A parameter configuration method, device and equipment for communication link simulation and a medium

Abstract

This application relates to a parameter configuration method, apparatus, device, and medium for communication link simulation. It includes: selecting a target simulation model from a default simulation model based on the basic signal characteristics of the model to be simulated in the link to be simulated, and obtaining the corresponding initial link simulation result based on the target simulation model; determining the parameter configuration information corresponding to the initial link simulation result, and determining the link simulation parameter configuration information based on the parameter configuration information; extracting feature indices from the link simulation parameter configuration information, calculating the margin space corresponding to the feature indices, and determining the corresponding gradient strength; sorting the gradient strengths, selecting the target gradient strength based on the sorting result, and determining the link simulation parameter configuration information corresponding to the target gradient strength as the target parameter configuration. This application can effectively remove redundant models, normalize model configuration, improve the efficiency of simulation tuning, and complete the evaluation of configuration parameters by solving for gradient strength values, ensuring the quality of simulation configuration.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a parameter configuration method, apparatus, device and medium for communication link simulation. Background Technology

[0002] In a communication link system, configuring software parameters can optimize the performance of the communication link, enabling it to transmit data efficiently. For example, by adjusting parameters such as data transmission rate and frame size, the transmission speed of data in the communication link can be optimized, thereby improving the efficiency of the entire communication system.

[0003] Currently, there are various methods for optimizing software parameter configuration in communication links, including simulating the communication link to optimize its configuration. The simulation optimization method is mature, the simulation results are controllable, and it can optimize the configuration in advance, saving the time window for later debugging. Moreover, the debugging of parameter samples is easier to control. The mainstream simulation models are the SPICE model and the IBIS model. However, the SPICE model has a large number of transistor parameters, making it difficult to quickly solve the results. The IBIS model has many types of models, many of which are redundant and difficult to normalize simply, resulting in a large number of simulation iterations when selecting the optimal configuration, which makes it difficult to promote and apply on a large scale.

[0004] Currently, no effective solution has been proposed for the problems of low simulation efficiency and poor results in existing scenarios for simulating and optimizing communication links. Summary of the Invention

[0005] Therefore, it is necessary to provide a parameter configuration method, apparatus, device, and medium for communication link simulation to address the aforementioned technical problems.

[0006] Firstly, this application provides a parameter configuration method for communication link simulation, the method comprising:

[0007] Based on the preset basic signal characteristics of the simulation model in the simulation link, a target simulation model corresponding to the simulation link is selected from the default simulation models, and the simulation link is simulated based on the target simulation model to obtain the corresponding initial simulation results of the link.

[0008] Determine all parameter configuration information of the target simulation model corresponding to the initial simulation results of the link, and determine at least one set of link simulation parameter configuration information based on the parameter configuration information;

[0009] Extract the feature indicators of the link simulation parameter configuration information, calculate the margin space corresponding to the feature indicators based on the data of the feature indicators and the preset expected indicators, and determine the gradient strength of the corresponding feature indicators according to the margin space.

[0010] The gradient intensities corresponding to the link simulation parameter configuration information are sorted, the target gradient intensities are selected based on the sorting results, and the link simulation parameter configuration information corresponding to the target gradient intensities is determined as the target parameter configuration.

[0011] In one embodiment, based on preset signal characteristics of the model to be simulated in the simulated link, a target simulation model corresponding to the simulated link is selected from the default simulation models, including:

[0012] Based on the basic characteristics of the signal, the corresponding initial simulation model is selected from the default simulation model, and the working condition information of each initial simulation model is extracted.

[0013] Based on the preset communication protocol corresponding to the link to be simulated, each initial simulation model is grouped to obtain at least one network group; wherein, a network group includes at least two model types;

[0014] The working condition information of the initial simulation models corresponding to each model type in the network group is matched pairwise. If a set of target working condition information is successfully matched in each model type, the target simulation model corresponding to the target working condition information is determined according to the matching result.

[0015] In one embodiment, the operating condition information of the initial simulation models corresponding to each model type in the network group is matched pairwise, including:

[0016] Get the preset window constraints;

[0017] Based on the time sequence of the operating condition information, dynamic programming calculations are performed pairwise on the operating condition information of different model types belonging to the same network group, according to window constraints. The similarity results between the operating condition information of different model types are obtained based on the dynamic programming calculation results. When it is detected that the similarity of the operating condition information of the initial simulation model is greater than the preset similarity threshold in each model type, the operating condition information of the initial simulation model corresponding to each model type is determined to be matched.

[0018] In one embodiment, after selecting the corresponding initial simulation model from the default simulation model based on the basic characteristics of the signal and extracting the operating condition information of each initial simulation model, the method further includes:

[0019] Based on the operating condition information, the input and output categories of each initial simulation model are divided to obtain the input category model, the output category model, and the input-output category model.

[0020] For each of the input category model, output category model, and input-output category model, the corresponding target simulation model is selected from the default simulation models.

[0021] In one embodiment, determining the gradient strength of the corresponding feature index based on the margin space includes:

[0022] Iterate through the feature metrics of each dimension and calculate the gradient value of each dimension respectively;

[0023] Convert the gradient values ​​for each dimension into norm form to obtain the gradient strength for each dimension.

[0024] In one embodiment, the gradient intensities corresponding to the link simulation parameter configuration information are sorted, and target gradient intensities are selected based on the gradient intensities, including:

[0025] The gradient intensities are sorted in ascending order, and the maximum value among the sorted gradient intensities is determined as the target gradient intensity.

