Bidirectional prediction method and device for circuit parameters and performance indexes, storage medium and electronic equipment
By using a bidirectional prediction model to efficiently generate circuit parameters and performance indicators, the problem of low circuit design efficiency in existing technologies is solved, and design efficiency is improved.
Patent Information
- Application Number
- CN202511843996.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-09
- Publication Date
- 2026-03-03
AI Technical Summary
In existing technologies, the setting of circuit parameters and performance indicators relies on multiple simulations and parameter adjustments, resulting in low circuit design efficiency.
A bidirectional prediction model is adopted, which uses a combination of encoder and decoder to achieve bidirectional prediction of circuit parameters and performance indicators based on mode flag bits and process tag information, thus avoiding the tedious parameter adjustment and simulation process.
It improves the efficiency of generating circuit parameters and performance indicators, reduces the consumption of computing resources and time, and enhances circuit design efficiency.
Smart Images
Figure CN121598875A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of integrated circuit technology, specifically to a method, apparatus, storage medium, and electronic device for bidirectional prediction of circuit parameters and performance indicators. Background Technology
[0002] In analog circuit design and optimization, the setting of circuit parameters and performance indicators is crucial. Circuit parameters (such as capacitors, resistors, and transistor sizes) determine the basic structure of the circuit, while performance indicators (such as bandwidth, gain, and power consumption) reflect the circuit's performance in practical applications. By properly setting circuit parameters, it is possible to ensure that the circuit meets functional requirements while optimizing energy consumption, improving stability, and enhancing reliability. Therefore, accurately setting circuit parameters and performance indicators is key to successful circuit design.
[0003] Currently, setting circuit parameters and performance indicators typically relies on multiple simulations and parameter adjustments. When optimizing a circuit, designers first set the circuit parameters and then calculate the corresponding performance indicators using simulation tools. If the performance does not meet the requirements, the parameters need to be readjusted, and the simulation and adjustment process repeated until the design goals are achieved. This process requires multiple iterations and repeated verifications, consuming significant computational resources and time, resulting in low efficiency in circuit design. Summary of the Invention
[0004] This application provides a method, apparatus, storage medium, and electronic device for bidirectional prediction of circuit parameters and performance indicators, which can improve the efficiency of circuit design.
[0005] In a first aspect, embodiments of this application provide a bidirectional prediction method for circuit parameters and performance indicators, including: Receive data to be processed and determine the mode condition vector of the data to be processed, the mode condition vector including mode flag bits and process tag information; The pattern condition vector and the data to be processed are concatenated to generate the input vector; The input vector is input to the bidirectional prediction model, which includes an encoder and a decoder. After receiving the input vector, the bidirectional prediction model determines whether to run a forward prediction mode or a reverse prediction mode based on the mode flag. When the bidirectional prediction model runs in forward prediction mode, it transmits the process label information and the data to be processed to the encoder to generate the corresponding target performance index. When the bidirectional prediction model runs in reverse prediction mode, it transmits the process tag information and the data to be processed to the decoder to generate the corresponding target circuit parameters.
[0006] In the bidirectional prediction method for circuit parameters and performance indicators provided in this application embodiment, the step of determining the mode condition vector of the data to be processed, wherein the mode condition vector includes mode flag bits and process tag information, includes: Set the mode flag bit according to the source of the data to be processed or the calling parameters of the calling interface; The process label information is determined based on the sample label or external parameter file of the data to be processed; The mode flag and the process label information are concatenated into a mode condition vector.
[0007] The bidirectional prediction method for circuit parameters and performance indicators provided in the embodiments of this application also includes constructing a bidirectional prediction model.
[0008] In the bidirectional prediction method for circuit parameters and performance indicators provided in the embodiments of this application, the construction of the bidirectional prediction model includes: Construct a basic two-way prediction model; A training dataset is collected, and the primary bidirectional prediction model is trained using the training dataset to form a bidirectional prediction model.
