Simulation methods and systems for computer-aided design simulation platforms for integrated circuits

By introducing intelligent experimental design modes and continuous range interval settings, the integrated circuit computer-aided design simulation platform solves the problem of low simulation efficiency and achieves more efficient parameter exploration and design optimization.

CN123117003APending Publication Date: 2026-11-17BEIJING SMARTCHIP MICROELECTRONICS TECHNOLOGY CO LTD +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611110174.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-24
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing computer-aided design simulation platforms for integrated circuits suffer from low simulation efficiency, huge consumption of computing resources, rigid parameter input methods, lack of automated optimization mechanisms, difficulty in systematically exploring complex coupling effects in multi-parameter space, and cumbersome operation, making them unable to adapt to design change requirements.

Method used

It introduces an intelligent experimental design mode, supports setting continuous range intervals as input parameters, performs automated experimental planning through a configurable algorithm engine, generates a set of simulation experimental points, and provides a full-factor simulation mode and an intelligent experimental design mode, including parameter screening, design space exploration, and performance index optimization.

Benefits of technology

It improves simulation efficiency, reduces manual discretization steps, enhances the agility and parameter flexibility of design exploration, and enables a more comprehensive exploration of the multi-parameter space to discover potential design optimization solutions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123117003A_ABST
    Figure CN123117003A_ABST
Patent Text Reader

Abstract

This disclosure relates to the field of computer-aided design technology for integrated circuits, specifically to a simulation method and system for an integrated circuit computer-aided design simulation platform. The simulation method includes: providing a simulation flow definition interface; receiving parameter types set by a user for input parameters of at least one simulation module; responding to a user's selection instruction for a simulation mode, invoking a preset algorithm engine corresponding to the selected simulation mode, and automatically generating a set of simulation experimental points based on the parameter types of the input parameters; executing the simulation tasks corresponding to the set of simulation experimental points, and outputting corresponding simulation results; wherein, when the selected simulation mode is an intelligent experimental design mode, at least one input parameter is set to a continuous range interval. The above technical solution improves simulation efficiency and provides a foundation for a more comprehensive exploration of the multi-parameter space, helping to discover potential design optimization schemes.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This disclosure relates to the field of computer-aided design technology for integrated circuits, and specifically to a simulation method and system for an integrated circuit computer-aided design simulation platform. Background Technology

[0002] In the field of semiconductor process and device design, integrated circuit computer-aided design simulation platforms are widely used as key tools in the development and optimization of RF, power and other semiconductor devices.

[0003] In existing technologies, platforms such as Synopsys' Sentaurus Workbench integrate modules for process simulation and device simulation through a graphical interface, allowing users to build simulation workflows and execute experiments. Users need to manually specify a discrete list of values ​​for each input parameter in a table interface, such as the specific values ​​for device width and length. The platform then uses a full factorial design method to generate a set of experimental points for all parameter combinations and runs the simulation.

[0004] However, this method suffers from several drawbacks: the number of simulations increases exponentially with the number of parameters and levels, leading to enormous computational resource consumption and lengthy simulation cycles, severely restricting design efficiency; users are forced to subjectively select a limited number of parameters for experiments, making it impossible to systematically explore the complex coupling effects in the multi-parameter space, easily overlooking globally optimal design points and drawing biased conclusions; parameter input only supports predefined discrete value lists, not direct definition of continuous range intervals, making the operation cumbersome and difficult to adapt to design change requirements. The key to these drawbacks is that the platform essentially functions only as an experimental executor, lacking a built-in automated optimization mechanism, and design optimization heavily relies on repeated trial and error based on human experience. These shortcomings result in significant deficiencies in the existing technology regarding simulation efficiency, experimental scientific rigor, operational flexibility, and intelligence level. Summary of the Invention

[0005] To address the problems in the related technologies, this disclosure provides a simulation method and system for an integrated circuit computer-aided design simulation platform.

[0006] In a first aspect, this disclosure provides a simulation method for an integrated circuit computer-aided design simulation platform, including:

[0007] A simulation process definition interface is provided for receiving user-defined simulation processes, wherein the simulation process includes at least one simulation module. Receive the parameter type set by the user for the input parameters of the at least one simulation module; In response to the user's command to select a simulation mode, the system calls the preset algorithm engine corresponding to the selected simulation mode and automatically generates a set of simulation experimental points based on the parameter type of the input parameters; the simulation modes include full factorial simulation mode and intelligent experimental design mode. Execute the simulation tasks corresponding to the set of simulation experiment points and output the corresponding simulation results; When the selected simulation mode is the intelligent experimental design mode, at least one input parameter is set to a continuous range interval. The intelligent experimental design mode takes the continuous range space as the operation object and performs automated experimental planning through a configurable algorithm engine to allow users to perform any one or more of the following operations: parameter screening, design space exploration, and performance index optimization.

[0008] In one embodiment of this disclosure, the intelligent experimental design mode includes a parameter filtering mode, and the algorithm engine uses a filtering design algorithm to generate a set of simulation experimental points within the continuous range. The simulation results corresponding to the output include: After executing the simulation task, calculate the influence factor of each input parameter on the specified output parameter and output it in the form of a sorted list of influence factors.

[0009] In one embodiment of this disclosure, the method is applied to the design optimization of radio frequency semiconductor devices; The specified output parameters include the cutoff frequency and the maximum oscillation frequency.

[0010] In one embodiment of this disclosure, the intelligent experimental design mode includes an intelligent DoE mode, wherein the algorithm engine uses a space filling algorithm to generate a uniformly distributed set of simulation experimental points within the continuous range.

[0011] In one embodiment of this disclosure, the method further includes: In the simulation process definition interface, the distribution of the set of simulation experimental points in the parameter space can be previewed in real time in the form of a two-dimensional or three-dimensional scatter plot.

[0012] In one embodiment of this disclosure, the method is used for process window analysis and verification in integrated circuit manufacturing; The continuous range interval refers to the process specification range of the key process parameters.

[0013] In one embodiment of this disclosure, the intelligent experimental design mode includes an intelligent optimization mode, and the method further includes: receiving a user-defined performance score function, wherein the performance score function is a weighted normalization function of at least two output parameters; The algorithm engine employs an optimization algorithm to iteratively generate a set of simulation test points based on the continuous range interval and the performance score function. The simulation results corresponding to the output include: Output a list of simulation results sorted in descending order of performance score.

[0014] In one embodiment of this disclosure, the optimization algorithm includes at least one of a genetic algorithm, a Bayesian optimization algorithm, a particle swarm optimization algorithm, or a surrogate model-based optimization algorithm.

[0015] In one embodiment of this disclosure, the method is applied to the design optimization of power semiconductor devices; The performance score function is constructed based on a weighted combination of on-resistance and switching loss.

[0016] In one embodiment of this disclosure, when the selected simulation mode is a full factorial simulation mode, at least one input parameter is set as a list of discrete values. The algorithm engine uses a full factorial design algorithm to generate a set of simulation test points with complete combinations of all discrete values.

[0017] In one embodiment of this disclosure, the parameter type set by the user for the input parameters of the at least one simulation module includes: Define a continuous range interval for the first type of input parameters; and, Set a list of fixed or discrete values ​​for the second type of input parameters; The automatic generation of the simulation experiment point set includes: generating the simulation experiment point set within a mixed parameter space defined by the continuous range of the first type of parameters and the list of fixed or discrete values ​​of the second type of parameters.

[0018] In one embodiment of this disclosure, the hybrid parameter space includes a continuous range of geometric dimensions and a discrete list of material types.

[0019] In one embodiment of this disclosure, providing a simulation process definition interface includes: providing a graphical flowchart interface, wherein at least one simulation module and its input / output ports are presented graphically and support connection.

[0020] In one embodiment of this disclosure, the method further includes: Receive user selection instructions for subsystems composed of some simulation modules in the simulation process; In response to the selection instruction, the selected subsystem is taken as the target simulation process; For the target simulation process, the steps are as follows: receiving the parameter type set by the user for the input parameters of the at least one simulation module, calling the preset algorithm engine corresponding to the selected simulation mode, executing the simulation task corresponding to the set of simulation experimental points, and outputting the corresponding simulation results.

[0021] In one embodiment of this disclosure, the subsystem includes a process simulation module and a device simulation module; The intelligent experimental design mode is used to optimize the parameters of the subsystem to reduce the performance dispersion of devices caused by process fluctuations.

[0022] Secondly, this disclosure provides an integrated circuit computer-aided design simulation platform, including: The simulation process definition interface module is configured to provide a simulation process definition interface for receiving user-defined simulation processes, wherein the simulation process includes at least one simulation module. The parameter receiving module is configured to receive the parameter type set by the user for the input parameters of the at least one simulation module; The simulation test point generation module is configured to respond to the user's command to select a simulation mode, call the preset algorithm engine corresponding to the selected simulation mode, and automatically generate a set of simulation test points based on the parameter type of the input parameters; the simulation modes include full factorial simulation mode and intelligent experimental design mode; The simulation execution module is configured to execute the simulation tasks corresponding to the set of simulation test points and output the corresponding simulation results; When the selected simulation mode is the intelligent experimental design mode, at least one input parameter is set to a continuous range interval. The intelligent experimental design mode takes the continuous range space as the operation object and performs automated experimental planning through a configurable algorithm engine to allow users to perform any one or more of the following operations: parameter screening, design space exploration, and performance index optimization.