[0026] In one embodiment, based on preset signal characteristics of the model to be simulated in the simulated link, a target simulation model corresponding to the simulated link is selected from the default simulation models, including:

[0027] Determine the basic signals corresponding to the model to be simulated;

[0028] Define the target signal set of the basic signal, where the target signal set includes the basic signal characteristics corresponding to the model to be simulated;

[0029] The target signal set is mapped and matched with the default signal characteristics corresponding to the default simulation model, and the target simulation model is obtained by filtering.

[0030] Secondly, this application also provides a parameter configuration device. The device includes:

[0031] The acquisition module is used to select the target simulation model corresponding to the link to be simulated from the default simulation models based on the preset basic signal characteristics of the model to be simulated in the link to be simulated, and to simulate the link to be simulated based on the target simulation model to obtain the corresponding initial simulation results of the link.

[0032] The calculation module is used to determine all parameter configuration information of the target simulation model corresponding to the initial simulation results of the link, and determine at least one set of link simulation parameter configuration information based on the parameter configuration information; extract the feature index of the link simulation parameter configuration information, calculate the margin space corresponding to the feature index based on the data of the feature index and the preset expected index, and determine the gradient strength of the corresponding feature index based on the margin space.

[0033] The generation module is used to sort the gradient intensities corresponding to the link simulation parameter configuration information, filter out the target gradient intensities based on the sorting results, and determine the link simulation parameter configuration information corresponding to the target gradient intensities as the target parameter configuration.

[0034] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to perform the following steps:

[0035] Based on the preset basic signal characteristics of the simulation model in the simulation link, a target simulation model corresponding to the simulation link is selected from the default simulation models, and the simulation link is simulated based on the target simulation model to obtain the corresponding initial simulation results of the link.

[0036] Determine all parameter configuration information of the target simulation model corresponding to the initial simulation results of the link, and determine at least one set of link simulation parameter configuration information based on the parameter configuration information;

[0037] Extract the feature indicators of the link simulation parameter configuration information, calculate the margin space corresponding to the feature indicators based on the data of the feature indicators and the preset expected indicators, and determine the gradient strength of the corresponding feature indicators according to the margin space.

[0038] The gradient intensities corresponding to the link simulation parameter configuration information are sorted, the target gradient intensities are selected based on the sorting results, and the link simulation parameter configuration information corresponding to the target gradient intensities is determined as the target parameter configuration.

[0039] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, performs the following steps:

[0040] Based on the preset basic signal characteristics of the simulation model in the simulation link, a target simulation model corresponding to the simulation link is selected from the default simulation models, and the simulation link is simulated based on the target simulation model to obtain the corresponding initial simulation results of the link.

[0041] Determine all parameter configuration information of the target simulation model corresponding to the initial simulation results of the link, and determine at least one set of link simulation parameter configuration information based on the parameter configuration information;

[0042] Extract the feature indicators of the link simulation parameter configuration information, calculate the margin space corresponding to the feature indicators based on the data of the feature indicators and the preset expected indicators, and determine the gradient strength of the corresponding feature indicators according to the margin space.

[0043] The gradient intensities corresponding to the link simulation parameter configuration information are sorted, the target gradient intensities are selected based on the sorting results, and the link simulation parameter configuration information corresponding to the target gradient intensities is determined as the target parameter configuration.

[0044] The aforementioned parameter configuration method, apparatus, device, and medium for communication link simulation first selects a target simulation model based on the basic signal characteristics of the model to be simulated, and obtains the corresponding initial link simulation results based on the target simulation model. Second, it determines all parameter configuration information corresponding to the initial link simulation results, and determines the link simulation parameter configuration information based on this information. Then, it extracts feature indices from the link simulation parameter configuration information, calculates the margin space corresponding to the feature indices based on the data and expected indices, determines the corresponding gradient strength based on the margin space, and finally sorts the gradient strengths corresponding to the link simulation parameter configuration information. Based on the sorting results, it selects the target gradient strength and determines the link simulation parameter configuration information corresponding to the target gradient strength as the target parameter configuration. This application can effectively remove redundant models, normalize model configuration, improve the efficiency of simulation tuning, and ensure the quality of simulation configuration by evaluating the configuration parameters through solving for the gradient strength values. Attached Figure Description

[0045] Figure 1 This is a flowchart illustrating a parameter configuration method in one embodiment;

[0046] Figure 2 This is a flowchart illustrating the parameter configuration method in a preferred embodiment;

[0047] Figure 3 This is a schematic diagram of the simulated waveform curve in one embodiment;

[0048] Figure 4 This is a structural block diagram of a parameter configuration device in one embodiment;

[0049] Figure 5 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0051] In one embodiment, such as Figure 1 As shown, a parameter configuration method for communication link simulation is provided, including:

[0052] Step S110: Based on the preset basic signal characteristics of the simulation model in the simulation link, select the target simulation model corresponding to the simulation link from the default simulation models, and simulate the simulation link based on the target simulation model to obtain the corresponding initial simulation results of the link.

[0053] Specifically, the aforementioned link to be simulated is the communication link that needs to be simulated. A complete link to be simulated typically includes multiple simulation models. Based on the basic signal characteristics of multiple simulation models, the corresponding target simulation model is selected from the default simulation models. In this embodiment, the default simulation model refers to the preset IBIS model (Input / output buffer information specification). The basic signal characteristics refer to the characteristics of several important signals of the simulation model, such as signal group, signal transmission type, and voltage level. The signal group is generally established for networks of the same type, such as DDR (Double Data Rate Synchronous Dynamic Random Access Memory) and RGMII (Reduced Gigabit Media Independent Interface), which includes data, address, and control signals. Different protocols have different classification specifications, that is, in practical applications, the signal group can be determined according to the communication protocol. The aforementioned signal transmission type is generally one of unidirectional and bidirectional transmission. The aforementioned voltage level mainly describes the voltage value of the circuit signal in different instantaneous states, and in some embodiments, it is also the maximum tolerance value of the level. Furthermore, the default simulation model mentioned above is an existing model related to the link to be simulated. In this embodiment, the default simulation model is an existing IBIS model. In summary, in this embodiment, since the basic signal characteristics are key items in the basic simulation parameters, and based on the characteristics of IBIS model simulation, it is necessary to identify the basic signal characteristics of the default simulation model and the model to be simulated one by one. The default simulation models with the same or similar basic signal characteristics are determined as the target simulation model corresponding to the link to be simulated, to ensure the accuracy of the model simulation and the normalization of the model configuration. That is, the model selection needs to meet the requirements of no conflict in signal transmission type, consistent voltage levels, and consistent model configuration within the signal group. Finally, the link to be simulated is simulated based on the target simulation model to obtain the initial simulation results of the link.

