Circuit wiring congestion prediction method, circuit wiring method, device and electronic equipment

Abstract

The present application discloses a circuit wiring congestion prediction method, a circuit wiring method, a device and an electronic device. The circuit wiring congestion prediction method comprises: obtaining register transfer level RTL code information and circuit layout strategy information; predicting circuit features associated with circuit wiring congestion in the RTL code information based on the circuit layout strategy information through a trained feature prediction model, wherein the circuit features are used to characterize the features associated with wiring congestion of the layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted, and the circuit features include the logic unit features of at least one circuit logic unit of the layout circuit; predicting the congestion features of the layout circuit based on the circuit features through the trained congestion prediction model, wherein the congestion features include the congestion values ​​corresponding to each circuit logic unit in the layout circuit, and the congestion values ​​are used to characterize the congestion degree of the circuit logic unit. The wiring efficiency can be improved.

Description

Technical Field

[0001] The present application belongs to the field of artificial intelligence technology, and in particular relates to a circuit wiring congestion prediction method, a circuit wiring method, a device and an electronic device. Background Art

[0002] In recent years, with the rapid development of FPGA (Field-Programmable Gate Array) technology, its internal structure and design complexity have continued to increase, and routing congestion has become a key factor affecting FPGA performance and design efficiency.

[0003] In the FPGA design process, routing congestion problems usually occur during the placement and routing stage, especially in high-density areas. The current solution is to perform post-routing analysis and obtain routing congestion reports to identify congestion hotspots. However, this method can only detect congestion problems after routing is completed, resulting in long design iteration cycles and low efficiency. Summary of the invention

[0004] The embodiments of the present application provide a circuit wiring congestion prediction method, a circuit wiring method, a device and an electronic device, which can effectively improve the efficiency of circuit wiring at least to a certain extent.

[0005] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present application.

[0006] According to a first aspect of an embodiment of the present application, a circuit routing congestion prediction method is provided, comprising:

[0007] Obtain register transfer level RTL code information and obtain circuit layout strategy information;

[0008] Predicting circuit features associated with circuit routing congestion in the RTL code information based on the circuit layout strategy information by using the trained feature prediction model, wherein the circuit features are used to characterize features associated with routing congestion of a layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted, and the circuit features include logic unit features of at least one circuit logic unit of the layout circuit;

[0009] The congestion prediction model after training is used to predict the congestion characteristics of the layout circuit based on the circuit characteristics, wherein the congestion characteristics include congestion values ​​corresponding to each circuit logic unit in the layout circuit, and the congestion values ​​are used to characterize the congestion degree of the circuit logic unit.

[0010] In some possible implementations of the present application, the feature prediction model is trained in the following manner:

[0011] Acquire a first sample training data set, wherein the first sample training data set includes multiple groups of first sample training data, and the first sample training data includes sample RTL code information, sample circuit layout strategy information, and sample circuit features of a sample layout circuit corresponding to the sample RTL code information when the sample circuit layout strategy information is adopted;

[0012] Based on the first sample training data set, at least one first training operation is performed on the initial feature prediction model until a first training end condition is met to obtain a feature prediction model.

[0013] In some possible implementations of the present application, the first training operation includes:

[0014] For each group of first sample training data, sample RTL code information and sample circuit layout strategy information in the first sample training data are input into an initial feature prediction model to obtain predicted circuit features;

[0015] Determine first difference information between the predicted circuit feature and the sample circuit feature in the first sample training data;

[0016] Determine a first training loss based on first difference information corresponding to each group of first sample training data;

[0017] The parameters of the initial feature prediction model are adjusted based on the first training loss, and the initial feature prediction model after the adjusted parameters is used as the initial feature prediction model corresponding to the next first training operation.

[0018] In some possible implementations of the present application, the congestion prediction model is trained in the following manner:

[0019] Acquire a second sample training data set, wherein the second sample training data set includes multiple sets of second sample training data, and the second sample training data includes sample circuit features and sample congestion features of sample layout circuits corresponding to the sample circuit features;

[0020] Based on the second sample training data set, at least one second training operation is performed on the initial congestion prediction model until a second training end condition is met, thereby obtaining a congestion prediction model.

[0021] In some possible implementations of the present application, the second training operation includes:

[0022] For each group of second sample training data, inputting sample circuit features in the second sample training data into an initial congestion prediction model to obtain first predicted congestion features;

[0023] Determining second difference information between the first predicted congestion feature and the sample congestion feature in the second sample training data;

[0024] Determine a second training loss based on the second difference information corresponding to each group of second sample training data;

[0025] The parameters of the initial congestion prediction model are adjusted based on the second training loss, and the initial congestion prediction model after the adjusted parameters is used as the initial congestion prediction model corresponding to the next second training operation.

[0026] In some possible implementations of the present application, the first sample training data set and the second sample training data set are obtained in the following manner:

[0027] Acquire multiple groups of initial sample data, each of which includes sample RTL code information and sample circuit layout strategy information;

[0028] For each initial sample data, generating a corresponding sample layout circuit based on sample RTL code information and sample circuit layout strategy information, and extracting sample circuit features of the sample layout circuit;

[0029] Obtaining a sample routing congestion report corresponding to a sample layout circuit, including generating a sample congestion feature for the sample layout circuit based on the sample routing congestion;

[0030] Generate an initial training data set based on each initial sample data, a sample circuit feature and a sample congestion feature respectively corresponding to each initial sample data;

[0031] A first sample training data set is extracted from the initial training data set, and a second sample training data set is extracted from the initial training data set.

[0032] In some possible implementations of the present application, the feature prediction model and the congestion prediction model are trained in the following manner:

[0033] Acquire a third sample training data set, wherein the third sample training data set includes multiple sets of third sample training data, and the third sample training data includes sample RTL code information, sample circuit layout strategy information, and sample congestion characteristics;

[0034] Based on the third sample training data set, at least one third training operation is performed on the initial feature prediction model and the initial congestion prediction model until a third training end condition is met, thereby obtaining a feature prediction model and a congestion prediction model.