[0009] In the bidirectional prediction method for circuit parameters and performance indicators provided in the embodiments of this application, the construction of the primary bidirectional prediction model includes: Define the network structure of the encoder and decoder. The encoder is used to generate corresponding performance indicators based on process label information and circuit parameters, and the decoder is used to generate corresponding circuit parameters based on process label information and performance indicators. A latent space layer is established between the encoder and the decoder to characterize the implicit relationship between the circuit parameters and the performance indicators; The input terminals of the encoder and the decoder are connected to the mode determination module to form a primary bidirectional prediction model. The mode determination module is used to switch between the forward prediction mode and the reverse prediction mode.
[0010] In the bidirectional prediction method for circuit parameters and performance indicators provided in the embodiments of this application, after receiving the data to be processed, the method further includes data preprocessing of the data to be processed.
[0011] Secondly, embodiments of this application provide a bidirectional prediction device for circuit parameters and performance indicators, comprising: A data receiving unit is used to receive data to be processed and determine the mode condition vector of the data to be processed, wherein the mode condition vector includes mode flag bits and process tag information; The vector concatenation unit is used to concatenate the pattern condition vector and the data to be processed to generate the input vector; A vector input unit is used to input the input vector to a bidirectional prediction model. The bidirectional prediction model includes an encoder and a decoder. After receiving the input vector, the bidirectional prediction model determines whether to run a forward prediction mode or a reverse prediction mode based on the mode flag bit. A forward prediction unit is used to transmit the process label information and the data to be processed to the encoder when the bidirectional prediction model is running in forward prediction mode, so as to generate the corresponding target performance index. The reverse prediction unit is used to transmit the process tag information and the data to be processed to the decoder when the bidirectional prediction model is running in reverse prediction mode, so as to generate the corresponding target circuit parameters.
[0012] The bidirectional prediction device for circuit parameters and performance indicators provided in the embodiments of this application further includes: Model building unit, used to build bidirectional prediction models.
[0013] Thirdly, this application provides a storage medium storing a plurality of instructions adapted for loading by a processor to execute the bidirectional prediction method for circuit parameters and performance indicators described in any of the preceding claims.
[0014] Fourthly, this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the bidirectional prediction method for circuit parameters and performance indicators as described in any of the preceding claims.
[0015] In summary, the bidirectional prediction method for circuit parameters and performance indicators provided in this application includes receiving data to be processed and determining a mode condition vector for the data to be processed, wherein the mode condition vector includes a mode flag and process tag information; concatenating the mode condition vector and the data to be processed to generate an input vector; inputting the input vector to a bidirectional prediction model, wherein the bidirectional prediction model includes an encoder and a decoder, wherein after receiving the input vector, the bidirectional prediction model determines whether to run a forward prediction mode or a reverse prediction mode based on the mode flag; when the bidirectional prediction model runs a forward prediction mode, transmitting the process tag information and the data to be processed to the encoder to generate the corresponding target performance indicator; when the bidirectional prediction model runs a reverse prediction mode, transmitting the process tag information and the data to be processed to the decoder to generate the corresponding target circuit parameters. In this application embodiment, when circuit parameters are input, the bidirectional prediction model can directly generate the corresponding performance indicator through the encoder, avoiding repeated simulations after each parameter adjustment. When performance indicators are input, the bidirectional prediction model can generate the corresponding circuit parameters through the decoder, thereby avoiding the tedious process of manual adjustment and re-simulation. Therefore, the embodiments of this application can improve the efficiency of generating circuit parameters and performance indicators, thereby improving the efficiency of circuit design. Attached Figure Description
[0016] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0017] Figure 1 This is a schematic diagram illustrating an application scenario of the bidirectional prediction method for circuit parameters and performance indicators provided in the embodiments of this application.
[0018] Figure 2 This is a flowchart illustrating the bidirectional prediction method for circuit parameters and performance indicators provided in the embodiments of this application.
[0019] Figure 3 This is a schematic diagram of the structure of the bidirectional prediction device for circuit parameters and performance indicators provided in the embodiments of this application.
[0020] Figure 4 This is another schematic diagram of the bidirectional prediction device for circuit parameters and performance indicators provided in the embodiments of this application.
[0021] Figure 5 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation
[0022] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.
[0023] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, components, features, and elements with the same names in different embodiments of this application may have the same meaning or different meanings, the specific meaning of which must be determined by its interpretation in that specific embodiment or further in conjunction with the context of that specific embodiment.