[0054] Step S120: Determine all parameter configuration information of the target simulation model corresponding to the initial simulation results of the link, and determine at least one set of link simulation parameter configuration information based on the parameter configuration information.

[0055] Specifically, since signal transmission methods mainly include single-drive-single-load and single-drive-multiple-load structures, and the configuration methods between single drives and between single loads and multiple loads are consistent, to ensure the adaptability of chip configurations in various practical applications, the configurations of various devices in a multiple load generally adopt the same configuration. Furthermore, to verify the simulation effect of the target simulation model, all possible parameter configurations of the target simulation model can be traversed, and the simulation effect of each set of parameter configurations can be calculated, thereby determining the set of parameter configurations with the best simulation effect. In summary, this embodiment first arranges and combines the selected target simulation models to obtain multiple initial simulation results corresponding to the link to be simulated. Then, it counts all parameter configuration information that each target simulation model can obtain in each initial simulation result of the link, and determines the corresponding link simulation parameter configuration information based on each set of parameter configuration information. The above-mentioned initial simulation results of the link are the arrangement results obtained after arranging and combining the above target simulation models. In practical applications, multiple initial simulation results of the link are generally obtained based on the target simulation model. Taking a basic CPU-memory system as an example, the target simulation model generally includes multiple types of CPUs and multiple types of memory modules. Therefore, after arranging and combining the various target simulation models, multiple initial simulation results of the link are generally obtained.

[0056] Step S130: Extract the feature indicators of the link simulation parameter configuration information, calculate the margin space corresponding to the feature indicators based on the data of the feature indicators and the preset expected indicators, and determine the gradient strength of the corresponding feature indicators according to the margin space.

[0057] Specifically, the characteristic indicators corresponding to the link simulation parameter configuration are first extracted. These characteristic indicators include, but are not limited to, indicators such as eye diagrams, bathtub curves, bit error rate, and signal-to-noise ratio. Different signal protocol types have different electrical specification requirements and template indicators. Therefore, in practical applications, the required characteristic indicators can be determined according to actual needs. In some embodiments, the corresponding time-domain waveform curve can be calculated based on the link simulation parameter configuration information applied to the target simulation model, and the corresponding characteristic indicators can be extracted based on the time-domain waveform curve.

[0058] Furthermore, based on the data of each feature indicator and the corresponding expected indicator, the corresponding margin space is calculated. This margin space is generally in vector form. In practical applications, there are usually multiple feature indicators, and in this case, the margin space of each feature indicator also needs to be calculated independently. The margin space for calculating the parameters of each feature indicator and the expected indicator is as follows:

[0059] y = [y1(x) - y2(x)]

[0060] Here, y represents the parameter of the feature index in the y-direction, y1(x) represents the feature index data obtained from simulation (e.g., an eye diagram, in which case the feature index is converted into eye diagram data), and y2(x) represents the corresponding preset expected index. Similarly, if the product parameters are sensitive to key feature indices, the feature index parameters can be decomposed and refined from the existing feature indices. Finally, based on the above expected index, gradient calculation is performed on the margin space corresponding to each feature index to obtain the gradient strength of each feature index.

[0061] Step S140: Sort the gradient intensities corresponding to the link simulation parameter configuration information, filter out the target gradient intensities based on the sorting results, and determine the link simulation parameter configuration information corresponding to the target gradient intensities as the target parameter configuration.

[0062] Specifically, the gradient intensities are sorted, the target gradient intensities are selected based on the sorting results, and the link simulation parameter configuration information corresponding to the target gradient intensities is determined as the target parameter configuration. In some preferred embodiments, the link simulation parameter configuration information corresponding to the largest gradient intensities is determined as the target parameter configuration.

[0063] Through steps S110 to S140, a suitable target simulation model is first selected from the default simulation model, and an evaluation method for the configuration information of each link simulation parameter to verify the initial simulation results of the link is determined. First, a margin space between the data of the feature index and the expected index is constructed. The gradient strength of each feature index is determined by parameter normalization and gradient calculation. Based on the gradient strength, the target gradient strength is selected, and the corresponding configuration is determined as the optimal target parameter configuration. This avoids the problem that it is difficult to obtain more reliable configuration recommendations due to the lack of a systematic evaluation method for simulation test results, and greatly improves the accuracy of the optimal configuration parameters.

[0064] In one embodiment, based on preset signal characteristics of the model to be simulated in the simulated link, a target simulation model corresponding to the simulated link is selected from the default simulation models, including:

[0065] Based on the basic characteristics of the signal, the corresponding initial simulation model is selected from the default simulation model, and the working condition information of each initial simulation model is extracted.

[0066] Based on the preset communication protocol corresponding to the link to be simulated, each initial simulation model is grouped to obtain at least one network group; wherein, a network group includes at least two model types;

[0067] The working condition information of the initial simulation models corresponding to each model type in the network group is matched pairwise. If a set of target working condition information is successfully matched in each model type, the target simulation model corresponding to the target working condition information is determined according to the matching result.