[0035] In some possible implementations of the present application, the third training operation includes:

[0036] For each set of third sample training data, sample RTL code information and sample circuit layout strategy information in the third sample training data are input into an initial feature prediction model to obtain predicted circuit features;

[0037] Inputting the predicted circuit feature into the initial congestion prediction model to obtain a second predicted congestion feature;

[0038] Determining third difference information between the second predicted congestion feature and the sample congestion feature in the third sample training data;

[0039] Determine a third training loss based on the third difference information corresponding to each group of third sample training data;

[0040] Based on the third training loss, the parameters of the initial congestion prediction model and the initial feature prediction model are adjusted, and the initial feature prediction model after the adjustment of the parameters is used as the initial feature prediction model corresponding to the next third training operation, and the initial congestion prediction model after the adjustment of the parameters is used as the initial congestion prediction model corresponding to the next third training operation.

[0041] According to a second aspect of an embodiment of the present application, a circuit wiring method is provided, comprising:

[0042] Acquire RTL code information, perform at least one congestion prediction operation based on the RTL code information, until the predicted congestion characteristics meet the preset congestion characteristics, and determine circuit layout strategy information that meets the preset congestion characteristics;

[0043] Generate a corresponding layout circuit based on the circuit layout strategy information and the RTL code information;

[0044] The congestion prediction operations include:

[0045] Obtain register transfer level RTL code information, and obtain circuit layout strategy information corresponding to the current congestion prediction operation;

[0046] Predicting circuit features associated with circuit routing congestion in the RTL code information based on the circuit layout strategy information by using the trained feature prediction model, wherein the circuit features are used to characterize features associated with routing congestion of a layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted, and the circuit features include logic unit features of at least one circuit logic unit of the layout circuit;

[0047] The trained congestion prediction model is used to predict the congestion characteristics of the layout circuit based on the circuit characteristics, and the circuit layout strategy information corresponding to the next congestion prediction operation is determined based on the circuit layout strategy information and congestion characteristics corresponding to the current congestion prediction operation, wherein the congestion characteristics include congestion values ​​corresponding to each circuit logic unit in the layout circuit, and the congestion values ​​are used to characterize the congestion degree of the circuit logic unit.

[0048] According to a third aspect of an embodiment of the present application, a circuit wiring congestion prediction device is provided, comprising:

[0049] A first information acquisition module is used to acquire register transfer level RTL code information and circuit layout strategy information;

[0050] A feature prediction module, configured to predict, by means of a trained feature prediction model, circuit features associated with circuit routing congestion in the RTL code information based on the circuit layout strategy information, wherein the circuit features are used to characterize features associated with routing congestion of a layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted, and the circuit features include logic unit features of at least one circuit logic unit of the layout circuit;

[0051] The congestion prediction module is used to predict the congestion characteristics of the layout circuit based on the circuit characteristics through the trained congestion prediction model, wherein the congestion characteristics include the congestion values ​​corresponding to each circuit logic unit in the layout circuit, and the congestion values ​​are used to characterize the congestion degree of the circuit logic unit.

[0052] In some possible implementations of the present application, a first training module is further included for:

[0053] Acquire a first sample training data set, wherein the first sample training data set includes multiple groups of first sample training data, and the first sample training data includes sample RTL code information, sample circuit layout strategy information, and sample circuit features of a sample layout circuit corresponding to the sample RTL code information when the sample circuit layout strategy information is adopted;

[0054] Based on the first sample training data set, at least one first training operation is performed on the initial feature prediction model until a first training end condition is met to obtain a feature prediction model.

[0055] In some possible implementations of the present application, when performing the first training operation, the first training module is specifically used to:

[0056] For each group of first sample training data, sample RTL code information and sample circuit layout strategy information in the first sample training data are input into an initial feature prediction model to obtain predicted circuit features;

[0057] Determine first difference information between the predicted circuit feature and the sample circuit feature in the first sample training data;

[0058] Determine a first training loss based on first difference information corresponding to each group of first sample training data;

[0059] The parameters of the initial feature prediction model are adjusted based on the first training loss, and the initial feature prediction model after the adjusted parameters is used as the initial feature prediction model corresponding to the next first training operation.

[0060] In some possible implementations of the present application, a second training module is further included for:

[0061] Acquire a second sample training data set, wherein the second sample training data set includes multiple sets of second sample training data, and the second sample training data includes sample circuit features and sample congestion features of sample layout circuits corresponding to the sample circuit features;

[0062] Based on the second sample training data set, at least one second training operation is performed on the initial congestion prediction model until a second training end condition is met, thereby obtaining a congestion prediction model.

[0063] In some possible implementations of the present application, when performing the second training operation, the second training module is used to:

[0064] For each group of second sample training data, inputting sample circuit features in the second sample training data into an initial congestion prediction model to obtain first predicted congestion features;

[0065] Determining second difference information between the first predicted congestion feature and the sample congestion feature in the second sample training data;

[0066] Determine a second training loss based on the second difference information corresponding to each group of second sample training data;

[0067] The parameters of the initial congestion prediction model are adjusted based on the second training loss, and the initial congestion prediction model after the adjusted parameters is used as the initial congestion prediction model corresponding to the next second training operation.

[0068] In some possible implementations of the present application, a training data acquisition module is further included, which is used to:

[0069] Acquire multiple groups of initial sample data, each of which includes sample RTL code information and sample circuit layout strategy information;

[0070] For each initial sample data, generating a corresponding sample layout circuit based on sample RTL code information and sample circuit layout strategy information, and extracting sample circuit features of the sample layout circuit;

[0071] Obtaining a sample routing congestion report corresponding to a sample layout circuit, including generating a sample congestion feature for the sample layout circuit based on the sample routing congestion;

[0072] Generate an initial training data set based on each initial sample data, a sample circuit feature and a sample congestion feature respectively corresponding to each initial sample data;

[0073] A first sample training data set is extracted from the initial training data set, and a second sample training data set is extracted from the initial training data set.