[0023] Thirdly, embodiments of this disclosure provide an electronic device including a memory and a processor, wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method as described in any of the first aspects.

[0024] Fourthly, this disclosure provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the method as described in any of the first aspects.

[0025] Fifthly, this disclosure provides a chip comprising: at least one processor for implementing the method as described in any of the first aspects.

[0026] The technical effects provided by the embodiments of this disclosure may include the following beneficial effects: According to the technical solution provided in this disclosure, a simulation method for an integrated circuit computer-aided design simulation platform includes: providing a simulation flow definition interface for receiving a user-defined simulation flow, the simulation flow including at least one simulation module; receiving parameter types set by the user for input parameters of the at least one simulation module; responding to the user's selection instruction for a simulation mode, calling a preset algorithm engine corresponding to the selected simulation mode, and automatically generating a set of simulation experimental points based on the parameter types of the input parameters; the simulation mode includes a full-factor simulation mode and an intelligent experimental design mode; executing the simulation task corresponding to the set of simulation experimental points, and outputting the corresponding simulation results; wherein, when the selected simulation mode is the intelligent experimental design mode, at least one input parameter is set as a continuous range interval, the intelligent experimental design mode takes the continuous range space as the operation object, and performs automated experimental planning through a configurable algorithm engine to allow the user to perform any one or more of the following operations: parameter screening, design space exploration, and performance index optimization. The above technical solution, by introducing an intelligent experimental design mode and supporting the setting of continuous range intervals as input parameters, overcomes the limitations of traditional integrated circuit computer-aided design simulation platforms in terms of simulation efficiency and parameter flexibility. This method allows users to define parameter spaces more conveniently, reducing the tedious steps of manual discretization and thus improving the agility of design exploration. By automatically generating experimental point sets by calling a preset algorithm engine, it effectively reduces the amount of simulation experiments, improves simulation efficiency, and provides a foundation for a more comprehensive exploration of multi-parameter spaces, helping to discover potential design optimization schemes.

[0027] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0028] Other features, objects, and advantages of this disclosure will become more apparent from the following detailed description of non-limiting embodiments, taken in conjunction with the accompanying drawings. The following is a description of the accompanying drawings.

[0029] Figure 1 A flowchart illustrating a simulation method for an integrated circuit computer-aided design simulation platform according to an embodiment of the present disclosure is shown.

[0030] Figure 2 A schematic diagram showing a list of simulation results according to embodiments of the present disclosure is provided.

[0031] Figure 3 A schematic diagram illustrating a simulation of a subsystem according to an embodiment of the present disclosure is shown.

[0032] Figure 4A structural block diagram of an integrated circuit computer-aided design simulation platform according to an embodiment of the present disclosure is shown.

[0033] Figure 5 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown.

[0034] Figure 6 A schematic diagram of the structure of a computer system suitable for implementing the method according to embodiments of the present disclosure is shown. Detailed Implementation

[0035] In the following, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings to enable those skilled in the art to readily implement them. Furthermore, for clarity, portions unrelated to the description of exemplary embodiments have been omitted from the drawings.

[0036] In this disclosure, it should be understood that terms such as “comprising” or “having” are intended to indicate the presence of features, figures, steps, behaviors, components, parts or combinations thereof disclosed in this specification, and are not intended to exclude the possibility of the presence or addition of one or more other features, figures, steps, behaviors, components, parts or combinations thereof.

[0037] It should also be noted that, unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other. This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0038] Traditional integrated circuit computer-aided design simulation platforms rely solely on the full factorial design method for simulation experiments, resulting in low simulation efficiency, high computational costs, and lengthy simulation cycles. Their parameter input methods are rigid, supporting only discrete value lists and lacking support for continuous ranges, thus limiting operational flexibility. Furthermore, existing platforms have low levels of automation and intelligence, making it difficult to reveal complex interaction effects, and their support for collaborative optimization of complex processes is weak, failing to achieve efficient closed-loop optimization.

[0039] In view of the above-mentioned shortcomings, the simulation method of the integrated circuit computer-aided design simulation platform provided in this disclosure overcomes the limitations of traditional integrated circuit computer-aided design simulation platforms in terms of simulation efficiency and parameter flexibility by introducing an intelligent experimental design mode and supporting continuous range intervals as input parameters. This method allows users to define the parameter space more conveniently, reducing the tedious steps of manual discretization and thus improving the agility of design exploration. By calling a preset algorithm engine to automatically generate experimental point sets, the amount of simulation experiments is effectively reduced, simulation efficiency is improved, and a foundation is provided for a more comprehensive exploration of the multi-parameter space, which helps to discover potential design optimization schemes.

[0040] For ease of understanding, the following explains some key terms in this embodiment: An integrated circuit computer-aided design simulation platform typically refers to a software system used for the design, verification, and optimization of integrated circuits. This platform integrates various simulation tools, such as process simulation, device simulation, and circuit simulation, to support users in predicting and analyzing the performance of integrated circuits.

[0041] The simulation process definition interface refers to the graphical or text-based interactive interface provided by the platform for users to build and configure simulation processes. Through this interface, users can intuitively define simulation steps, module connections, and parameter settings.

[0042] A simulation process refers to a series of simulation operations or steps arranged in a predetermined order and logically connected to achieve a specific simulation objective. This process can consist of one or more independent simulation modules, with defined data flow and control flow.

[0043] A simulation module is an independent functional unit in the simulation process, such as a process simulator, a device simulator, or a parameter extraction tool. Each simulation module has clearly defined input and output ports, and can receive parameters and generate results.

[0044] Simulation modes refer to the different strategies or methods provided by the platform for generating sets of simulation experiment points. In this embodiment, the modes include a full-factor simulation mode and an intelligent experimental design mode to adapt to different simulation needs and optimization objectives.

[0045] An algorithm engine is a software component within the platform used to perform specific computational tasks. During the generation of simulation test points, the algorithm engine applies appropriate mathematical algorithms to determine the distribution of test points based on the selected simulation mode.

[0046] The set of simulation test points refers to a specific combination of input parameters generated by the algorithm engine. Each test point represents an independent simulation run, the purpose of which is to explore the parameter space and obtain the corresponding simulation results.

[0047] A continuous range refers to the continuous range of values ​​that an input parameter can take, such as all real numbers from the minimum to the maximum value. This setting allows the algorithm to explore a wide parameter space, rather than being limited to preset discrete points.

[0048] Figure 1 A flowchart illustrating a simulation method for an integrated circuit computer-aided design simulation platform according to an embodiment of the present disclosure is shown.

[0049] like Figure 1 As shown, the simulation method of the integrated circuit computer-aided design simulation platform includes the following steps S110-S140: In step S110, a simulation process definition interface is provided for receiving user-defined simulation processes, wherein the simulation process includes at least one simulation module. In step S120, the parameter type set by the user for the input parameters of the at least one simulation module is received; In step S130, in response to the user's command to select a simulation mode, a preset algorithm engine corresponding to the selected simulation mode is invoked to automatically generate a set of simulation test points based on the parameter type of the input parameters. In step S140, the simulation task corresponding to the set of simulation experiment points is executed, and the corresponding simulation results are output.

[0050] In this disclosure, the simulation process definition interface in step S110 can be a graphical flowchart interface, allowing users to construct and modify the simulation process through drag-and-drop operations; or it can be a text editor, allowing users to describe the various steps and modules of the simulation process by writing scripts or configuration files. The simulation process consists of at least one simulation module, each module performing a specific simulation function, such as a process simulator or a device simulator.

[0051] In step S120, the user can specify the parameter type for the input parameters of each simulation module, such as numerical parameters or strings. For numerical parameters, the user can manually input fixed values, a series of discrete values, or specify a minimum and a maximum value to represent a continuous range of values.

[0052] In step S130, the user can select the simulation mode by setting a flag in the configuration file, such as "full factorial simulation mode" or "intelligent experimental design mode". When full factorial simulation mode is selected, the algorithm engine will generate all possible parameter combinations as experimental points based on the list of discrete values ​​set by the user for each parameter. When intelligent experimental design mode is selected, the algorithm engine will generate a set of experimental points based on the parameter type set by the user, such as a continuous range interval. For example, for a parameter set as a continuous range interval, the algorithm engine can randomly select several points within that interval as experimental points.

[0053] In step S140, the platform will sequentially start the simulation module and run the simulation according to the parameter combinations defined for each generated experimental point. After each simulation task is completed, the generated raw data or log files will be collected and stored. For example, the simulation results can be written to a separate text file or a simple table for subsequent manual analysis by the user.

