Hierarchical static time sequence analysis method and system for FPGA (Field Programmable Gate Array) chip

Through the hierarchical static timing analysis method, the problems of mismatch and low efficiency in FPGA chip design are solved, efficient and accurate timing analysis is achieved, adapting to complex design needs, and improving the flexibility and adaptability of analysis.

CN119026542APending Publication Date: 2024-11-26SHANGHAI ANLOGIC INFOTECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411033713.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-30
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

The static timing analysis method of traditional FPGA chips has problems such as timing information mismatch, risk of human error, low work efficiency, difficulty in verification, insufficient analysis efficiency, and insufficient adaptability, and it is difficult to meet the needs of complex FPGA chip design.

Method used

Using a hierarchical static timing analysis method, by generating files containing timing model information and delay information, marking keywords for each functional module, obtaining netlist information to find matching timing arcs, and selecting and building according to the reference clock position The model method performs timing modeling, generates timing analysis reports, and uses standard format Liberty files and keyword tags to improve analysis accuracy and efficiency.

Benefits of technology

The results of FPGA timing analysis and mainstream ASIC timing analysis engines in the industry are achieved, avoiding mismatch between timing model and arc delay information, significantly improving the accuracy and efficiency of analysis, simplifying the engineering practice process, and enhancing the analysis Flexibility and adaptability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119026542A_ABST
    Figure CN119026542A_ABST
Patent Text Reader

Abstract

The invention relates to the field of integrated circuit design, and discloses a hierarchical static time sequence analysis method and system of an FPGA (Field Programmable Gate Array) chip. The method comprises the following steps: generating a Liberty file containing time sequence model information and delay information for different working modes of each functional module in the FPGA chip; marking keywords for the time sequence arc of each functional module in the file; obtaining netlist information of the FPGA chip, and searching a matched time sequence arc; selecting a time sequence diagram modeling or input / output port delay modeling method according to the reference clock position of the time sequence arc; performing static time sequence analysis based on the constructed time sequence model; and generating a time sequence analysis report. According to the method, by adopting a standard Liberty file format and a flexible modeling strategy, the consistency of FPGA time sequence analysis and an industrial standard tool is realized, the analysis accuracy and efficiency are improved, the engineering practice process is simplified, and the adaptability to complex FPGA design is enhanced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of integrated circuit design, and in particular to static timing analysis technology for FPGA chips. Background Art

[0002] Static timing analysis is a key technology for evaluating performance and ensuring mass production yield in FPGA chip design. Mainstream FPGA chip companies such as Xilinx, Altera, and Anlogic all provide timing analysis engines that match their chips for timing analysis and sign-off of user designs. With the increasing complexity of PGA chip functions and the continuous improvement of integration, how to accurately and efficiently perform static timing analysis on these complex systems has become an urgent problem to be solved.

[0003] Traditional FPGA timing analysis methods have many shortcomings, for example:

[0004] Consistency issues of timing information: In traditional methods, timing model files and timing arc delay files are usually provided and maintained independently, which easily leads to information mismatch between the two, thus affecting the accuracy of timing analysis.

[0005] Risk of human error: Timing model files and timing arc delay files are often manually written by designers based on their understanding of the design. This process can easily introduce a large number of unknown errors, reducing the reliability of timing analysis.

[0006] Low work efficiency: With the increase in the number of integrated modules in FPGA chips and the complexity of working modes, designers need to spend a lot of time to describe timing models and timing arc delays, which seriously affects work efficiency.

[0007] Increased verification difficulty: Complex FPGA chip design brings more challenges to verification work. Verifiers need to invest a lot of time to learn the design and verify the accuracy of the model, which is not only time-consuming and labor-intensive, but may also affect the comprehensiveness and accuracy of the verification.

[0008] Insufficient analysis efficiency: As the complexity of FPGA chips increases, traditional timing analysis methods are less efficient when dealing with large-scale designs and cannot meet the needs of rapid iterative development.

[0009] Insufficient adaptability to different working modes: Existing methods have limitations in processing the differences in timing information of FPGA modules in different working modes, making it difficult to fully and accurately reflect the performance of the chip under various actual working conditions.

[0010] Insufficient dynamic adaptability during the design process: Traditional methods are unable to effectively support the dynamic requirements of FPGA chip design and are unable to provide the most suitable timing analysis results at different stages of the design.

[0011] In the face of these technical challenges, a new static timing analysis method for FPGA chips is urgently needed to improve analysis accuracy, enhance efficiency, simplify engineering practice processes, enhance analysis flexibility and adaptability, and support the dynamic needs of the design process. In addition, the new method also needs to ensure consistency with the analysis results of industry standard tools to promote consistency and reliability of design verification. Summary of the invention

[0012] The purpose of this application is to provide a hierarchical static timing analysis method and system for FPGA chips to solve the problems raised in the above background technology.

[0013] The present application discloses a hierarchical static timing analysis method for an FPGA chip, comprising the following steps:

[0014] Step A: Generate a file containing timing model information and delay information for different working modes of each functional module in the FPGA chip;

[0015] Step B: in the file, marking a keyword for the timing arc of each functional module;

[0016] Step C: Obtaining the netlist information of the FPGA chip, and searching for a timing arc that matches the functional module and the working mode according to the netlist information;

[0017] Step D: For each timing arc found, select a modeling method according to its reference clock position to perform timing modeling;

[0018] Step E: Based on the constructed timing model, static timing analysis is performed on the FPGA chip;

[0019] Step F: Generate a timing analysis report to guide the optimization and verification of FPGA chip design.

[0020] In a preferred example, in step A, a hardware timing signoff tool is used to generate a Liberty file based on a timing signoff database.

[0021] In a preferred example, in step B, all timing arcs of each functional module are marked with the module keyword, and the timing arcs in each working mode are marked with the when keyword.

[0022] In a preferred example, the netlist information in step C is used to determine the connection relationship and current working mode of each functional module.

[0023] In a preferred embodiment, the modeling method in step D includes:

[0024] a) When the reference clock is outside the functional module currently being analyzed, the timing diagram modeling method is used;

[0025] b) When the reference clock is located inside the functional module currently being analyzed, the input and output port delay modeling method is used.

[0026] In a preferred embodiment, when the timing diagram modeling method is adopted, the timing arc delay information is directly marked into the timing diagram.

[0027] In a preferred example, when using the input and output port delay modeling method, the input port of the functional module is regarded as the output port of the chip top-level module, and the output port of the functional module is regarded as the input port of the chip top-level module, and the delay information is used to set the corresponding input and output delays.

[0028] In a preferred example, in step E, a FPGA timing analysis engine is used to perform static timing analysis, and the FPGA timing analysis engine uses the keywords marked in step B to identify and process the timing information of different functional modules in different working modes.

[0029] The present application also discloses a hierarchical static timing analysis system for FPGA chips, including:

[0030] A file generation module, used for generating files containing timing model information and delay information for different working modes of each functional module in the FPGA chip;

[0031] A keyword marking module, used for marking a keyword for a timing arc of each functional module in the file;

[0032] A timing arc matching module is used to obtain the netlist information of the FPGA chip and search for a timing arc that matches the functional module and the working mode according to the netlist information;

[0033] A timing modeling module, for selecting a modeling method for each found timing arc according to its reference clock position to perform timing modeling;

[0034] Static timing analysis module, used to perform static timing analysis on FPGA chips based on the constructed timing model;

[0035] The report generation module is used to generate timing analysis reports to guide the optimization and verification of FPGA chip design.

[0036] The present application also discloses a hierarchical static timing analysis system for FPGA chips, including:

[0037] a memory for storing computer executable instructions; and,

[0038] A processor is used to implement the steps in the method described above when executing the computer executable instructions.

[0039] This application proposes a hierarchical static timing analysis method for FPGA chips, which uses a standard format Liberty file to transmit and process timing information, thereby achieving efficient and accurate timing analysis. The specific technical effects are as follows:

[0040] 1) Combining FPGA timing analysis with the timing analysis process of mainstream ASIC timing analysis engines in the industry: By using the Liberty file generated by the mainstream ASIC timing analysis engine in the industry, this application achieves consistency in the timing analysis results of the FPGA timing analysis engine and the mainstream ASIC timing analysis engine in the industry. The standard format of the Liberty file can describe both the timing model information and the delay information, ensuring the accuracy and reliability of the FPGA timing analysis results.

[0041] 2) Application of the keywords "module" and "when": This application first proposed the use of the keywords "module" and "when" to distinguish the timing information of different modules and their working modes. This marking method effectively avoids the problem of mismatch between the timing model and the timing arc delay information, and improves the systematization and orderliness of timing information management.

