Tap parameter adjustment method, device, medium and product for decision feedback equalization

Abstract

The present invention discloses a method, device, medium and product for adjusting the tap parameters of decision feedback equalization, which relates to the field of signal processing and is used to solve the problem that the tap parameters are not used or the tap parameters depend on the default values ​​of chip manufacturers. The method obtains the structural parameters of a printed circuit board, determines and simulates the topological model of the printed circuit board, and obtains simulation results including eye diagram parameters. According to these simulation results, the tap parameters of the decision feedback equalization are adjusted to determine the optimal tap parameters and optimize the data signal input to the printed circuit board. It can be seen that the present application can determine the optimal tap parameters through precise simulation and optimization adjustment, so that when the data signal is processed using the tap parameters, the bandwidth and quality of the signal can be improved to meet higher rate requirements.

Description

Technical Field

[0001] The present invention relates to the field of signal processing, and in particular to a method, device, medium and product for adjusting tap parameters of decision feedback equalization. Background Art

[0002] As a core component in modern digital systems, memory is widely used in consumer electronics, commercial industrial equipment, terminal products and data centers for data processing by the CPU (Central Processing Unit). The significant advances in bandwidth, performance and power consumption of each generation of DDR (Double Data Rate) technology have greatly promoted the improvement of computing performance. At present, DDR technology has developed to DDR5 with a higher rate, but the higher rate also brings greater channel attenuation. Compared with DDR4, DDR5 introduces DFE (Decision Feedback Equalization) in the architecture to overcome channel attenuation and reflection problems.

[0003] However, in actual applications, DDR5 signal tuning mainly relies on adjusting the driving resistance of the transmitting end and the ODT (On-Die Termination) parameters of the receiving end, and the DFE parameters are not effectively used; or the DFE parameters provided by the chip manufacturers are used. However, the DFE parameters provided by the chip manufacturers do not significantly improve the DDR5 signal quality. On the contrary, in some cases, they deteriorate the eye diagram, resulting in the DDR5 signal being unable to reach a higher rate, affecting the signal bandwidth. Summary of the invention

[0004] The purpose of the present invention is to provide a tap parameter adjustment method, device, medium and product for decision feedback equalization. Through precise simulation and optimization adjustment, the optimal tap parameters can be determined, so that when the tap parameters are used to process the data signal, the bandwidth and quality of the signal can be improved to adapt to higher rate requirements.

[0005] On the one hand, the present application provides a method for adjusting tap parameters of decision feedback equalization, including: obtaining structural parameters of a printed circuit board, and determining a simulation topology model of the printed circuit board based on the structural parameters; simulating the simulation topology model to determine simulation results, wherein the simulation results include eye diagram parameters; adjusting the tap parameters of the decision feedback equalization based on the simulation results to determine optimal tap parameters, wherein the optimal tap parameters are used to optimize processing of data signals input to the printed circuit board.

[0006] Among them, the tap parameters of the decision feedback equalizer are adjusted according to the simulation results to determine the optimal tap parameters, including: determining the influence weight of each tap parameter of the decision feedback equalizer on the eye diagram parameters; determining the corresponding preset adjustment method according to the influence weight of each tap parameter; adjusting each corresponding tap parameter according to the determined preset adjustment method to determine the optimal value of each tap parameter.

[0007] Among them, determining the influence weight of each tap parameter of the decision feedback equalization on the eye diagram parameter includes: for each current tap parameter, adjusting the current tap parameter within a preset range with a first preset step size, and controlling the remaining tap parameters except the current tap parameter to remain unchanged; whenever the current tap parameter is adjusted, simulating the simulation topology model according to the adjusted current tap parameter to obtain the corresponding test eye diagram parameter; determining the influence weight of the current tap parameter on the eye diagram parameter according to the first preset step size of the current tap parameter and the test eye diagram parameter.

[0008] Among them, determining the influence weight of the current tap parameter on the eye diagram parameter according to the first preset step size of the current tap parameter and the test eye diagram parameter includes: calculating the change amount between the test eye diagram parameter of the current adjustment of the current tap parameter and the test eye diagram parameter of the last adjustment of the current tap parameter; determining the influence weight of the current tap parameter on the eye diagram parameter according to the change amount and the first preset step size.

[0009] Among them, determining the influence weight of the current tap parameter on the eye diagram parameter according to the change amount and the first preset step size includes: summarizing the change amount of the test eye diagram parameter corresponding to each adjustment of the current tap parameter within the entire preset range; calculating the average change amount corresponding to the current tap parameter according to the change amount; and determining the influence weight of the current tap parameter on the eye diagram parameter according to the average change amount.