[0054] In the intelligent experimental design mode, when the selected simulation mode is the intelligent experimental design mode, at least one input parameter is set as a continuous range. This means that users do not need to manually discretize the parameter space, but can directly specify the effective value range of the parameter. For example, for the width parameter of a device, the user can directly input "10nm-100nm" as its continuous range, without having to list a series of discrete values ​​such as 10nm, 20nm, 30nm, etc. This setting allows the algorithm engine to explore within a wider parameter space, thereby improving the flexibility of design exploration. The intelligent experimental design mode uses a continuous range space as the operation object and automates experimental planning through a configurable algorithm engine to allow users to perform any one or more of the following operations: parameter screening, design space exploration, and performance index optimization. Specifically, the intelligent experimental design mode can include parameter screening mode, intelligent DoE mode, and intelligent optimization mode. The parameter screening mode corresponds to the parameter screening operation, the intelligent DoE mode corresponds to the design space exploration operation, and the intelligent optimization mode corresponds to the performance index optimization operation. Specific examples are provided below for illustration.

[0055] In one embodiment of this disclosure, the intelligent experimental design mode includes a parameter filtering mode, and the algorithm engine uses a filtering design algorithm to generate a set of simulation experimental points within the continuous range. The simulation results corresponding to the output include: After executing the simulation task, calculate the influence factor of each input parameter on the specified output parameter and output it in the form of a sorted list of influence factors.

[0056] During the operation of an integrated circuit computer-aided design simulation platform, when the user selects the intelligent experimental design mode and sets the input parameters to a continuous range, although a set of simulation experimental points can be generated, for complex systems, the key challenge is how to efficiently identify the critical parameters that significantly affect specific output parameters from numerous continuously changing input parameters, and quantify their degree of influence. Without an effective parameter selection mechanism, a large amount of simulation computing resources may be consumed in exploring non-critical parameters, thereby reducing the efficiency and relevance of design optimization.

[0057] This disclosure introduces a parameter screening mode as a specific implementation of an intelligent experimental design pattern. Its aim is to efficiently identify key input parameters that significantly impact integrated circuit performance in the early stages of the design process through a systematic approach. By introducing this mode, designers can quickly focus on the most important parameters when faced with a large number of potential influencing factors, thereby optimizing subsequent design and simulation work.

[0058] In parameter screening mode, the algorithm engine is configured to use a screening design algorithm to generate a set of simulation experimental points. Screening design algorithms are a class of statistical methods specifically designed to identify key factors in multi-factor experiments. Their core idea is to minimize the number of experiments required while ensuring the assessment of major effects. When input parameters are set as continuous ranges, the screening design algorithm selects representative combinations of discrete points within these ranges according to a pre-defined strategy to effectively cover the parameter space and evaluate the effects of each parameter. For example, for a continuous parameter varying within the range [min, max], the screening design algorithm might select its minimum, maximum, or median value as the experimental level and combine it with the levels of other parameters to construct a set of experimental points that reveals the main effects of the parameter. This algorithm ensures that the ability to identify parameter effects is maximized within limited simulation resources.

[0059] After executing the simulation tasks corresponding to the set of simulation experimental points generated by the screening design algorithm, the system further calculates the influence factors of each input parameter on the specified output parameter and outputs them in the form of a ranked list of influence factors. The influence factor is an indicator that quantifies the degree to which changes in input parameters affect a specified output parameter; it reflects the sensitivity or importance of the parameter. Calculating influence factors typically involves statistical analysis of the simulation results, such as using regression analysis or analysis of variance (ANOVA) to quantify the contribution of each input parameter to the specified output parameter. Specifically, after collecting all combinations of input parameters and their corresponding output parameter values ​​for all experimental points, the algorithm engine uses a statistical model to evaluate the effect size of each input parameter. These calculated influence factors are then compiled into a list and sorted in descending order of their numerical values. This ranked list visually displays the relative magnitude of the influence of all input parameters on a specific output parameter, from most important to least important, allowing users to easily identify key parameters and their degree of influence.

[0060] Through the above technical solution, this application can efficiently explore the parameter space with fewer simulation experiments. Furthermore, by calculating the influence factors of each input parameter on a specified output parameter and outputting them in a sorted list, this application provides users with an intuitive and quantitative assessment of parameter importance. This enables designers to quickly identify the key parameters that have the greatest impact on integrated circuit performance, thus focusing limited simulation resources and optimization efforts on these key parameters, avoiding unnecessary exploration of non-critical parameters, significantly improving the efficiency and focus of integrated circuit design optimization, and providing a clear direction for subsequent refined design and optimization.

[0061] In some embodiments, the parameter screening mode method can be applied to the design optimization of radio frequency semiconductor devices; The specified output parameters include the cutoff frequency and the maximum oscillation frequency.

[0062] Specifically, radio frequency (RF) semiconductor devices refer to semiconductor devices operating in the radio frequency band, such as RF transistors and RF diodes, which play a crucial role in wireless communication, radar, satellite navigation, and other fields. Design optimization refers to the process of adjusting the structure, materials, and process parameters of a device to achieve or exceed preset performance indicators (such as high-frequency characteristics, low noise, and high power efficiency). Applying parameter screening mode methods to the design optimization of such devices means that the platform can handle simulation models and parameters unique to RF devices, such as high-frequency effect models and parasitic parameter models, thereby systematically exploring the parameter space of RF devices, identifying key factors affecting their performance, and accelerating design iteration. For example, the geometric dimensions and process parameters of RF transistors, such as gate length, gate width, and doping concentration, can be screened to optimize their high-frequency characteristics.

[0063] Simultaneously, specifying the cutoff frequency and maximum oscillation frequency as output parameters is a core requirement for performance evaluation of RF semiconductor devices. The cutoff frequency (fT) is the frequency at which the current gain of a bipolar transistor or field-effect transistor drops to unit (0dB). It characterizes the highest operating frequency of the device and is a key indicator of its high-frequency current gain capability. The maximum oscillation frequency (fmax) is the frequency at which the single-sided power gain of a transistor drops to unit (0dB). It characterizes the highest frequency the device can generate in an oscillating circuit and is an important indicator of its high-frequency power gain capability. By setting these key indicators as specified output parameters, the simulation platform can accurately calculate and extract these specific high-frequency performance data after executing the simulation task. This allows designers to intuitively understand which input parameters have the greatest impact on fT and fmax, thus enabling targeted design improvements. For example, in parameter filtering mode, the system calculates the impact factor of each input parameter on fT and fmax and outputs a ranking list, guiding designers to prioritize adjusting parameters that significantly affect fT and fmax.

[0064] In one embodiment of this disclosure, the intelligent experimental design mode includes an intelligent DoE mode, wherein the algorithm engine uses a space filling algorithm to generate a uniformly distributed set of simulation experimental points within the continuous range.

[0065] In the simulation methods of computer-aided design simulation platforms for integrated circuits, when users select the intelligent experimental design mode and set the input parameters to a continuous range, how to efficiently and comprehensively explore this continuous parameter space to ensure that the generated set of simulation experimental points can uniformly cover the entire design space, thereby avoiding the omission of potential design areas or key performance points, is a problem that needs to be solved.

[0066] This disclosure introduces an intelligent DoE (Design of Experiments) model as a specific implementation of an intelligent experimental design pattern. Its core lies in optimizing the selection of experimental points through intelligent algorithms to explore the design space more efficiently. This model aims to find the optimal design, understand the influence of parameters, or construct surrogate models in a multi-dimensional parameter space. Compared with traditional full or partial factorial designs, it places greater emphasis on efficient exploration in a continuous parameter space.

[0067] In Intelligent DoE mode, the algorithm engine is configured to generate a uniformly distributed set of simulation test points within the continuous range using a space-filling algorithm. Common space-filling algorithms include Latin hypercube sampling (LHS), which divides the range of each parameter into equally probable intervals and randomly selects a point within each interval, ensuring that each interval for each parameter is sampled only once, thus achieving good uniformity of sample points in each dimension. Furthermore, low-discrepancy sequences such as Sobol or Halton sequences are designed to generate point sets with even better uniformity in multidimensional space, and they fill the space faster than pseudo-random numbers. These algorithms can effectively generate representative simulation test points in a multidimensional space composed of multiple continuous input parameters, thereby providing a comprehensive understanding of the device performance behavior as a function of parameters within a limited number of simulations.

[0068] Through the above technical solution, this application effectively addresses the problem of how to efficiently and comprehensively explore a continuous parameter space under intelligent experimental design mode. Specifically, when the intelligent experimental design mode is set to intelligent DoE mode, the algorithm engine uses a space-filling algorithm to generate a uniformly distributed set of simulation experimental points within a continuous range. The space-filling algorithm ensures that the generated experimental points have good coverage and uniformity in the multidimensional parameter space, avoiding local clustering or sparsity that may be caused by traditional random sampling, and also avoiding the problem of point explosion in the continuous space of full factorial design. This allows for a more comprehensive understanding of the performance variation trend of integrated circuits or devices throughout the entire design space with limited simulation resources, providing a high-quality and representative data foundation for subsequent performance analysis, sensitivity studies, or surrogate model construction, thereby improving the efficiency and accuracy of design exploration.