[0042] 3) Timing modeling method based on Liberty file: Two timing modeling methods based on Liberty file are proposed: timing diagram modeling and input / output port delay modeling. The timing diagram modeling method integrates timing information and delay information into the timing diagram, avoiding the need to repeatedly search for timing arc delay information and significantly reducing the running time of timing analysis. The input / output port delay modeling method is suitable for situations where the internal clock generation circuit is complex, and simplifies the timing analysis process through equivalent processing.

[0043] 4) Modular design and black box processing: Through the hierarchical timing analysis method, the top-level module and functional module of the FPGA chip are processed separately, which not only ensures the accuracy of the overall timing performance of the chip, but also reduces the complexity of the timing diagram and analysis time. The black box processing method allows timing analysis without knowing the specific implementation inside the module, which promotes modular design and IP reuse.

[0044] 5) Automated Liberty file extraction and timing modeling: This application uses fully automated Liberty file extraction and timing modeling of the FPGA timing engine, which reduces the errors caused by designers manually writing timing models and timing arc delay files, saves time costs for design and verification, and greatly simplifies the verification of timing models.

[0045] 6) Unified integration of delay information: When constructing the module timing diagram, the timing arc delay information is directly marked into the timing diagram, eliminating the need to repeatedly search for the corresponding delay during the delay calculation process, significantly improving the efficiency of timing analysis.

[0046] In summary, this application realizes efficient, accurate and systematic hierarchical static timing analysis of FPGA chips through standardized Liberty files and keyword tagging methods, significantly improves the accuracy and efficiency of timing analysis, and has important technical advantages and application value.

[0047] A large number of technical features are recorded in the specification of this application, which are distributed in various technical solutions. If all possible combinations of technical features of this application (i.e., technical solutions) are to be listed, the specification will be too long. In order to avoid this problem, the various technical features disclosed in the above-mentioned invention content of this application, the various technical features disclosed in the various embodiments and examples below, and the various technical features disclosed in the accompanying drawings can be freely combined with each other to form various new technical solutions (these technical solutions are all deemed to have been recorded in this specification), unless the combination of such technical features is technically infeasible. For example, in one example, feature A+B+C is disclosed, and in another example, feature A+B+D+E is disclosed, and features C and D are equivalent technical means that play the same role. Technically, only one can be used, and it is impossible to use them at the same time. Feature E can be combined with feature C technically. Then, the solution of A+B+C+D should not be deemed to have been recorded because it is technically infeasible, and the solution of A+B+C+E should be deemed to have been recorded. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a schematic diagram of the workflow of the hierarchical static timing analysis method for FPGA chips according to the first embodiment of the present application.

[0049] Figure 2 It is a schematic diagram of the main information stored in each timing arc in the hierarchical static timing analysis method of the FPGA chip according to the first embodiment of the present application.

[0050] Figure 3 It is a schematic diagram of a timing diagram modeling method in a hierarchical static timing analysis method for an FPGA chip according to the first embodiment of the present application, showing a situation where a reference clock is located outside a module.

[0051] Figure 4 It is a schematic diagram of the input and output port delay modeling method in the hierarchical static timing analysis method of the FPGA chip according to the first embodiment of the present application, showing the situation where the reference clock is located inside the module.

[0052] Figure 5It is a flowchart of a hierarchical static timing analysis method for FPGA chips according to the first embodiment of the present application.

[0053] Figure 6 It is a structural diagram of a hierarchical static timing analysis system for FPGA chips according to the second embodiment of the present application. DETAILED DESCRIPTION

[0054] In the following description, many technical details are provided to help readers better understand the present application. However, those skilled in the art can understand that the technical solution claimed in the present application can be implemented even without these technical details and various changes and modifications based on the following embodiments.

[0055] Description of some concepts:

[0056] FPGA (Field Programmable Gate Array): Field programmable gate array is an integrated circuit that can be programmed to achieve specific functions and is widely used in communications, computing, automation and other fields.

[0057] STA (Static Timing Analysis): Static timing analysis is a method of checking and verifying the timing of circuit design to ensure that the circuit can meet the expected timing requirements under various operating conditions.

[0058] ATE (Anlogic Timing Engine): Anlogic static timing analysis engine is a timing analysis tool provided by Anlogic for its FPGA chips. It is used to perform timing analysis and sign-off on user designs.

[0059] Liberty file: An interactive file in the field of computer-aided design that describes the timing model and delay information of a design or circuit unit. It is used to describe the timing model and delay information of a circuit. Liberty files can contain both timing model information and delay information and are widely used in static timing analysis.

[0060] Timing arc: In timing analysis, it represents the delay path of a signal from one timing point (such as an input port or register) to another timing point. The information of the timing arc includes the target port, reference port, timing arc type, and delay value.

[0061] Parallel arc: Indicates that there are multiple possible timing paths physically between two specific pins of a unit. Only one of them is activated in a specific operating mode.

[0062] Clock Tree: In synchronous digital circuits, a network structure used to distribute clock signals from the source to each trigger. The design of the clock tree directly affects the performance and power consumption of the chip.

[0063] Timing Sign-off: In the chip design process, the process of confirming that the design meets all timing requirements through comprehensive static timing analysis.

[0064] Timing diagram modeling: A timing modeling method that intuitively represents the timing characteristics of a circuit by constructing a timing diagram that contains all the timing and delay information.

[0065] Input / output port delay modeling: A timing modeling method that is applicable to the case where the target port reference clock is located inside the module. Timing analysis is performed by treating the module's input / output port delays as equivalent to the input / output delays of the chip's top-level module.

[0066] Black box processing: In timing analysis, some complex modules are treated as "black boxes", and only the timing information of their input and output ports is focused on without analyzing their internal implementation in detail. This method helps to simplify the timing analysis process and improve analysis efficiency.

[0067] Module keyword (module): used to mark the timing information in the Liberty file and distinguish the timing arcs of different functional modules.

[0068] Operating Mode: The operating state of an FPGA chip or its internal modules under different conditions. Different operating modes may affect the timing characteristics of the circuit.

[0069] Working mode keyword (when): used to mark the timing information in the Liberty file and distinguish the timing arcs of the same module in different working modes.

[0070] The following is a brief description of the technical concept and overall technical solution of this application:

[0071] In order to meet the above technical challenges, the inventor of this application has proposed an innovative static timing analysis method for FPGA chips after in-depth research and analysis. This method not only solves the problems existing in traditional methods, but also introduces some new technical means to further improve the accuracy and efficiency of timing analysis.

[0072] There are two main types of static timing analysis methods for chips: flatten timing analysis and hierarchical timing analysis. The main difference between these two analysis methods lies in the different timing modeling methods.

[0073] Flattened timing analysis refers to treating the entire FPGA chip as a whole, and integrating all timing models into a complete timing diagram in the form of a timing diagram during the timing modeling process. This method can comprehensively analyze the timing of the entire chip, but for complex chip designs, the timing diagram may become very large, resulting in a long analysis process.

[0074] In contrast, hierarchical timing analysis uses a different approach. It treats the top layer of the chip as a whole, describes the top layer as a timing diagram through timing modeling, and integrates some modules in the chip into the top layer timing diagram by building a timing diagram, while treating other IP modules as black boxes. IP treated as a black box only focuses on the timing information at the port. The timing arcs of the black box IP will not be mapped to the timing diagram, but will only be marked as port delays in the way of input and output delays.

[0075] When designing FPGA chips, the timing results inside the black box IP are guaranteed to converge by the chip designer through the ASIC timing sign-off tool, and the user of the FPGA chip as a black box IP only needs to perform timing analysis on the IP port. This method can not only ensure the normal operation of the chip, but also effectively reduce the size of the timing diagram, thereby saving the running time of the timing analysis.

[0076] Hierarchical timing analysis treats the entire IP as a black box and only analyzes the timing of the black box IP port, which has significant advantages. It not only ensures the normal operation of the chip, but also greatly reduces the scale of the timing diagram that needs to be analyzed, thereby improving the analysis efficiency. This method is particularly suitable for the timing analysis of complex functional modules, and can complete the analysis task faster and better.

[0077] In general, the hierarchical timing analysis method significantly improves the analysis efficiency while ensuring the accuracy of the analysis by reasonably dividing the analysis levels and scope, and provides an effective static timing analysis solution for complex FPGA chip design.

