Test method and system for chip simulation model

Abstract

The application relates to the technical field of chip testing, in particular to a chip simulation model testing method and system. The method comprises the following steps: collecting a target chip simulation model and performing timing compensation to generate jitter compensation data, then performing eye diagram reconstruction to obtain waveform eye diagram data, performing amplitude normalization and timing phase correction on the waveform data to obtain purified alignment data, generating main mode data through orthogonal mapping processing and mode projection analysis, constructing kernel space data and performing manifold cognitive aggregation to generate test coverage features, performing boundary analysis and density distribution analysis on the features to obtain clustering boundary data, and finally optimizing boundary precision and executing an adaptive test sequence to generate a test case result. The application realizes a more efficient chip simulation model testing method.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of chip testing, in particular to a chip simulation model testing method and system. BACKGROUND

[0002] With the rapid development of semiconductor technology, the complexity and integration of chip design are continuously improving, and simulation models play an increasingly important role in the chip development process. However, existing chip simulation technology faces problems such as timing jitter and signal interference, which affects the accuracy of simulation results. These problems not only increase the design cycle, but also cause performance instability in actual application, affecting the market competitiveness of products. In the process of chip testing, the processing and analysis of waveform data are often key links. The existing technology still has certain limitations in waveform amplitude normalization, timing phase correction, etc. Especially in processing high-frequency signals and complex data, the traditional method is difficult to effectively extract important features, resulting in insufficient test coverage and inability to comprehensively evaluate the performance and reliability of the chip, further increasing the risk in the development process and affecting the quality and stability of the final product. SUMMARY

[0003] Therefore, it is necessary to provide a chip simulation model testing method and system to solve at least one of the above technical problems.

[0004] To achieve the above purpose, the chip simulation model testing method comprises the following steps:

[0005] Step S1: Collecting a target chip simulation model; performing timing integrity compensation on the target chip simulation model to generate jitter compensation data; performing eye diagram reconstruction on the jitter compensation data to obtain waveform eye diagram data;

[0006] Step S2: Normalizing the waveform eye diagram data to generate normalized data; performing timing phase correction on the normalized data to generate purified alignment data;

[0007] Step S3: Generating orthogonal basis data by performing orthogonal mapping processing on the purified alignment data; generating principal modal data by performing modal projection analysis based on the orthogonal basis data;

[0008] Step S4: Constructing kernel space data by performing vector mapping driving on the principal modal data; generating test coverage features by performing manifold cognitive intelligent aggregation on the kernel space data;

[0009] Step S5: Generating density distribution data by performing boundary driving analysis on the test coverage features; accurately calibrating stable features based on the density distribution data to obtain clustering boundary data;

[0010] Step S6: boundary accuracy tuning is performed on the clustering boundary data to generate compensation result data; and adaptive test sequence execution is performed on the target chip simulation model based on the compensation result data to generate test case results.

[0011] The present application ensures the stability of the signal through timing complete compensation of the target chip simulation model, improves the accuracy of the waveform through generation of the jitter compensation data, provides an intuitive basis for signal quality evaluation through acquisition of the waveform eye diagram data, enhances the consistency of the data through implementation of the waveform amplitude normalization, ensures the phase accuracy of the signal through generation of the purified alignment data, provides an effective mathematical basis for subsequent analysis through implementation of the orthogonal mapping processing, can effectively extract the main characteristics of the signal through generation of the main modal data, provides support for in-depth analysis of the signal characteristics through construction of the kernel space data, enhances the comprehensiveness of the test coverage through development of the manifold cognitive intelligence aggregation, provides a quantitative basis for stability analysis of the characteristics through generation of the density distribution data, can effectively identify the characteristic distribution of the signal through formation of the clustering boundary data, ensures the reliability of the test results through implementation of the boundary accuracy tuning, provides a basis for execution of the adaptive test sequence through generation of the compensation result data, and can provide comprehensive support for performance verification of the target chip through formation of the test case results, thereby improving the scientificity and efficiency of the chip simulation model test as a whole and providing an effective technical means and solution for chip development and optimization.

[0012] Preferably, step S1 comprises the following steps:

[0013] Step S11: collecting the target chip simulation model; and performing waveform injection scanning on the target chip simulation model to obtain original waveform data;

[0014] Step S12: reorganizing the timing structure of the original waveform data to obtain timing deconstruction data; and performing edge jitter detection on the timing deconstruction data to obtain edge jitter data;

[0015] Step S13: performing jitter compensation on the timing deconstruction data according to the edge jitter data to generate jitter compensation data;

[0016] Step S14: performing multi-domain sampling remapping on the jitter compensation data to obtain remapping data; and performing clock recovery calibration on the remapping data to obtain calibration data;

[0017] Step S15: performing crosstalk elimination processing on the calibration data to generate waveform denoising data; and performing eye diagram reconstruction on the waveform denoising data to obtain waveform eye diagram data.

[0018] The application provides basic data for signal analysis through waveform injection scanning of the target chip simulation model, ensures comprehensiveness of signal characteristics through acquisition of original waveform data, enhances analyzability of the data through reorganization of the timing structure, provides an effective basis for subsequent processing through generation of timing deconstruction data, can accurately identify jitter problems in the signal through edge jitter detection, provides important parameters for jitter compensation through generation of edge jitter data, improves stability and accuracy of the signal through formation of jitter compensation data, provides support for multi-dimensional analysis of the signal through implementation of multi-domain sampling remapping, ensures integrity of the signal through acquisition of remapping data, enhances time accuracy of the signal through clock recovery calibration, provides a guarantee for subsequent processing through generation of calibration data, effectively eliminates noise interference through crosstalk elimination processing, improves clarity of the signal through formation of waveform noise reduction data, provides a basis for intuitive evaluation of signal quality through successful implementation of eye diagram reconstruction, and improves scientificity and effectiveness of chip simulation model testing as a whole, and provides reliable support for chip performance verification and optimization.

[0019] Preferably, step S2 comprises the following steps:

[0020] Step S21: performing distortion compensation processing on the waveform eye diagram data to obtain compensated eye diagram data;

[0021] Step S22: performing amplitude normalization on the compensated eye diagram data to generate normalized data;

[0022] Step S23: performing phase alignment calibration on the normalized data to obtain phase alignment eye diagram data;

[0023] Step S24: performing inter-symbol deinterleaving processing on the phase alignment eye diagram data to generate purified alignment data.

[0024] The application improves accuracy and readability of the signal through distortion compensation processing on the waveform eye diagram data, provides a basis for subsequent analysis through generation of the compensated eye diagram data, ensures consistency of the data through implementation of amplitude normalization, enhances effectiveness of the analysis through acquisition of the normalized data, improves time accuracy of the signal through execution of phase alignment calibration, can better reflect signal characteristics through formation of the phase alignment eye diagram data, effectively eliminates signal interference through inter-symbol deinterleaving processing, ensures clarity and usability of the signal through generation of the purified alignment data, and improves scientificity and accuracy of chip simulation model testing as a whole, thereby providing reliable support for chip performance evaluation and optimization.

[0025] Preferably, step S3 comprises the following steps:

[0026] Step S31: performing modal decomposition processing on the purified alignment data to obtain modal feature data;

[0027] Step S32: eigenvector extraction is performed on the modal characteristic data to obtain vector group data; orthogonalization processing is performed on the vector group data to generate orthogonal basis data;

[0028] Step S33: feature projection simulation is performed based on the orthogonal basis data to generate a simulation projection matrix;

[0029] Step S34: modal correlation analysis is performed on the orthogonal basis data based on the simulation projection matrix to obtain modal correlation data; and modal screening is performed based on the modal correlation data to generate main modal data.

[0030] The present application discloses the internal characteristics of the signal through the modal decomposition processing of the purified alignment data, and the generation of the modal characteristic data provides a basis for subsequent analysis. The eigenvector extraction ensures the comprehensiveness of the signal characteristics, the formation of the vector group data provides support for subsequent processing, the orthogonalization processing enhances the independence and effectiveness of the data, the generation of the orthogonal basis data provides a reliable basis for the feature projection simulation, the generation of the simulation projection matrix can effectively reflect the distribution of the signal characteristics, the implementation of the modal correlation analysis reveals the correlation between the signal characteristics, the formation of the modal correlation data provides a basis for modal screening, and the generation of the main modal data can effectively extract the main characteristics of the signal, thereby improving the accuracy and scientificity of the chip simulation model test as a whole, and providing an effective technical means and support for chip performance evaluation and optimization.

[0031] Preferably, step S4 comprises the following steps:

[0032] Step S41: test vector mapping is performed on the main modal data to obtain test sequence data;

[0033] Step S42: kernel function transformation is performed on the test sequence data to generate kernel space data;

[0034] Step S43: manifold learning reconstruction is performed on the kernel space data to generate manifold structure data;

[0035] Step S44: coverage depth evaluation is performed on the manifold structure data to generate depth index data;

[0036] Step S45: feature fusion processing is performed on the depth index data to generate test coverage features.