[0024] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.
[0025] In the following description, the use of suffixes such as "module," "part," or "unit" to denote elements is solely for the purpose of illustrative purposes and has no specific meaning in itself. Therefore, "module," "part," or "unit" may be used interchangeably.
[0026] In the description of this application, it should be noted that the terms "upper," "lower," "left," "right," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. In addition, terms such as "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.
[0027] Currently, setting circuit parameters and performance indicators typically relies on multiple simulations and parameter adjustments. When optimizing a circuit, designers first set the circuit parameters and then calculate the corresponding performance indicators using simulation tools. If the performance does not meet the requirements, the parameters need to be readjusted, and the simulation and adjustment process repeated until the design goals are achieved. This process requires multiple iterations and repeated verifications, consuming significant computational resources and time, resulting in low efficiency in circuit design.
[0028] Based on this, embodiments of this application provide a method, apparatus, storage medium, and electronic device for bidirectional prediction of circuit parameters and performance indicators. Specifically, the bidirectional prediction apparatus for circuit parameters and performance indicators can be integrated into an electronic device, which can be a server or a terminal, etc. The terminal can include tablet computers, laptops, and personal computers (PCs), etc. The server can be a single server or a server cluster composed of multiple servers, and can be a physical server or a virtual server.
[0029] For example, such as Figure 1 As shown, the electronic device can receive data to be processed and determine the mode condition vector of the data to be processed. The mode condition vector includes a mode flag and process tag information. The mode condition vector and the data to be processed are concatenated to generate an input vector. The input vector is then input to a bidirectional prediction model, which includes an encoder and a decoder. After receiving the input vector, the bidirectional prediction model determines whether to run a forward prediction mode or a reverse prediction mode based on the mode flag. When the bidirectional prediction model runs in forward prediction mode, the process tag information and the data to be processed are transmitted to the encoder to generate the corresponding target performance index. When the bidirectional prediction model runs in reverse prediction mode, the process tag information and the data to be processed are transmitted to the decoder to generate the corresponding target circuit parameters.
[0030] The technical solutions shown in this application will be described in detail below through specific embodiments. It should be noted that the order of description of the following embodiments is not intended to limit the priority of the embodiments.
[0031] Please see Figure 2 , Figure 2 This is a flowchart illustrating the bidirectional prediction method for circuit parameters and performance indicators provided in this application. The specific flow of this bidirectional prediction method for circuit parameters and performance indicators can be as follows: 101. Receive the data to be processed and determine the mode condition vector of the data to be processed. The mode condition vector includes mode flag bits and process tag information.
[0032] In this embodiment of the application, the data to be processed refers to the input information in the circuit design, which may be circuit parameters or performance indicators.
[0033] In some embodiments, after obtaining the data to be processed, data preprocessing can be performed on the data. This data preprocessing may include one or more of the following methods: data cleaning and outlier handling, data normalization, data standardization, and feature encoding. Data preprocessing can eliminate differences in units and numerical ranges between different features, effectively accelerate model convergence, and significantly improve prediction accuracy and numerical stability.
[0034] Circuit parameters refer to physical quantities or design variables used to define the physical properties, topology, and bias conditions of various components in an analog circuit. These circuit parameters collectively determine the final performance indicators of the analog circuit. Specifically, these circuit parameters include, but are not limited to, discrete component parameters (e.g., resistor-related parameters, capacitor-related parameters, inductor-related parameters, etc.), active device size parameters (e.g., transistor width, transistor length, transistor index, etc.), bias and operating point parameters (e.g., bias current value, bias voltage value, etc.), topology selection parameters (e.g., selector switch state, etc.), and process-related model parameters (e.g., threshold voltage, mobility, etc.).