[0078] The hierarchical timing analysis method for FPGA chips mainly includes two key parts: first, using the hardware timing signoff tool to extract the Liberty timing files, and second, the FPGA timing analysis engine to identify and model these Liberty timing files. These two parts are closely combined to form a complete hierarchical timing analysis process.

[0079] Secondly, considering that FPGA chip design is a dynamic process, a strategy of providing Liberty files at different design stages is proposed. For example, in the early stage of module design, Liberty files can be extracted based on the phased results of logic synthesis. These files are mainly used for preliminary testing of FPGA chip timing analysis, which can detect potential timing problems early. After the module design is completed and the timing is signed off, more accurate Liberty files can be extracted based on the timing sign-off database. These files will be used for the final timing analysis of mass-produced FPGA chips to ensure that the timing performance of the chip in actual applications meets the requirements.

[0080] Finally, an important innovation of this application is the strategy of extracting independent Liberty files according to modules and their different working modes. This means that each module will have a corresponding Liberty file in different working modes to describe the timing model and timing arc delay information in that mode. This fine-grained classification method enables timing analysis to more accurately reflect the performance of the chip under various actual working conditions, greatly improving the accuracy and reliability of the analysis.

[0081] Through this multi-level, multi-stage Liberty file extraction method, this application provides a comprehensive and detailed timing information foundation for hierarchical timing analysis of FPGA chips. This not only improves the accuracy of timing analysis, but also enhances the credibility of analysis results, providing strong support for performance optimization and quality assurance of FPGA chips.

[0082] In the hierarchical timing analysis method of FPGA chips, the storage and timing modeling of Liberty files are key links. The embodiments of the present application adopt a series of innovative strategies to effectively manage and utilize the timing information in Liberty files.

[0083] First, for the Liberty information of different modules, the timing analysis engine introduced the "Module" keyword as a distinguishing identifier. This means that all timing arcs of each module share an independent Module keyword. This method enables the timing analysis engine to quickly identify and distinguish timing information from different modules, effectively improving the efficiency of data management and retrieval.

[0084] Secondly, in order to process the Liberty information of the same module in different working modes, the timing analysis engine cleverly uses the "when" keyword and the concept of parallel timing arcs. Specifically, all timing arcs in a specific working mode share the same when keyword. This design allows the timing analysis engine to accurately identify and analyze the timing characteristics of the module in different working states, thereby providing more detailed and accurate timing analysis results.

[0085] In terms of timing modeling, the FPGA timing analysis engine mainly uses two modeling methods based on Liberty information: timing diagram modeling and port constraint modeling. The choice of modeling method mainly depends on the characteristics of the module. If the module contains clock tree delay information, port constraint modeling is required. This method can better handle complex clock tree structures. For modules that do not contain clock tree delay information, timing diagram modeling can be used to directly integrate timing information and delay information into the timing diagram. This method is usually more intuitive and efficient.

[0086] When the timing analysis engine performs timing modeling, it also models the timing arcs separately according to their types. There are four main types: combinational logic, sequential logic, check logic, and reset logic. This classification modeling method enables the timing analysis to more accurately reflect the timing characteristics of different types of logic circuits, thereby providing more comprehensive and accurate analysis results.

[0087] Finally, the embodiment of the present application adopts an important optimization strategy in the process of constructing the timing diagram: while constructing the timing diagram, the timing arc delay information is directly marked into the timing diagram. This approach avoids the need to repeatedly search for timing arcs in the subsequent delay calculation process, significantly reducing the overhead of data access and processing. This innovation greatly improves the operating efficiency of the FPGA timing analysis engine and effectively shortens the overall running time of the timing analysis.

[0088] Through these carefully designed storage and modeling strategies, the embodiments of the present application not only improve the management efficiency of Liberty file information, but also significantly enhance the accuracy and speed of timing analysis. This provides strong technical support for the timing analysis of complex FPGA chips and helps to improve the overall performance and reliability of FPGA chips.

[0089] Figure 1 A hierarchical timing analysis workflow for FPGA chips is presented in detail. This workflow fully reflects the hierarchical structure and modular design concept of FPGA chips, and provides an efficient and flexible method for timing analysis of complex FPGA systems.

[0090] In this FPGA chip design, the entire chip consists of three main parts: a top-level module (marked as top) and two functional modules (marked as a and b respectively). This structural design reflects the typical architecture of modern FPGA chips, which is to achieve complex functions through modular design while maintaining the clarity and manageability of the overall structure.

[0091] It is worth noting that these two functional modules have different working modes, which reflects the versatility and flexibility of FPGA chips in practical applications. Specifically, module a is designed with two different working modes, while module b has three working modes. This multi-mode design enables FPGA chips to adapt to different application scenarios and work requirements, fully demonstrating the advantages of FPGA.

[0092] Although this modular and multi-mode design enhances the functionality of FPGA chips, it also brings challenges to timing analysis. Each working mode may have different timing characteristics and needs to be analyzed and verified separately. Figure 1 The hierarchical timing analysis workflow shown is particularly important, as it provides a systematic approach to handling this complexity and ensuring that the chip maintains correct timing behavior in various operating modes.

[0093] Extracting Liberty files is a key step in the hierarchical timing analysis process of FPGA chips. This process fully demonstrates the advantages of modular design and teamwork, and also shows how to handle multiple working modes of different modules in complex FPGA chips.

[0094] First of all, it is worth noting that the timing information of modules a and b is provided by different chip design engineers. This division of labor not only improves work efficiency, but also ensures that each module can be handled professionally and targeted. This reflects the importance of teamwork in modern FPGA chip design, where each engineer can focus on his or her own professional field, thereby improving the overall design quality.

[0095] For module a, after completing the timing sign-off, the design engineer used the hardware timing sign-off tool to generate Liberty files for its two different working modes based on the timing sign-off database. The two files are named a_1.lib and a_2.lib, corresponding to the two working modes of module a. In order to accurately identify and process these files in subsequent analysis, the designer adopted a specific keyword marking method. Specifically, all timing arcs of module a are marked with the keyword "module: a", so that all timing information belonging to module a can be easily identified. In addition, the timing arcs in a_1.lib are marked with the keyword "when: 1", while the timing arcs in a_2.lib are marked with the keyword "when: 2". This marking method enables the system to accurately distinguish the timing characteristics under different working modes.

[0096] Similarly, after completing the timing sign-off, module b also uses the hardware timing sign-off tool to extract the Liberty file based on the timing sign-off database. Since module b has three working modes, three Liberty files are generated: b_1.lib, b_2.lib, and b_3.lib. Similarly, all timing arcs of module b are marked with the keyword "module: b" for easy identification. The timing arcs in each Liberty file are marked with the corresponding working mode: b_1.lib is marked with "when: 1", b_2.lib is marked with "when: 2", and b_3.lib is marked with "when: 3".

[0097] This sophisticated file generation and marking method provides strong support for subsequent timing analysis. It not only ensures that the timing characteristics of each module in different operating modes can be accurately captured, but also provides clear structured data for the timing analysis tool, making the timing analysis of complex FPGA chips more manageable and efficient. The flexibility and scalability of this method also means that it can adapt to more complex FPGA design requirements that may arise in the future.

[0098] In the FPGA chip hierarchical timing analysis method of this application, the storage and timing modeling of Liberty files is a core link. This process uses a carefully designed data structure and tagging system to ensure the integrity, accuracy and traceability of timing information.

[0099] The timing information in the Liberty file is stored in the basic unit of timing arc. Figure 2 As shown in the figure, each timing arc contains rich information, which together describe the timing characteristics of a specific path in the FPGA chip. Specifically, each timing arc contains the following key information: target port, reference port, timing arc type, timing arc delay, module keyword module, and working mode keyword when.

[0100] Among them, the module keyword is an important innovation of this application. It is used to distinguish different modules, allowing the timing analysis system to quickly identify and classify timing information from different functional modules. This module-level distinction is crucial for the management and analysis of large FPGA designs.

[0101] Target port and reference port are the basic elements to describe the timing path. The target port refers to the specific port that the timing analysis focuses on, while the reference port is the timing reference point of the target port. These two pieces of information together define a complete timing path.

[0102] The when keyword is another important innovation, which is used to distinguish the timing information of the same module in different operating modes. This design allows the FPGA to maintain optimal performance under different operating conditions, while also providing greater flexibility for designers.

[0103] The timing arc delay information records the delay values ​​under different process angles. This multi-angle delay information enables the timing analysis to take into account various possible manufacturing and operating conditions, thereby improving the reliability of the analysis results.