[0069] In some embodiments, in Smart DoE mode, the method further includes: In the simulation process definition interface, the distribution of the set of simulation experimental points in the parameter space can be previewed in real time in the form of a two-dimensional or three-dimensional scatter plot.

[0070] In this disclosure, the simulation process definition interface is the core interface for user interaction with the simulation platform. It is typically presented graphically, allowing users to configure simulation modules and set parameters. Integrating visualization functionality into this interface enables users to directly observe the generation of the experimental point set while defining the simulation process, providing a unified operational experience. Two-dimensional or three-dimensional scatter plots are a commonly used data visualization method, intuitively displaying the distribution of data points in a coordinate system. When the parameter space has a high dimensionality, dimensionality reduction techniques (such as Principal Component Analysis (PCA), t-SNE, etc.) can be used to map high-dimensional data to a two-dimensional or three-dimensional space for visualization, or the two or three most important parameter dimensions can be projected for display. This visualization clearly reveals the uniformity, coverage, and presence of clustered or sparse regions in the experimental point set. By visualizing the distribution of the experimental point set in this space, users can evaluate whether the selected algorithm (such as a space-filling algorithm) effectively covers the entire design space and whether the experimental points are distributed uniformly as expected, thereby ensuring the representativeness and reliability of subsequent simulation results.

[0071] In some embodiments, the intelligent DoE mode method can be used for process window analysis and verification in integrated circuit manufacturing; The continuous range interval refers to the process specification range of the key process parameters.

[0072] Specifically, process window analysis and verification is a core step in integrated circuit manufacturing. Its purpose is to determine and optimize the allowable range of manufacturing process parameters to ensure the stability, reliability, and high yield of device performance. By simulating the integrated circuit manufacturing process in a simulation platform, using manufacturing process parameters as input and device performance indicators as output, this method can systematically explore and analyze the impact of these parameters on device performance. For example, it can simulate key process steps such as photolithography, etching, thin film deposition, and ion implantation, and evaluate their impact on the threshold voltage, transconductance, leakage current, and other electrical characteristics of transistors. This allows for the prediction of device performance under different combinations of process parameters before actual manufacturing, effectively reducing trial-and-error costs and development cycles.

[0073] Meanwhile, key process parameters refer to manufacturing process variables that significantly affect the performance and yield of integrated circuit devices, such as photoresist thickness, exposure dose, etching time, ion implantation energy and dose, and annealing temperature. The process specification range refers to the upper and lower limits of fluctuations allowed for these key process parameters in actual production. Setting the continuous range interval in the simulation as the process specification range of these key process parameters ensures that the generation of the simulation experimental point set is based on practically feasible manufacturing conditions. In the simulation flow definition interface, when the user selects the intelligent DoE mode, the system will prompt the user to enter the corresponding process specification range for input parameters set as continuous range intervals. For example, for a photoresist thickness parameter, the user can set its continuous range interval to 100nm to 120nm; this 100nm to 120nm is the process specification range for that photoresist thickness. The algorithm engine will generate uniformly distributed simulation experimental points within this range to comprehensively cover possible process fluctuations in actual production.

[0074] In one embodiment of this disclosure, the intelligent experimental design mode includes an intelligent optimization mode, and the method further includes: receiving a user-defined performance score function, wherein the performance score function is a weighted normalization function of at least two output parameters; The algorithm engine employs an optimization algorithm to iteratively generate a set of simulation test points based on the continuous range interval and the performance score function. The simulation results corresponding to the output include: Output a list of simulation results sorted in descending order of performance score.

[0075] In the simulation methods of computer-aided design simulation platforms for integrated circuits, although a set of simulation experimental points can be generated and simulation tasks can be executed to obtain simulation results based on the user-defined simulation process and input parameter types in the intelligent experimental design mode, when the design needs to be optimized, especially when multiple output parameters are involved and there may be trade-offs between these parameters, it is difficult to directly obtain the optimal design scheme by simply generating and executing the set of experimental points. Users need to spend a lot of effort on result analysis and decision-making.

[0076] This disclosure introduces an intelligent optimization model as a specific implementation of intelligent experimental design patterns. Its core lies in systematically exploring the design space through iteration and feedback mechanisms to find the optimal or near-optimal design point that satisfies a specific optimization objective. Unlike simple parameter exploration or screening, the intelligent optimization model aims to actively converge to the optimal solution, rather than simply generating a set of experimental points for analysis. It typically involves an objective function or evaluation criterion to guide the search process.

[0077] The user-defined performance score function refers to an interface provided by the simulation platform that allows users to define a mathematical expression based on their design goals and the importance they place on different output parameters. This function takes at least two output parameters generated by the simulation task as input, and through weighting and normalization, integrates multiple output parameters into a single, quantifiable "performance score." For example, a user can specify that the importance of one output parameter is twice that of another, reflected by a weighting coefficient. Normalization ensures that output parameters with different dimensions or ranges participate fairly in the total score calculation, preventing any single parameter from dominating the performance score due to its large numerical range. This function serves as a guideline for the optimization process, used to evaluate the performance of each simulation experiment. The performance score function is designed as a weighted normalization function of at least two output parameters, meaning it can integrate multiple simulation outputs that may have different dimensions and importance into a unified evaluation metric. Through weighting, users can assign different importance to different output parameters according to actual needs; through normalization, output parameters with different ranges can be mapped to a uniform scale, thereby preventing parameters with large numerical ranges from dominating the performance score calculation and ensuring the fairness and accuracy of the evaluation.

[0078] The algorithm engine employs an optimization algorithm that intelligently generates a new set of simulation test points iteratively based on the continuous range of input parameters set by the user and a defined performance score function. This means that the algorithm does not generate all test points at once, but rather through a series of iterations. In each iteration, it adjusts the search strategy based on the simulation results of the previous iteration (especially the performance score), generating test points that are more likely to be close to the optimal solution. For example, if a previous iteration achieved a high performance score in a certain region, the optimization algorithm may generate more test points in that region for a more refined search. This iterative process allows the algorithm to efficiently converge to the optimal region of the design space, thereby finding the parameter combination with the highest performance score.

[0079] The output list of simulation results, sorted in descending order of performance score, means that after the optimization process, the simulation platform sorts all executed simulation test points and their corresponding simulation results according to their calculated performance scores from highest to lowest, and presents them to the user in list format. This output method allows users to intuitively see which combinations of design parameters achieved the best overall performance, as well as the next best performing design solutions. By viewing this list, users can quickly identify the optimal design solution and analyze the performance differences between different design solutions, thus providing strong support for the final design decision.

[0080] For example, in simulating a power device, the key parameters are the saturation current Idsat and the equivalent resistance Rdson, and their performance score can be set to 0.4. Norm(Idsat) + 0.3 Norm(1 / Rdson). Where Norm represents normalization. The simulation results list output by the platform is as follows: Figure 2 As shown, the apparent scores of 8.62, 8.6, and 8.56 are illustrative.

[0081] In one embodiment of this disclosure, the optimization algorithm includes at least one of a genetic algorithm, a Bayesian optimization algorithm, a particle swarm optimization algorithm, or a surrogate model-based optimization algorithm.

[0082] Among them, surrogate model-based optimization algorithms are optimization methods that replace expensive simulation models by constructing easily evaluable surrogate models (such as response surface models, radial basis functions, or neural networks). This method first trains the surrogate model using a small amount of real-world simulation data, then performs rapid optimization on the surrogate model to find potential optimal solutions, and finally verifies and updates the surrogate model through actual simulations. This method has significant advantages in integrated circuit design optimization, where simulation costs are extremely high, and can significantly shorten the optimization cycle and reduce computational resource consumption.

[0083] For example, in the first iteration of the intelligent optimization mode, the intelligent DoE mode is invoked to generate an initial set of experimental points, which serves as training data for the surrogate model. This initial set of experimental points is the result of invoking the intelligent DoE mode; it consists of a carefully selected, representative set of simulation experimental points in the parameter space. Using this initial set of experimental points as training data for the surrogate model can improve the accuracy and generalization ability of the surrogate model for optimization algorithms based on it.

[0084] In some embodiments, the intelligent optimization mode method is applied to the design optimization of power semiconductor devices; The performance score function is constructed based on a weighted combination of on-resistance and switching loss.

[0085] Specifically, power semiconductor devices are core components in the field of power electronics, widely used in power management, motor drives, and new energy sources. Their design optimization aims to improve the electrical performance (such as conduction losses, switching losses, and breakdown voltage), thermal performance, and reliability of these devices. Intelligent optimization modes can systematically explore the design parameter space to find the optimal design solution that meets or exceeds specific performance indicators. For example, when designing new IGBTs or MOSFETs, it is necessary to balance multiple interdependent parameters such as on-resistance, switching speed, and breakdown voltage; intelligent optimization modes provide an efficient optimization framework.