[0074] In some possible implementations of the present application, a third training module is further included for:

[0075] Acquire a third sample training data set, wherein the third sample training data set includes multiple sets of third sample training data, and the third sample training data includes sample RTL code information, sample circuit layout strategy information, and sample congestion characteristics;

[0076] Based on the third sample training data set, at least one third training operation is performed on the initial feature prediction model and the initial congestion prediction model until a third training end condition is met, thereby obtaining a feature prediction model and a congestion prediction model.

[0077] In some possible implementations of the present application, when the third training module performs the third training operation, it is specifically used to:

[0078] For each set of third sample training data, sample RTL code information and sample circuit layout strategy information in the third sample training data are input into an initial feature prediction model to obtain predicted circuit features;

[0079] Inputting the predicted circuit feature into the initial congestion prediction model to obtain a second predicted congestion feature;

[0080] Determining third difference information between the second predicted congestion feature and the sample congestion feature in the third sample training data;

[0081] Determine a third training loss based on the third difference information corresponding to each group of third sample training data;

[0082] Based on the third training loss, the parameters of the initial congestion prediction model and the initial feature prediction model are adjusted, and the initial feature prediction model after the adjustment of the parameters is used as the initial feature prediction model corresponding to the next third training operation, and the initial congestion prediction model after the adjustment of the parameters is used as the initial congestion prediction model corresponding to the next third training operation.

[0083] According to a fourth aspect of an embodiment of the present application, there is provided a circuit wiring device, comprising:

[0084] A second information acquisition module is used to acquire RTL code information, perform at least one congestion prediction operation based on the RTL code information, until the predicted congestion characteristics meet the preset congestion characteristics, and determine the circuit layout strategy information that meets the preset congestion characteristics;

[0085] A circuit generation module, used to generate a corresponding layout circuit based on the circuit layout strategy information and the RTL code information;

[0086] The congestion prediction operations include:

[0087] Obtain register transfer level RTL code information, and obtain circuit layout strategy information corresponding to the current congestion prediction operation;

[0088] Predicting circuit features associated with circuit routing congestion in the RTL code information based on the circuit layout strategy information by using the trained feature prediction model, wherein the circuit features are used to characterize features associated with routing congestion of a layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted, and the circuit features include logic unit features of at least one circuit logic unit of the layout circuit;

[0089] The trained congestion prediction model is used to predict the congestion characteristics of the layout circuit based on the circuit characteristics, and the circuit layout strategy information corresponding to the next congestion prediction operation is determined based on the circuit layout strategy information and congestion characteristics corresponding to the current congestion prediction operation, wherein the congestion characteristics include congestion values ​​corresponding to each circuit logic unit in the layout circuit, and the congestion values ​​are used to characterize the congestion degree of the circuit logic unit.

[0090] According to a fifth aspect of an embodiment of the present application, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method of the above embodiment.

[0091] According to a sixth aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored, characterized in that when the computer program is executed by a processor, the steps of the method of the above embodiment are implemented.

[0092] According to a seventh aspect of the embodiments of the present application, a computer program product is provided, including a computer program, which implements the steps of the above-mentioned embodiment method when the computer program is executed by a processor.

[0093] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application.

[0094] The beneficial effects of the technical solution provided by the embodiment of the present application are:

[0095] Through the feature prediction model, directly based on the circuit layout strategy information and the RTL code information, the features associated with routing congestion of the layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted are predicted, and then through the trained congestion prediction model, the congestion features of the layout circuit are predicted based on the features associated with routing congestion of the layout circuit. The congestion information corresponding to the circuit layout strategy information can be predicted before routing, so as to determine whether the circuit layout strategy information needs to be adjusted, which can effectively reduce the design iteration cycle of circuit routing and improve circuit routing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present application, and together with the specification, are used to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0097] Figure 1 A schematic diagram of a flow chart of a circuit wiring congestion prediction method provided in an embodiment of the present application;

[0098] Figure 2 A schematic diagram of the structure of a feature prediction network provided for an example of this application;

[0099] Figure 3 A schematic diagram of a congestion prediction network structure provided as an example of this application;

[0100] Figure 4 A schematic diagram of features of a sample circuit logic unit provided for one example of the present application;

[0101] Figure 5 A schematic diagram of a flow chart of a circuit wiring method provided in an embodiment of the present application;

[0102] Figure 6 A schematic diagram of the structure of a circuit wiring congestion prediction device provided in an embodiment of the present application;

[0103] Figure 7 A schematic diagram of the structure of a circuit wiring device provided in an embodiment of the present application;

[0104] Figure 8 A schematic diagram of the structure of an electronic device for predicting circuit wiring congestion provided in an embodiment of the present application. DETAILED DESCRIPTION

[0105] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0106] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present application. However, those skilled in the art will appreciate that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, known methods, devices, realizations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0107] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0108] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0109] It should also be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the objects used in this way can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those shown or described.

[0110] The following describes several exemplary embodiments to illustrate the technical solutions of the embodiments of the present application and the technical effects produced by the technical solutions of the present application. It should be noted that the following embodiments can refer to, draw on or combine with each other, and the same terms, similar features and similar implementation steps in different embodiments will not be described repeatedly.

[0111] The model training method of the present application can be executed by any electronic device, and the electronic device may include a server or a terminal.

[0112] Technical personnel in this technical field can understand that the server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server or server cluster that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), as well as big data and artificial intelligence platforms.

[0113] The terminal may be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a laptop computer, a digital broadcast receiver, a MID (Mobile Internet Devices), a PDA (Personal Digital Assistant), a desktop computer, a smart home appliance, a vehicle-mounted terminal (such as a vehicle-mounted navigation terminal, a vehicle-mounted computer, etc.), a smart speaker, a smart watch, etc. The terminal and the server may be directly or indirectly connected via wired or wireless communication, but are not limited thereto. The embodiments of the present invention may be applied to various scenarios, including but not limited to cloud technology, artificial intelligence, smart transportation, assisted driving, etc. The specific requirements may also be determined based on the actual application scenario requirements, and are not limited here. The terminal (also referred to as a user terminal or user device) may be a smart phone, a tablet computer, a laptop computer, a desktop computer, an intelligent voice interaction device (such as a smart speaker), a wearable electronic device (such as a smart watch), a vehicle-mounted terminal, a smart home appliance (such as a smart TV), an AR / VR device, an aircraft, etc., but are not limited thereto.