[0104] The distinction between timing arc types (combinational logic, sequential logic, check logic, and reset logic) reflects the characteristics of different types of circuits inside the FPGA. This classification allows timing analysis to use the most appropriate analysis strategy for different types of logic.

[0105] In the specific implementation, the timing arcs of modules a and b are marked in the above way. All timing arcs of module a are marked with "module: a", and its two working modes are marked with "when: 1" and "when: 2" in a_1.lib and a_2.lib respectively. Similarly, the timing arcs of module b are marked with "module: b", and its three working modes are marked with "when: 1", "when: 2" and "when: 3" in b_1.lib, b_2.lib and b_3.lib respectively.

[0106] This sophisticated information organization and labeling method provides a solid foundation for hierarchical timing analysis of FPGA chips. It not only makes the management of timing information more systematic and clear, but also provides rich contextual information for subsequent timing analysis and optimization. The flexibility and scalability of this method also means that it can adapt to more complex FPGA design requirements that may appear in the future.

[0107] In the hierarchical timing analysis method of FPGA chips, timing modeling for modules is a key step. This process is mainly divided into two important stages: finding timing arcs and describing timing models. This two-step method not only makes the modeling process more systematic, but also improves the accuracy and efficiency of timing analysis.

[0108] First of all, finding the timing arc is the basis of the entire modeling process. The main purpose of this step is to find the timing arc corresponding to the current working mode for each module in the netlist. This process can be further divided into two sub-steps:

[0109] 1) Module identification: The system first needs to find the corresponding module data in the Liberty file based on the module information in the netlist. This step uses the module keyword mentioned above to quickly locate the timing information of a specific module. This module-level matching ensures that timing analysis can be accurately performed on specific functional modules.

[0110] 2) Working mode matching: After determining the module, the system will select the corresponding timing arc according to the current working mode of the module. The when keyword is used here to accurately identify the timing arc that is valid in a specific working mode. This fine-grained matching allows the timing analysis to take into account the timing differences that may exist in different working states of the module.

[0111] The advantage of this two-step search method is that it can effectively handle multi-module and multi-mode situations in complex FPGA designs. By first locating the module and then refining it to a specific operating mode, the system can quickly and accurately find the required information in the huge amount of timing data.

[0112] The process of finding the timing arc actually lays the foundation for the subsequent timing model description. Through this process, the system has obtained all relevant timing information of a specific module in a specific working mode. This information includes the definition of the timing path (target port and reference port), timing characteristics (such as delay value) and logic type.

[0113] This precise timing arc search method not only improves the accuracy of timing analysis, but also greatly improves the analysis efficiency. It allows the timing analysis tool to use the most relevant and accurate timing information for each part of the FPGA chip, thereby obtaining more reliable analysis results. At the same time, the modular nature of this method also makes it scalable and can easily cope with the increasingly complex FPGA design requirements.

[0114] In general, this method of finding timing arcs reflects the innovative thinking of the present application in dealing with complex FPGA timing analysis problems, and provides a solid foundation for achieving efficient and accurate hierarchical timing analysis.

[0115] In the hierarchical timing analysis method of FPGA chips, describing the timing model based on the information in the Liberty file is a key step. This application proposes two different modeling methods: timing diagram modeling and input and output port delay modeling. The choice of these two methods depends on the characteristics of the timing information in the Liberty file, especially the reference clock position of the target port.

[0116] First, the timing diagram modeling method is explained. This method is applicable to Figure 3In the case shown, the reference clocks of the target ports in the Liberty file are all located at the external ports of the module. This structure makes it possible to build a complete, self-contained timing diagram to describe the timing characteristics of the module.

[0117] The process of timing diagram modeling can be divided into the following steps:

[0118] 1) Timing arc identification: The system first searches for each timing analysis target port (such as Figure 3 This step makes use of the module and when keywords discussed earlier to ensure that the found timing arcs match the current module and operating mode.

[0119] 2) Timing arc classification: After finding the relevant timing arcs, the system will classify them according to the type of timing arc. There are four main types: combinational logic, sequential logic, check logic, and reset logic. This classification is crucial for accurately describing the timing characteristics of different types of logic circuits.

[0120] 3) Timing diagram construction: The system marks the classified timing information and delay information into the timing diagram.

[0121] Specifically:

[0122] For combinational logic, the timing diagram establishes a direct path from input to output and annotates the corresponding delays.

[0123] For sequential logic, the timing diagram includes register elements with setup and hold times noted.

[0124] Check logic is often used to describe signal validity checks, such as setup and hold time checks.

[0125] Reset logic describes the timing characteristics of the reset signal.

[0126] 4) Delay information marking: When constructing the timing diagram, the system will directly mark the delay information of each timing arc on the corresponding path in the diagram. This approach avoids repeated search for delay information in subsequent analysis, thereby improving analysis efficiency.

[0127] A major advantage of the timing diagram modeling approach is that it provides an intuitive and complete representation of the timing behavior of a module. By integrating all relevant timing information into a unified graphical model, designers can more easily understand and analyze the timing characteristics of a module. This approach is particularly suitable for modules whose timing relationships are relatively simple and whose reference clocks are external.

[0128] In addition, this modeling method has good scalability. As the complexity of FPGA design increases, more complex timing relationships can be described by simply adding new nodes and paths in the timing diagram. This makes the timing diagram modeling method adaptable to FPGA design requirements of various complexities.

[0129] In general, the timing diagram modeling method provides a powerful and flexible tool for hierarchical timing analysis of FPGA chips. It can not only accurately capture the timing characteristics of the module, but also present this information in an intuitive and easy-to-understand way, thus laying a solid foundation for subsequent timing analysis and optimization.

[0130] The following describes a timing modeling method for a specific situation, namely input and output port delay modeling. This method is suitable for the case where the reference clock of the target port in the Liberty file is located inside the module, such as Figure 4 As shown, the principles and steps of this method are as follows:

[0131] 1) Applicable situation: When the reference clock of the target port described in the Liberty file is inside the module (such as Figure 4 The traditional timing diagram modeling method may not be able to accurately capture the delay of the internal clock tree. In this case, the input and output port delay modeling method is required.

[0132] 2) Modeling principle: The core idea of ​​this method is to transform the complex internal timing relationship into a simple input-output delay relationship, thereby treating the module as a "black box".

[0133] 3) Modeling steps: a) Timing arc identification: According to the module name and working mode in the netlist, find the relevant timing arcs for each timing analysis target port (such as in and out) in the Liberty file. b) Input port processing: Figure 4 In the output port, the output port of the top module of the chip is regarded as the output port of the top module of the chip. For these ports, the check logic is equivalent to applying output delay (set_output_delay). This can simulate the time requirement for data to be transmitted from the top module to the input port of the module. c) Output port processing: The output port of the module (such as Figure 4d) Internal clock processing: For internally generated clocks (such as gen_clk), create it as a virtual port inside the black box. This virtual port is used to mark the delay from the external clock port (clk) to the internal generated clock (gen_clk). e) Reference clock setting: Set the internally generated clock (gen_clk) as the reference clock port of the target port (in, out). This ensures that the delay of the internal clock tree is included in the overall clock path.

[0134] This approach has the following benefits: 1) Simplify complexity: Simplify complex internal timing relationships into input and output delays, making analysis more intuitive. 2) Accuracy: By including internal clock tree delays in the clock path, the accuracy of timing analysis is improved. 3) Black box processing: Allows timing analysis without knowing the specific implementation of the module, which is conducive to modular design and IP reuse.

[0135] This approach is particularly useful for modules that contain complex internal clock generation circuits, such as Figure 4 Design shown with a DFFA (possibly a clock divider or phase-locked loop).

[0136] This input and output port delay modeling method allows designers to effectively handle the complexities of internal clock generation while maintaining the accuracy and efficiency of timing analysis. This method complements the timing diagram modeling method discussed earlier and together constitutes a comprehensive FPGA chip hierarchical timing analysis solution.

[0137] The main innovations of this application are further summarized below:

[0138] This application proposes an innovative FPGA chip hierarchical timing analysis method, which integrates multiple technical innovations and aims to improve the accuracy, efficiency and reliability of FPGA chip timing analysis. The following is a detailed description of the main innovations of this application:

[0139] First, this application combines FPGA chip timing analysis with the industry-standard hardware timing sign-off tool timing analysis process, and uses the standard format Liberty file for timing information interaction. Liberty files are a widely used standard format that can describe both timing model information and delay information. By using Liberty files, the embodiments of this application ensure the consistency of timing models and delay information, greatly improving the accuracy of timing analysis. This standardized approach not only simplifies the management of timing information, but also improves the compatibility between different tools and processes.