[0037] The application effectively generates a targeted test sequence through the test vector mapping of the principal modal data, the formation of the test sequence data provides specific basis for signal evaluation, the implementation of the kernel function transformation improves the representation ability of the data in the high-dimensional space, the generation of the kernel space data lays a foundation for in-depth analysis of signal characteristics, the execution of the manifold learning reconstruction can effectively reveal the internal structure of the signal, the generation of the manifold structure data provides an intuitive view for subsequent analysis, the implementation of the coverage depth evaluation quantifies the comprehensiveness of the test, the formation of the depth index data provides data support for the identification of test characteristics, the successful implementation of the feature fusion processing enhances the expression ability of the test coverage characteristics, and overall, the scientificity and effectiveness of the chip simulation model test are improved, which provides a reliable technical means and solution for chip performance verification and optimization.

[0038] Preferably, step S5 comprises the following steps:

[0039] Step S51: limit condition inference is performed on the test coverage characteristics to obtain limit coverage parameters;

[0040] Step S52: density estimation clustering is performed on the limit coverage parameters to generate density distribution data;

[0041] Step S53: inflection point positioning is performed based on the density distribution data to obtain inflection point characteristic data; boundary positioning is performed according to the inflection point characteristic data to generate boundary characteristic data;

[0042] Step S54: stability evaluation is performed based on the boundary characteristic data to generate stability index parameters; clustering optimization is performed on the stability index parameters to obtain clustering boundary data.

[0043] The application provides boundary parameters of signal testing through limit condition inference of test coverage characteristics, the generation of limit coverage parameters provides a basis for limit performance analysis of signals, the implementation of density estimation clustering effectively reveals the distribution characteristics of signal characteristics, the formation of density distribution data provides support for subsequent analysis, the execution of inflection point positioning can identify the key change point of signal characteristics, the generation of inflection point characteristic data provides a basis for boundary analysis, the formation of boundary characteristic data ensures the accurate positioning of signal characteristics, the implementation of stability evaluation quantifies the reliability of the signal, the generation of stability index parameters provides data support for signal characteristic analysis, the successful implementation of clustering optimization improves the accuracy and effectiveness of the boundary data, and overall, the scientificity and accuracy of the chip simulation model test are improved, which provides an effective technical means and solution for chip performance verification and optimization.

[0044] Preferably, step S53 comprises the following steps:

[0045] Gradient mapping data is obtained by performing gradient transformation on the density distribution data; and extreme point set data is generated by performing local extreme value detection on the gradient mapping data;

[0046] Curvature characteristic parameters are obtained by performing curvature calculation on the extreme point set data; and candidate inflection point data is generated by performing threshold screening on the curvature characteristic parameters based on a preset inflection point curvature threshold value;

[0047] Window characteristic data is obtained by performing local window segmentation on the candidate inflection point data; and inflection point morphological features are generated by performing morphological processing on the candidate inflection point data based on the window characteristic data;

[0048] Inflection point characteristic data is generated by performing inflection point center positioning according to the inflection point morphological features;

[0049] Simulated expansion regions are obtained by performing neighborhood expansion simulation on the inflection point characteristic data; and direction characteristic data is generated by performing directionality analysis on the simulated expansion regions;

[0050] Continuity index data is obtained by performing continuity recognition on the direction characteristic data; and initial boundary data is generated by performing boundary tracking based on the continuity index data;

[0051] Boundary smoothness parameters are obtained by performing smoothness evaluation on the initial boundary data; and boundary characteristic data is generated by performing accurate repositioning on the initial boundary data based on the boundary smoothness parameters.

[0052] The application improves the visualization effect of signal characteristics through gradient transformation processing of density distribution data, and the generation of gradient mapping data provides a basis for subsequent analysis, the local extreme value detection effectively identifies the key change points in the signal, the formation of the extreme point set data provides important information for further analysis, the implementation of curvature calculation reveals the shape change of signal characteristics, the generation of curvature feature parameters provides a basis for inflection point identification, the application of the preset inflection point curvature threshold ensures the accuracy of the candidate inflection point data, the generation of window feature data enhances the fineness of inflection point analysis, the morphological processing improves the feature expression ability of the candidate inflection point data, the successful implementation of the inflection point center positioning ensures the accuracy of the inflection point feature data, the neighborhood expansion simulation provides a wider perspective for subsequent analysis, the directional analysis reveals the change trend of signal characteristics, the generation of directional feature data enhances the depth of signal analysis, the implementation of continuity recognition quantifies the stability of signal characteristics, the formation of continuity index data provides support for boundary tracking, the generation of initial boundary data lays a foundation for comprehensive analysis of signal characteristics, the implementation of smoothness evaluation ensures the refinement of boundary features, the generation of boundary smoothness parameters improves the accuracy and reliability of the initial boundary data, the successful implementation of the accurate repositioning ensures the high quality of the final boundary feature data, and overall, the scientificity and effectiveness of the chip simulation model test are improved, which provides strong support for chip performance verification and optimization.

[0053] Preferably, step S6 comprises the following steps:

[0054] Step S61: Similarity calculation is performed on the clustering boundary data to obtain boundary similarity data;

[0055] Step S62: Consistency comparison is performed on the boundary similarity data to generate comparison result parameters;

[0056] Step S63: Deviation analysis correction is performed based on the comparison result parameters to generate compensation result data; based on the compensation result data, candidate test allocation is performed to generate supplementary test data;

[0057] Step S64: Adaptive test sequence execution is performed on the target chip simulation model according to the supplementary test data and the compensation result data to generate test case results.

[0058] The application effectively reveals the correlation between the boundary characteristics by similarity calculation of the clustered boundary data, the generation of the boundary similarity data provides a basis for subsequent consistency comparison, the implementation of the consistency comparison ensures the accuracy of data analysis, the generation of the comparison result parameter provides a basis for deviation analysis, the execution of the deviation analysis correction improves the reliability of the data, the formation of the compensation result data can effectively guide the candidate test allocation, the generation of the supplementary test data ensures the comprehensiveness of the test, the execution of the adaptive test sequence provides dynamic adjustment capability for the target chip simulation model, the generation of the test case result can effectively reflect the signal characteristics and performance, and overall, the scientificity and effectiveness of the chip simulation model test are improved, and strong technical support is provided for chip performance verification and optimization.

[0059] Preferably, step S63 comprises the following steps:

[0060] The comparison result parameter is subjected to difference mapping processing to obtain difference distribution data; and the difference distribution data is subjected to regional segmentation processing to obtain regional characteristic data;

[0061] The regional characteristic data is subjected to priority sorting based on the difference distribution data to generate deviation level data;

[0062] The deviation level data is subjected to threshold classification according to a preset deviation level threshold to generate classification mark data; and the classification mark data is subjected to compensation rule matching to generate a matching compensation rule;

[0063] The classification mark data is subjected to compensation calculation based on the matching compensation rule to generate compensation result data;

[0064] Residual error characteristics are obtained by performing residual error calculation on the compensation result data; and convergence index data are generated by performing convergence analysis on the residual error characteristics;

[0065] Compensation coverage characteristics are generated by performing coverage degree analysis on the convergence index data; and a mapping scene matrix is generated by performing test scene mapping based on the compensation coverage characteristics;

[0066] Demand item data are obtained by performing test requirement decomposition according to the mapping scene matrix; and a test condition set is generated by performing test condition supplementation on the demand item data;

[0067] A test task sequence is generated by performing test case construction based on the test condition set; and task dependency relationships are obtained by performing dependency relationship analysis on the test task sequence;

[0068] Supplementary test data are generated by performing arrangement and allocation on the test task sequence according to the task dependency relationships.

[0069] The application discloses a chip simulation model test method and system, and relates to the technical field of chip simulation model test. The chip simulation model test method comprises the following steps: collecting a target chip simulation model; performing time sequence complete compensation on the target chip simulation model to generate jitter compensation data; performing eye diagram reconstruction on the jitter compensation data to obtain waveform eye diagram data; performing waveform amplitude normalization on the waveform eye diagram data to generate normalized data; performing time sequence phase correction on the normalized data to generate purified alignment data; performing orthogonal mapping processing on the purified alignment data to generate orthogonal basis data; performing modal projection analysis based on the orthogonal basis data to generate main modal data; performing vector mapping driving construction on the main modal data to generate kernel space data; performing manifold cognitive intelligent aggregation on the kernel space data to generate test coverage features; performing boundary driving analysis on the test coverage features to generate density distribution data; performing stable feature accurate calibration based on the density distribution data to obtain clustering boundary data; performing mapping scene matrix formation based on the clustering boundary data to generate demand item data; supplementing a test condition set based on the demand item data to generate a test case; performing task dependency relationship analysis on the test case to generate a test task sequence; and performing a test on the target chip simulation model based on the test task sequence.