[0068] Specifically, based on the basic characteristics of the signal, the corresponding initial simulation model is selected from the default simulation model, and the operating condition information of each initial simulation model is extracted. The basic characteristics of the signal mainly include the signal group, signal transmission type, voltage level, etc. mentioned above. The operating condition information of the initial simulation model reflects the basic parameter information of the initial simulation model when it is working. This operating condition information can be determined by the input-output curve of the model. In some embodiments, the operating condition information can be defined as an x-dimensional and y-dimensional matrix, where the y-dimensional matrix is ​​a three-dimensional matrix, corresponding to the matrix information of the typical value, minimum value, and maximum value, respectively. These three dimensions correspond to the parameter information under different operating conditions. Table 1-1 below shows the V / I curve information of one of the simulation models, i.e., the operating condition information table. Here, Voltage is the voltage information of the simulation model, I(typ) is the typical current value information of the simulation model under a certain operating voltage, I(min) is the minimum current value information of the simulation model under a certain operating voltage, and I(max) is the maximum current value information of the simulation model under a certain operating voltage.

[0069] Table 1-1

[0070]

[0071] In this embodiment, no restrictions are placed on the screening method for the initial simulation model. Any screening method based on the basic characteristics of the signal should fall within the protection scope of this application.

[0072] Furthermore, based on the preset communication protocol corresponding to the link to be simulated, each initial simulation model is grouped to obtain multiple network groups, and each network group generally includes multiple network types. Taking an ideal CPU-memory system as an example, according to the existing communication protocol, it is assumed that the system is divided into 8 or 9 data network groups and 1 address command control group. Each data signal group is divided into three types of models, including an eight-data-line model, a two-clock-signal model, and a one-mask-signal model.

[0073] Finally, the operating condition information of the initial simulation models corresponding to each model type in the network group is matched pairwise to ensure that the operating condition information of the models of different types in the same network group is consistent to meet the timing requirements. If a set of target operating condition information is successfully matched in each model type, that is, if a model with consistent operating condition information is detected in each model type, the target simulation model corresponding to the target operating condition information is determined according to the matching result. Similarly, models that fail to match are directly deleted. Taking the CPU-memory system mentioned above as an example, the initial simulation models that can be taken as data cable models are abcd. For the clock signal model, the viable initial simulation models are ABCDE; for the mask signal model, the viable initial simulation models are 12345. It is crucial to ensure that the operating conditions of the data line model, mask signal model, and clock signal model are consistent under the same configuration. That is, if the data line model, mask signal model, and clock signal model are found to have consistent operating conditions under the same configuration, then a successful match is determined. Assuming the successfully matched target operating conditions are aA1, bB2, cC3, dD4, and eE5, then model f in the data line model fails to match. Therefore, model f is deleted and not configured or simulated. The aforementioned target simulation models are:

[0074]

[0075] Among them, ST mi_s ST represents an ordered list of models representing all signals in a network group, that is, a list of the various target simulation models in the aforementioned network group. kmi_s It represents a single network ordered model, and there are relationships between models in the same network group. Their configuration attributes and operating conditions must be consistent. The types of network group models are generally equivalent to the types of single network models in the network group.

[0076] This embodiment ensures that the parameters of the signal network models in the same group are consistent based on the differences in the input and output curve parameters of the initial simulation model, thereby further realizing the normalization of signal configuration.

[0077] In one embodiment, the operating condition information of the initial simulation models corresponding to each model type in the network group is matched pairwise, including:

[0078] Get the preset window constraints;

[0079] Based on the time sequence of the operating condition information, dynamic programming calculations are performed pairwise on the operating condition information of different model types belonging to the same network group, according to window constraints. The similarity results between the operating condition information of different model types are obtained based on the dynamic programming calculation results. When it is detected that the similarity of the operating condition information of the initial simulation model is greater than the preset similarity threshold in each model type, the operating condition information of the initial simulation model corresponding to each model type is determined to be matched.

[0080] Specifically, a window constraint is first preset. In the standard Dynamic Time Warping (DTW) algorithm, all possible cross-sequence point pairs are considered, which may lead to high computational complexity and high redundancy. Therefore, the above-mentioned window constraint is introduced in this embodiment. The window constraint is obtained from the VI / VT curve. The length of the window constraint is a dynamic length. The constraint is based on the obtained length. Taking Table 1-1 as an example, according to the three-dimensional information shown in Table 1-1, the row number represents the constraint length of the window. The window length of signal models from different manufacturers will be inconsistent, but the window length of the same type is basically consistent.

[0081] Furthermore, based on the time sequence of the operating condition information and the aforementioned window constraints, dynamic programming calculations are performed pairwise between operating condition information of different model types belonging to the same network group. Specifically, this includes analyzing the x-dimensional matrix information of the operating condition information, filling in the dynamic programming table, and calculating all possible alignment paths and difference values ​​between two operating condition information according to the properties of optimal substructure and overlapping subproblems. Here, the alignment path refers to the data in the first column of Table 1-1 above, i.e., the voltage value; the difference value represents the similarity result between the operating condition information of each model type. When it is detected that the similarity between the operating condition information of the initial simulation model is greater than the similarity threshold in each model type, it is determined that the operating condition information of the initial simulation model corresponding to each model type matches. In summary, this embodiment uses the curve dynamic time warping algorithm to store intermediate results, avoiding repeated calculations and realizing the redundancy removal of the model.