[0140] Secondly, this application first proposed the use of the keywords module and when to solve the problem of different timing information of FPGA modules in different working modes. The module keyword is used to identify a specific functional module, while the when keyword is used to distinguish the timing characteristics of the same module in different working modes. This innovative labeling method enables the timing analysis tool to accurately identify and process the timing information of different modules in various working modes, significantly improving the flexibility and accuracy of the analysis.

[0141] Third, this application proposes two methods for building timing models based on Liberty files: timing diagram modeling and input and output port delay modeling. These two methods complement each other and can adapt to different timing analysis scenarios. Timing diagram modeling is suitable for situations where the reference clock is outside the module, and it provides an intuitive representation of the timing relationship. Input and output port delay modeling is suitable for situations where the reference clock is inside the module, and can effectively handle internal clock tree delays. The combination of these two methods enables this application to accurately analyze various complex FPGA module structures, including those that contain internal clock generation circuits, thereby improving the comprehensiveness and accuracy of timing analysis.

[0142] Finally, in view of the characteristics of FPGA timing analysis, this application proposes for the first time to directly mark the module's timing arc delay into the timing diagram when constructing the module timing diagram. This innovative method significantly improves the efficiency of timing analysis. By embedding the delay information directly into the timing diagram, there is no need to look up the corresponding delay value again in the subsequent delay calculation process, thereby effectively reducing the running time and speeding up the overall analysis process.

[0143] In general, the combination of these innovative points makes the FPGA chip hierarchical timing analysis method of the present application have significant advantages. It not only improves the accuracy and efficiency of timing analysis, but also simplifies the workflow of engineers and reduces the possibility of human error. By combining with the timing analysis process of the hardware timing sign-off tool, the embodiment of the present application ensures the consistency of the FPGA timing analysis engine with industry standard tools and enhances the credibility of the analysis results. In addition, the flexibility of the embodiment of the present application enables it to adapt to various complex FPGA designs, including multi-module and multi-working mode situations, providing strong support for the design and verification of modern FPGA chips.

[0144] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below in conjunction with the accompanying drawings.

[0145] The first embodiment of the present application relates to a hierarchical static timing analysis method for an FPGA chip, the process of which is as follows: Figure 5 As shown, the method comprises the following steps:

[0146] Step 100: Use a hardware timing signoff tool to generate Liberty files for different working modes of each functional module in the FPGA chip based on a timing signoff database. The Liberty files contain both timing model information and delay information.

[0147] Step 200: In the Liberty file, all timing arcs of each functional module are marked with the module keyword, and the timing arcs in each working mode are marked with the when keyword.

[0148] Step 300: Obtain the netlist information of the FPGA chip, which is used to determine the connection relationship and current working mode of each functional module, and according to the functional module name and current working mode in the netlist information, search for the timing arc that matches the functional module and the current working mode in the corresponding Liberty file generated in step 100.

[0149] Step 400: For each timing arc found in step 300, one of the following two modeling methods is selected to perform timing modeling according to its reference clock position:

[0150] a) When the reference clock is outside the functional module currently being analyzed, a timing diagram modeling method is used, and while constructing the timing diagram, the timing arc delay information contained in the Liberty file in step 100 is directly marked into the timing diagram;

[0151] b) When the reference clock is located inside the functional module currently being analyzed, the input and output port delay modeling method is adopted, the input port of the functional module is regarded as the output port of the chip top-level module, and the output port of the functional module is regarded as the input port of the chip top-level module, and the delay information contained in the Liberty file in step 100 is used to set the corresponding input and output delays.

[0152] Step 500: Based on the timing model constructed in step 400, use the FPGA timing analysis engine to perform static timing analysis on the FPGA chip, wherein the FPGA timing analysis engine uses the module keyword and when keyword marked in step 200 to identify and process the timing information of different functional modules in different working modes.

[0153] Step 600: Generate a timing analysis report to guide the optimization and verification of the FPGA chip design and ensure that the chip can meet the timing requirements in different working modes.

[0154] The above steps are further described in detail below.

[0155] Step 100 involves using a hardware timing sign-off tool to generate Liberty files for different operating modes of each functional module in the FPGA chip based on a timing sign-off database. Specifically, hardware timing sign-off tools are widely used in timing analysis and verification of integrated circuit design. The software can extract detailed timing model information and delay information from the timing sign-off database to generate a standard-compliant Liberty file.

[0156] Specifically, there are a variety of hardware timing signoff tools available on the market, such as PrimeTime from Synopsys, Tempus Timing Signoff Solution from Cadence, and Questa TimeQuest Timing Analyzer from Mentor Graphics. These tools can meet the needs of generating Liberty files, and designers can choose the appropriate tool based on specific project requirements and team familiarity.

[0157] In the process of generating a Liberty file, you first need to divide the various functional modules in the FPGA chip and determine the different working modes of each functional module. For example, a functional module may have different timing characteristics in different working modes, and these characteristics need to be described and analyzed separately. To achieve this goal, the hardware timing signoff tool will extract the timing model and delay information of each functional module in different working modes based on the information in the timing signoff database, and integrate this information into the corresponding Liberty file.

[0158] Each Liberty file contains complete timing model information and delay information. The timing model information describes the timing relationship of the functional modules in different working modes, including the input and output delays, setup time, and hold time of each port. The delay information records in detail the specific delay time of the signal during transmission. These time values ​​are usually based on the measurement results under different process corners. By integrating this information into the Liberty file, it can be ensured that the timing characteristics of each functional module can be accurately captured and described in the subsequent timing analysis process.

[0159] In summary, step 100 lays a solid foundation for hierarchical static timing analysis of FPGA chips by using the hardware timing sign-off tool to generate a Liberty file containing timing models and delay information. This method not only improves the accuracy and efficiency of timing analysis, but also ensures the consistency and reliability of the analysis results, which helps to optimize the design and verification process of FPGA chips.

[0160] Step 200 involves marking all timing arcs of each functional module with the module keyword and marking the timing arcs in each working mode with the when keyword in the generated Liberty file. The details of this step are as follows:

[0161] First of all, Liberty files are a standard file format used to describe the timing characteristics of integrated circuits. They contain detailed timing model information and delay information. In these files, the timing information of each functional module needs to be clearly marked so that it can be accurately identified and processed in the subsequent timing analysis process.

[0162] The module keyword is used to distinguish different functional modules. In an FPGA chip, each functional module may be responsible for different functions and have unique timing characteristics. By marking the module keyword for the timing arcs of each functional module in the Liberty file, it can be clearly indicated to which functional module these timing arcs belong. This marking method not only makes the timing information management systematic, but also improves the efficiency of the timing analysis engine when processing different modules. For example, if a functional module is marked as module: A, then all timing arcs belonging to this module will be marked with module: A, and the timing analysis engine can quickly locate and process these timing arcs based on this mark.

[0163] The marking of the when keyword is to distinguish the timing characteristics of the same functional module in different working modes. Some functional modules in the FPGA chip may exhibit different timing behaviors in different working modes. For example, a module may have different timing characteristics in normal operating mode and power saving mode. In order to accurately describe these differences, the timing arcs in each working mode are marked with the when keyword in the Liberty file. In this way, the timing analysis engine can process the corresponding timing arc information separately according to the different working modes. For example, if the timing arc of a functional module in normal mode is marked as when: normal, and the timing arc in power saving mode is marked as when: power_saving, then the timing analysis engine can select the correct timing arc for analysis according to the current working mode when performing timing analysis.

[0164] Through these keyword markings in step 200, the timing information in the Liberty file can be made clearer and more organized. The combined use of the module keyword and the when keyword can not only accurately distinguish the timing characteristics of different modules and different working modes, but also significantly improve the efficiency and accuracy of the timing analysis process. This method provides more accurate basic data for the timing analysis of FPGA chips, which helps to improve the overall quality of chip design and verification.

[0165] Step 300 involves obtaining the netlist information of the FPGA chip, and using the information to determine the connection relationship and current working mode of each functional module, and then searching the corresponding Liberty file generated in step 100 for the timing arc that matches the functional module and the current working mode. The specific contents are as follows:

[0166] Netlist information is an important file that describes the internal circuit connection relationship of the FPGA chip. The netlist contains detailed connection information of all functional modules in the chip, including the input and output ports of each module, the connection method, and the connection relationship between modules. In the FPGA design process, netlist information is an important basic data for timing analysis.

