial model to obtain the trained power simulation dialogue large model.

4. The power integrity simulation optimization method according to claim 3, wherein The preprocessing of the training data set based on the data preprocessing rules to obtain a processed training data set includes: Clean the training data set based on preset data cleaning rules to obtain a cleaned training data set; the preset data cleaning rules include data deduplication processing and abnormal data filtering processing; Convert the data format of each piece of training data in the cleaned training dataset into the target data format recognizable by the initial model to obtain the processed training dataset.

5. The power integrity simulation optimization method according to claim 3, characterized in that, The step of inputting the current simulation result into a pre-trained power supply simulation dialogue large model to obtain the current optimization suggestions output by the power supply simulation dialogue large model includes: Extract feature information from the current simulation result and construct a current problem description; Input the current problem description, the feature information, the preset simulation parameters, and the preset constraint conditions into a pre-trained power supply simulation dialogue large model to obtain the current optimization suggestions output by the power supply simulation dialogue large model.

6. The power integrity simulation optimization method according to claim 1, characterized in that The current optimization suggestions include at least one of widening the power plane, increasing the number of decoupling capacitors, adjusting the capacitor position, increasing the number of vias, adjusting the trace width, and modifying the PCB stack-up design.

7. The power integrity simulation optimization method according to any one of claims 1 to 6, characterized in that, After re-importing the adjusted PCB design file into the simulation software tool for simulation analysis, it further includes: Obtain the simulation analysis result corresponding to the adjusted PCB design file; If the simulation analysis result indicates that the simulation test fails, use the simulation analysis result as the new current simulation result, and then jump back to the step of inputting the current simulation result into the pre-trained power supply simulation dialogue large model until the simulation test passes.

8. A power integrity simulation optimization device, characterized in that, It includes: A file import module for importing a PCB design file to be simulated into a preset simulation software tool; A result acquisition module for setting preset simulation parameters and preset constraint conditions in the simulation software tool and running the simulation software tool to perform power integrity simulation analysis to obtain the current simulation result; An optimization suggestion acquisition module for, if the current simulation result indicates that the simulation test fails, inputting the current simulation result into a pre-trained power supply simulation dialogue large model to obtain the current optimization suggestions output by the power supply simulation dialogue large model; wherein, the power supply simulation dialogue large model is trained based on a training dataset, and the training dataset includes historical simulation results and corresponding historical optimization adjustment measures; An optimization adjustment module for adjusting the PCB design file based on the current optimization suggestions, re-importing the adjusted PCB design file into the simulation software tool for simulation analysis, and adding the current simulation result and the current optimization adjustment measures obtained based on the current optimization suggestions to the training dataset after the simulation test passes.

9. An electronic device, characterized in that, It includes: A memory for storing a computer program; A processor for executing the computer program to implement the steps of the power integrity simulation optimization method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, For storing a computer program; wherein, when the computer program is executed by the processor, it implements the steps of the power integrity simulation optimization method according to any one of claims 1 to 7.

Images