[0070] The application further provides a chip simulation model test system for performing the chip simulation model test method as described above, and the chip simulation model test system comprises:

[0071] an eye diagram reconstruction module configured to collect a target chip simulation model, perform time sequence complete compensation on the target chip simulation model to generate jitter compensation data, and perform eye diagram reconstruction on the jitter compensation data to obtain waveform eye diagram data;

[0072] a phase correction module configured to perform waveform amplitude normalization on the waveform eye diagram data to generate normalized data, and perform time sequence phase correction on the normalized data to generate purified alignment data;

[0073] a modal analysis module configured to perform orthogonal mapping processing on the purified alignment data to generate orthogonal basis data, and perform modal projection analysis based on the orthogonal basis data to generate main modal data;

[0074] a feature mapping module configured to perform vector mapping driving construction on the main modal data to generate kernel space data, and perform manifold cognitive intelligent aggregation on the kernel space data to generate test coverage features;

[0075] a boundary analysis module configured to perform boundary driving analysis on the test coverage features to generate density distribution data, and perform stable feature accurate calibration based on the density distribution data to obtain clustering boundary data;

[0076] The test optimization module is used for boundary accuracy tuning of the clustering boundary data to generate compensation result data; and adaptive test sequence execution is performed on the target chip simulation model based on the compensation result data to generate test case results.

[0077] The eye diagram reconstruction module provides a basis for real-time monitoring of the target chip signal, the generation of the jitter compensation data improves the stability and reliability of the signal, the acquisition of the waveform eye diagram data can intuitively show the signal quality and characteristics, the function of the phase correction module ensures the consistency of the signal amplitude, the generation of the normalized data enhances the accuracy of subsequent analysis, the acquisition of the purified alignment data provides a guarantee for the correction of the signal phase, the orthogonal mapping processing of the modal analysis module provides an effective method for signal feature extraction, the generation of the main modal data can effectively identify the main features and trends of the signal, the construction of the kernel space data of the feature mapping module provides support for deep analysis of signal features, the implementation of the manifold cognitive intelligent aggregation improves the comprehensiveness of the test coverage, the density distribution data of the boundary analysis module provides a quantitative basis for the stability analysis of the features, the generation of the clustering boundary data effectively identifies the distribution and changes of the signal features, the boundary accuracy tuning ensures the reliability of the test results, the generation of the compensation result data provides a basis for the execution of the adaptive test sequence, the formation of the test case results provides comprehensive support for the performance verification of the target chip, and overall, the scientificity and efficiency of the chip simulation model test are improved, which provides an effective technical means and solution for chip development and optimization. BRIEF DESCRIPTION OF DRAWINGS

[0078] Fig. 1 It is a step flowchart of a chip simulation model test method;

[0079] Fig. 2 It is a detailed implementation step flowchart of step S2;

[0080] Fig. 3 It is a detailed implementation step flowchart of step S3.

[0081] The implementation of the present application, functional features and advantages will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0082] The technical method of the present application will be described clearly and completely below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0083] Furthermore, the accompanying drawings are included to provide a further understanding of the present application, and are incorporated in and constitute a part of this specification. The drawings illustrate embodiments of the present application and, together with the description, serve to explain the principles of the present application. In the drawings:

[0084] It is to be understood that, although terms such as "first", "second", and so on can be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, without departing from the scope of the exemplary embodiments, a first element can be referred to as a second element, and similarly a second element can be referred to as a first element. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0085] To achieve the above object, there is provided Figs. 1 to 3 A chip simulation model testing method, comprising the following steps:

[0086] Step S1: collecting a target chip simulation model; performing timing complete compensation on the target chip simulation model to generate jitter compensation data; performing eye diagram reconstruction on the jitter compensation data to obtain waveform eye diagram data;

[0087] Step S2: performing waveform amplitude normalization on the waveform eye diagram data to generate normalized data; performing timing phase correction on the normalized data to generate purifying alignment data;

[0088] Step S3: performing orthogonal mapping processing on the purifying alignment data to generate orthogonal basis data; performing modal projection analysis based on the orthogonal basis data to generate principal modal data;

[0089] Step S4: performing vector mapping driving construction on the principal modal data to generate kernel space data; performing manifold cognitive intelligent aggregation on the kernel space data to generate test coverage features;

[0090] Step S5: performing boundary driving analysis on the test coverage features to generate density distribution data; performing stable feature precision calibration based on the density distribution data to obtain clustering boundary data;

[0091] Step S6: performing boundary precision tuning on the clustering boundary data to generate compensation result data; performing adaptive test sequence execution on the target chip simulation model based on the compensation result data to generate test case results.

[0092] The application ensures the stability of the signal through the timing complete compensation of the target chip simulation model, improves the accuracy of the waveform through the generation of the jitter compensation data, provides an intuitive basis for the signal quality evaluation through the acquisition of the waveform eye diagram data, enhances the consistency of the data through the implementation of the waveform amplitude normalization, ensures the phase accuracy of the signal through the generation of the purification alignment data, provides an effective mathematical basis for the subsequent analysis through the implementation of the orthogonal mapping processing, can effectively extract the main characteristics of the signal through the generation of the main modal data, provides support for the in-depth analysis of the signal characteristics through the construction of the kernel space data, enhances the comprehensiveness of the test coverage through the development of the manifold cognitive intelligence aggregation, provides a quantitative basis for the stability analysis of the characteristics through the generation of the density distribution data, can effectively identify the characteristic distribution of the signal through the formation of the clustering boundary data, ensures the reliability of the test result through the implementation of the boundary accuracy optimization, provides a basis for the execution of the adaptive test sequence through the generation of the compensation result data, can provide comprehensive support for the performance verification of the target chip through the formation of the test case result, and overall improves the scientificity and efficiency of the chip simulation model test, and provides an effective technical means and solution for chip development and optimization.

[0093] In the embodiment of the application, the test method of the chip simulation model comprises the following steps:

[0094] Step S1: collecting a target chip simulation model; performing timing complete compensation on the target chip simulation model to generate jitter compensation data; performing eye diagram reconstruction on the jitter compensation data to obtain waveform eye diagram data;

[0095] In the process of collecting the target chip simulation model, high-precision data acquisition equipment is used to comprehensively sample the physical parameters of the target chip, the sampling signals include clock signals, data signals and power signals, lossless compression algorithm is used for storing the sampling data to ensure the integrity of the data, after the sampling is completed, an adaptive timing compensation algorithm is used to perform timing complete compensation on the sampling data, a time domain cross correction method is used in the compensation process, a dynamically adjusted compensation node is inserted in the original data to keep the signal jitter amplitude at the nanosecond level, after the compensation is completed, the timing corrected data is reconstructed into an eye diagram, the sliding window average method is used to segment the sampling data, the peak value, level, overshoot and other characteristic parameters are extracted in each time window, and an eye diagram matrix is constructed based on the statistical characteristics to generate complete waveform eye diagram data.

[0096] Step S2: performing waveform amplitude normalization on the waveform eye diagram data to generate normalized data; performing timing phase correction on the normalized data to generate purification alignment data;

[0097] In the embodiment, in the process of waveform amplitude normalization of the waveform eye diagram data, first, the maximum amplitude and the minimum amplitude of the signal are calculated based on the amplitude distribution histogram, the amplitude value is adjusted to a normalized range by using an amplitude stretching transformation method, in the normalization process, an interval scaling algorithm is used to linearly map the amplitude data to the interval [0, 1], after the normalization is completed, timing phase correction is performed, the correction method uses a least square fitting method, the phase offset of the data is adjusted by calculating the phase deviation of the eye diagram data and the reference clock, so that the zero crossing point of the signal is consistent with the ideal clock signal, thereby generating purified alignment data.

[0098] Step S3: performing orthogonal mapping processing on the purified alignment data to generate orthogonal basis data; performing modal projection analysis based on the orthogonal basis data to generate main modal data;

[0099] In the embodiment, in the process of orthogonal mapping processing on the purified alignment data, first, a singular value decomposition (SVD) method is used to decompose the data matrix, the main feature vectors are extracted as orthogonal basis vectors, a principal component analysis (PCA) method is used to reduce the dimension of the orthogonal basis vectors, and more than 95% of the variance information is retained, after the mapping is completed, a modal projection analysis method is used to perform modal projection on the orthogonal basis data, a weighted projection algorithm is used for weight distribution of different modal characteristics, and finally the main modal data is obtained.

[0100] Step S4: vector mapping driven construction of the main modal data to generate kernel space data; manifold cognitive intelligent aggregation of the kernel space data to generate test coverage features;

[0101] In the embodiment, in the process of vector mapping driven construction of the main modal data, a multilayer perceptron (MLP) is used to map the modal characteristics, in the mapping process, a weight sharing mechanism is used to optimize the computing resources, after the mapping is completed, a kernel function method is used to construct the kernel space data, a radial basis function (RBF) is used as the kernel function, the mapping relationship of each modal vector in the high-dimensional space is calculated, after the kernel space data is generated, an intelligent aggregation method based on manifold learning (Manifold Learning) is used to intelligently aggregate the data, in the aggregation process, a manifold dimension reduction algorithm is used to construct a nonlinear low-dimensional space, and the test coverage features are obtained.