[0167] First, by obtaining the netlist information of the FPGA chip, the connection relationship of each functional module can be determined. This step is the basis of timing analysis, because only by clarifying the connection relationship between each module can the transmission path and delay of the signal in the chip be further analyzed. The netlist information describes the port connection of each functional module in detail, and this information will be used to identify and match the timing arcs in the Liberty file.

[0168] Secondly, the netlist information is also used to determine the current working mode of each functional module. Some modules in the FPGA chip may have multiple working modes, and the timing characteristics of the modules in different working modes will be different. By analyzing the netlist information, the working mode of the current chip can be determined, so as to select the corresponding timing arc for analysis.

[0169] After obtaining the netlist information, the next step is to search for matching timing arcs in the corresponding Liberty file generated in step 100 according to the function module name and the current working mode in the netlist information.

[0170] The specific operations are as follows:

[0171] 1) Determine the name of the functional module: Identify the name of each functional module by parsing the netlist information. Each functional module has a corresponding timing model and delay information in the Liberty file, which has been marked with the module keyword.

[0172] 2) Determine the current working mode: Also by parsing the netlist information, identify the current working mode of each functional module. The timing information under different working modes is marked in the Liberty file by the when keyword.

[0173] 3) Find matching timing arcs: According to the function module name and the current working mode, find the corresponding timing arcs in the Liberty file. Specifically, first find the timing arcs with the corresponding module keyword, and then further filter out the timing arcs with the current working mode when keyword from these timing arcs. These matching timing arcs contain the timing characteristics of the current function module in a specific working mode, and are the basic data for static timing analysis.

[0174] Through the above steps, the timing arcs matching each functional module and the current working mode can be accurately extracted, providing necessary data support for subsequent timing modeling and analysis. The detailed operation of step 300 ensures the accuracy and reliability of the timing analysis process, so that the timing characteristics of the FPGA chip in different working modes can be accurately described and analyzed, thereby improving the quality and efficiency of design verification.

[0175] Step 400 involves performing timing modeling on each timing arc found in step 300. One of the following two modeling methods is selected to perform timing modeling according to the position of its reference clock to ensure the accuracy and efficiency of timing analysis.

[0176] When the reference clock is outside the functional module being analyzed, the timing diagram modeling method is used. This method is suitable for the case where the reference clock signal of the module comes from outside the module. When building the timing diagram, first determine the timing relationship of each signal path in the module based on the netlist information and the timing arcs found in the Liberty file. The starting point and end point of each signal path (i.e., the input port and output port) are represented in the timing diagram to form a complete timing path. While building the timing diagram, the timing arc delay information contained in the Liberty file is directly marked into the timing diagram. This includes the setup time, hold time, and signal propagation delay of each signal path. By marking the delay information directly in the timing diagram, repeated searches for delay values ​​can be avoided in subsequent analysis processes, thereby improving analysis efficiency. The timing diagram modeling method provides an intuitive representation of the timing behavior of the module, allowing designers to clearly see the timing relationship and delay distribution of each signal path. This method is particularly suitable for situations where the timing relationship is relatively simple and the reference clock is located outside the module.

[0177] When the reference clock is inside the functional module currently being analyzed, the input and output port delay modeling method is used. This method is applicable to the case where the reference clock signal is generated inside the module. First, the input port of the functional module is regarded as the output port of the chip's top-level module, which means that in the timing analysis, the check logic of the input port is equivalent to applying an output delay (set_output_delay). This processing method can simulate the time requirement for data to be transmitted from the top-level module to the input port of the functional module. Then, the output port of the functional module is regarded as the input port of the chip's top-level module, and in the timing analysis, the timing logic of the output port is equivalent to applying an input delay. This processing method simulates the time requirement for data to be transmitted from the output port of the functional module to the top-level module. For an internally generated reference clock (such as gen_clk), a virtual port is created inside the module to mark the delay from the external clock port (such as clk) to the internally generated clock, ensuring that the delay of the internal clock tree is included in the overall clock path so that the total delay of signal transmission can be accurately calculated.

[0178] The input and output port delay modeling method converts complex internal timing relationships into simple input and output delays, making timing analysis more intuitive and efficient. This method is particularly suitable for modules with complex internal clock generation circuits.

[0179] Through the above two modeling methods, step 400 ensures accurate modeling of each timing arc, regardless of whether its reference clock is located inside or outside the module. Timing diagram modeling provides an intuitive timing relationship diagram, while input and output port delay modeling simplifies the complexity of internal clock generation. The two methods complement each other and together constitute a comprehensive FPGA chip static timing analysis solution, thereby improving the accuracy and efficiency of timing analysis.

[0180] Step 500 involves using an FPGA timing analysis engine to perform static timing analysis on the FPGA chip based on the timing model constructed in step 400. During the analysis process, the FPGA timing analysis engine uses the module keyword and when keyword marked in step 200 to identify and process the timing information of different functional modules in different working modes to ensure the accuracy and comprehensiveness of the analysis.

[0181] First, based on the timing model constructed in step 400, the FPGA timing analysis engine starts to perform static timing analysis on the FPGA chip. Step 400 provides two modeling methods, namely, timing diagram modeling and input and output port delay modeling. These models describe in detail the timing characteristics and signal transmission paths of each functional module in different working modes. Through these models, the timing analysis engine can fully understand the transmission delay and timing relationship of each signal inside the FPGA chip.

[0182] During the analysis process, the FPGA timing analysis engine will first use the module keyword to identify different functional modules. Each timing arc in the Liberty file has the module keyword, which marks the functional module to which each timing arc belongs. For example, a timing arc marked as module: A belongs to functional module A. By identifying these keywords, the timing analysis engine can accurately locate and process the timing information of each functional module, ensuring the systematic and orderly analysis process.

[0183] At the same time, the FPGA timing analysis engine also uses the when keyword to identify and process timing information in different operating modes. Each functional module may have different timing characteristics in different operating modes, and these characteristics are marked in the Liberty file by the when keyword. For example, the timing arc of a functional module in normal operating mode may be marked as when: normal, while the timing arc in power saving mode is marked as when: power_saving. By identifying these keywords, the timing analysis engine can distinguish and process timing information in different operating modes, ensuring that the timing analysis results in various operating modes are accurate.

[0184] In the specific static timing analysis process, the FPGA timing analysis engine will analyze the delay information and timing relationship of each timing arc one by one. This includes calculating the transmission delay of the signal from the input port to the output port, checking whether the signal meets the setup time and hold time requirements, and verifying the timing performance of each signal path under different process corners. Through detailed analysis of these timing characteristics, the FPGA timing analysis engine can generate a comprehensive timing analysis report, pointing out possible timing violations and performance bottlenecks, and providing a basis for design optimization.

[0185] In summary, step 500 ensures that the FPGA timing analysis engine can accurately identify and process the timing information of different functional modules in different working modes by using the module keyword and when keyword marked in step 200. Based on the detailed timing model constructed in step 400, the FPGA timing analysis engine can perform static timing analysis comprehensively and efficiently, generate accurate and reliable timing analysis reports, and provide important support for the design and verification of FPGA chips.

[0186] Step 600 involves generating a timing analysis report to guide the optimization and verification of the FPGA chip design and ensure that the chip can meet the timing requirements in different working modes. The specific contents are as follows:

[0187] After completing the previous static timing analysis, the FPGA timing analysis engine will generate a detailed timing analysis report based on the analysis results. This report is an important tool for designers to optimize and verify FPGA chip designs.

[0188] First, the timing analysis report will list the timing characteristics of each functional module in detail, including key timing parameters such as signal transmission delay, setup time and hold time. These parameters are based on the timing model constructed in step 400 and the static timing analysis results in step 500. Through these detailed timing data, designers can fully understand the timing performance of each functional module in different working modes.

[0189] Secondly, the report will point out any timing violations and performance bottlenecks found during the timing analysis process. For example, some signal paths may have problems such as too long transmission delays, setup time or hold time that does not meet requirements. These violations and bottleneck information will be listed in detail in the report, along with the corresponding timing paths and specific problem descriptions. With this information, designers can optimize the design in a targeted manner, solve timing problems, and improve chip performance.

[0190] In addition, the timing analysis report will also provide the timing verification results of each functional module in different working modes. Since FPGA chips may operate in multiple working modes, it is very important to ensure that the timing performance in each mode meets the requirements. The report will list the timing analysis results in each working mode separately to ensure that the timing performance of the chip in different modes can meet the design standards. These verification results include information such as key timing parameters, timing violations, and performance bottlenecks in each mode.