[0035] Performance metrics refer to a series of electrical characteristic parameters used to quantitatively evaluate the function, efficiency, and quality of analog circuits. Performance metrics are the ultimate goal of circuit design and the standard for measuring whether circuit parameters are optimized. Specifically, these performance metrics include, but are not limited to, gain and accuracy-related metrics (such as DC gain, gain-bandwidth product, etc.), power consumption and efficiency-related metrics (such as static power consumption, dynamic power consumption, etc.), frequency response and speed-related metrics (such as bandwidth, slew rate, etc.), linearity and distortion-related metrics (such as total harmonic distortion, input / output second-order intermodulation point, etc.), noise-related metrics (such as input reference noise, signal-to-noise ratio, noise figure, etc.), and other general metrics (such as input / output common-mode range, common-mode rejection ratio, power supply rejection ratio, etc.).
[0036] In some embodiments, a mode flag can be set according to the source of the data to be processed or the calling parameters of the calling interface; and process label information can be determined according to the sample label of the data to be processed or the external parameter file; and the mode flag and process label information can be concatenated into a mode condition vector.
[0037] The mode flag can be set by checking the source of the data to be processed or the parameters of the called interface. It determines whether the bidirectional prediction model performs forward prediction (predicting performance indicators based on circuit parameters) or backward prediction (predicting circuit parameters based on performance indicators). For example, if the input data is circuit parameters, the mode flag is 0, and the bidirectional prediction model performs forward prediction; if the input data is circuit performance indicators, the mode flag is 1, and the bidirectional prediction model performs backward prediction.
[0038] Process labeling information is an identifier used to indicate the semiconductor manufacturing process conditions or background upon which the data (circuit parameters or performance metrics) being processed depends. This process labeling information enables bidirectional prediction models to make accurate, process-related predictions within the different process contexts covered by their training data. In some embodiments, the process labeling information may include information such as process corners and process nodes.
[0039] During the training and prediction process of the bidirectional prediction model, the process label information is encoded into a fixed-length vector (such as a one-hot vector or a dense vector learned through an embedding layer), and concatenated with the pattern flag and the data to be processed, together serving as the input to the bidirectional prediction model.
[0040] 102. Concatenate the pattern condition vector and the data to be processed to generate the input vector.
[0041] In this embodiment, the mode condition vector obtained in step 101 can be concatenated with the data to be processed (whether it is a circuit parameter vector or a performance index vector).
[0042] For example, suppose the data to be processed (circuit parameters) is a 10-dimensional vector and the mode condition vector is a 6-dimensional vector, then the resulting input vector after concatenation is a 16-dimensional vector.
[0043] This splicing operation is equivalent to attaching a clear "instruction header" to the data to be processed, which can indicate in what mode (forward prediction / backward prediction) and under what process background the bidirectional prediction model should process the subsequent data.
[0044] 103. Input the vector to be input into the bidirectional prediction model. The bidirectional prediction model includes an encoder and a decoder. After receiving the vector to be input, the bidirectional prediction model determines whether to run the forward prediction mode or the backward prediction mode based on the mode flag.
[0045] It is understood that the bidirectional prediction model is pre-built. In the embodiments of this application, the bidirectional prediction model is a neural network model based on a conditional autoencoder architecture, and its capabilities derive from a systematic process that includes model building and model training.
[0046] In some embodiments, a primary bidirectional prediction model can be constructed first; then a training dataset can be collected, and the primary bidirectional prediction model can be trained using the training dataset to form a bidirectional prediction model.
[0047] Specifically, the network structures of the encoder and decoder can be defined first. The encoder can consist of an input layer, several fully connected hidden layers, and an output layer. Its purpose is to learn a complex nonlinear mapping from the "circuit parameter space" to the "performance index space." That is, the encoder's role is to generate corresponding performance indices based on process label information and circuit parameters. In some embodiments, the number of nodes in its input layer equals the sum of the circuit parameter dimension and the process label information vector; the number of nodes in the output layer equals the dimension of the target performance index.
[0048] The decoder can consist of an input layer, several fully connected hidden layers, and an output layer. Its purpose is to learn the inverse mapping from the "performance index space" to the "circuit parameter space." That is, the decoder's function is to generate corresponding circuit parameters based on process label information and performance indicators. In some embodiments, the number of nodes in its input layer can be equal to the sum of the performance index dimension and the process label information dimension; the number of nodes in the output layer is equal to the dimension of the target circuit parameters.