[0102] Step S5: boundary driven analysis of the test coverage features to generate density distribution data; stable feature accurate calibration based on the density distribution data to obtain clustering boundary data;

[0103] In the process of boundary driving analysis of the test coverage feature in this embodiment, first, the density peak clustering algorithm (Density Peak Clustering, DPC) is used to cluster the test features, the local density of each feature point is calculated and the boundary point position is marked, the density distribution function is used to fit the feature boundary in the analysis process, and finally the density distribution data is generated. Based on the density distribution data, the minimum entropy optimization algorithm is used for stable feature precision calibration. In the calibration process, Kalman filter is used to smooth the feature data to obtain the clustering boundary data.

[0104] Step S6: boundary accuracy optimization is performed on the clustering boundary data to generate compensation result data; and adaptive test sequence execution is performed on the target chip simulation model based on the compensation result data to generate test case results.

[0105] In the process of boundary accuracy optimization of the clustering boundary data in this embodiment, the Iterative Closest Point (ICP) algorithm is used to match the boundary data with the ideal model, the boundary error is optimized through iteration, and the boundary error is converged to a set threshold. After the accuracy optimization is completed, adaptive test sequence execution is performed on the target chip simulation model based on the compensation result data. The test sequence uses a dynamic random sequence generation algorithm to dynamically adjust the input sequence according to the actual operating environment of the chip. In the execution process, high-precision measuring instruments are used to collect test results, and finally test case results are generated.

[0106] Preferably, step S1 comprises the following steps:

[0107] Step S11: collecting a target chip simulation model; performing waveform injection scanning on the target chip simulation model to obtain original waveform data;

[0108] Step S12: reorganizing the time sequence structure of the original waveform data to obtain time sequence deconstruction data; and performing edge jitter detection on the time sequence deconstruction data to obtain edge jitter data;

[0109] Step S13: performing jitter compensation on the time sequence deconstruction data according to the edge jitter data to generate jitter compensation data;

[0110] Step S14: performing multi-domain sampling remapping on the jitter compensation data to obtain remapping data; and performing clock recovery calibration on the remapping data to obtain calibration data;

[0111] Step S15: performing crosstalk elimination processing on the calibration data to generate waveform noise reduction data; and performing eye diagram reconstruction on the waveform noise reduction data to obtain waveform eye diagram data.

[0112] In this embodiment, the register transfer level (RTL) model of the target chip is loaded through the simulation tool Cadence Virtuoso, the working clock frequency is configured as 500 MHz in the simulation environment, the voltage is set as 1.2 V, the signal source generator Keysight M8195A is connected to the input port of the chip, the PRBS31 pseudo-random code sequence is used for waveform injection, the oscilloscope Tektronix DPO73304D is used for synchronous acquisition of the output end signal data, the acquisition depth is set as 100 M points, the time domain resolution is set as 10 ps, and after the data acquisition is completed, the original waveform data is exported as.The collected data is preliminarily visually analyzed by using a signal analysis tool MATLAB, the data integrity is verified, and it is ensured that the sampling range covers the entire test time window; a timing analysis tool Synopsys PrimeTime is used to perform timing analysis on the original waveform data, the collected waveform is divided into multiple clock cycle windows according to the period of the simulation clock signal, the rising edge and the falling edge of the waveform are identified by using a zero-crossing detection algorithm, after the complete timing structure is reconstructed, the signal edge features are extracted by using a Gaussian weighted moving average filtering method, the jitter amplitude of each signal edge is calculated through a timing edge detection module, the detection threshold is set to ±5ps, the edge jitter data is generated after the abnormal points are removed, the timing deconstruction data is input into a timing correction module, the signal waveform is phase-corrected by using a Lagrange interpolation method in combination with the edge jitter data, the jitter amplitude of each cycle is adjusted by using a window smoothing method, the correction step is set to 2ps, the non-linear compensation is performed according to the distribution characteristics of the edge jitter data, after the compensation is completed, the comparison and verification are performed by using a standard reference clock signal, it is ensured that the jitter-corrected data meets the timing constraints of the target chip, and finally the compensated signal waveform data is output; the jitter compensation data is loaded into a frequency domain conversion module, a Fast Fourier Transform (FFT) is used to perform frequency domain analysis on the time domain signal, the sampling rate is set to 10GS / s, multiple frequency bands are divided and a remapping operation is performed, a band-pass filter is used to extract the baseband and harmonic components, after the frequency domain reconstruction is completed, a Phase-Locked Loop (PLL) is used to realize clock recovery and calibration, the phase deviation is adjusted to be within ±1ps, and finally the calibration data with accurate clock recovery is generated; the calibration data is input into a signal integrity analysis tool Ansys SIwave, a spatial proximity analysis method is used to identify the crosstalk interference source between signals, an adaptive noise suppression algorithm is used for crosstalk elimination processing, the crosstalk suppression threshold is set to-30dB, after the crosstalk elimination is completed, a sample-and-hold circuit is used for waveform stabilization processing, the waveform is loaded into an eye diagram analysis module, the data is aligned according to a UI (Unit Interval), a clear waveform eye diagram is generated by using 1000 times of repeated sampling superposition, and key parameters such as eye height, eye width, jitter margin, etc. are calculated.

[0113] Preferably, step S2 comprises the following steps:

[0114] Step S21: performing distortion compensation processing on the waveform eye diagram data to obtain compensated eye diagram data;

[0115] Step S22: performing amplitude normalization on the compensated eye diagram data to generate normalized data;

[0116] Step S23: performing phase alignment calibration on the normalized data to obtain phase alignment eye diagram data;

[0117] Step S24: Inter-symbol de-disruption processing is performed on the phase-aligned eye diagram data to generate purified alignment data.

[0118] In this embodiment, the waveform eye diagram data is input into a signal distortion analysis module, the frequency domain characteristics of the waveform are analyzed using a Band-Limited Model, based on distortion types such as amplitude distortion, phase distortion, and nonlinear distortion, an optimization algorithm based on least squares is used to perform reverse compensation on the waveform data, the step size of distortion compensation is set to 0.5 ns, and the compensation coefficient is adjusted in combination with actual test data, finally the compensated eye diagram data is obtained, the eye diagram display tool is used to compare the eye diagrams before and after compensation, the compensated eye diagram data is input into an amplitude normalization module, the maximum and minimum values of the waveform are standardized using a normalization algorithm, the normalization range is set to 0 to 1, the amplitude ratio is adjusted through dynamic normalization, in the specific operation process, first the maximum and minimum values of the waveform data are calculated, then the scaling ratio is calculated according to the normalization formula, and finally the normalized data is obtained, the time domain data is converted into frequency domain data through Fourier transform, the phase deviation of the signal is analyzed, then a phase adjustment algorithm is used, the phase alignment target is set to an error range of 0.1 ps, in the signal phase correction process, the Bisection method is used to refine the adjustment step size, the correction error is gradually reduced, after adjustment, the data is converted back to the time domain and phase-aligned, the accurately aligned eye diagram data is obtained, the oscilloscope is used to compare the eye diagram data before and after calibration, the data is divided into time domain windows, the sliding window method is used to process the data locally, the noise with a frequency lower than the signal bandwidth is filtered out through a high-pass filter, the cutoff frequency of the filter is set to 5 GHz, the Kalman Filtering algorithm is used to further eliminate the interference between symbols, the purified signal data is generated after correction, the Recursive Least Squares (RLS) algorithm is used to estimate and correct the noise in real time in the data processing, and finally the processed purified data is verified, and the final purified eye diagram data is generated through the eye diagram tool.

[0119] Preferably, step S3 comprises the following steps:

[0120] Step S31: modal decomposition processing is performed on the purified alignment data to obtain modal feature data;

[0121] Step S32: feature vector extraction is performed on the modal feature data to obtain vector group data; orthogonalization processing is performed on the vector group data to generate orthogonal basis data;

[0122] Step S33: feature projection simulation is performed based on the orthogonal basis data to generate a simulated projection matrix;

[0123] Step S34: Perform modal correlation analysis on the orthogonal basis data according to the simulated projection matrix to obtain modal correlation data; perform modal screening based on the modal correlation data to generate main modal data.