[0191] The report also provides optimization suggestions to help designers further improve chip design. Based on the timing analysis results, the report may suggest adjusting the routing of certain signal paths, optimizing the timing logic of functional modules, or modifying design parameters. These suggestions are designed to solve timing problems and improve the overall performance and reliability of the chip.

[0192] In order to better understand the technical solution of the present application, a specific example is provided below for illustration. The details listed in the example are mainly for ease of understanding and are not intended to limit the scope of protection of the present application.

[0193] Before explaining this example, it is necessary to understand the basic concepts and methods of static timing analysis. This will help to better understand the innovative technical solutions proposed in this application. The static timing analysis methods of chips are divided into flattened timing analysis and hierarchical timing analysis. The main difference between the two analysis methods is the different methods of timing modeling. Flattened timing analysis refers to treating the entire FPGA chip as a whole, and integrating all timing models into a complete timing diagram in the form of a timing diagram during the timing modeling process. Hierarchical timing analysis refers to treating the top layer of the chip as a whole, describing the top layer as a timing diagram through timing modeling, integrating some modules in the chip into the top-level timing diagram by building a timing diagram, and treating some IP modules as black boxes. The IP treated as a black box only focuses on the timing information at the port. The timing arc of the black box IP will not be mapped to the timing diagram, and will only be marked as a port delay by the method of input and output delay. When designing FPGA chips, the timing results inside the black box IP are ensured to converge by the designer through the hardware timing sign-off tool, and the user of the FPGA chip as a black box IP only needs to perform timing analysis on the IP port. Hierarchical timing analysis treats the entire IP as a black box and only analyzes the timing of the black box IP port. This method can ensure the normal operation of the chip and effectively reduce the size of the timing diagram, effectively saving the running time of the timing analysis. The hierarchical timing analysis method can complete the timing analysis of complex functional modules faster and better.

[0194] The hierarchical timing analysis method of FPGA chips is mainly divided into two parts: the hardware timing sign-off tool extracts the Liberty timing file and the FPGA timing analysis engine recognizes and models the Liberty timing file.

[0195] Extraction of Liberty files:

[0196] The Liberty file is provided by the hierarchical timing analysis liberty extraction command of the hardware timing signoff tool. You can use either the ETM (Extract Timing Model) process or the ILM (Interface Logic Model) process provided by the hardware timing signoff tool.

[0197] Liberty files can be provided at different design stages. For example, in the early stage of module design, Liberty files can be extracted based on the front-end database for timing analysis testing of FPGA chips. After the module design is completed and the timing is signed off, Liberty files can be extracted based on the timing sign-off database for timing analysis of mass-produced FPGA chips.

[0198] The method proposes to extract independent Liberty files according to modules and different working modes of modules to describe the timing models and timing arc delay information of different modules in different working modes;

[0199] Storage and timing modeling of Liberty files:

[0200] The timing analysis engine uses the keyword Module to distinguish the Liberty information of different modules, that is, all timing arcs of each module share an independent Module keyword;

[0201] For the Liberty information of different working modes of the same module, the timing analysis engine uses the keyword when and parallel timing arcs to solve it. That is, all timing arcs corresponding to a certain independent working mode share the keyword when;

[0202] There are two main modeling methods for FPGA timing analysis engine based on Liberty information modeling: timing diagram modeling and port constraint modeling. If the module contains clock tree delay information, port constraint modeling is required. Otherwise, timing diagram modeling can be used to integrate timing information and delay into the timing diagram.

[0203] When the timing analysis engine performs timing modeling, it models the timing arcs according to their types: combinational logic, sequential logic, check logic, and reset logic.

[0204] When the timing diagram is constructed, the timing arc delay information will be marked in the timing diagram at the same time, which avoids repeated search for timing arcs during delay calculation, effectively reducing the timing analysis running time of the FPGA timing analysis engine;

[0205] Figure 1 A hierarchical timing analysis workflow for FPGA chips is introduced in detail. The FPGA chip consists of a top layer (top) and two functional modules (a, b). Module a has two working modes, and module b has three working modes.

[0206] Extraction of Liberty files:

[0207] The timing information of modules a and b is provided by different chip design engineers.

[0208] After the a module completes the timing sign-off, the hardware timing sign-off tool is used to generate Liberty files a_1.lib and a_2.lib for different working modes (1, 2) based on the timing sign-off database. Among them, all the timing arcs of the a module are marked with the keyword module (module: a), the timing arcs in a_1.lib are marked with the keyword when (when: 1), and the timing arcs in a_2.lib are marked with the keyword when (when: 2).

[0209] After the b module completes the timing sign-off, the hardware timing sign-off tool is used to extract the liberty files b_1.lib, b_2.lib, and b_3.lib for different working modes (1, 2, and 3) based on the timing sign-off database. Among them, all the timing arcs of the b module are marked with the keyword module (module: b), the timing arcs in b_1.lib are marked with the keyword when (when: 1), the timing arcs in b_2.lib are marked with the keyword when (when: 2), and the timing arcs in b_3.lib are marked with the keyword when (when: 3).

[0210] Storage and timing modeling of Liberty files:

[0211] The timing information in Liberty is stored in units of timing arcs, such as Figure 2 Each timing arc mainly stores the following information: target port, reference port, timing arc type, timing arc delay, module keyword module and working mode keyword when.

[0212] Among them, the module keyword is unique to this application and is used to distinguish modules;

[0213] The target port refers to the port for timing analysis, and the reference port refers to the timing reference port of the target port;

[0214] The when keyword is used to distinguish the timing information in different working modes;

[0215] The delay values ​​under different process angles are recorded in the timing arc delay;

[0216] The timing arc type is used to distinguish between combinational logic, sequential logic, check logic, and reset logic;

[0217] All the timing arcs of module a are marked with the keyword module(module:a), including the timing arcs in a_1.lib marked with the keyword when(when:1), and the timing arcs in a_2.lib marked with the keyword when(when:2). The timing arcs of module b are marked with the keyword module(module:b), including the timing arcs in b_1.lib marked with the keyword when(when:1), the timing arcs in b_2.lib marked with the keyword when(when:2), and the timing arcs in b_3.lib marked with the keyword when(when:3).

[0218] The process of modeling module timing is mainly divided into two steps: finding timing arcs and describing timing models. The purpose of finding timing arcs is to find the corresponding timing arcs for the modules in the netlist according to the working mode. First, find the Liberty data corresponding to the module according to the module in the netlist, and then select the corresponding timing arc according to the current working mode of the module.

[0219] The following describes two methods for describing timing models based on Liberty information: timing diagram modeling and input and output port delay modeling.

[0220] The reference clocks of the target ports in Liberty are all external ports of the module (such as Figure 3 ), you can use the timing diagram modeling method to perform timing modeling. According to the module name and module working mode in the netlist, find the timing arc of each timing analysis target port (in, out), and mark the timing information and delay information into the timing diagram according to the type of timing arc (combinatorial logic, sequential logic, check logic, reset logic).

[0221] The reference clock of the target port in Liberty is internal to the model (such as Figure 4 ), the internal clock tree delay needs to be calculated during timing analysis. At this time, the input and output port delay method can be used for timing modeling. According to the module name and module working mode in the netlist, the timing arc of each timing analysis target port (in, out) is found, the input port (in) is regarded as the output port of the chip top module (top), and the port check logic is equivalent to applying output delay (set_output_delay), the output port (out) is regarded as the input port of the chip top module (top), and the port timing logic is equivalent to applying input delay. This solution treats the module as a black box, and gen_clk will be created as a virtual port inside the black box. The virtual port is used to mark the delay from port clk to gen_clk, and port gen_clk is the reference clock port of the target port (in, out). The internal clock tree delay will be included in the clock path.

[0222] The FPGA chip hierarchical static timing analysis method proposed in the above embodiments and examples has the following significant technical effects:

[0223] 1) Improve the accuracy of timing analysis: By using the standard format Liberty file for timing information exchange, the above embodiments and examples ensure the consistency of the timing model and delay information. Combining the FPGA chip timing analysis with the hardware timing sign-off tool timing analysis process further ensures the accuracy and reliability of the analysis results. This highly accurate timing analysis helps improve the performance and reliability of FPGA chips and reduce chip failures caused by timing problems.

[0224] 2) Enhanced flexibility of timing analysis: The innovative use of the module and when keywords solves the problem of different timing information of FPGA modules in different working modes. This enables the timing analysis tool to accurately identify and process timing information in various complex scenarios, greatly enhancing the flexibility of analysis. This flexibility enables the above embodiments and examples to adapt to the increasingly complex multi-module, multi-mode structure in modern FPGA design.