[0086] Meanwhile, the performance score function is constructed based on a weighted combination of on-resistance and switching losses. On-resistance is the resistance of a power semiconductor device in the on-state, directly affecting the device's conduction loss and is one of the key indicators for measuring device efficiency; its value is generally preferred to be as low as possible. Switching losses refer to the energy loss generated by the device during the switching process (from on to off or from off to on). It is related to factors such as switching frequency, voltage, and current change rate, and is also an important indicator affecting device efficiency and heat generation; its value is also preferred to be as low as possible. In the design optimization of power semiconductor devices, on-resistance and switching losses often have a mutually restrictive relationship. For example, reducing on-resistance may lead to an increase in switching losses, and vice versa. Therefore, by weighting and combining on-resistance and switching losses, a comprehensive performance score function can be constructed to quantify the overall performance of the device. This weighted combination allows designers to adjust the relative importance of on-resistance and switching losses according to the actual application scenario. For example, in low-frequency applications, more emphasis may be placed on on-resistance, while in high-frequency applications, more emphasis may be placed on switching losses.

[0087] In one embodiment of this disclosure, when the selected simulation mode is a full factorial simulation mode, at least one input parameter is set as a list of discrete values. The algorithm engine uses a full factorial design algorithm to generate a set of simulation test points with complete combinations of all discrete values.

[0088] Specifically, when a user selects the full-factor simulation mode, this mode aims to comprehensively and thoroughly explore the discrete levels of all selected input parameters. In this mode, at least one input parameter is set as a list of discrete values, meaning that the parameter is no longer a continuously varying quantity, but is limited to a series of predefined, discontinuous, specific values. For example, an input parameter could be the width of a transistor, whose list of discrete values ​​might include {0.1µm, 0.2µm, 0.5µm}; or a material parameter, whose list of discrete values ​​might include {material A, material B, material C}.

[0089] In one embodiment of this disclosure, the parameter type set by the user for the input parameters of the at least one simulation module includes: Define a continuous range interval for the first type of input parameters; and, Set a list of fixed or discrete values ​​for the second type of input parameters; The automatic generation of the simulation experiment point set includes: generating the simulation experiment point set within a mixed parameter space defined by the continuous range of the first type of parameters and the list of fixed or discrete values ​​of the second type of parameters.

[0090] In simulation methods of integrated circuit computer-aided design simulation platforms, it is typically necessary to receive the parameter types set by the user for the input parameters of the simulation module and automatically generate a set of simulation experimental points based on these types. However, in actual integrated circuit design optimization processes, input parameters are often not of a single type, and may simultaneously include continuously changing parameters and discretely selected parameters. Failure to effectively handle this mixed parameter type may result in an incomplete or inefficient generation of the simulation experimental point set, failing to fully explore the complex design space and thus affecting the design optimization effect.

[0091] In this disclosure, the first type of input parameter refers to a parameter whose value can continuously vary between a preset minimum and maximum value in integrated circuit design simulation. Examples include geometrical parameters such as channel width, length, and oxide layer thickness of semiconductor devices, or electrical or thermal parameters such as operating voltage, current, and temperature. Users can specify a clear lower and upper limit value for these parameters in the simulation flow definition interface using graphical interface elements (such as text input boxes, sliders, range selectors, etc.). The system will identify and store the continuous range attributes of these parameters so that the subsequent simulation experimental point generation algorithm can effectively sample within this continuous interval.

[0092] The second type of input parameter refers to parameters in integrated circuit design simulation whose values ​​can only be selected from a finite predefined set, or are set to a fixed value. Examples include different material types (such as silicon, germanium, gallium arsenide), different process options (such as different doping concentration levels, different etching schemes), or different device structure types. Users can select one or more discrete values ​​for these parameters, or set a fixed value, through drop-down menus, radio buttons, checkboxes, or by directly entering a fixed value. The system will identify and store the discrete or fixed attributes of these parameters, ensuring that when simulation experimental points are generated, these parameters only take values ​​from their allowed finite set.

[0093] The hybrid parameter space is a multidimensional space composed of the continuous range of the first type of parameters and the list of fixed or discrete values ​​of the second type of parameters. Generating a set of simulation experimental points within this space means that the algorithm engine needs to consider both the sampling strategy for continuous parameters and the combination strategy for discrete parameters. Specifically, the algorithm engine samples within a specified range of continuous parameters based on the selected simulation mode (e.g., parameter selection mode, intelligent DoE mode, or intelligent optimization mode in intelligent experimental design mode) and its corresponding algorithm (e.g., selection design algorithm, space filling algorithm, or optimization algorithm), and combines this with all possible or selected combinations of discrete parameters to construct a set of simulation experimental points covering the entire hybrid parameter space. For example, for each specific combination of discrete parameters, a set of experimental points is generated in the continuous parameter space to ensure a comprehensive exploration of all key parameter combinations.

[0094] By clearly distinguishing and defining the continuous range of the first type of input parameters and the list of fixed or discrete values ​​for the second type of input parameters, and generating a set of simulation experimental points within the mixed parameter space defined by these different types of parameters, this application can more accurately and comprehensively characterize the complex parameter combinations present in actual integrated circuit designs. This approach avoids the undersampling or redundancy caused by treating all parameters indiscriminately, thus significantly improving the efficiency and effectiveness of generating the set of simulation experimental points. Designers can more fully explore the complex design space containing both continuous and discrete variables, which helps to discover better design solutions and accelerate the design optimization process of integrated circuits, especially suitable for design scenarios that involve multiple parameters such as geometry, material selection, or process options.

[0095] In some embodiments, the hybrid parameter space includes a continuous range of geometric dimensions and a discrete list of material types.

[0096] In this disclosure, the continuous range of geometric dimensions refers to parameters describing the size, shape, and other characteristics of the physical structure of an integrated circuit or semiconductor device. These parameters can continuously take values ​​between a preset minimum and maximum value. For example, the channel length, width, oxide layer thickness, interconnect linewidth, and spacing of a transistor can be defined by the user in the simulation flow definition interface. For instance, the channel length can be set to [0.1μm, 1μm], and the width can be set to [0.5μm, 10μm]. The simulation platform will select experimental points within this range, allowing designers to explore the impact of continuous parameter variations on device performance and perform refined design optimization. The discrete value list of material types refers to the types of materials constituting different parts of an integrated circuit or semiconductor device. These materials are predefined, finite, and non-continuously variable options. For example, for gate dielectric materials (such as silicon dioxide and high-dielectric-constant materials), substrate materials (such as silicon, silicon carbide, and gallium nitride), and metal interconnect materials (such as copper and aluminum), users can provide a discrete list of options for these material parameters in the simulation flow definition interface. For example, the gate dielectric material can be selected as {silicon dioxide, high-dielectric-constant materials, and aluminum oxide}. The simulation platform will select one or more of these discrete options for combination, allowing designers to evaluate the impact of different material choices on device performance, especially in scenarios involving the introduction or substitution of new materials.

[0097] In one embodiment of this disclosure, providing a simulation process definition interface includes: providing a graphical flowchart interface, wherein at least one simulation module and its input / output ports are presented graphically and support connection.

[0098] Specifically, the graphical flowchart interface provides users with an intuitive workspace where they can build and modify simulation processes through drag-and-drop operations. This visual presentation greatly reduces the cognitive burden on users to understand and operate complex processes. Simulation modules, such as device simulation modules, circuit simulation modules, or post-processing modules, are represented on the interface as independent graphical elements, such as rectangles or specific icons, each clearly indicating its simulation function. Simultaneously, the input and output ports of each simulation module are also presented graphically, for example, represented by small dots or squares on the module edges, and can be distinguished by different colors or shapes according to port type (such as data input, control signal output, etc.), making the functional boundaries and data flow of the modules immediately clear. Furthermore, the graphical flowchart interface also supports connection functionality, allowing users to establish data or control flow relationships by drawing lines between the input and output ports of different modules. For example, a user can drag a line from the output port of one module to the input port of another module to specify data transfer from the former to the latter. The system can perform real-time verification of these connections to ensure data type or interface compatibility, thereby avoiding simulation failures caused by manual configuration errors.

[0099] In one embodiment of this disclosure, the method further includes: Receive user selection instructions for subsystems composed of some simulation modules in the simulation process; In response to the selection instruction, the selected subsystem is taken as the target simulation process; For the target simulation process, the steps are as follows: receiving the parameter type set by the user for the input parameters of the at least one simulation module, calling the preset algorithm engine corresponding to the selected simulation mode, executing the simulation task corresponding to the set of simulation experimental points, and outputting the corresponding simulation results.

[0100] In the simulation methods of integrated circuit computer-aided design simulation platforms, users typically need to define a complete simulation flow, which may contain multiple interrelated simulation modules. However, in actual design optimization, users may only focus on a specific part or subsystem of the entire flow. If the entire complex simulation flow is configured with parameters, generated experimental points, and executed for each iteration, it will lead to unnecessary consumption of computational resources and time overhead, reducing the efficiency of design iteration.