[0082] In one embodiment, after selecting the corresponding initial simulation model from the default simulation model based on the basic characteristics of the signal and extracting the operating condition information of each initial simulation model, the method further includes:

[0083] Based on the operating condition information, the input and output categories of each initial simulation model are divided to obtain the input category model, the output category model, and the input-output category model.

[0084] For each of the input category model, output category model, and input-output category model, the corresponding target simulation model is selected from the default simulation models.

[0085] Specifically, after selecting the initial simulation models and before grouping them according to the preset communication protocol, this application needs to first divide the initial simulation models according to their input and output categories to obtain input category models, output category models, and input / output category models. In practical applications, the network signal corresponds to the transceiver chip model type, which mainly includes three types: input, output, and I / O. Input represents the receiving state, and its correlation curves are power clamping and ground clamping curves. That is, for input type models, attention should be paid to the operating condition information of its power clamping and ground clamping curves. Output represents the transmitting state, and its correlation curves are pull-up / pull-down curves and rising / falling edge curves. That is, for output type models, attention should be paid to the operating condition information of its pull-up / pull-down curves and rising / falling edge curves. Similarly, I / O has both transmitting and receiving states. When the model type is I / O, it is necessary to determine the input or output state according to the actual transmission type TT of the signal and then detect the corresponding correlation curves.

[0086] After classifying the input category model, output category model, and input-output category model, dynamic programming calculations are performed separately for each of these three types to match the corresponding working condition information, thereby obtaining the target simulation models mentioned above.

[0087] In this embodiment, before screening the target simulation model, it is preferable to classify the input signal, output signal, and input-output signal models. This is because the specific transmission properties of the signal cannot be accurately determined from the IBIS model. It is necessary to obtain the actual transmission type of the signal through database information to achieve the effect of being consistent with the actual physical transmission of the signal.

[0088] In one embodiment, determining the gradient strength of the corresponding feature index based on the margin space includes:

[0089] Iterate through the feature metrics of each dimension and calculate the gradient value of each dimension respectively;

[0090] Convert the gradient values ​​for each dimension into norm form to obtain the gradient strength for each dimension.

[0091] Specifically, the gradient of each feature index is obtained through gradient calculation, and the gradient calculation satisfies the following partial derivative relationship:

[0092]

[0093] Where x and y represent the horizontal and vertical coordinate parameters of each feature index, respectively, and z represents the normalized overall feature index parameter.

[0094] After calculating the gradient value for each dimension using the method described above, the gradient values ​​for each dimension are converted into norm form to obtain the gradient strength for each dimension:

[0095]

[0096] Among them, ||z|| 2i represents the L2 norm of each feature parameter, i.e. the gradient strength of each dimension, i represents the index of the feature parameter, and t represents the number of feature parameters.

[0097] In one embodiment, the gradient intensities corresponding to the link simulation parameter configuration information are sorted, and target gradient intensities are selected based on the gradient intensities, including:

[0098] The gradient intensities are sorted in ascending order, and the maximum value among the sorted gradient intensities is determined as the target gradient intensity.

[0099] Specifically, define the gradient strength ||z||' of the multi-feature index parameters. 2i For ||z|| 2i A sorted list, ordered from smallest to largest, and sorted by ||z||' 2i Create a new list L j , for L j Sort the data to obtain the optimal parameters. The list and the optimal parameter selection satisfy the following:

[0100]

[0101] Among them, L j This represents a list of gradient strengths for specific indicators under various configurations, where j is the index of the active device optimization parameters, cn represents the total types of active device optimization parameters, and L... opt For optimal parameter selection, choose the independent variable that has the largest value after sorting the list.

[0102] In this embodiment, the gradient with the largest gradient strength is determined as the target gradient strength, thereby quantifying the evaluation method of the simulation device and effectively improving the optimization accuracy and efficiency of the simulation device configuration.

[0103] In one embodiment, based on preset signal characteristics of the model to be simulated in the simulated link, a target simulation model corresponding to the simulated link is selected from the default simulation models, including:

[0104] Determine the basic signals corresponding to the model to be simulated;

[0105] Define the target signal set of the basic signal, where the target signal set includes the basic signal characteristics corresponding to the model to be simulated;

[0106] The target signal set is mapped and matched with the default signal characteristics corresponding to the default simulation model, and the target simulation model is obtained by filtering.

[0107] Specifically, establishing a simulation signal database based on material codes can effectively identify high-speed signals in design data, providing data support for automated simulation design. This database includes the default simulation model described in this embodiment. In electronic product design, the aforementioned basic signals, i.e., high-speed signals, can be identified through network signal naming. However, due to the lack of standardized naming criteria, it is difficult to accurately extract these basic signals. These basic signals are important signals of the model to be simulated, such as DQS0_N, DQS0_P, DQ0-DQ7, DMO, etc. In this embodiment, signal naming can be determined based on the pin numbers of the material model during load testing. The component part number and pins can be determined as fixed network information. For high-speed signals, the main control device, such as an ASIC or CPU, can be obtained through mapping its network relationship with the load device. The pin information of FPGAs and CPLDs, since they may have different uses in different projects, can be identified through the relationship between the FPGA and the load device.

[0108] Specifically, the simulation signal database defining the device materials mainly includes the following information, as shown in the example below:

[0109]

[0110] In this configuration, "New Part Number" represents the material code; "PIN Number" represents the pin number; "XNetName" represents the network name, used to indicate the original network name of the pin and the network name matching the actual project; "Net Type" represents the network type, which can be used to determine the subtype of the signal, and thus apply different electrical templates to judge the simulation results; "Edgesampling" represents edge sampling, which can be used to determine the actual operating speed of the signal; and "SignalType" represents the signal protocol type, which can be used to determine the signal type of this network. The design of these fields will vary depending on the simulation type. For different types of signal and power supply simulations, classification fields for active device models can be added. The above mainly provides an example of signal link parameter configuration.