[0225] 3) Improve the comprehensiveness of timing analysis: By introducing two methods, timing diagram modeling and input and output port delay modeling, this application can fully cover various FPGA module structures, including those complex modules with internal clock generation circuits. This comprehensiveness ensures that the timing analysis will not miss any critical paths, thereby improving the overall quality of FPGA chip design.

[0226] 4) Significantly improve analysis efficiency: When constructing a module timing diagram, the timing arc delay is directly marked in the timing diagram. This innovation greatly improves the efficiency of timing analysis. By reducing repeated delay search operations, the above embodiments and examples significantly shorten the running time of timing analysis, allowing engineers to perform design iterations and optimizations faster.

[0227] 5) Simplify the workflow and reduce human errors: The above embodiments and examples greatly simplify the workflow of engineers through automated processing and standardized processes. This not only improves work efficiency, but also significantly reduces the possibility of human errors and improves the reliability of the design process.

[0228] 6) Enhanced design verifiability: By generating detailed timing analysis reports, the above embodiments and examples provide a reliable basis for the optimization and verification of FPGA chip design. This helps engineers quickly identify and resolve potential timing issues and improve the verifiability and reliability of the design.

[0229] 7) Promote modular design and IP reuse: The above embodiments and examples support timing analysis of independent modules. This modular analysis method is conducive to promoting modular development and IP reuse in FPGA design, thereby accelerating the overall chip design process.

[0230] 8) Adapting to future design requirements: As the complexity of FPGA chips continues to increase, the flexibility and scalability of the above-mentioned embodiments and examples enable them to adapt to more complex FPGA design requirements that may arise in the future, providing strong support for the long-term development of FPGA technology.

[0231] In summary, the hierarchical static timing analysis method for FPGA chips in this application not only improves the accuracy, efficiency and comprehensiveness of timing analysis, but also simplifies the design process, reduces errors, and enhances the verifiability and reliability of the design. These technical effects work together to significantly improve the design quality and performance of FPGA chips, laying a solid foundation for the further development of FPGA technology.

[0232] The second embodiment of the present application relates to a hierarchical static timing analysis system for FPGA chips, the structure of which is as follows: Figure 6 As shown, the hierarchical static timing analysis system of the FPGA chip includes:

[0233] The file generation module is used to generate files containing timing model information and delay information for different working modes of each functional module in the FPGA chip.

[0234] The keyword marking module is used to mark the timing arc of each functional module with a keyword in the file.

[0235] The timing arc matching module is used to obtain the netlist information of the FPGA chip and search for the timing arc matching the functional module and the working mode according to the netlist information.

[0236] The timing modeling module is used to select a modeling method for each found timing arc according to its reference clock position to perform timing modeling.

[0237] The static timing analysis module is used to perform static timing analysis on the FPGA chip based on the constructed timing model.

[0238] The report generation module is used to generate timing analysis reports to guide the optimization and verification of FPGA chip design.

[0239] The first implementation manner is a method implementation manner corresponding to the present implementation manner. The technical details in the first implementation manner can be applied to the present implementation manner, and the technical details in the present implementation manner can also be applied to the first implementation manner.

[0240] It should be noted that those skilled in the art should understand that the implementation functions of each module shown in the implementation of the hierarchical static timing analysis system of the above-mentioned FPGA chip can be understood by referring to the relevant description of the hierarchical static timing analysis method of the above-mentioned FPGA chip. The functions of each module shown in the implementation of the hierarchical static timing analysis system of the above-mentioned FPGA chip can be realized by a program (executable instruction) running on a processor, or by a specific logic circuit. If the hierarchical static timing analysis system of the FPGA chip in the embodiment of the present application is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, ReadOnly Memory), disk or optical disk and other media that can store program code. In this way, the embodiment of the present application is not limited to any specific combination of hardware and software.

[0241] Accordingly, an embodiment of the present application further provides a computer storage medium, in which computer executable instructions are stored. When the computer executable instructions are executed by a processor, the various method embodiments of the present application are implemented.

[0242] In addition, the embodiment of the present application also provides a hierarchical static timing analysis system for an FPGA chip, which includes a memory for storing computer executable instructions, and a processor; the processor is used to implement the steps in the above-mentioned method implementations when executing the computer executable instructions in the memory. Among them, the processor can be a central processing unit (Central Processing Unit, referred to as "CPU"), or other general-purpose processors, digital signal processors (Digital Signal Processor, referred to as "DSP"), Application Specific Integrated Circuit (Application Specific Integrated Circuit, referred to as "ASIC"), etc. The aforementioned memory can be a read-only memory (read-only memory, referred to as "ROM"), a random access memory (random access memory, referred to as "RAM"), a flash memory (Flash), a hard disk or a solid-state hard disk, etc. The steps of the method disclosed in each embodiment of the present application can be directly embodied as a hardware processor execution, or a combination of hardware and software modules in the processor.

[0243] It should be noted that in the application documents of this patent, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "including one" do not exclude the existence of other identical elements in the process, method, article or device including the elements. In the application documents of this patent, if it is mentioned that an action is performed according to an element, it means that the action is performed at least according to the element, which includes two situations: performing the action only according to the element, and performing the action according to the element and other elements. Expressions such as multiple, multiple, and multiple include 2, 2 times, 2 kinds, and more than 2, more than 2 times, and more than 2 kinds.

[0244] All documents mentioned in this application are considered to be included in the disclosure of this application as a whole, so that they can be used as the basis for modification when necessary. In addition, it should be understood that after reading the above disclosure of this application, those skilled in the art can make various changes or modifications to this application, and these equivalent forms also fall within the scope of protection claimed in this application.

Claims

1. A hierarchical static timing analysis method for FPGA chips, characterized in that: The following steps are involved: Step A: Generate a file containing timing model information and delay information for different working modes of each functional module in the FPGA chip; Step B: in the file, marking a keyword for the timing arc of each functional module; Step C: Obtaining the netlist information of the FPGA chip, and searching for a timing arc that matches the functional module and the working mode according to the netlist information; Step D: For each timing arc found, select a modeling method according to its reference clock position to perform timing modeling; Step E: Based on the constructed timing model, perform static timing analysis on the FPGA chip; Step F: Generate a timing analysis report to guide the optimization and verification of FPGA chip design.

2. The method according to claim 1, characterized in that In step A, use the hardware timing signoff tool to generate a Liberty file based on the timing signoff database.

3. The method according to claim 1, characterized in that: In step B, all timing arcs of each functional module are marked with the module keyword, and the timing arcs in each working mode are marked with the when keyword.

4. The method according to claim 1, characterized in that The netlist information in step C is used to determine the connection relationship and current working mode of each functional module.

5. The method according to claim 1, characterized in that: The modeling method in step D includes: a) When the reference clock is outside the functional module currently being analyzed, the timing diagram modeling method is used; b) When the reference clock is located inside the functional module currently being analyzed, the input and output port delay modeling method is used.

6. The method according to claim 5, characterized in that When the timing diagram modeling method is adopted, the timing arc delay information is directly marked into the timing diagram.

7. The method according to claim 5, characterized in that When using the input and output port delay modeling method, the input port of the functional module is regarded as the output port of the chip top-level module, and the output port of the functional module is regarded as the input port of the chip top-level module, and the delay information is used to set the corresponding input and output delays.

8. The method according to claim 1, characterized in that In step E, a static timing analysis is performed using an FPGA timing analysis engine, wherein the FPGA timing analysis engine utilizes the keywords marked in step B to identify and process timing information of different functional modules in different operating modes.

9. A hierarchical static timing analysis system for FPGA chips, characterized in that: include: A file generation module, used for generating files containing timing model information and delay information for different working modes of each functional module in the FPGA chip; A keyword marking module, used for marking a keyword for a timing arc of each functional module in the file; A timing arc matching module is used to obtain the netlist information of the FPGA chip and search for a timing arc that matches the functional module and the working mode according to the netlist information; A timing modeling module, for selecting a modeling method for each found timing arc according to its reference clock position to perform timing modeling; Static timing analysis module, used to perform static timing analysis on FPGA chips based on the constructed timing model; The report generation module is used to generate timing analysis reports to guide the optimization and verification of FPGA chip design.

10. A hierarchical static timing analysis system for FPGA chips, characterized in that: include: A memory for storing computer executable instructions; as well as, A processor, configured to implement the steps in the method according to any one of claims 1 to 8 when executing the computer executable instructions.