[0114] like Figure 1 As shown, in some possible implementations, the present application embodiment provides a circuit wiring congestion prediction method. Taking the execution subject as a server as an example, the method may include the following steps:

[0115] Step S101, obtaining register transfer level RTL code information and circuit layout strategy information.

[0116] The RTL (Register Transfer Level) code information is a project design code associated with chip design; and the circuit layout strategy information is strategy information for guiding the generation of layout circuits based on the RTL code information.

[0117] Step S102 : predicting circuit features associated with circuit routing congestion in the RTL code information based on the circuit layout strategy information using the trained feature prediction model.

[0118] The circuit characteristics are used to characterize the characteristics associated with routing congestion of the layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted, and the circuit characteristics include logic unit characteristics of at least one circuit logic unit of the layout circuit.

[0119] Among them, the circuit logic unit is the basic unit structure that constitutes the layout circuit.

[0120] Specifically, circuit characteristics may include the number of main memory units in each circuit logic unit, the total number of main memory units in the circuit logic unit, the number of pins used, and other characteristics that are highly related to wiring congestion. For example, circuit characteristics may include the number of registers, the number of logic gates, the clock frequency, etc.

[0121] In some embodiments of the present application, the feature prediction model may be a deep neural network (DNN), such as Figure 2 As shown, the structure of the feature prediction model may include multiple cascaded FCs (fully connected layers). Circuit layout strategy information and RTL code information are used as input features and input into the feature prediction model to obtain predicted circuit features, that is, features associated with routing congestion of the layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted.

[0122] In some other possible embodiments of the present application, the feature prediction model may also be a recurrent neural network (RNN), a graph neural network (GNN), etc.

[0123] Step S103 , predicting the congestion characteristics of the layout circuit based on the circuit characteristics by using the trained congestion prediction model.

[0124] The congestion feature includes a congestion value corresponding to each circuit logic unit in the layout circuit, and the congestion value is used to characterize the congestion degree of the circuit logic unit.

[0125] Among them, the congestion prediction model can also be CNN, such as Figure 3 As shown, the congestion prediction model may also include multiple cascaded FC layers. The circuit features are used as input features and input into the congestion prediction model to predict the congestion features of the layout circuit.

[0126] Similarly, the congestion prediction model can also be a recurrent neural network, a graph neural network, etc.

[0127] In the above embodiment, through the feature prediction model, the features associated with routing congestion of the layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted are predicted directly based on the circuit layout strategy information and the RTL code information, and then the congestion features of the layout circuit are predicted based on the features associated with routing congestion of the layout circuit through the trained congestion prediction model. The congestion information corresponding to the circuit layout strategy information can be predicted before routing, so as to determine whether the circuit layout strategy information needs to be adjusted, which can effectively reduce the design iteration cycle of circuit routing and improve circuit routing efficiency.

[0128] The training process of the feature prediction model and the congestion prediction model will be further described below in conjunction with the embodiments.

[0129] In some possible implementations of the present application, the feature prediction model is trained in the following manner:

[0130] (1) Obtain the first sample training data set.

[0131] The first sample training data set includes multiple groups of first sample training data, and the first sample training data includes sample RTL code information, sample circuit layout strategy information, and sample circuit features of the sample layout circuit corresponding to the sample RTL code information when the sample circuit layout strategy information is adopted.

[0132] (2) Performing at least one first training operation on the initial feature prediction model based on the first sample training data set until a first training end condition is met to obtain a feature prediction model.

[0133] In some possible implementations of the present application, the first training operation includes:

[0134] ① For each group of first sample training data, the sample RTL code information and the sample circuit layout strategy information in the first sample training data are input into the initial feature prediction model to obtain the predicted circuit features;

[0135] ② Determine first difference information between the predicted circuit feature and the sample circuit feature in the first sample training data;

[0136] ③ Determine the first training loss based on the first difference information corresponding to each group of first sample training data;

[0137] ④ Adjust the parameters of the initial feature prediction model based on the first training loss, and use the initial feature prediction model after adjusting the parameters as the initial feature prediction model corresponding to the next first training operation.

[0138] The first training end condition may include at least one of the following: the number of first training operations reaches a first preset number, the first training loss converges, and the first training loss is less than or equal to a first preset threshold.

[0139] Specifically, if the first training loss is determined and the first training end condition is not met, the parameters of the initial feature prediction model can be adjusted based on the first training loss, and the initial feature prediction model after the adjusted parameters is used as the initial feature prediction model corresponding to the next first training operation, that is, the next first training operation is repeated.

[0140] In some possible implementations of the present application, the congestion prediction model is trained in the following manner:

[0141] (1) Obtain the second sample training data set.

[0142] The second sample training data set includes multiple groups of second sample training data, and the second sample training data includes sample circuit features and sample congestion features of sample layout circuits corresponding to the sample circuit features.

[0143] (2) Based on the second sample training data set, perform at least one second training operation on the initial congestion prediction model until the second training end condition is met, thereby obtaining the congestion prediction model.

[0144] In some possible implementations of the present application, the second training operation includes:

[0145] ① For each group of second sample training data, input the sample circuit features in the second sample training data into the initial congestion prediction model to obtain the first predicted congestion features;

[0146] ② Determine second difference information between the first predicted congestion feature and the sample congestion feature in the second sample training data;

[0147] ③ Determine the second training loss based on the second difference information corresponding to each group of second sample training data;

[0148] ④ Adjust the parameters of the initial congestion prediction model based on the second training loss, and use the initial congestion prediction model after the adjustment of the parameters as the initial congestion prediction model corresponding to the next second training operation.

[0149] Similarly, the second training end condition may include at least one of the following: the number of second training operations reaches a second preset number, the second training loss converges, and the second training loss is less than or equal to a second preset threshold.