[0124] In this embodiment, the purified alignment data is input into the modal decomposition module, the principal component analysis (PCA) method is used for time domain decomposition of the signal, the number of principal components of modal decomposition is set to 5, the principal component coefficients of each signal are calculated by using the PCA algorithm, and the principal components with contribution degree higher than 85% are selected for analysis, the principal components are used to construct modal feature data, and finally the modal data containing significant features is obtained, the effectiveness of the modal data is further verified through frequency domain analysis, the frequency domain reconstruction is carried out by using the toolbox of MATLAB, the modal feature data is input into the characteristic vector extraction module, the singular value decomposition (SVD) method is used for matrix decomposition of the modal feature data, the precision threshold of matrix decomposition is set to 0.01, a plurality of singular values and corresponding characteristic vectors are generated in the decomposition process, the first 5 characteristic vectors in the characteristic vector matrix are selected, the vector group data is generated, the “svd” function in R language is used for characteristic vector extraction, the contribution degrees of the characteristic vectors are compared to ensure that the selected characteristic vectors cover most of the features of the data, the extracted characteristic vector data is input into the orthogonalization processing module, the Gram-Schmidt orthogonalization method is used for orthogonalization of the vector group data, the error threshold of each vector in the orthogonalization process is set to 10^-4, the original vector is first normalized, and then the inner product values between vectors are adjusted step by step to make the orthogonal error between vectors reach the set threshold, the orthogonal basis data is obtained, the operation is realized by using the NumPy library in Python, the included angle between vectors is calculated to ensure that the orthogonal basis data completely meets the orthogonal standard, the orthogonal basis data is input into the characteristic projection simulation module, each characteristic vector is projected into a new space by using the projection analysis algorithm, the dimension of the projection space is set to 3, the projection matrix is fitted by using the least square method to obtain the simulation projection matrix, in the projection process, the projection error is set to 0.001, the projection angle and the offset are adjusted step by step to ensure that the projection matrix can accurately reflect the characteristics of the orthogonal basis data, the “proj” function in Matlab is used for data projection, and the proximity of each projection point to the actual data point is checked, the simulation projection matrix and the orthogonal basis data are input into the modal correlation analysis module, the Pearson correlation coefficient method is used to calculate the correlation between each modal, the correlation threshold is set to 0.8, the correlation coefficient of each pair of modal data is calculated, the modal pairs with correlation greater than 0.8 are selected as effective modes by analyzing the correlation between modes, the modal correlation data is obtained, the “cor” function in R language is used for correlation calculation of the modal data, the modal correlation data is input into the modal screening module, the threshold-based screening algorithm is used, and the correlation threshold is set to 0.85, the modalities with correlation greater than the threshold value are screened out, and finally the main modal data is generated, in the screening process, the correlation score of each modality is set to be between 0 and 1, the selection of the main modal data is determined by regression analysis on the screened data, the screening process is processed using the "sklearn" library in Python, and finally the accurate main modal data is obtained.

[0125] Preferably, step S4 comprises the following steps:

[0126] Step S41: mapping the main modal data to a test vector to obtain test sequence data;

[0127] Step S42: performing kernel function transformation on the test sequence data to generate kernel space data;

[0128] Step S43: performing manifold learning reconstruction on the kernel space data to generate manifold structure data;

[0129] Step S44: performing coverage depth evaluation on the manifold structure data to generate depth index data;

[0130] Step S45: performing feature fusion processing on the depth index data to generate test coverage features.

[0131] In this embodiment, the principal modal data is input into the test vector mapping module, and the support vector machine (SVM) method is used for mapping, and the kernel function of the support vector machine is set to radial basis function (RBF). In the mapping process, the mapping dimension is selected as 6, the mapping value of the principal modal data in the high-dimensional space is calculated by the SVM algorithm, the mapping is carried out by using the "svm" toolbox of MATLAB, and in the mapping process, the classification accuracy is controlled by setting the C value as 1 and the γ value as 0.5. The test sequence data is obtained by post-processing the test vector mapping result, which contains the test vector of each group of mapped principal modal data and the corresponding output category information. The test sequence data is input into the kernel function transformation module, the Gaussian kernel function (Gaussian Kernel) is selected as the transformation function, the parameter σ of the kernel function is set as 0.8, the "kernel_transform" function in the "sklearn" library of Python is used to perform Gaussian kernel transformation on the test sequence data, and the kernel space data is obtained by mapping the original data to the high-dimensional space. In the kernel function transformation process, the similarity between the input data points and the mapping points is calculated to generate a new data representation, so that the data points are more easily distinguished in the high-dimensional space, and the transformed data has higher classification characteristics. The kernel space data is input into the manifold learning reconstruction module, the Laplacian Eigenmaps (LE) algorithm is used for manifold learning reconstruction of the data, the neighborhood size is set as 10, the "laplacian_eigenmaps" function in the R language is used for reconstruction, the adjacency matrix between data points is constructed, the weight function is set as Gaussian function, the feature decomposition is carried out based on the Laplacian matrix, and the low-dimensional representation of the data is obtained. The local structure characteristics of the data can be effectively preserved by visualizing the manifold reconstruction result. In the reconstruction process, the reconstruction error threshold is set as 0.005, the manifold structure data is input into the coverage depth evaluation module, the K nearest neighbors (KNN) method is used for coverage depth evaluation, the K value is set as 5, the local density and coverage depth of each data point in the manifold structure are calculated, and the depth index data is obtained. The "KNN" function in Matlab is used to realize the operation, the coverage depth of each point is determined by analyzing the distance relationship between the data points, and in the evaluation process, the depth evaluation threshold is set as 0.02 to ensure the accuracy of the depth evaluation. The final depth index data is generated by continuously adjusting the K value and the depth threshold. The depth index data is input into the feature fusion processing module, and different features are fused by using the weighted fusion method. The weighted factors are set as 0.6 and 0.4, respectively represent the weights of the depth index data and the test sequence data, the weighted sum of each feature is calculated to obtain the fused test coverage feature, the weighted average function in the "numpy" library in Python is used for feature fusion, the fusion result is analyzed to ensure that the test coverage feature can reflect the comprehensive information of the test sequence data and the depth index data at the same time, and finally the test coverage feature data is generated.

[0132] Preferably, step S5 comprises the following steps:

[0133] Step S51: limit condition inference is performed on the test coverage feature to obtain limit coverage parameters;

[0134] Step S52: density estimation clustering is performed on the limit coverage parameters to generate density distribution data;

[0135] Step S53: inflection point positioning is performed based on the density distribution data to obtain inflection point feature data; boundary positioning is performed according to the inflection point feature data to generate boundary feature data;

[0136] Step S54: stability evaluation is performed based on the boundary feature data to generate stability index parameters; the stability index parameters are clustered and optimized to obtain clustered boundary data.

[0137] In this embodiment, the test coverage feature data is input into the limit condition inference module, the extreme value theory (EVT) is used for data analysis, the threshold is set as the 95% quantile point of the data, the maximum value clustering is performed on the test coverage feature data, the behavior of the data under the limit condition is inferred, the limit condition inference is performed using the "evm" toolbox in Matlab, the tail characteristics of the data are modeled in combination with the extreme value distribution, the limit coverage parameter is determined, the inference result will include the response limit of the coverage feature under the limit condition, the statistical characteristics of the limit coverage parameter are obtained by calculation, the tolerance of the parameter is set as 0.1, the limit coverage parameter is input into the density estimation clustering module, the kernel density estimation (KDE) algorithm is used for density estimation of the limit coverage parameter, the kernel function is set as Gaussian kernel function, the bandwidth is selected as 0.5, the "density" function in R language is used for density estimation clustering, the local density of each limit coverage parameter point is calculated in the estimation process, the probability density function (PDF) is formed, the density distribution data is generated by integrating the density function, the density distribution data will reflect the distribution characteristics of the limit coverage parameter, the density distribution data is input into the inflection point positioning module, the second derivative method is used for inflection point analysis of the data, the "find_peaks" function in the "scipy" library of Python is used for inflection point positioning, the threshold for inflection point detection is set as 0.02, the second derivative calculation is performed on the density distribution curve, the local maximum and minimum values of the curve are detected and calibrated, the inflection points will be extracted as key feature points, the inflection point feature data is obtained, the inflection point data will include the position and the corresponding density value, the inflection point feature data is input into the boundary positioning module, the boundary segmentation algorithm is used for processing the inflection point data, the minimum distance for boundary recognition is set as 0.3. The inflection points are connected by a dynamic programming (DP) based method to form the boundary interval of the data. The boundary feature data is generated by this method. The boundary positioning process determines the start and end points of the boundary by analyzing the relative position and density changes between the inflection points. The generated boundary feature data includes the specific position of the boundary and its corresponding density value. The boundary feature data is input into the stability evaluation module. The Lyapunov index method is used to evaluate the stability of the boundary. The evaluation window size is set to 5. The Lyapunov index is calculated using the "lyapunov" function in Python. The dynamic changes of the boundary feature data in different time windows are analyzed to obtain the stability index parameters. In the stability evaluation process, the volatility and uncertainty of the data are considered. The Lyapunov index obtained by calculation reflects the stability of the system. If the index is negative, the system is stable. The stability index parameters generated include Lyapunov index value, standard deviation, mean value, etc. The stability index parameters are input into the clustering optimization module. The K-means clustering algorithm (K-Means Clustering) is used to optimize the stability index data. The K value is set to 3. The "sklearn.cluster.KMeans" function in Python is used for clustering processing. In the optimization process, the K value is adjusted to achieve the optimal clustering effect. The clustering boundary data is generated. The boundary data generated after clustering is grouped according to the distribution of the stability index to obtain the boundary position and stability characteristics of each group. Finally, the clustering boundary data is generated.

[0138] Preferably, step S53 comprises the following steps:

[0139] Gradient transformation is performed on the density distribution data to obtain gradient mapping data. Local extreme value detection is performed on the gradient mapping data to generate extreme point set data.