[0049] Next, a latent space layer is established between the encoder and decoder to characterize the implicit relationships between circuit parameters and performance metrics. This latent space layer is a bottleneck layer with a dimension much lower than the input or output data. It is used to learn and compress the most essential and abstract implicit relationships between circuit parameters and performance metrics, and serves as a bridge connecting forward prediction and backward prediction.
[0050] Finally, the inputs of the encoder and decoder are connected to the mode determination module to form a primary bidirectional prediction model. This mode determination module is not an independent physical module, but rather a conditional control flow built into the forward propagation logic of the bidirectional prediction model.
[0051] The function of this mode determination module is to parse the mode flag bits in the input vector. Based on their values (e.g., 0 or 1), this logic controls the data flow to either the encoder (forward prediction) or the decoder (backward prediction), and blocks the other path, thereby enabling the switching between forward prediction mode and backward prediction mode.
[0052] Understandably, after the initial bidirectional prediction model is built, it needs to be trained using a training dataset to form a usable bidirectional prediction model.
[0053] Specifically, a large number of data samples can be collected first to form a training dataset. Each data sample includes corresponding process label information, circuit parameters, and performance indicators.
[0054] It should be noted that the circuit parameters in the training samples can be derived from systematic sampling within the design space via computer scripts (such as Latin hypercube sampling) or historical design databases, thereby generating a large number of combinations of component dimensions and bias conditions covering different possibilities. Performance indicators can be automatically extracted by inputting each set of circuit parameters into a SPICE simulator (such as Cadence Spectre) and performing precise DC, AC, and transient analyses, ensuring the accuracy and authority of the data. Process label information can be directly derived from the process design kits provided by the wafer fab. By calling models of different process corners (such as TT, FF, SS) under the same set of circuit parameters for multiple simulations, a complete sample closely linking circuit parameters, performance indicators, and specific process conditions is generated, laying the data foundation for the model to learn complex mapping relationships under multiple processes.
[0055] During training, supervised learning methods from deep learning can be employed. The bidirectional prediction model optimizes its weights by minimizing the loss function (such as mean squared error or cross-entropy) between the predicted and true values, gradually improving prediction accuracy. The general training steps are as follows: ① Data preprocessing: This can include steps such as data cleaning and outlier handling, data normalization, and data standardization to ensure data quality and consistency.
[0056] ② Training and test set partitioning: The training dataset can be divided into a training set, a validation set, and a test set. Typically, this is done in a 70 / 15 / 15 ratio.
[0057] ③ Model training: The loss function is calculated through forward propagation, and the weight parameters of the model are optimized through backpropagation algorithms (such as stochastic gradient descent, Adam algorithm, etc.) to minimize the defined loss function (such as the combination function of mean squared error and relative error).
[0058] ④ Validation and Evaluation: Evaluate the model's performance on the validation set to check its accuracy and generalization ability. Afterwards, if the model's performance on the test set is unsatisfactory, it can be optimized by adjusting the network structure, increasing training data, and optimizing hyperparameters.
[0059] The specific model training process is not described in detail in the embodiments of this application. Those skilled in the art can use conventional neural network training methods to implement it according to actual needs.
[0060] In this embodiment, after receiving the input vector, the bidirectional prediction model can determine the value of the mode flag bit (e.g., 0 or 1) through the mode determination module. When the value of the mode flag bit is 0, it indicates that the mode flag bit indicates a forward prediction mode, and the bidirectional prediction model activates the encoder path. When the value of the mode flag bit is 1, it indicates that the mode flag bit indicates a backward prediction mode, and the bidirectional prediction model activates the decoder path.
[0061] 104. When the bidirectional prediction model is running in forward prediction mode, the process tag information and the data to be processed are transmitted to the encoder to generate the corresponding target performance indicators.
[0062] At this point, the data to be processed (circuit parameters) and process label information are fed into the encoder. Based on the mapping relationship learned during training, the encoder directly outputs the predicted target performance indicators. For example, the encoder can predict the bandwidth, gain, power consumption, and other performance characteristics of a circuit based on parameters such as circuit size and materials.
[0063] 105. When the bidirectional prediction model runs in reverse prediction mode, the process tag information and the data to be processed are transmitted to the decoder to generate the corresponding target circuit parameters.