[0111] "BUS NAME" represents the network group name, used for timing analysis of the network group in simulation. Since different signal types exist in the same network group, i.e., different signal models, consistency in the configuration of various different types of models needs to be ensured during actual simulation optimization and product configuration to guarantee timing consistency. "TransmissionType" represents the signal transmission type. This field determines the signal transmission method, is verified against the "Model_type" field in MI, and further classifies and identifies model features based on model input and output characteristics. "Level Voltage" represents the signal level voltage. This field determines the identification and filtering based on "[Voltage Range]" under multiple model conditions.

[0112] In summary, this embodiment determines the basic signals of the model to be simulated and thus the accurate name of the IBIS model corresponding to the model to be simulated, namely 1.2.03.01.03XXX in the table above, so as to further obtain the model information related to the IBIS model corresponding to the model to be simulated.

[0113] In this embodiment, the preset basic signals corresponding to the model to be simulated are first determined. This can be achieved by defining a set of basic signals and subsets, that is, by defining a target set of basic signals:

[0114]

[0115] Where S represents the set of high-frequency, high-speed network signals, i.e., the basic signal set mentioned above, k represents any network group, n is the total number of networks in the network group, and N k This represents the set of parameters for each high-frequency network signal. Sk represents a single high-speed network signal and its associated information. Skr is the associated device reference number, Skp is the associated device part number, Skn is the model name of the associated device, Skpp is the device pin information, Ntt represents the transmission type of the signal in the database association (including bidirectional and unidirectional), Nlv represents the voltage level of the network in the database association, and Nbn represents the signal group to which the network belongs. The basic characteristics of the above signals generally include the transmission type Ntt of the signal in the database association, the voltage level Nlv of the network in the database association, and the signal group Nbn to which the network belongs.

[0116] The target signal set and the default signal characteristics corresponding to the default simulation model are mapped and matched to extract the high-speed network signal set S. The information of the corresponding model is obtained through the network associated pin information, and the network-based model parameter information is defined as follows:

[0117]

[0118] Where STkpii represents the pin information of the model associated with the network, STkmi represents the model or model list associated with the network, T[Skp][Skn][Cms] corresponds to the model list, T[Skp][Skn][Cmd] corresponds to a single model, STkvr represents the voltage level information associated with the network, STkmt represents the transmit / receive characteristics associated with the network, and STmi represents the model list of all signals in the group to which the network belongs. The single network model is an unordered structure and there is no correlation between networks. The total number of network models is the product of the number of network models of each individual signal.

[0119] Mapping and matching are performed based on the fundamental and default signal characteristics Ntt and STkmt, Nlv and STkvr, and Nbn and STmi from the two sets mentioned above. This involves mapping and matching the signal transmission type, voltage level, and signal group of the network database data with the network-associated main control chip model, thereby filtering the target simulation model. In practical applications, since some IBIS model files are large and require a lot of information, a multi-threaded, block-based reading method can be used to obtain key voltage range information and achieve device model filtering based on signal level.

[0120] In summary, this embodiment first establishes a target signal set for basic signals (which can be obtained through design projects in practical applications) and establishes a target signal set corresponding to the basic signals. The target signal set includes preset basic signal characteristics (transmission type, voltage level, and signal group). Then, based on the pin information of the device corresponding to the network, a default simulation model is determined, along with all model information corresponding to the default simulation model. The target signal set is then compared with the default signal characteristics of the default simulation model (which are also the transmission type, voltage level, and signal group) to complete the selection of the target simulation model.

[0121] This application also provides a preferred embodiment of a parameter configuration method for communication link simulation. Figure 2 This is a flowchart illustrating a parameter configuration method in one embodiment.

[0122] Step S210: Establish the parameters of the default simulation model. In this embodiment, the default simulation model is the IBIS model. The IBIS model has a file format specification that standardizes the data format of each component. The specific process for establishing the IBIS model structure parameters is as follows: First, define the IBIS structure as a dictionary structure:

[0123]

[0124] Wherein, PartNum is the device material code, Component is the device attribute field information index of the IBIS model, CIi is the device information, which includes at least: device name CCN, package parameter CPAI (i.e., RLC parasitic parameter), pin information CPII, model selection information CMS, and model information CM; Model is the model field information index of the IBIS model, MIj is the model information, which includes at least: voltage level MVR, model transceiver characteristics MMT, C_comp parameter MCC, rising edge MRI, falling edge MFA, pull-up curve MPU, pull-down curve MPD, power clamping curve MGC, and ground clamping curve GC.

[0125] Specifically, these include:

[0126] i. RLC parasitic parameters: chip package electrical parameters

[0127] ii.Ccomp: Transistor junction capacitance

[0128] iii. Pull-up / Pull-down VI Curves: The VI curve represents the relationship between voltage and current. This curve involves two aspects. First, the magnitude of the transistor's drive current, which is determined by the chip's own circuitry and is provided in some manufacturers' datasheets. The other aspect is the load resistance setting. Because the pull-up curve has three values—maximum, typical, and minimum—it is based on the load resistance setting, which is also specified in some datasheets.

[0129] iv. Rising and Falling Edge Curves: The VT curve represents the relationship between voltage and time, showing the instant a signal changes from a low level to a high level. In digital circuits and communication systems, the rising edge generally indicates the process of a signal state changing from 0 (low level) to 1 (high level), which is the beginning edge of the signal. The falling edge curve is the opposite of the rising edge curve.

[0130] v. Power supply and ground clamping curve: The VI curve is the relationship between voltage and current, representing the voltage and current relationship between the power supply and ground of the clamping diode.