[0140] Curvature calculation is performed on the extreme point set data to obtain curvature feature parameters. Threshold filtering is performed on the curvature feature parameters based on a preset inflection point curvature threshold to generate candidate inflection point data.

[0141] Local window segmentation is performed on the candidate inflection point data to obtain window feature data. Morphological processing is performed on the candidate inflection point data based on the window feature data to generate inflection point morphological features.

[0142] Inflection point center positioning is performed based on the inflection point morphological features to generate inflection point feature data.

[0143] Neighborhood expansion simulation is performed on the inflection point feature data to obtain a simulated expansion region. Directional analysis is performed on the simulated expansion region to generate direction feature data.

[0144] The continuity of the direction feature data is identified to obtain continuity index data; boundary tracking is performed based on the continuity index data to generate initial boundary data;

[0145] The smoothness of the initial boundary data is evaluated to obtain a boundary smoothness parameter; the initial boundary data is accurately relocated based on the boundary smoothness parameter to generate boundary feature data.

[0146] In this embodiment, the density distribution data is input into the gradient transformation processing module, the gradient of each data point is calculated using the Central Difference Method, the window size is set to 3, the gradient of each data point and its neighborhood data is calculated to obtain gradient mapping data, in the gradient transformation processing, the local change rate of the density distribution is calculated to identify the significant area of data change, the obtained data contains the gradient amplitude and direction of each point, the process is realized using the “gradient” function in the NumPy library of Python, the gradient mapping data is input into the local extreme value detection module, the second derivative method is used to detect the local extreme value, the extreme value detection threshold is set to 0.1, the “findpeaks” function in Matlab is used for detection, the second derivative of each data point and its neighborhood point is analyzed to find the local extreme point, the extreme point set data is generated, which contains the position of each extreme point and the corresponding gradient value, the extreme point is a key change point extracted from the gradient mapping data, representing the significant turning point of the density distribution data, the extreme point set data is input into the curvature calculation module, the curve fitting method is used to calculate the curvature, the fitting function is set to a quadratic polynomial, the “scipy.optimize.curve_fit” function of Python is used for data fitting, the second derivative of the curve is calculated according to the fitting result, and then the curvature characteristic parameter is obtained, in the curvature calculation process, the convexity and concavity of the point are judged by the change speed of the curve, the generated curvature characteristic parameter includes the curvature value of each point, representing the bending degree of the point on the curve, the curvature characteristic parameter is input into the inflection point screening module, the curvature threshold is set to 0.05, the threshold screening is performed on the curvature characteristic parameter, and the points with curvature value greater than the threshold are screened out as candidate inflection point data, the “numpy.where” function in Python is used for threshold screening, and the generated candidate inflection point data contains the points with curvature greater than the set threshold and their corresponding curvature values, these points represent the turning position of the curve, the candidate inflection point data is input into the local window segmentation module, the sliding window method is used to segment the data, the window size is set to 5, the candidate inflection points are divided by moving the window, and it is ensured that the data in each window contains at least one inflection point, in each window, the maximum and minimum values in the window are extracted to generate window feature data, the sliding window method divides the candidate inflection point data into multiple local regions for analysis to facilitate subsequent processing, the generated window feature data includes the extreme point in each window and its change trend, the window feature data is input into the morphological processing module, the Erosion and Dilation operation is used for morphological processing of the data, the structure element is set to a 3x3 matrix, and the “cv2.erode” and “cv2.dilate” functions in OpenCV are used for processing.The inflection point morphological feature data generated by the "dilate" function through the expansion and contraction of the data generates the morphological features of the inflection points. The morphological processing can remove noise and extract the main structure of the data. The final generated inflection point morphological feature data contains the shape of each inflection point and its spatial position. The inflection point morphological feature is input into the center positioning module, and the centroid method (Centroid Method) is used to calculate the center position of the inflection point. The positioning accuracy is set to 0.01. The centroid of the inflection point morphological data is calculated, and the "regionprops" function in Matlab is used to calculate the centroid. The position of the inflection point center is obtained, and the inflection point feature data is generated. The inflection point center positioning process determines the geometric center of the data points, and the generated inflection point feature data contains the accurate position of the inflection point and its corresponding attributes. The inflection point feature data is input into the neighborhood expansion simulation module, and the Voronoi diagram (Voronoi Diagram) algorithm is used for neighborhood expansion. The neighborhood expansion radius is set to 0.1. The "scipy.spatial.Voronoi" function in Python is used to calculate the neighborhood area of each inflection point. The expansion process generates an expansion area for each inflection point based on its position, and the simulated expansion area is obtained. The results of the neighborhood expansion simulation contain the influence range of each inflection point, representing the influence of the inflection point on the surrounding area. The simulated expansion area is input into the directionality analysis module, and the gradient direction method (Gradient Direction Method) is used to analyze the expansion area. The analysis accuracy is set to 0.02. The gradient direction of the data in each region is calculated to obtain the directionality features of the expansion area. The "gradient" function in Matlab is used for gradient direction analysis. The generated direction feature data includes the main direction and gradient value of each region, representing the main flow direction of each expansion area. The direction feature data is input into the continuity recognition module, and the time series analysis method is used for continuity recognition. The time window is set to 10. The "statsmodels.tsa" library in Python is used for time series modeling to analyze the trend of each data point and obtain the continuity index data. The continuity recognition identifies the stable changing areas by analyzing the time sequence of the data. The generated continuity index data includes the trend and stability of each data point. The continuity index data is input into the boundary tracking module, and the boundary tracking algorithm (Boundary Tracking Algorithm) is used to track the data. The tracking threshold is set to 0.05. The "skimage.measure" library in Python is used to track the boundary.The "label" function marks the regions of the data, identifies the continuous regions of the boundary, generates initial boundary data, and the boundary tracking process analyzes the trend of the continuity index data, generates initial boundary data containing the shape and position of the boundary, inputs the initial boundary data into the smoothness evaluation module, uses the least squares method to calculate the smoothness of the boundary data, sets the fitting accuracy to 0.01, performs boundary fitting through the "scipy.optimize.curve_fit" function in Python, calculates the smoothness parameters of the boundary, and the generated smoothness parameters will contain the fitting curve and residual of the boundary, indicating the smoothness of the boundary. The boundary smoothness parameters are input into the accurate relocation module, and the minimum distance relocation method is used to adjust the boundary, the relocation accuracy is set to 0.01, the boundary points are adjusted point by point, the "interpolate" function in Matlab is used for data interpolation, and the boundary is relocated to generate boundary feature data. The relocated boundary feature data will contain more accurate position and shape.

[0147] Preferably, step S6 comprises the following steps:

[0148] Step S61: Similarity calculation is performed on the clustered boundary data to obtain boundary similarity data;

[0149] Step S62: Consistency comparison is performed on the boundary similarity data to generate comparison result parameters;

[0150] Step S63: Deviation analysis and correction are performed based on the comparison result parameters to generate compensation result data; based on the compensation result data, candidate test allocation is performed to generate supplementary test data;

[0151] Step S64: Adaptive test sequence execution is performed on the target chip simulation model according to the supplementary test data and the compensation result data to generate test case results.

[0152] In this embodiment, the Euclidean distance is used as a measurement index, and the similarity between each pair of boundary data points is calculated by comparing each point with all other points. The similarity matrix is obtained by calculating the shortest distance between each point and all other points. The similarity matrix is used as the basis data for subsequent consistency comparison. The boundary similarity data obtained in the previous step is input into the consistency analysis algorithm. The algorithm uses the matching degree algorithm to compare each group of data. If the similarity between two groups of boundaries exceeds the threshold, it is determined to be consistent, and the comparison result parameter is generated. The consistency of the boundary data is determined by scanning the boundary data step by step based on the similarity value sorting. The result including the matching degree of each group of data is output. The clustering analysis or K-means algorithm is usually used to sort and compare the matching degree. The comparison result parameter is represented by a binary value, 1 represents consistency, and 0 represents inconsistency. The inconsistent data pairs are selected from the matching result and used as the object of deviation analysis. The deviation analysis adjusts the deviation between the two groups of boundary data. The inconsistent boundary data is corrected using the weighted average method. The correction coefficient of each boundary data is calculated during the correction process. The compensation result data is generated after correction. This compensation result data plays an important role in subsequent test allocation. The compensation result data is mainly composed of corrected boundary parameters, including the compensated position, direction, and shape characteristics. After the compensation result data is generated, it is distributed through the test allocation algorithm. First, the candidate test data is generated based on the characteristics of the compensation result data. The candidate test allocation is adjusted based on the parameters in the compensation result. The genetic algorithm is used for optimal allocation. During the optimization process, the error range of the compensation result data is calculated, and the test task is dynamically allocated to generate a set of candidate test data. These candidate data are effectively expanded based on the compensation data to ensure that the test data and the target chip simulation model have the maximum matching degree. The candidate test data contains the corrected value and the optimized test path parameter. Based on the supplementary test data and the compensation result data, the adaptive test sequence of the target chip simulation model is executed. First, according to the rules of the test sequence, for each test data point, the test order and execution path are adjusted based on the previous compensation result data. The automatic test platform is used to simulate and execute each test case. The execution result of the test case is generated through the simulation software (such as Simulink).The execution of test cases is dynamically adjusted based on previously supplemented test data. During execution, test data feedback is acquired in real time, and the test data is updated based on the feedback results. Finally, the adaptive test sequence is executed, and the generated test case results are used for subsequent chip performance evaluation.