[0150] Specifically, if the second training loss is determined and the second training end condition is not met, the parameters of the initial congestion prediction model can be adjusted based on the second training loss, and the initial congestion prediction model after the adjusted parameters is used as the initial congestion prediction model corresponding to the next second training operation, that is, the next second training operation is repeated.

[0151] In some possible implementations of the present application, the first sample training data set and the second sample training data set are obtained in the following manner:

[0152] Acquire multiple groups of initial sample data, each of which includes sample RTL code information and sample circuit layout strategy information;

[0153] For each initial sample data, generating a corresponding sample layout circuit based on sample RTL code information and sample circuit layout strategy information, and extracting sample circuit features of the sample layout circuit;

[0154] Obtaining a sample routing congestion report corresponding to a sample layout circuit, including generating a sample congestion feature for the sample layout circuit based on the sample routing congestion;

[0155] Generate an initial training data set based on each initial sample data, a sample circuit feature and a sample congestion feature respectively corresponding to each initial sample data;

[0156] A first sample training data set is extracted from the initial training data set, and a second sample training data set is extracted from the initial training data set.

[0157] Specifically, circuit wiring may be performed according to the initial sample data to obtain a sample layout circuit, and then sample circuit features may be acquired based on the sample layout circuit.

[0158] Specifically, Figure 4 As shown, after obtaining the sample layout circuit, the features of each sample circuit logic unit in the sample layout circuit, that is, the features in the feature list, can be extracted to obtain the sample circuit features, and the sample congestion features can be generated based on the sample routing congestion report of the sample layout circuit.

[0159] In the above embodiments, the feature prediction model and the congestion prediction model are trained separately. In other implementations, the feature prediction model and the congestion prediction model may be trained jointly.

[0160] In some possible implementations of the present application, the feature prediction model and the congestion prediction model are trained in the following manner:

[0161] (1) Obtain the third sample training data set.

[0162] The third sample training data set includes multiple sets of third sample training data, and the third sample training data includes sample RTL code information, sample circuit layout strategy information and sample congestion characteristics;

[0163] (2) Based on the third sample training data set, perform at least one third training operation on the initial feature prediction model and the initial congestion prediction model until the third training end condition is met, thereby obtaining the feature prediction model and the congestion prediction model.

[0164] In some possible implementations of the present application, the third training operation includes:

[0165] ① For each group of third sample training data, the sample RTL code information and the sample circuit layout strategy information in the third sample training data are input into the initial feature prediction model to obtain predicted circuit features;

[0166] ② Input the predicted circuit features into the initial congestion prediction model to obtain the second predicted congestion features;

[0167] ③ Determine third difference information between the second predicted congestion feature and the sample congestion feature in the third sample training data;

[0168] ④ Determine the third training loss based on the third difference information corresponding to each group of third sample training data;

[0169] ⑤ Adjust the parameters of the initial congestion prediction model and the initial feature prediction model based on the third training loss, and use the initial feature prediction model after adjusting the parameters as the initial feature prediction model corresponding to the next third training operation, and use the initial congestion prediction model after adjusting the parameters as the initial congestion prediction model corresponding to the next third training operation.

[0170] It can be seen that in this embodiment, when the initial congestion prediction model and the initial feature prediction model are jointly trained, the input features of the initial congestion prediction model are the output features of the initial feature prediction model; and when the initial congestion prediction model and the initial feature prediction model are separately trained, the input features of the initial congestion prediction model are the sample circuit features.

[0171] Similarly, the third training end condition may include at least one of the following: the number of third training operations reaches a third preset number, the third training loss converges, and the third training loss is less than or equal to a third preset threshold.

[0172] Specifically, if the third training loss is determined and the third training end condition is not met, the parameters of the initial feature prediction model and the initial congestion prediction model can be adjusted based on the third training loss, and the initial feature prediction model after the adjustment of the parameters can be used as the initial feature prediction model corresponding to the next third training operation, and the initial congestion prediction model after the adjustment of the parameters can be used as the initial congestion prediction model corresponding to the next third training operation, that is, continue to repeat the next third training operation.

[0173] The above-mentioned circuit wiring congestion prediction method, through the feature prediction model, directly predicts the features associated with wiring congestion of the layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted based on the circuit layout strategy information and the RTL code information, and then predicts the congestion features of the layout circuit based on the features associated with wiring congestion of the layout circuit through the trained congestion prediction model. The congestion information corresponding to the circuit layout strategy information can be predicted before wiring, so as to determine whether the circuit layout strategy information needs to be adjusted, which can effectively reduce the design iteration cycle of circuit wiring and improve circuit wiring efficiency.

[0174] like Figure 5 As shown, in some possible implementations of the present application, a circuit wiring method is provided, including:

[0175] Step S501, obtaining RTL code information, performing at least one congestion prediction operation based on the RTL code information, until the predicted congestion characteristics meet the preset congestion characteristics, and determining circuit layout strategy information that meets the preset congestion characteristics;

[0176] Step S502 : generating a corresponding layout circuit based on the circuit layout strategy information and the RTL code information.

[0177] The congestion prediction operations include:

[0178] (1) Obtain register transfer level RTL code information and obtain circuit layout strategy information corresponding to the current congestion prediction operation;

[0179] (2) predicting circuit features associated with circuit routing congestion in RTL code information based on circuit layout strategy information using the trained feature prediction model;

[0180] (3) Predicting the congestion characteristics of the layout circuit based on the circuit characteristics through the trained congestion prediction model, and determining the circuit layout strategy information corresponding to the next congestion prediction operation based on the circuit layout strategy information and congestion characteristics corresponding to the current congestion prediction operation.

[0181] The circuit feature is used to characterize the features associated with routing congestion of the layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted, and the circuit feature includes a logic unit feature of at least one circuit logic unit of the layout circuit;

[0182] The congestion feature includes a congestion value corresponding to each circuit logic unit in the layout circuit, and the congestion value is used to characterize the congestion degree of the circuit logic unit.

[0183] Specifically, the specific training process of the congestion prediction model and the feature prediction model has been described in detail in the above embodiments and will not be repeated here.