[0101] In this disclosed method, users can intuitively see the complete simulation process consisting of multiple simulation modules in the simulation process definition interface. To enable focused analysis or optimization of specific parts, the platform provides interactive functionality, allowing users to select one or more continuous or discontinuous simulation modules in the simulation process through a graphical interface, such as by dragging, selecting boxes, or clicking. These selected modules are then combined and defined as a "subsystem." This selection instruction is received by the system, clarifying the scope of subsequent operations the user wishes to perform.

[0102] Once the user has selected a subsystem, the system will respond immediately. At this point, the system will no longer treat the entire original simulation process as the object of operation, but will instead treat the user-selected subsystem as an independent, temporary "target simulation process." This means that all subsequent parameter settings, experimental design, and simulation execution will be limited to the selected subsystem, and its input / output interfaces will be automatically determined based on the subsystem boundaries, thereby effectively isolating the influence of non-target modules.

[0103] After determining the target simulation process (i.e., the selected subsystem), the system will repeatedly execute the core operational steps of the integrated circuit computer-aided design simulation platform for that subsystem. Specifically, the user can set the parameter types for the input parameters of the simulation modules within the subsystem, such as continuous range intervals or discrete value lists. Subsequently, the system will call the corresponding preset algorithm engine based on the user-selected simulation mode, such as full-factor simulation mode or intelligent experimental design mode, to automatically generate a set of simulation experimental points for the subsystem based on these parameter types. Finally, the platform will execute the simulation tasks within the subsystem only for these experimental points and output the simulation results related to that subsystem.

[0104] In one embodiment of this disclosure, the subsystem includes a process simulation module and a device simulation module; The intelligent experimental design mode is used to optimize the parameters of the subsystem to reduce the performance dispersion of devices caused by process fluctuations.

[0105] Specifically, the process simulation module simulates the manufacturing process of semiconductor devices, such as ion implantation, diffusion, thin film deposition, and etching, to predict the physical structure and doping distribution of the devices. This module can use the Technology Computer-Aided Design (TCAD) tool, inputting process parameters (such as implantation dose, annealing temperature, and etching time) to output the device's geometry and doping distribution files. The device simulation module, based on the physical structure obtained from the process simulation, simulates the device's electrical and thermal behavior, such as IV characteristics, CV characteristics, switching speed, and power consumption. This module can also use the TCAD tool, receiving the output from the process simulation module as input, and combining it with electrical and physical models to output the device's electrical characteristic curves and parameters. These two modules can exchange data through a file interface or API, forming an automated simulation chain.

[0106] In the above embodiments, the intelligent experimental design modes include parameter selection mode, intelligent DoE mode, and intelligent optimization mode. Depending on the actual situation, during a simulation, one, two, or all three modes can be selected for execution. Alternatively, any two or three modes can be executed sequentially once or multiple times. This disclosure does not impose any limitations on this.

[0107] For example, in some embodiments, the following steps are executed sequentially: First, a parameter filtering mode is applied to determine key input parameters; then, an intelligent optimization mode is applied only to the key input parameters. Here, the sequential execution steps refer to a phased, sequential execution process, where the output or conclusion of the previous stage serves as the input or constraint for the next stage. This sequential execution mode aims to improve the efficiency and accuracy of the overall process through gradual focusing and refinement.

[0108] Specifically, in the first stage, a parameter filtering model is applied to determine key input parameters. During implementation, the algorithm engine employs filtering design algorithms, such as Plackett-Burman design and fractional factorial design, to generate a set of simulation test points within a continuous range of the input parameters. By executing these simulation tasks and analyzing the results, the influence factor of each input parameter on the specified output parameter is calculated. These influence factors are typically output in the form of a sorted list, and users can select parameters with higher influence factors from the list as key input parameters based on preset thresholds or experience. The purpose of this step is to focus attention on truly important parameters in subsequent optimization stages, thereby reducing computational load and improving optimization efficiency.

[0109] In the second stage, after the key input parameters are determined, the intelligent optimization mode is specifically applied to these key parameters. This means that in the subsequent generation and optimization iteration of simulation test points, only parameters identified as key will be explored and adjusted within their continuous range, while non-key parameters can be fixed at a certain value (e.g., their nominal value or a value that performs well in screening experiments). The intelligent optimization mode typically involves receiving a user-defined performance score function, which quantifies the quality of simulation results. The algorithm engine uses optimization algorithms (such as genetic algorithms, Bayesian optimization, particle swarm optimization, etc.) to iteratively generate a new set of simulation test points based on the continuous range of key parameters and the performance score function, and sorts the simulation results according to the performance score to find the optimal design point.

[0110] The simulation method of the integrated circuit computer-aided design simulation platform disclosed herein, by introducing a parameter definition method that supports continuous range intervals and constructing an architecture that can automatically schedule corresponding algorithm engines in response to various intelligent experimental design modes, can reduce the number of simulations required by several orders of magnitude while ensuring comprehensive exploration, thereby significantly saving computing resources and shortening the design cycle. At the same time, its automated experimental design based on preset algorithm engines avoids the bias caused by subjective parameter selection, and can systematically reveal complex coupling effects in high-dimensional parameter space, significantly improving the scientificity and reliability of the design. In addition, the fully automated execution transforms the platform from a passive simulation tool into an active design optimization partner, greatly reducing the dependence on human experience, and realizing efficient, intelligent, and collaborative optimization of complex integrated circuit design tasks, fundamentally changing the traditional simulation design paradigm.

[0111] The following example will provide a more detailed explanation of the above technical solution: In an integrated circuit computer-aided design simulation platform, a design engineer (User A) needs to optimize the design of a radio frequency semiconductor device to improve its cutoff frequency and maximum oscillation frequency.

[0112] First, the platform provides a simulation workflow definition interface. In this interface, User A uses a graphical flowchart to build a simulation workflow that includes a process simulation module, a device simulation module, and an electrical characteristic extraction module. These simulation modules and their input / output ports are presented graphically and can be connected by User A, clearly defining the execution sequence of the entire simulation task.

[0113] Next, User A sets the parameter type for the input parameters of at least one simulation module in the simulation flow. For example, for device geometry (such as gate length and gate width) and process parameters (such as doping concentration and oxide thickness), User A sets them to a continuous range (e.g., gate length from 10 nm to 50 nm, doping concentration from 1e18 to 5e18 cm⁻¹). -3 Meanwhile, for some discrete parameters, such as material type, user A can set them as a list of discrete values. This defines a hybrid parameter space consisting of a continuous range and a list of discrete values.

[0114] Subsequently, User A selected different simulation modes based on the design goals.

[0115] In the initial exploration phase, User A selected the "Intelligent DoE Mode" within the Intelligent Experiment Design Mode. The platform responded to this selection by invoking the preset algorithm engine corresponding to the Intelligent DoE Mode. This algorithm engine employs a space-filling algorithm, automatically generating a uniformly distributed set of simulation experimental points within the mixed parameter space based on the continuous range of input parameters set by User A. In the simulation flow definition interface, User A can preview the distribution of these simulation experimental points in the parameter space in real time as a two-dimensional or three-dimensional scatter plot, intuitively evaluating the coverage quality of the experimental points.

[0116] To further focus on key parameters, User A can switch to parameter filtering mode. In this mode, the algorithm engine employs a filtering design algorithm to generate a new set of simulation test points within a continuous range. After executing these simulation tasks, the platform calculates the influence factor of each input parameter on a specified output parameter (e.g., cutoff frequency and maximum oscillation frequency) and outputs it as a sorted list of influence factors. Using this list, User A can quickly identify the key input parameters that have the greatest impact on device performance.

[0117] After determining the key parameters, User A selects the intelligent optimization mode. User A first defines a performance score function, a weighted normalized function of the cutoff frequency and the maximum oscillation frequency, designed to comprehensively evaluate device performance. The algorithm engine employs an optimization algorithm (e.g., a genetic algorithm) to iteratively generate a set of simulation test points based on the continuous range of key parameters and the performance score function defined by User A. In the first iteration of the intelligent optimization mode, the platform can invoke the intelligent DoE mode to generate an initial set of test points as the initial population for the optimization algorithm. The platform executes these simulation tasks and outputs a list of simulation results sorted in descending order of performance score. User A can directly obtain the optimal design scheme that meets the performance requirements from the list.

[0118] Furthermore, User A can select a subsystem within the simulation workflow, comprised of a process simulation module and a device simulation module, as the target simulation process. For this subsystem, the platform executes the aforementioned steps of parameter setting, algorithm engine invocation, simulation task execution, and result output, thereby optimizing the subsystem parameters to reduce device performance dispersion caused by process fluctuations. For example... Figure 3 As shown, the subsystem consists of the process simulation module sProcess and the device simulation module sDevice (within the dashed box). User A can first use the "parameter filtering mode" on the sProcess module to determine key process parameters, such as the process parameters for chemical vapor deposition (CVD) and reactive ion etching (RIE). Then, the "intelligent optimization mode" on the sDevice module is used to optimize device performance. Based on the design performance score of saturation current Idsat and equivalent resistance Rdson, the device width W and length L are determined. Finally, the entire process is analyzed using the "intelligent DoE mode" to evaluate the yield.