[0153] Preferably, step S63 includes the following steps:

[0154] The results parameters are compared and processed by difference mapping to obtain difference distribution data; based on the difference distribution data, regional segmentation processing is performed to obtain regional feature data.

[0155] Based on the differential distribution data, the regional characteristic data are prioritized and sorted to generate deviation level data;

[0156] The deviation level data is classified according to the preset deviation level threshold to generate classification label data; the classification label data is then matched with compensation rules to generate matching compensation rules.

[0157] Compensation calculations are performed on the classification label data based on matching compensation rules to generate compensation result data;

[0158] The residuals of the compensation results are calculated to obtain the characteristics of the compensation residuals; the convergence of the compensation residual characteristics is analyzed to generate convergence index data.

[0159] Perform coverage analysis on the convergence index data to generate compensation coverage features; map test scenarios based on the compensation coverage features to generate a mapping scenario matrix;

[0160] Test requirements are decomposed based on the mapping scenario matrix to obtain requirement item data; test conditions are added to the requirement item data to generate a test condition set.

[0161] Test cases are constructed based on test condition sets to generate test task sequences; dependency analysis is performed on the test task sequences to obtain task dependencies.

[0162] The test task sequence is arranged and allocated according to task dependencies to generate supplementary test data.

[0163] In this embodiment, the comparison result parameter data set is obtained, the comparison result parameters and the data of the target chip simulation model are compared item by item using the difference mapping algorithm, the difference values between the parameters are converted into difference mapping data, the difference values are quantified by the Euclidean distance, the quantized values of each difference are mapped to the specified area, and complete difference distribution data is obtained, wherein the difference values of each area reflect the deviation of the chip model under different operating conditions during the test. When the area segmentation processing is performed, the K-means clustering algorithm is used to segment the data based on the difference distribution data, the areas with large difference values are marked as highlights, the average difference value of each area is calculated as the representative characteristic value of the area, and the area characteristic data obtained by this method includes the main difference characteristics and relative positions of each area. The segmented area data provides a basis for subsequent deviation level generation. In the area segmentation process, the difference distribution data input and the number of clustering centers of the K-means algorithm are optimized through experimental data. When the area characteristic data is prioritized based on the difference distribution data, the characteristics of each area are calculated first, and the priority is sorted according to the difference values of the area characteristics. The priority is calculated according to the absolute size of the difference value and the range of the difference mapping. The area characteristic data is sorted according to the weight factor sorting rule. The deviation level data obtained includes the deviation level of each area. The area with a larger numerical value has a higher deviation level. When the deviation level data is threshold classified according to the preset deviation level threshold, the deviation level data is classified into different categories according to the predetermined standard. Multiple thresholds are set, such as high, medium and low levels. The high level indicates a large difference, which needs to be corrected first. The medium and low levels are corrected according to the difference degree. The data is divided into different categories according to the difference range of each level during the classification. The generated classification mark data is stored in the form of category number. When the compensation rule matching is performed, the data of each category is matched according to the classification mark data. In the matching process, the pre-defined compensation rule library is used. The rule library includes data compensation strategies for different deviation levels. The corresponding compensation rule is applied to each classification mark data through the algorithm. The compensation rule is matched according to the characteristics of the deviation level data. The matched compensation rule is applied to the related data to generate a matching compensation rule. When the classification mark data is compensated based on the matching compensation rule, the classification mark data is compensated using the matched compensation rule. In the calculation process, the deviation value of each data is corrected one by one through the compensation algorithm according to the parameters in the compensation rule. The compensation calculation uses the weighted average method to correct each data, ensuring that each data point is appropriately corrected according to its deviation level. The compensation result data is obtained. In the calculation process of the compensation residual characteristics, the compensation result data and the original data are compared, and the residual value of each data point is calculated.The residual value is the difference before and after compensation. The compensation residual feature is obtained in this way. In the residual calculation process, the compensation residual of each data point is quantified using the mean squared error (MSE), and a data set of the compensation residual feature is obtained. When performing convergence analysis, the data is analyzed based on the compensation residual feature. In the analysis process, a convergence detection algorithm is used, such as the commonly used convergence criterion, to check whether the data residual is stable. When the residual value is lower than the preset convergence threshold, it is determined to be converged, and the convergence index data is generated based on the trend of the residual value, recording the convergence process of the residual value. When performing compensation coverage feature analysis, the coverage degree is analyzed in detail based on the convergence index data. The coverage degree analysis uses a graphical algorithm to calculate the coverage range of the data points. In the analysis process, the covered area is calculated according to the distribution range of the data points, and the coverage feature data is generated. The coverage feature data reflects the distribution of the compensated data points in the area. The coverage area of each compensation point is displayed graphically to generate the compensation coverage feature. When mapping the test scene based on the compensation coverage feature, the compensation coverage feature is mapped to the specified test scene. When mapping, the compensation coverage area is mapped to the parameters of the test scene through the mapping algorithm to ensure that the compensated data can cover the area required for actual testing. The mapping scene matrix is generated. The mapping scene matrix records the corresponding position and test requirements of each compensation point in the test scene in detail. When decomposing the test requirements according to the mapping scene matrix, the test scene in the mapping scene matrix is first decomposed, and the test conditions are allocated as needed. The generated requirement item data includes the specific test requirements of each test point. The decomposition process ensures that each test requirement is independently identified and listed in detail. When supplementing the test conditions for the requirement item data, the relevant test conditions are supplemented according to the characteristics of the requirement item data. The test conditions are expanded according to the test requirements. After supplementation, the test condition set is generated. The test condition set includes the detailed conditions of all requirement items. When constructing test cases based on the test condition set, detailed test cases are generated by combining the data in the test condition set. The test cases are constructed based on the compensation result data collected earlier to ensure that the test cases can cover all test requirements and dynamically generate test task sequences according to different test conditions. The generation of the task sequence is optimized according to the dependency relationship between the test cases to ensure that each test case can be executed in a specific order. When analyzing the dependency relationship of the test task sequence, the dependency relationship of all test cases is combed, and the order of each test task is analyzed to ensure that the test tasks can be allocated according to the dependency relationship. The task dependency relationship is obtained. The dependency relationship analysis ensures that there is no conflict or repeated execution during the execution of the test cases. Finally, a complete test task sequence is generated. When arranging and allocating the test task sequence according to the task dependency relationship, the test tasks are first sorted according to the task dependency relationship.Ensure the execution order of the test task conforms to the dependency relationship, allocate the test task through an algorithm, finally generate the supplementary test data, and reasonably allocate the supplementary test data according to the test task sequence, so as to ensure the efficient execution of the test work.

[0164] The application further provides a chip simulation model test system for executing the chip simulation model test method.

[0165] An eye diagram reconstruction module is configured to collect the target chip simulation model, perform timing complete compensation on the target chip simulation model to generate jitter compensation data, and perform eye diagram reconstruction on the jitter compensation data to obtain waveform eye diagram data.

[0166] A phase correction module is configured to perform waveform amplitude normalization on the waveform eye diagram data to generate normalized data, and perform timing phase correction on the normalized data to generate purified alignment data.

[0167] A modal analysis module is configured to perform orthogonal mapping processing on the purified alignment data to generate orthogonal basis data, and perform modal projection analysis based on the orthogonal basis data to generate main modal data.

[0168] A feature mapping module is configured to perform vector mapping driving construction on the main modal data to generate kernel space data, and perform manifold cognitive intelligent aggregation on the kernel space data to generate test coverage features.

[0169] A boundary analysis module is configured to perform boundary driving analysis on the test coverage features to generate density distribution data, and perform stable feature accurate calibration based on the density distribution data to obtain clustering boundary data.

[0170] A test optimization module is configured to perform boundary precision tuning on the clustering boundary data to generate compensation result data, and perform adaptive test sequence execution on the target chip simulation model based on the compensation result data to generate test case results.

[0171] The application provides a basis for real-time monitoring of the target chip signal through the implementation of the eye diagram reconstruction module, the generation of the jitter compensation data improves the stability and reliability of the signal, the acquisition of the waveform eye diagram data can intuitively show the signal quality and characteristics, the function of the phase correction module ensures the consistency of the signal amplitude, the generation of the normalized data enhances the accuracy of subsequent analysis, the acquisition of the purified alignment data provides a guarantee for the correction of the signal phase, the orthogonal mapping processing of the modal analysis module provides an effective method for signal feature extraction, the generation of the main modal data can effectively identify the main features and trends of the signal, the construction of the kernel space data of the feature mapping module provides support for deep analysis of signal features, the implementation of the manifold cognitive intelligent aggregation improves the comprehensiveness of test coverage, the density distribution data of the boundary analysis module provides a quantitative basis for stability analysis of features, the generation of the clustering boundary data effectively identifies the distribution and change of signal features, boundary accuracy tuning ensures the reliability of the test result, the generation of the compensation result data provides a basis for the execution of the adaptive test sequence, the formation of the test case result provides comprehensive support for performance verification of the target chip, and overall, the scientificity and efficiency of the chip simulation model test are improved, which provides an effective technical means and solution for chip development and optimization.