[0184] The circuit wiring method of the present application performs at least one congestion prediction operation. Each time the congestion prediction operation is performed, the corresponding circuit layout strategy information is selected, and the circuit characteristics are predicted based on the RTL code information and the circuit layout strategy information. The congestion characteristics are then predicted based on the circuit characteristics. That is, each time the circuit layout strategy information is selected, the corresponding congestion information can be predicted, thereby reducing the iteration cycle of circuit wiring and effectively improving the circuit wiring efficiency.

[0185] like Figure 6 As shown, in some possible implementations of the present application, a circuit wiring congestion prediction device 60 is provided, comprising:

[0186] A first information acquisition module 601 is used to acquire register transfer level RTL code information and circuit layout strategy information;

[0187] A feature prediction module 602 is used to predict circuit features associated with circuit routing congestion in the RTL code information based on the circuit layout strategy information by using a trained feature prediction model, wherein the circuit features are used to characterize features associated with routing congestion of a layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted, and the circuit features include logic unit features of at least one circuit logic unit of the layout circuit;

[0188] The congestion prediction module 603 is used to predict the congestion characteristics of the layout circuit based on the circuit characteristics through the trained congestion prediction model, wherein the congestion characteristics include the congestion values ​​corresponding to each circuit logic unit in the layout circuit, and the congestion values ​​are used to characterize the congestion degree of the circuit logic unit.

[0189] In some possible implementations of the present application, a first training module is further included for:

[0190] Acquire a first sample training data set, wherein the first sample training data set includes multiple groups of first sample training data, and the first sample training data includes sample RTL code information, sample circuit layout strategy information, and sample circuit features of a sample layout circuit corresponding to the sample RTL code information when the sample circuit layout strategy information is adopted;

[0191] Based on the first sample training data set, at least one first training operation is performed on the initial feature prediction model until a first training end condition is met to obtain a feature prediction model.

[0192] In some possible implementations of the present application, when performing the first training operation, the first training module is specifically used to:

[0193] For each group of first sample training data, sample RTL code information and sample circuit layout strategy information in the first sample training data are input into an initial feature prediction model to obtain predicted circuit features;

[0194] Determine first difference information between the predicted circuit feature and the sample circuit feature in the first sample training data;

[0195] Determine a first training loss based on first difference information corresponding to each group of first sample training data;

[0196] The parameters of the initial feature prediction model are adjusted based on the first training loss, and the initial feature prediction model after the adjusted parameters is used as the initial feature prediction model corresponding to the next first training operation.

[0197] In some possible implementations of the present application, a second training module is further included for:

[0198] Acquire a second sample training data set, wherein the second sample training data set includes multiple sets of second sample training data, and the second sample training data includes sample circuit features and sample congestion features of sample layout circuits corresponding to the sample circuit features;

[0199] Based on the second sample training data set, at least one second training operation is performed on the initial congestion prediction model until a second training end condition is met, thereby obtaining a congestion prediction model.

[0200] In some possible implementations of the present application, when performing the second training operation, the second training module is used to:

[0201] For each group of second sample training data, inputting sample circuit features in the second sample training data into an initial congestion prediction model to obtain first predicted congestion features;

[0202] Determining second difference information between the first predicted congestion feature and the sample congestion feature in the second sample training data;

[0203] Determine a second training loss based on the second difference information corresponding to each group of second sample training data;

[0204] The parameters of the initial congestion prediction model are adjusted based on the second training loss, and the initial congestion prediction model after the adjusted parameters is used as the initial congestion prediction model corresponding to the next second training operation.

[0205] In some possible implementations of the present application, a training data acquisition module is further included, which is used to:

[0206] Acquire multiple groups of initial sample data, each of which includes sample RTL code information and sample circuit layout strategy information;

[0207] For each initial sample data, generating a corresponding sample layout circuit based on sample RTL code information and sample circuit layout strategy information, and extracting sample circuit features of the sample layout circuit;

[0208] Obtaining a sample routing congestion report corresponding to a sample layout circuit, including generating a sample congestion feature for the sample layout circuit based on the sample routing congestion;

[0209] Generate an initial training data set based on each initial sample data, a sample circuit feature and a sample congestion feature respectively corresponding to each initial sample data;

[0210] A first sample training data set is extracted from the initial training data set, and a second sample training data set is extracted from the initial training data set.

[0211] In some possible implementations of the present application, a third training module is further included for:

[0212] Acquire a third sample training data set, wherein the third sample training data set includes multiple sets of third sample training data, and the third sample training data includes sample RTL code information, sample circuit layout strategy information, and sample congestion characteristics;

[0213] Based on the third sample training data set, at least one third training operation is performed on the initial feature prediction model and the initial congestion prediction model until a third training end condition is met, thereby obtaining a feature prediction model and a congestion prediction model.

[0214] In some possible implementations of the present application, when the third training module performs the third training operation, it is specifically used to:

[0215] For each set of third sample training data, sample RTL code information and sample circuit layout strategy information in the third sample training data are input into an initial feature prediction model to obtain predicted circuit features;

[0216] Inputting the predicted circuit feature into the initial congestion prediction model to obtain a second predicted congestion feature;

[0217] Determining third difference information between the second predicted congestion feature and the sample congestion feature in the third sample training data;

[0218] Determine a third training loss based on the third difference information corresponding to each group of third sample training data;

[0219] Based on the third training loss, the parameters of the initial congestion prediction model and the initial feature prediction model are adjusted, and the initial feature prediction model after the adjustment of the parameters is used as the initial feature prediction model corresponding to the next third training operation, and the initial congestion prediction model after the adjustment of the parameters is used as the initial congestion prediction model corresponding to the next third training operation.

[0220] The above-mentioned circuit wiring congestion prediction device, through a feature prediction model, directly predicts the features associated with wiring congestion of the layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted based on the circuit layout strategy information and the RTL code information, and then predicts the congestion features of the layout circuit based on the features associated with wiring congestion of the layout circuit through the trained congestion prediction model. The congestion information corresponding to the circuit layout strategy information can be predicted before wiring, so as to determine whether the circuit layout strategy information needs to be adjusted, which can effectively reduce the design iteration cycle of circuit wiring and improve circuit wiring efficiency.