[0064] At this point, the data to be processed (performance metrics) and process tag information are fed into the decoder. Based on the inverse mapping relationship learned during training, the decoder outputs target circuit parameters that meet the target performance. For example, the decoder can generate appropriate design parameters such as circuit size and component selection based on the circuit's power consumption and bandwidth requirements.
[0065] In summary, the bidirectional prediction method for circuit parameters and performance indicators provided in this application includes receiving data to be processed and determining a mode condition vector for the data to be processed. The mode condition vector includes a mode flag and process tag information. The mode condition vector and the data to be processed are concatenated to generate an input vector. The input vector is then input to a bidirectional prediction model, which includes an encoder and a decoder. After receiving the input vector, the bidirectional prediction model determines whether to run a forward prediction mode or a reverse prediction mode based on the mode flag. When the bidirectional prediction model runs a forward prediction mode, the process tag information and the data to be processed are transmitted to the encoder to generate the corresponding target performance indicator. When the bidirectional prediction model runs a reverse prediction mode, the process tag information and the data to be processed are transmitted to the decoder to generate the corresponding target circuit parameters. In this application embodiment, when circuit parameters are input, the bidirectional prediction model can directly generate the corresponding performance indicator through the encoder, avoiding repeated simulations after each parameter adjustment. When performance indicators are input, the bidirectional prediction model can generate the corresponding circuit parameters through the decoder, thereby avoiding the tedious process of manual adjustment and re-simulation. Therefore, this application embodiment can improve the generation efficiency of circuit parameters and performance indicators, thereby improving the efficiency of circuit design.
[0066] To facilitate better implementation of the bidirectional prediction method for circuit parameters and performance indicators provided in the embodiments of this application, the embodiments of this application also provide a bidirectional prediction device for circuit parameters and performance indicators. The meanings of the terms used are the same as in the bidirectional prediction method for circuit parameters and performance indicators described above, and specific implementation details can be found in the descriptions in the method embodiments.
[0067] Please see Figure 3 , Figure 3 This is a schematic diagram of the structure of a bidirectional prediction device for circuit parameters and performance indicators provided in an embodiment of this application. The bidirectional prediction device for circuit parameters and performance indicators may include a data receiving unit 201, a vector splicing unit 202, a vector input unit 203, a forward prediction unit 204, and a backward prediction unit 205. The data receiving unit 201 is used to receive data to be processed and determine the mode condition vector of the data to be processed. The mode condition vector includes mode flag bits and process tag information. The vector concatenation unit 202 is used to concatenate the pattern condition vector and the data to be processed to generate the input vector; The vector input unit 203 is used to input the vector to be input into the bidirectional prediction model. The bidirectional prediction model includes an encoder and a decoder. After receiving the vector to be input, the bidirectional prediction model determines whether to run the forward prediction mode or the backward prediction mode according to the mode flag bit. The forward prediction unit 204 is used to transmit process tag information and data to be processed to the encoder when the bidirectional prediction model is running in forward prediction mode, so as to generate the corresponding target performance index. The reverse prediction unit 205 is used to transmit process tag information and data to be processed to the decoder when the bidirectional prediction model is running in reverse prediction mode, so as to generate the corresponding target circuit parameters.
[0068] like Figure 4 As shown, in some embodiments, the bidirectional prediction device for circuit parameters and performance indicators may further include a model building unit 206, which is used to build a bidirectional prediction model.
[0069] For specific implementation methods of each of the above units, please refer to the embodiments of the bidirectional prediction method of circuit parameters and performance indicators described above, which will not be repeated here.