[0119] Figure 4 A structural block diagram of an integrated circuit computer-aided design simulation platform according to an embodiment of the present disclosure is shown. This device can be implemented as part or all of an electronic device through software, hardware, or a combination of both.

[0120] like Figure 4 As shown, the integrated circuit computer-aided design simulation platform 400 includes: The simulation process definition interface module 410 is configured to provide a simulation process definition interface for receiving a user-defined simulation process, wherein the simulation process includes at least one simulation module. The parameter receiving module 420 is configured to receive the parameter type set by the user for the input parameters of the at least one simulation module; The simulation test point generation module 430 is configured to respond to the user's command to select a simulation mode, call the preset algorithm engine corresponding to the selected simulation mode, and automatically generate a set of simulation test points based on the parameter type of the input parameters; the simulation modes include full factorial simulation mode and intelligent experimental design mode; The simulation execution module 440 is configured to execute the simulation tasks corresponding to the set of simulation experiment points and output the corresponding simulation results; When the selected simulation mode is the intelligent experimental design mode, at least one input parameter is set to a continuous range interval. The intelligent experimental design mode takes the continuous range space as the operation object and performs automated experimental planning through a configurable algorithm engine to allow users to perform any one or more of the following operations: parameter screening, design space exploration, and performance index optimization.

[0121] According to the technical solution provided in this disclosure, by introducing an intelligent experimental design mode and supporting continuous range intervals as input parameter settings, the limitations of traditional integrated circuit computer-aided design simulation platforms in terms of simulation efficiency and parameter flexibility are overcome. This platform allows users to define parameter spaces more conveniently, reducing the tedious steps of manual discretization and thus improving the agility of design exploration. By automatically generating experimental point sets by calling a preset algorithm engine, the amount of simulation experiments is effectively reduced, simulation efficiency is improved, and a foundation is provided for a more comprehensive exploration of multi-parameter spaces, which helps to discover potential design optimization schemes.

[0122] In one embodiment of this disclosure, the intelligent experimental design mode includes a parameter filtering mode, and the algorithm engine uses a filtering design algorithm to generate a set of simulation experimental points within the continuous range. The simulation results corresponding to the output include: After executing the simulation task, calculate the influence factor of each input parameter on the specified output parameter and output it in the form of a sorted list of influence factors.

[0123] In one embodiment of this disclosure, the platform is applied to the design optimization of radio frequency semiconductor devices; The specified output parameters include the cutoff frequency and the maximum oscillation frequency.

[0124] In one embodiment of this disclosure, the intelligent experimental design mode includes an intelligent DoE mode, wherein the algorithm engine uses a space filling algorithm to generate a uniformly distributed set of simulation experimental points within the continuous range.

[0125] In one embodiment of this disclosure, the simulation process definition interface module includes: The preview submodule is configured to preview the distribution of the set of simulation experimental points in the parameter space in real time in the form of a two-dimensional or three-dimensional scatter plot in the simulation process definition interface.

[0126] In one embodiment of this disclosure, the platform is used for process window analysis and verification in integrated circuit manufacturing; The continuous range interval refers to the process specification range of the key process parameters.

[0127] In one embodiment of this disclosure, the intelligent experimental design mode includes an intelligent optimization mode, and the platform further includes: a performance score function receiving module, configured to receive a user-defined performance score function, wherein the performance score function is a weighted normalization function of at least two output parameters; The algorithm engine employs an optimization algorithm to iteratively generate a set of simulation test points based on the continuous range interval and the performance score function. The simulation results corresponding to the output include: Output a list of simulation results sorted in descending order of performance score.

[0128] In one embodiment of this disclosure, the optimization algorithm includes at least one of a genetic algorithm, a Bayesian optimization algorithm, a particle swarm optimization algorithm, or a surrogate model-based optimization algorithm.

[0129] In one embodiment of this disclosure, the platform is applied to the design optimization of power semiconductor devices; The performance score function is constructed based on a weighted combination of on-resistance and switching loss.

[0130] In one embodiment of this disclosure, when the selected simulation mode is a full factorial simulation mode, at least one input parameter is set as a list of discrete values. The algorithm engine uses a full factorial design algorithm to generate a set of simulation test points with complete combinations of all discrete values.

[0131] In one embodiment of this disclosure, the parameter type set by the user for the input parameters of the at least one simulation module includes: Define a continuous range interval for the first type of input parameters; and, Set a list of fixed or discrete values ​​for the second type of input parameters; The automatic generation of the simulation experiment point set includes: generating the simulation experiment point set within a mixed parameter space defined by the continuous range of the first type of parameters and the list of fixed or discrete values ​​of the second type of parameters.

[0132] In one embodiment of this disclosure, the hybrid parameter space includes a continuous range of geometric dimensions and a discrete list of material types.

[0133] In one embodiment of this disclosure, providing a simulation process definition interface includes: providing a graphical flowchart interface, wherein at least one simulation module and its input / output ports are presented graphically and support connection.

[0134] In one embodiment of this disclosure, the simulation process definition interface module is further configured to receive a user's selection instruction for a subsystem composed of some simulation modules in the simulation process; and, in response to the selection instruction, to use the selected subsystem as the target simulation process.

[0135] In one embodiment of this disclosure, the subsystem includes a process simulation module and a device simulation module; The intelligent experimental design mode is used to optimize the parameters of the subsystem to reduce the performance dispersion of devices caused by process fluctuations.

[0136] This disclosure also discloses an electronic device. Figure 5 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown.

[0137] like Figure 5 As shown, the electronic device includes a memory and a processor, wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method according to embodiments of the present disclosure.

[0138] The simulation method of the integrated circuit computer-aided design simulation platform includes: A simulation process definition interface is provided for receiving user-defined simulation processes, wherein the simulation process includes at least one simulation module. Receive the parameter type set by the user for the input parameters of the at least one simulation module; In response to the user's command to select a simulation mode, the system calls the preset algorithm engine corresponding to the selected simulation mode and automatically generates a set of simulation experimental points based on the parameter type of the input parameters; the simulation modes include full factorial simulation mode and intelligent experimental design mode. Execute the simulation tasks corresponding to the set of simulation experiment points and output the corresponding simulation results; When the selected simulation mode is the intelligent experimental design mode, at least one input parameter is set to a continuous range interval. The intelligent experimental design mode takes the continuous range space as the operation object and performs automated experimental planning through a configurable algorithm engine to allow users to perform any one or more of the following operations: parameter screening, design space exploration, and performance index optimization.

[0139] In one embodiment of this disclosure, the intelligent experimental design mode includes a parameter filtering mode, and the algorithm engine uses a filtering design algorithm to generate a set of simulation experimental points within the continuous range. The simulation results corresponding to the output include: After executing the simulation task, calculate the influence factor of each input parameter on the specified output parameter and output it in the form of a sorted list of influence factors.

[0140] In one embodiment of this disclosure, the method is applied to the design optimization of radio frequency semiconductor devices; The specified output parameters include the cutoff frequency and the maximum oscillation frequency.

[0141] In one embodiment of this disclosure, the intelligent experimental design mode includes an intelligent DoE mode, wherein the algorithm engine uses a space filling algorithm to generate a uniformly distributed set of simulation experimental points within the continuous range.

[0142] In one embodiment of this disclosure, the method further includes: In the simulation process definition interface, the distribution of the set of simulation experimental points in the parameter space can be previewed in real time in the form of a two-dimensional or three-dimensional scatter plot.

[0143] In one embodiment of this disclosure, the method is used for process window analysis and verification in integrated circuit manufacturing; The continuous range interval refers to the process specification range of the key process parameters.

[0144] In one embodiment of this disclosure, the intelligent experimental design mode includes an intelligent optimization mode, and the method further includes: receiving a user-defined performance score function, wherein the performance score function is a weighted normalization function of at least two output parameters; The algorithm engine employs an optimization algorithm to iteratively generate a set of simulation test points based on the continuous range interval and the performance score function. The simulation results corresponding to the output include: Output a list of simulation results sorted in descending order of performance score.

[0145] In one embodiment of this disclosure, the optimization algorithm includes at least one of a genetic algorithm, a Bayesian optimization algorithm, a particle swarm optimization algorithm, or a surrogate model-based optimization algorithm.

[0146] In one embodiment of this disclosure, the method is applied to the design optimization of power semiconductor devices; The performance score function is constructed based on a weighted combination of on-resistance and switching loss.

[0147] In one embodiment of this disclosure, when the selected simulation mode is a full factorial simulation mode, at least one input parameter is set as a list of discrete values. The algorithm engine uses a full factorial design algorithm to generate a set of simulation test points with complete combinations of all discrete values.

[0148] In one embodiment of this disclosure, the parameter type set by the user for the input parameters of the at least one simulation module includes: Define a continuous range interval for the first type of input parameters; and, Set a list of fixed or discrete values ​​for the second type of input parameters; The automatic generation of the simulation experiment point set includes: generating the simulation experiment point set within a mixed parameter space defined by the continuous range of the first type of parameters and the list of fixed or discrete values ​​of the second type of parameters.