[0131] Step S220: Based on the basic signals corresponding to the model to be simulated, determine the target signal set of the basic signals, and map and match the basic signal set with the default signal features corresponding to the default simulation model to obtain the initial simulation model. The basic signal set mainly includes three characteristics: signal transmission type, voltage level, and signal group.

[0132] Step S230: The initial simulation model is divided into input category model, output category model, and input-output category model. Network groups are then formed for each of these three types of models. Pairwise matching of the operating condition information between the initial simulation models corresponding to each model type within the network group is performed. If a successful match of target operating condition information is detected, the target simulation model corresponding to the target operating condition information is determined based on the matching result. Before performing pairwise matching, if an excessive number of models of one or more types are detected, redundant models with the same configuration must be identified and removed.

[0133] Step S240: Determine the list of target simulation models after screening, and apply Cartesian formula to perform a direct product on the model configuration list to obtain the link simulation parameter configuration information. Specifically, since the signal transmission methods mainly include single-drive-single-load and single-drive-multiple-load structures, and the configuration methods between single drives and between single loads and multiple loads are consistent, and to ensure the adaptability of chip configurations across multiple projects, the configurations of various devices in multiple loads generally adopt the same configuration. In this embodiment, the simulation signal data information can be obtained through the load device part number, and the network signal type can be determined according to the pin signal type of the load device to determine the signal link transmission and reception method. The configuration of each active chip satisfies the Cartesian product formula, which satisfies:

[0134]

[0135] Where: Ai represents the ordered model of the transmitting device of the signal network group, and Bi represents the ordered model of the receiving device of the signal network group; when i=0, A0 represents the configuration of the main control chip during forward transmission, and B0 represents the configuration of the load chip during forward transmission; when i=1, A1 represents the configuration of the main control chip during reverse transmission, and B1 represents the configuration of the load chip during reverse transmission.

[0136] Step S250: Calculate the simulation waveform based on the above link simulation parameter configuration information. This simulation waveform is generally a time-domain simulation waveform, and the time-domain simulation waveform curve is as follows: Figure 3 As shown. Figure 3 The figure shows a simulation waveform of the target simulation model selected above and the corresponding link simulation parameter configuration information. As can be seen from the figure, it is a voltage value time series diagram, that is, the change of voltage value in the time dimension. Those skilled in the art will understand that calculating the simulation waveform corresponding to the target simulation model and the link simulation parameter configuration information is an existing technology. For example, the predicted simulation waveform can be obtained based on simulation software; or the output time domain waveform curve of the signal link can be calculated based on the voltage and current relationship of the target simulation model, that is, the simulation waveform mentioned above.

[0137] Step S260: Extract feature indices based on the above time-domain simulation waveform, and calculate the margin space of the feature indices. Among them, multi-dimensional feature indices are as follows: Si = [EP BT BER SNR …] i=1,2,...t;

[0138] Where: EP represents eye diagram features, BT represents bathtub curve features, BER represents bit error rate, and SNR represents signal-to-noise ratio features. S is a multi-dimensional parameter feature, which is mostly four-dimensional parameter data in practical applications. The feature indicators can be increased or decreased according to the actual working conditions. When it is necessary to comprehensively evaluate multiple feature indicators, a weighted average can be calculated.

[0139] Step S270: Perform gradient calculation on the margin space, sort the gradient strengths corresponding to the link simulation parameter configuration information, filter out the target gradient strengths, and determine the link simulation parameter configuration information corresponding to the target gradient strengths as the target parameter configurations.

[0140] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0141] Based on the same inventive concept, this application also provides a parameter configuration apparatus for implementing the parameter configuration method described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more parameter configuration apparatus embodiments provided below can be found in the limitations of the parameter configuration method described above, and will not be repeated here.

[0142] In one embodiment, such as Figure 4 As shown, a parameter configuration device is provided, including: an acquisition module 41, a calculation module 42, and a generation module 43, wherein:

[0143] The acquisition module 41 is used to select the target simulation model corresponding to the link to be simulated from the default simulation models based on the preset basic signal characteristics of the model to be simulated in the link to be simulated, and to simulate the link to be simulated based on the target simulation model to obtain the corresponding initial simulation results of the link.

[0144] The calculation module 42 is used to determine all parameter configuration information of the target simulation model corresponding to the initial simulation result of the link, and determine at least one set of link simulation parameter configuration information based on the parameter configuration information; extract the feature index of the link simulation parameter configuration information, calculate the margin space corresponding to the feature index based on the data of the feature index and the preset expected index, and determine the gradient strength of the corresponding feature index based on the margin space.

[0145] The generation module 43 is used to sort the gradient intensities corresponding to the link simulation parameter configuration information, filter out the target gradient intensities based on the sorting results, and determine the link simulation parameter configuration information corresponding to the target gradient intensities as the target parameter configuration.

[0146] Each module in the above parameter configuration device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0147] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 5 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores parameter configuration-related data. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements a parameter configuration method.

[0148] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0149] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0150] Based on the preset basic signal characteristics of the simulation model in the simulation link, a target simulation model corresponding to the simulation link is selected from the default simulation models, and the simulation link is simulated based on the target simulation model to obtain the corresponding initial simulation results of the link.

[0151] Determine all parameter configuration information of the target simulation model corresponding to the initial simulation results of the link, and determine at least one set of link simulation parameter configuration information based on the parameter configuration information;

[0152] Extract the feature indicators of the link simulation parameter configuration information, calculate the margin space corresponding to the feature indicators based on the data of the feature indicators and the preset expected indicators, and determine the gradient strength of the corresponding feature indicators according to the margin space.