[0070] In summary, the bidirectional prediction device for circuit parameters and performance indicators provided in this application embodiment can receive data to be processed through the data receiving unit 201 and determine the mode condition vector of the data to be processed. The mode condition vector includes a mode flag bit and process tag information. The vector splicing unit 202 splices the mode condition vector and the data to be processed to generate an input vector. The vector input unit 203 inputs the input vector to the bidirectional prediction model, which includes an encoder and a decoder. After receiving the input vector, the bidirectional prediction model determines whether to run a forward prediction mode or a reverse prediction mode based on the mode flag bit. When the bidirectional prediction model runs in forward prediction mode, the forward prediction unit 204 transmits the process tag information and the data to be processed to the encoder to generate the corresponding target performance indicator. When the bidirectional prediction model runs in reverse prediction mode, the reverse prediction unit 205 transmits the process tag information and the data to be processed to the decoder to generate the corresponding target circuit parameters. In this application embodiment, when inputting circuit parameters, the bidirectional prediction model can directly generate the corresponding performance indicator through the encoder, avoiding repeated simulations after each parameter adjustment. When performance metrics are input, the bidirectional predictive model can generate corresponding circuit parameters through a decoder, thus avoiding the tedious process of manual adjustment and re-simulation. Therefore, the embodiments of this application can improve the efficiency of generating circuit parameters and performance metrics, thereby improving the efficiency of circuit design.
[0071] This application also provides an electronic device that can integrate a bidirectional prediction device for circuit parameters and performance indicators, as described in this application. Figure 5 As shown, it illustrates a structural schematic diagram of the electronic device involved in the embodiments of this application, specifically: The electronic device may include components such as a processor 301 with one or more processing cores and a memory 302 with one or more computer-readable storage media. Those skilled in the art will understand that... Figure 5 The electronic device structure shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, or combine certain components, or have different component arrangements. Wherein: The processor 301 is the control center of the electronic device. It connects various parts of the electronic device via various interfaces and lines. By running or executing software programs stored in the memory 302 and / or this application, and by calling data stored in the memory 302, it performs various functions and processes data, thereby providing overall monitoring of the electronic device. Optionally, the processor 301 may include one or more processing cores; preferably, the processor 301 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operation of the storage medium, user interface, and application programs, while the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into the processor 301.
[0072] The memory 302 can be used to store software programs and this application. The processor 301 executes various functional applications and data processing by running the software programs and this application stored in the memory 302. The memory 302 may mainly include a program storage area and a data storage area. The program storage area may store applications required for operating the storage medium and at least one function; the data storage area may store data created based on the use of the electronic device. In addition, the memory 302 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 302 may also include a memory controller to provide the processor 301 with access to the memory 302.
[0073] Although not shown, the electronic device may also include a display unit, an input unit, and a power supply, etc., which will not be described in detail here. Specifically, in this embodiment, the processor 301 in the electronic device loads the executable files corresponding to the processes of one or more application programs into the memory 302 according to the following instructions, and the processor 301 runs the application programs stored in the memory 302 to realize various functions, as follows: Receive the data to be processed and determine the mode condition vector of the data to be processed. The mode condition vector includes mode flag bits and process tag information. The pattern condition vector and the data to be processed are concatenated to generate the input vector; The input vector is fed into the bidirectional prediction model, which includes an encoder and a decoder. After receiving the input vector, the bidirectional prediction model determines whether to run the forward prediction mode or the backward prediction mode based on the mode flag. When the bidirectional prediction model is running in forward prediction mode, it transmits process tag information and data to be processed to the encoder to generate the corresponding target performance index. When the bidirectional prediction model runs in reverse prediction mode, it transmits the process tag information and the data to be processed to the decoder to generate the corresponding target circuit parameters.
[0074] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.
[0075] Therefore, embodiments of this application provide a storage medium storing a plurality of instructions that can be loaded by a processor to execute steps in any of the methods provided in embodiments of this application. For example, the instructions can execute the following steps: Receive the data to be processed and determine the mode condition vector of the data to be processed. The mode condition vector includes mode flag bits and process tag information. The pattern condition vector and the data to be processed are concatenated to generate the input vector; The input vector is fed into the bidirectional prediction model, which includes an encoder and a decoder. After receiving the input vector, the bidirectional prediction model determines whether to run the forward prediction mode or the backward prediction mode based on the mode flag. When the bidirectional prediction model is running in forward prediction mode, it transmits process tag information and data to be processed to the encoder to generate the corresponding target performance index. When the bidirectional prediction model runs in reverse prediction mode, it transmits the process tag information and the data to be processed to the decoder to generate the corresponding target circuit parameters.
[0076] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.
[0077] The storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.