[0149] In one embodiment of this disclosure, the hybrid parameter space includes a continuous range of geometric dimensions and a discrete list of material types.

[0150] In one embodiment of this disclosure, providing a simulation process definition interface includes: providing a graphical flowchart interface, wherein at least one simulation module and its input / output ports are presented graphically and support connection.

[0151] In one embodiment of this disclosure, the method further includes: Receive user selection instructions for subsystems composed of some simulation modules in the simulation process; In response to the selection instruction, the selected subsystem is taken as the target simulation process; For the target simulation process, the steps are as follows: receiving the parameter type set by the user for the input parameters of the at least one simulation module, calling the preset algorithm engine corresponding to the selected simulation mode, executing the simulation task corresponding to the set of simulation experimental points, and outputting the corresponding simulation results.

[0152] In one embodiment of this disclosure, the subsystem includes a process simulation module and a device simulation module; The intelligent experimental design mode is used to optimize the parameters of the subsystem to reduce the performance dispersion of devices caused by process fluctuations.

[0153] Figure 6 A schematic diagram of the structure of a computer system suitable for implementing the method according to embodiments of the present disclosure is shown.

[0154] like Figure 6 As shown, the computer system includes a processing unit that can execute various methods described above based on a program stored in a read-only memory (ROM) or a program loaded from a storage portion into a random access memory (RAM). The RAM also stores various programs and data required for the operation of the computer system. The processing unit, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0155] The following components are connected to the I / O interface: input sections including keyboards, mice, etc.; output sections including cathode ray tubes (CRTs), liquid crystal displays (LCDs), and speakers; storage sections including hard disks; and communication sections including network interface cards such as LAN cards and modems. The communication section performs communication processes via a network such as the Internet. Drives are also connected to the I / O interface as needed. Removable media, such as disks, optical disks, magneto-optical disks, semiconductor memories, etc., are installed on the drive as needed so that computer programs read from them can be installed into the storage section as required. The processing unit can be implemented as a CPU, GPU, TPU, FPGA, NPU, etc.

[0156] In particular, according to embodiments of this disclosure, the methods described above can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program containing program code for performing the methods described above. In such embodiments, the computer program can be downloaded and installed from a network via a communication component, and / or installed from a removable medium.

[0157] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0158] The units or modules described in the embodiments of this disclosure can be implemented in software or programmable hardware. The described units or modules can also be located in a processor, and the names of these units or modules do not necessarily constitute a limitation on the unit or module itself.

[0159] In another aspect, this disclosure also provides a chip including at least one processor that can be used to implement the methods involved in the above system embodiments.

[0160] In one possible design, the chip also includes a memory for storing program instructions and data, which may be located inside or outside the processor.

[0161] In another aspect, this disclosure also provides a computer-readable storage medium, which may be a computer-readable storage medium included in the electronic device or computer system described above; or it may be a standalone computer-readable storage medium not assembled into a device. The computer-readable storage medium stores one or more programs, which are used by one or more processors to perform the methods described in this disclosure.

[0162] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

Claims

1. A simulation method of an integrated circuit computer-aided design simulation platform, characterized by, include: A simulation process definition interface is provided for receiving user-defined simulation processes, wherein the simulation process includes at least one simulation module. Receive the parameter type set by the user for the input parameters of the at least one simulation module; In response to the user's command to select a simulation mode, the preset algorithm engine corresponding to the selected simulation mode is invoked, and a set of simulation experimental points is automatically generated based on the parameter type of the input parameters; the simulation modes include full factorial simulation mode and intelligent experimental design mode. Execute the simulation tasks corresponding to the set of simulation experiment points and output the corresponding simulation results; When the selected simulation mode is the intelligent experimental design mode, at least one input parameter is set to a continuous range interval. The intelligent experimental design mode takes the continuous range space as the operation object and performs automated experimental planning through a configurable algorithm engine to allow users to perform any one or more of the following operations: parameter screening, design space exploration, and performance index optimization.

2. The method of claim 1, wherein, The intelligent experimental design mode includes a parameter filtering mode, and the algorithm engine uses a filtering design algorithm to generate a set of simulation experimental points within the continuous range. The simulation results corresponding to the output include: After executing the simulation task, calculate the influence factor of each input parameter on the specified output parameter and output it in the form of a sorted list of influence factors.

3. The method of claim 2, wherein, The method is applied to the design optimization of radio frequency semiconductor devices; The specified output parameters include the cutoff frequency and the maximum oscillation frequency.

4. The method according to claim 1, characterized in that, The intelligent experimental design mode includes an intelligent DoE mode, and the algorithm engine uses a space filling algorithm to generate a uniformly distributed set of simulation experimental points within the continuous range.

5. The method according to claim 4, characterized in that, The method further includes: In the simulation process definition interface, the distribution of the set of simulation experimental points in the parameter space can be previewed in real time in the form of a two-dimensional or three-dimensional scatter plot.

6. The method according to claim 4, characterized in that, The method is used for process window analysis and verification in integrated circuit manufacturing; The continuous range interval refers to the process specification range of the key process parameters.

7. The method according to claim 1, characterized in that, The intelligent experimental design mode includes an intelligent optimization mode, and the method further includes: receiving a user-defined performance score function, wherein the performance score function is a weighted normalization function of at least two output parameters; The algorithm engine employs an optimization algorithm to iteratively generate a set of simulation test points based on the continuous range interval and the performance score function. The simulation results corresponding to the output include: Output a list of simulation results sorted in descending order of performance score.

8. The method according to claim 7, characterized in that, The optimization algorithm includes at least one of the following: genetic algorithm, Bayesian optimization algorithm, particle swarm optimization algorithm, or surrogate model-based optimization algorithm.

9. The method according to claim 7, characterized in that, The method is applied to the design optimization of power semiconductor devices; The performance score function is constructed based on a weighted combination of on-resistance and switching loss.

10. The method according to claim 1, characterized in that, When the selected simulation mode is full factorial simulation mode, at least one input parameter is set as a list of discrete values; The algorithm engine uses a full factorial design algorithm to generate a set of simulation test points with complete combinations of all discrete values.

11. The method according to claim 1, characterized in that, The parameter types that the user sets for the input parameters of the at least one simulation module include: Define a continuous range interval for the first type of input parameters; and, Set a list of fixed or discrete values ​​for the second type of input parameters; The automatic generation of the simulation experiment point set includes: generating the simulation experiment point set within a mixed parameter space defined by the continuous range of the first type of parameters and the list of fixed or discrete values ​​of the second type of parameters.

12. The method according to claim 11, characterized in that, The hybrid parameter space includes a continuous range of geometric dimensions and a discrete list of material types.

13. The method according to claim 1, characterized in that, The provision of the simulation process definition interface includes: providing a graphical flowchart interface, wherein at least one simulation module and its input / output ports are presented graphically and support connection.

14. The method according to claim 1, characterized in that, The method further includes: Receive user selection instructions for subsystems composed of some simulation modules in the simulation process; In response to the selection instruction, the selected subsystem is taken as the target simulation process; For the target simulation process, the steps are as follows: receiving the parameter type set by the user for the input parameters of the at least one simulation module, calling the preset algorithm engine corresponding to the selected simulation mode, executing the simulation task corresponding to the set of simulation experimental points, and outputting the corresponding simulation results.

15. The method according to claim 14, characterized in that, The subsystem includes a process simulation module and a device simulation module; The intelligent experimental design mode is used to optimize the parameters of the subsystem to reduce the performance dispersion of devices caused by process fluctuations.

16. A computer-aided design simulation platform for integrated circuits, characterized in that, include: The simulation process definition interface module is configured to provide a simulation process definition interface for receiving user-defined simulation processes, wherein the simulation process includes at least one simulation module. The parameter receiving module is configured to receive the parameter type set by the user for the input parameters of the at least one simulation module; The simulation test point generation module is configured to respond to the user's command to select a simulation mode, call the preset algorithm engine corresponding to the selected simulation mode, and automatically generate a set of simulation test points based on the parameter type of the input parameters; the simulation modes include full factorial simulation mode and intelligent experimental design mode; The simulation execution module is configured to execute the simulation tasks corresponding to the set of simulation test points and output the corresponding simulation results; When the selected simulation mode is the intelligent experimental design mode, at least one input parameter is set to a continuous range interval. The intelligent experimental design mode takes the continuous range space as the operation object and performs automated experimental planning through a configurable algorithm engine to allow users to perform any one or more of the following operations: parameter screening, design space exploration, and performance index optimization.

17. An electronic device, characterized in that, The method includes a memory and a processor; wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method according to any one of claims 1-15.

18. A computer-readable storage medium storing computer instructions thereon, characterized in that, When executed by a processor, the computer instructions implement the method described in any one of claims 1-15.

19. A chip, characterized in that, include: At least one processor for implementing the method as described in any one of claims 1-15.