[0153] The gradient intensities corresponding to the link simulation parameter configuration information are sorted, the target gradient intensities are selected based on the sorting results, and the link simulation parameter configuration information corresponding to the target gradient intensities is determined as the target parameter configuration.

[0154] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0155] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0156] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0157] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A parameter configuration method for communication link simulation, characterized in that, The method comprises: Based on the preset signal basic characteristics of the to-be-simulated model in the to-be-simulated link, a target simulation model corresponding to the to-be-simulated link is selected from a default simulation model, and the to-be-simulated link is simulated based on the target simulation model to obtain a corresponding link initial simulation result; including, based on the signal basic characteristics, selecting a corresponding initial simulation model from the default simulation model, and extracting the working condition information of each initial simulation model; the working condition information of the initial simulation model reflects the basic parameter information of the initial simulation model when working, and the working condition information is determined by the input-output curve of the initial simulation model; According to the preset communication protocol corresponding to the to-be-simulated link, each initial simulation model is grouped to obtain at least one network group; wherein the network group includes at least two model types; A preset window constraint is obtained; based on the window constraint, the working condition information of different model types belonging to the same network group is calculated by dynamic programming, and the similarity result between the working condition information of different model types is obtained according to the dynamic programming calculation result; when it is detected that the similarity of the working condition information of the initial simulation model between each model type is greater than a preset similarity threshold, it is determined that the working condition information of the initial simulation model corresponding to each model type is matched; If it is detected that a group of target working condition information in each model type is matched successfully, the target simulation model corresponding to the target working condition information is determined according to the matching result; All parameter configuration information of the target simulation model corresponding to the link initial simulation result is determined, and at least one group of link simulation parameter configuration information is determined according to the parameter configuration information; The feature index of the link simulation parameter configuration information is extracted, the margin space corresponding to the feature index is calculated based on the data of the feature index and a preset expected index, and the gradient intensity of the corresponding feature index is determined according to the margin space; The gradient intensity corresponding to the link simulation parameter configuration information is sorted, the target gradient intensity is selected based on the sorting result, and the link simulation parameter configuration information corresponding to the target gradient intensity is determined as the target parameter configuration.

2. The method of claim 1, wherein, After the corresponding initial simulation model is selected from the default simulation model based on the signal basic characteristics, and the working condition information of each initial simulation model is extracted, and before each initial simulation model is grouped according to the preset communication protocol corresponding to the to-be-simulated link, the method further comprises: Based on the working condition information, the input-output categories of each initial simulation model are divided to obtain input category models, output category models and input-output category models; After the division of the input category model, the output category model and the input-output category model is completed, dynamic programming calculation is performed on the input category model, the output category model and the input-output category model respectively, the matching of the corresponding working condition information is completed, and the corresponding target simulation model is determined.

3. The method of claim 1, wherein, The gradient intensity of the feature index corresponding to the margin space is determined, including: The gradient value of each dimension is calculated by traversing the feature index of each dimension; The gradient value corresponding to each dimension is converted into a norm form to obtain the gradient intensity corresponding to each dimension.

4. The method of claim 1, wherein, The gradient intensity corresponding to the link simulation parameter configuration information is sorted, and a target gradient intensity is screened out based on the gradient intensity, including: Each gradient intensity is sorted in descending order, and the maximum value in the sorted gradient intensity is determined as the target gradient intensity.

5. The method of claim 1, wherein, The target simulation model corresponding to the target link is screened out from the default simulation model based on the preset signal basic characteristics of the simulation model in the target link, including: The basic signal corresponding to the simulation model is determined; The target signal set of the basic signal is defined, wherein the target signal set includes the signal basic characteristics corresponding to the simulation model; The target signal set is mapped and matched with the default signal characteristics corresponding to the default simulation model, and the target simulation model is screened out.

6. A parameter configuration apparatus characterized by comprising: The device includes: The acquisition module is configured to screen out a target simulation model corresponding to the target link from a default simulation model based on preset signal basic characteristics of a simulation model in the target link, and simulate the target link based on the target simulation model to obtain a corresponding link initial simulation result. The signal basic characteristics are used to screen out corresponding initial simulation models from the default simulation model, and working condition information of each initial simulation model is extracted. The working condition information of the initial simulation model reflects the basic parameter information of the initial simulation model when it is working, and is determined by the input-output curve of the initial simulation model. According to the preset communication protocol corresponding to the target link, each initial simulation model is grouped to obtain at least one network group. Each network group includes at least two model types. A preset window constraint is obtained. Based on the window constraint, dynamic programming calculation is performed between each two of the working condition information of different model types belonging to the same network group in the time sequence of the working condition information. The similarity result between the working condition information of different model types is obtained according to the dynamic programming calculation result. When it is detected that the similarity of the working condition information of the initial simulation model between each model type is greater than a preset similarity threshold, it is determined that the working condition information of the initial simulation model corresponding to each model type is matched. If it is detected that a group of target working condition information in each model type is matched successfully, the target simulation model corresponding to the target working condition information is determined according to the matching result. The computing module is configured to determine all parameter configuration information of the target simulation model corresponding to the link initial simulation result, and determine at least one set of link simulation parameter configuration information according to the parameter configuration information; extract a feature index of the link simulation parameter configuration information, calculate a margin space corresponding to the feature index based on data of the feature index and a preset expected index, and determine a gradient intensity corresponding to the feature index according to the margin space; The generating module is configured to sort the gradient intensity corresponding to the link simulation parameter configuration information, filter out a target gradient intensity based on a sorting result, and determine the link simulation parameter configuration information corresponding to the target gradient intensity as a target parameter configuration. 7.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-6 when the computer program is executed by the processor. The processor executes the computer program to implement the steps of the method in any one of claims 1 to 5.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 5.

Images