[0078] Since the instructions stored in the storage medium can execute the steps of any method provided in the embodiments of this application, the beneficial effects that any method provided in the embodiments of this application can achieve can be realized. For details, please refer to the previous embodiments, which will not be repeated here.
[0079] The above provides a detailed description of the bidirectional prediction method, apparatus, storage medium, and electronic device for circuit parameters and performance indicators provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.
Claims
1. A bidirectional prediction method for circuit parameters and performance indicators, characterized in that, include: Receive data to be processed and determine the mode condition vector of the data to be processed, the mode condition vector including mode flag bits and process tag information; The pattern condition vector and the data to be processed are concatenated to generate the input vector; The input vector is input to the bidirectional prediction model, which includes an encoder and a decoder. After receiving the input vector, the bidirectional prediction model determines whether to run a forward prediction mode or a reverse prediction mode based on the mode flag. When the bidirectional prediction model runs in forward prediction mode, it transmits the process label information and the data to be processed to the encoder to generate the corresponding target performance index. When the bidirectional prediction model runs in reverse prediction mode, it transmits the process tag information and the data to be processed to the decoder to generate the corresponding target circuit parameters.
2. The bidirectional prediction method for circuit parameters and performance indicators as described in claim 1, characterized in that, The process of determining the pattern condition vector of the data to be processed, wherein the pattern condition vector includes pattern flag bits and process label information, including: Set the mode flag bit according to the source of the data to be processed or the calling parameters of the calling interface; The process label information is determined based on the sample label or external parameter file of the data to be processed; The mode flag and the process label information are concatenated into a mode condition vector.
3. The bidirectional prediction method for circuit parameters and performance indicators as described in claim 1, characterized in that, It also includes building a two-way prediction model.
4. The bidirectional prediction method for circuit parameters and performance indicators as described in claim 3, characterized in that, The construction of the bidirectional prediction model includes: Construct a basic two-way prediction model; A training dataset is collected, and the primary bidirectional prediction model is trained using the training dataset to form a bidirectional prediction model.
5. The bidirectional prediction method for circuit parameters and performance indicators as described in claim 4, characterized in that, The construction of the primary bidirectional prediction model includes: Define the network structure of the encoder and decoder. The encoder is used to generate corresponding performance indicators based on process label information and circuit parameters, and the decoder is used to generate corresponding circuit parameters based on process label information and performance indicators. A latent space layer is established between the encoder and the decoder to characterize the implicit relationship between the circuit parameters and the performance indicators; The input terminals of the encoder and the decoder are connected to the mode determination module to form a primary bidirectional prediction model. The mode determination module is used to switch between the forward prediction mode and the reverse prediction mode.
6. The bidirectional prediction method for circuit parameters and performance indicators as described in claim 1, characterized in that, After receiving the data to be processed, the process also includes data preprocessing of the data to be processed.
7. A bidirectional prediction device for circuit parameters and performance indicators, characterized in that, include: A data receiving unit is used to receive data to be processed and determine the mode condition vector of the data to be processed, wherein the mode condition vector includes mode flag bits and process tag information; The vector concatenation unit is used to concatenate the pattern condition vector and the data to be processed to generate the input vector; A vector input unit is used to input the input vector to a bidirectional prediction model. The bidirectional prediction model includes an encoder and a decoder. After receiving the input vector, the bidirectional prediction model determines whether to run a forward prediction mode or a reverse prediction mode based on the mode flag bit. A forward prediction unit is used to transmit the process label information and the data to be processed to the encoder when the bidirectional prediction model is running in forward prediction mode, so as to generate the corresponding target performance index. The reverse prediction unit is used to transmit the process tag information and the data to be processed to the decoder when the bidirectional prediction model is running in reverse prediction mode, so as to generate the corresponding target circuit parameters.
8. The bidirectional prediction device for circuit parameters and performance indicators as described in claim 7, characterized in that, Also includes: Model building unit, used to build bidirectional prediction models.
9. A storage medium, characterized in that, The storage medium stores multiple instructions, which are adapted for loading by a processor to execute the bidirectional prediction method for circuit parameters and performance indicators as described in any one of claims 1-5.
10. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the bidirectional prediction method for circuit parameters and performance indicators as described in any one of claims 1-5.