[0221] In some possible implementations of the present application, such as Figure 7 As shown, a circuit wiring device 70 is provided, comprising:

[0222] A second information acquisition module 701 is used to acquire RTL code information, perform at least one congestion prediction operation based on the RTL code information, until the predicted congestion characteristics meet the preset congestion characteristics, and determine the circuit layout strategy information that meets the preset congestion characteristics;

[0223] A circuit generation module 702, for generating a corresponding layout circuit based on the circuit layout strategy information and the RTL code information;

[0224] The congestion prediction operations include:

[0225] Obtain register transfer level RTL code information, and obtain circuit layout strategy information corresponding to the current congestion prediction operation;

[0226] Predicting circuit features associated with circuit routing congestion in the RTL code information based on the circuit layout strategy information by using the trained feature prediction model, wherein the circuit features are used to characterize features associated with routing congestion of a layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted, and the circuit features include logic unit features of at least one circuit logic unit of the layout circuit;

[0227] The trained congestion prediction model is used to predict the congestion characteristics of the layout circuit based on the circuit characteristics, and the circuit layout strategy information corresponding to the next congestion prediction operation is determined based on the circuit layout strategy information and congestion characteristics corresponding to the current congestion prediction operation, wherein the congestion characteristics include congestion values ​​corresponding to each circuit logic unit in the layout circuit, and the congestion values ​​are used to characterize the congestion degree of the circuit logic unit.

[0228] The circuit wiring device of the present application performs at least one congestion prediction operation. Each time the congestion prediction operation is performed, the corresponding circuit layout strategy information is selected, and circuit characteristics are predicted based on the RTL code information and the circuit layout strategy information. The congestion characteristics are then predicted based on the circuit characteristics. That is, each time the circuit layout strategy information is selected, the corresponding congestion information can be predicted, thereby reducing the iteration cycle of circuit wiring and effectively improving the circuit wiring efficiency.

[0229] In an alternative embodiment, an electronic device is provided, such as Figure 8 As shown, Figure 8 The electronic device 4000 shown includes: a processor 4001 and a memory 4003. The processor 4001 and the memory 4003 are connected, such as through a bus 4002. Optionally, the electronic device 4000 may also include a transceiver 4004, which may be used for data interaction between the electronic device and other electronic devices, such as data transmission and / or data reception. It should be noted that in actual applications, the transceiver 4004 is not limited to one, and the structure of the electronic device 4000 does not constitute a limitation on the embodiments of the present application.

[0230] Processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. Processor 4001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0231] The bus 4002 may include a path to transmit information between the above components. The bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 4002 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 8 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0232] The memory 4003 may be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), magnetic disk storage media, other magnetic storage devices, or any other medium that can be used to carry or store computer programs and can be read by a computer, without limitation herein.

[0233] The memory 4003 is used to store the computer program for executing the embodiment of the present application, and the execution is controlled by the processor 4001. The processor 4001 is used to execute the computer program stored in the memory 4003 to implement the steps shown in the above method embodiment.

[0234] An embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps and corresponding contents of the aforementioned method embodiment can be implemented.

[0235] The embodiment of the present application also provides a computer program product, including a computer program, which can implement the steps and corresponding contents of the aforementioned method embodiment when executed by a processor.

[0236] It should be understood that, although each operation step is indicated by arrows in the flowchart of the embodiment of the present application, the implementation order of these steps is not limited to the order indicated by the arrows. Unless clearly stated herein, in some implementation scenarios of the embodiment of the present application, the implementation steps in each flowchart can be performed in other orders according to demand. In addition, some or all of the steps in each flowchart may include multiple sub-steps or multiple stages based on actual implementation scenarios. Some or all of these sub-steps or stages may be executed at the same time, and each sub-step or stage in these sub-steps or stages may also be executed at different times respectively. In different scenarios of execution time, the execution order of these sub-steps or stages may be flexibly configured according to demand, and the embodiment of the present application does not limit this.

[0237] The above are only optional implementation methods for some implementation scenarios of the present application. It should be pointed out that for ordinary technicians in this technical field, without departing from the technical concept of the scheme of the present application, other similar implementation methods based on the technical ideas of the present application are also within the protection scope of the embodiments of the present application.

Claims

1. A circuit wiring congestion prediction method, characterized in that: include: Obtain register transfer level RTL code information and obtain circuit layout strategy information; Predicting, by means of a trained feature prediction model, circuit features associated with circuit routing congestion in the RTL code information based on the circuit layout strategy information, wherein the circuit features are used to characterize features associated with routing congestion of a layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted, and the circuit features include logic unit features of at least one circuit logic unit of the layout circuit; The congestion characteristics of the layout circuit are predicted based on the circuit characteristics through the trained congestion prediction model, wherein the congestion characteristics include congestion values ​​corresponding to each circuit logic unit in the layout circuit, and the congestion values ​​are used to characterize the congestion degree of the circuit logic unit.

2. The method according to claim 1, characterized in that The feature prediction model is trained in the following way: Acquire a first sample training data set, wherein the first sample training data set includes multiple groups of first sample training data, and the first sample training data includes sample RTL code information, sample circuit layout strategy information, and sample circuit features of a sample layout circuit corresponding to the sample RTL code information when the sample circuit layout strategy information is adopted; Based on the first sample training data set, at least one first training operation is performed on the initial feature prediction model until a first training end condition is met to obtain a feature prediction model.

3. The method according to claim 2, characterized in that The first training operation includes: For each group of first sample training data, sample RTL code information and sample circuit layout strategy information in the first sample training data are input into an initial feature prediction model to obtain predicted circuit features; Determining first difference information between the predicted circuit feature and the sample circuit feature in the first sample training data; Determine a first training loss based on first difference information corresponding to each group of first sample training data; The parameters of the initial feature prediction model are adjusted based on the first training loss, and the initial feature prediction model after the adjusted parameters is used as the initial feature prediction model corresponding to the next first training operation.