[0172] Therefore, embodiments should be considered in all respects as illustrative and non-restrictive, the scope of the application being defined by the appended claims rather than the above description, and all changes falling within the meaning and range of equivalence of the essential elements of the application are intended to be embraced therein.

[0173] The above description is merely one specific implementation of the application, which enables those skilled in the art to understand or implement the application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the application. Therefore, the application will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method of testing a chip emulation model, characterized by, The method comprises the following steps: Step S1: collecting a target chip simulation model; performing timing integrity compensation on the target chip simulation model to generate jitter compensation data; Performing eye diagram reconstruction on the jitter compensation data to obtain waveform eye diagram data; Step S2: performing waveform amplitude normalization on the waveform eye diagram data to generate normalized data; performing timing phase correction on the normalized data to generate purified alignment data; Step S3: performing orthogonal mapping processing on the purified alignment data to generate orthogonal basis data; performing modal projection analysis based on the orthogonal basis data to generate main modal data; Step S4: performing vector mapping driving construction on the main modal data to generate kernel space data; performing manifold cognitive intelligent aggregation on the kernel space data to generate test coverage features; Step S5: performing boundary driving analysis on the test coverage features to generate density distribution data; performing stable feature accurate calibration based on the density distribution data to obtain clustering boundary data; Step S6: performing boundary precision optimization on the clustering boundary data to generate compensation result data; Performing adaptive test sequence execution on the target chip simulation model based on the compensation result data to generate test case results.

2. The method of testing a chip emulation model of claim 1, wherein, Step S1 comprises the following steps: Step S11: collecting a target chip simulation model; performing waveform injection scanning on the target chip simulation model to obtain original waveform data; Step S12: performing timing structure reorganization on the original waveform data to obtain timing deconstruction data; performing edge jitter detection on the timing deconstruction data to obtain edge jitter data; Step S13: performing jitter compensation on the timing deconstruction data according to the edge jitter data to generate jitter compensation data; Step S14: performing multi-domain sampling remapping on the jitter compensation data to obtain remapping data; performing clock recovery calibration on the remapping data to obtain calibration data; Step S15: performing crosstalk elimination processing on the calibration data to generate waveform noise reduction data; performing eye diagram reconstruction on the waveform noise reduction data to obtain waveform eye diagram data.

3. The method of claim 1, wherein, Step S2 comprises the following steps: Step S21: performing distortion compensation processing on the waveform eye diagram data to obtain compensated eye diagram data; Step S22: performing amplitude normalization on the compensated eye diagram data to generate normalized data; Step S23: performing phase alignment calibration on the normalized data to obtain phase alignment eye diagram data; Step S24: performing inter-symbol de-scrambling processing on the phase alignment eye diagram data to generate purified alignment data.

4. The method of claim 1, wherein, Step S3 comprises the following steps: Step S31: performing modal decomposition processing on the purified alignment data to obtain modal feature data; Step S32: performing feature vector extraction on the modal feature data to obtain vector group data; performing orthogonalization processing on the vector group data to generate orthogonal basis data; Step S33: performing feature projection simulation based on the orthogonal basis data to generate simulated projection matrix; Step S34: performing modal correlation analysis on the orthogonal basis data according to the simulated projection matrix to obtain modal correlation data; performing modal screening based on the modal correlation data to generate main modal data.

5. The method of claim 1, wherein, Step S4 comprises the following steps: Step S41: performing test vector mapping on the main modal data to obtain test sequence data; Step S42: Kernel function transformation is performed on the test sequence data to generate kernel space data; Step S43: Manifold learning reconstruction is performed on the kernel space data to generate manifold structure data; Step S44: Coverage depth evaluation is performed on the manifold structure data to generate depth index data; Step S45: Feature fusion processing is performed on the depth index data to generate test coverage features.

6. The method of testing a chip emulation model of claim 1, wherein, Step S5 includes the following steps: Step S51: Limit condition inference is performed on the test coverage features to obtain limit coverage parameters; Step S52: Density estimation clustering is performed on the limit coverage parameters to generate density distribution data; Step S53: Inflection point positioning is performed based on the density distribution data to obtain inflection point feature data; boundary positioning is performed according to the inflection point feature data to generate boundary feature data; Step S54: Stability evaluation is performed based on the boundary feature data to generate stability index parameters; clustering optimization is performed on the stability index parameters to obtain clustered boundary data.

7. The method of testing a chip emulation model of claim 6, wherein, Step S53 includes the following steps: Gradient transformation processing is performed on the density distribution data to obtain gradient mapping data; local extreme value detection is performed on the gradient mapping data to generate extreme point set data; Curvature calculation is performed on the extreme point set data to obtain curvature feature parameters; threshold filtering is performed on the curvature feature parameters based on a preset inflection point curvature threshold to generate candidate inflection point data; Local window segmentation is performed on the candidate inflection point data to obtain window feature data; morphological processing is performed on the candidate inflection point data based on the window feature data to generate inflection point morphological features; Inflection point center positioning is performed according to the inflection point morphological features to generate inflection point feature data; Neighborhood expansion simulation is performed on the inflection point feature data to obtain simulated expansion regions; directionality analysis is performed on the simulated expansion regions to generate direction feature data; Continuity recognition is performed on the direction feature data to obtain continuity index data; boundary tracking is performed based on the continuity index data to generate initial boundary data; Smoothness evaluation is performed on the initial boundary data to obtain boundary smoothness parameters; accurate repositioning is performed on the initial boundary data based on the boundary smoothness parameters to generate boundary feature data.

8. The method of testing a chip emulation model of claim 1, wherein, Step S6 includes the following steps: Step S61: Similarity calculation is performed on the clustered boundary data to obtain boundary similarity data; Step S62: Consistency comparison is performed on the boundary similarity data to generate comparison result parameters; Step S63: Deviation analysis correction is performed based on the comparison result parameters to generate compensation result data; candidate test allocation is performed based on the compensation result data to generate supplementary test data; Step S64: Adaptive test sequence execution is performed on the target chip simulation model according to the supplementary test data and the compensation result data to generate test case results.

9. The method of testing a chip emulation model of claim 8, wherein, Step S63 includes the following steps: Difference mapping processing is performed on the comparison result parameters to obtain difference distribution data; region segmentation processing is performed based on the difference distribution data to obtain region feature data; Priority sorting is performed on the region feature data based on the difference distribution data to generate deviation level data; The deviation level data is threshold classified according to a preset deviation level threshold to generate classification mark data; the classification mark data is matched with a compensation rule to generate a matching compensation rule; The classification mark data is compensated based on the matching compensation rule to generate compensation result data; Residuals of the compensation result data are calculated to obtain compensation residual features; convergence analysis is performed on the compensation residual features to generate convergence index data; Coverage degree analysis is performed on the convergence index data to generate compensation coverage features; test scene mapping is performed based on the compensation coverage features to generate a mapping scene matrix; Test requirement decomposition is performed according to the mapping scene matrix to obtain requirement item data; test condition supplement is performed on the requirement item data to generate a test condition set; Test case construction is performed based on the test condition set to generate a test task sequence; dependency relationship analysis is performed on the test task sequence to obtain task dependency relationships; The test task sequence is arranged and distributed according to the task dependency relationships to generate supplementary test data.

10. A test system for a chip emulation model, characterized by, A test method for executing a chip simulation model as claimed in claim 1, the test system of the chip simulation model comprising: An eye diagram reconstruction module for collecting a target chip simulation model; performing timing complete compensation on the target chip simulation model to generate jitter compensation data; and performing eye diagram reconstruction on the jitter compensation data to obtain waveform eye diagram data; A phase correction module for performing waveform amplitude normalization on the waveform eye diagram data to generate normalized data; and performing timing phase correction on the normalized data to generate purified alignment data; A modal analysis module for performing orthogonal mapping processing on the purified alignment data to generate orthogonal basis data; and performing modal projection analysis based on the orthogonal basis data to generate main modal data; A feature mapping module for performing vector mapping driving construction on the main modal data to generate kernel space data; and performing manifold cognitive intelligent aggregation on the kernel space data to generate test coverage features; A boundary analysis module for performing boundary driving analysis on the test coverage features to generate density distribution data; and performing stable feature precision calibration based on the density distribution data to obtain clustering boundary data; A test optimization module for performing boundary precision tuning on the clustering boundary data to generate compensation result data; and performing adaptive test sequence execution on the target chip simulation model based on the compensation result data to generate test case results.

Images