4. The method according to claim 2, characterized in that: The congestion prediction model is trained in the following way: Acquire a second sample training data set, wherein the second sample training data set includes multiple groups of second sample training data, and the second sample training data includes sample circuit features and sample congestion features of sample layout circuits corresponding to the sample circuit features; Based on the second sample training data set, at least one second training operation is performed on the initial congestion prediction model until a second training end condition is met to obtain a congestion prediction model.

5. The method according to claim 4, characterized in that The second training operation includes: For each group of second sample training data, inputting sample circuit features in the second sample training data into an initial congestion prediction model to obtain first predicted congestion features; Determining second difference information between the first predicted congestion feature and the sample congestion feature in the second sample training data; Determine a second training loss based on the second difference information corresponding to each group of second sample training data; The parameters of the initial congestion prediction model are adjusted based on the second training loss, and the initial congestion prediction model after the adjusted parameters is used as the initial congestion prediction model corresponding to the next second training operation.

6. The method according to claim 4, characterized in that The first sample training data set and the second sample training data set are obtained in the following manner: Acquire multiple groups of initial sample data, each of which includes sample RTL code information and sample circuit layout strategy information; For each of the initial sample data, generating a corresponding sample layout circuit based on the sample RTL code information and the sample circuit layout strategy information, and extracting sample circuit features of the sample layout circuit; Obtaining a sample routing congestion report corresponding to the sample layout circuit, including generating a sample congestion feature for the sample layout circuit based on the sample routing congestion; Generate an initial training data set based on each of the initial sample data, and the sample circuit features and sample congestion features respectively corresponding to each of the initial sample data; A first sample training data set is extracted from the initial training data set, and a second sample training data set is extracted from the initial training data set.

7. The method according to claim 1, characterized in that The feature prediction model and the congestion prediction model are trained in the following manner: Acquire a third sample training data set, wherein the third sample training data set includes multiple sets of third sample training data, and the third sample training data includes sample RTL code information, sample circuit layout strategy information, and sample congestion characteristics; Based on the third sample training data set, at least one third training operation is performed on the initial feature prediction model and the initial congestion prediction model until a third training end condition is met, thereby obtaining a feature prediction model and a congestion prediction model.

8. The method according to claim 7, characterized in that The third training operation includes: For each set of third sample training data, inputting sample RTL code information and sample circuit layout strategy information in the third sample training data into an initial feature prediction model to obtain predicted circuit features; Inputting the predicted circuit feature into an initial congestion prediction model to obtain a second predicted congestion feature; Determining third difference information between the second predicted congestion feature and the sample congestion feature in the third sample training data; Determine a third training loss based on the third difference information corresponding to each group of third sample training data; Based on the third training loss, the parameters of the initial congestion prediction model and the initial feature prediction model are adjusted, and the initial feature prediction model after the adjustment of the parameters is used as the initial feature prediction model corresponding to the next third training operation, and the initial congestion prediction model after the adjustment of the parameters is used as the initial congestion prediction model corresponding to the next third training operation.

9. A circuit wiring method, characterized in that: include: Acquire RTL code information, perform at least one congestion prediction operation based on the RTL code information until the predicted congestion characteristics meet the preset congestion characteristics, and determine circuit layout strategy information that meets the preset congestion characteristics; Based on the circuit layout strategy information and the RTL code information, generating a corresponding layout circuit; The congestion prediction operation includes: Obtain register transfer level RTL code information, and obtain circuit layout strategy information corresponding to the current congestion prediction operation; Predicting, by means of a trained feature prediction model, circuit features associated with circuit routing congestion in the RTL code information based on the circuit layout strategy information, wherein the circuit features are used to characterize features associated with routing congestion of a layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted, and the circuit features include logic unit features of at least one circuit logic unit of the layout circuit; The congestion characteristics of the layout circuit are predicted based on the circuit characteristics through the trained congestion prediction model, and the circuit layout strategy information corresponding to the next congestion prediction operation is determined based on the circuit layout strategy information corresponding to the current congestion prediction operation and the congestion characteristics, wherein the congestion characteristics include congestion values ​​corresponding to each circuit logic unit in the layout circuit, and the congestion values ​​are used to characterize the congestion degree of the circuit logic unit.

10. A circuit wiring congestion prediction device, characterized in that: include: A first information acquisition module is used to acquire register transfer level RTL code information and circuit layout strategy information; a feature prediction module, configured to predict, by means of a trained feature prediction model, circuit features associated with circuit routing congestion in the RTL code information based on the circuit layout strategy information, wherein the circuit features are used to characterize features associated with routing congestion of a layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted, and the circuit features include logic unit features of at least one circuit logic unit of the layout circuit; A congestion prediction module is used to predict the congestion characteristics of the layout circuit based on the circuit characteristics through a trained congestion prediction model, wherein the congestion characteristics include congestion values ​​corresponding to each circuit logic unit in the layout circuit, and the congestion values ​​are used to characterize the congestion degree of the circuit logic unit.

11. A circuit wiring device, characterized in that: include: A second information acquisition module is used to acquire RTL code information, perform at least one congestion prediction operation based on the RTL code information, until the predicted congestion characteristics meet the preset congestion characteristics, and determine the circuit layout strategy information that meets the preset congestion characteristics; A circuit generation module, used to generate a corresponding layout circuit based on the circuit layout strategy information and the RTL code information; The congestion prediction operation includes: Obtain register transfer level RTL code information, and obtain circuit layout strategy information corresponding to the current congestion prediction operation; Predicting, by means of a trained feature prediction model, circuit features associated with circuit routing congestion in the RTL code information based on the circuit layout strategy information, wherein the circuit features are used to characterize features associated with routing congestion of a layout circuit corresponding to the RTL code information when the circuit layout strategy information is adopted, and the circuit features include logic unit features of at least one circuit logic unit of the layout circuit; The congestion characteristics of the layout circuit are predicted based on the circuit characteristics through the trained congestion prediction model, wherein the congestion characteristics include congestion values ​​corresponding to each circuit logic unit in the layout circuit, and the congestion values ​​are used to characterize the congestion degree of the circuit logic unit.

12. An electronic device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 9.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 9 are implemented.

14. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 9 are implemented.

Images