[0010] Among them, after determining the influence weight of the current tap parameter on the eye diagram parameter according to the first preset step size of the current tap parameter and the test eye diagram parameter, it also includes: sorting the influence weight according to the influence weight of each tap parameter on the eye diagram parameter; determining the corresponding preset adjustment method according to the influence weight of each tap parameter, including: determining the adjustment method of the first N tap parameters after the influence weight is sorted from large to small as the first adjustment method; determining the adjustment method of the last M tap parameters after the influence weight is sorted from large to small as the second adjustment method, wherein N and M are both integers, and the sum of N and M is the total number of the tap parameters.

[0011] Among them, the first adjustment method is: for each current tap parameter, the current tap parameter is adjusted within a preset range with a second preset step size, and the remaining tap parameters except the current tap parameter are controlled to remain unchanged; whenever the current tap parameter is adjusted, the simulation topology model is simulated according to the adjusted current tap parameter to obtain the corresponding first optimal eye diagram parameter; the optimal range of the current tap parameter is determined according to each of the first optimal eye diagram parameters; for each of the current tap parameters, the current tap parameter is adjusted within the optimal range with a third preset step size, and the remaining tap parameters except the current tap parameter are controlled to remain unchanged, and the third preset step size is smaller than the second preset step size; whenever the current tap parameter is adjusted, the simulation topology model is simulated according to the adjusted current tap parameter to obtain the corresponding second optimal eye diagram parameter; the optimal value of the current tap parameter is determined according to each of the second optimal eye diagram parameters.

[0012] Among them, the second adjustment method is: for each current tap parameter, the current tap parameter is adjusted within a preset range with a second preset step size, and the remaining tap parameters except the current tap parameter are controlled to remain unchanged; whenever the current tap parameter is adjusted, the simulation topology model is simulated according to the adjusted current tap parameter to obtain the corresponding first optimal eye diagram parameter; and the optimal value of the current tap parameter is determined according to each of the first optimal eye diagram parameters.

[0013] Wherein, obtaining the structural parameters of the printed circuit board includes: obtaining the scattering parameter of the longest line and the scattering parameter of the shortest line in the printed circuit board.

[0014] Among them, obtaining structural parameters of a printed circuit board and determining a simulation topology model of the printed circuit board according to the structural parameters include: obtaining a first scattering parameter of the printed circuit board; obtaining a second scattering parameter of other components associated with the printed circuit board, the other components including a central processing unit package, a central processing unit socket, a memory slot, and a memory module package; determining a third scattering parameter of an overall memory link according to the first scattering parameter and the second scattering parameter; the overall memory link includes the printed circuit board and the other components; and determining a passive simulation model of the overall memory link according to the third scattering parameter of the overall memory link.

[0015] Among them, after determining the passive simulation model of the overall memory link according to the third scattering parameter of the overall memory link, it also includes: acquiring the simulation model of the central processing unit and the simulation model of the dynamic random access memory particles; determining the active simulation model according to the simulation model of the central processing unit, the simulation model of the dynamic random access memory particles and the passive simulation model of the overall memory link.

[0016] The method further includes: setting the driving resistance of the central processing unit and the terminal resistance parameters of the dynamic random access memory particles to fixed values.

[0017] On the other hand, the present application provides a tap parameter adjustment device for decision feedback equalization, comprising: a memory for storing a computer program; a processor for implementing the steps of the tap parameter adjustment method for decision feedback equalization described above when executing the computer program.

[0018] On the other hand, the present application provides a non-volatile storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the tap parameter adjustment method for decision feedback equalization described above are implemented.

[0019] On the other hand, the present application provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the tap parameter adjustment method of the decision feedback equalization described above.

[0020] The present invention provides a method, device, medium and product for adjusting the tap parameters of decision feedback equalization, which relates to the field of signal processing and is used to solve the problem that the tap parameters are not used or the tap parameters depend on the default values ​​of the chip manufacturer. The method determines and simulates the topological model of the printed circuit board by obtaining the structural parameters of the printed circuit board to obtain simulation results including eye diagram parameters. According to these simulation results, the tap parameters of the decision feedback equalization are adjusted to determine the optimal tap parameters and optimize the data signal input to the printed circuit board. It can be seen that the present application can determine the optimal tap parameters through precise simulation and optimization adjustment, so that when the data signal is processed using the tap parameters, the bandwidth and quality of the signal can be improved to adapt to higher rate requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 A flow chart of a tap parameter adjustment method for decision feedback equalization provided by the present invention;

[0023] Figure 2 A schematic diagram of an original data signal waveform provided by the present invention;

[0024] Figure 3A schematic diagram of a data signal waveform after ISI processing provided by the present invention;

[0025] Figure 4 A schematic diagram of the data signal waveform after DFE processing provided in this application;

[0026] Figure 5 A schematic diagram of an active simulation model provided in this application;

[0027] Figure 6 Insertion loss diagram of scattering parameters of a passive simulation model provided in this application;

[0028] Figure 7 Insertion loss diagram of scattering parameters of another passive simulation model provided in this application;

[0029] Figure 8 A return loss diagram of scattering parameters of a passive simulation model provided in this application;

[0030] Fig. 9 A return loss diagram of scattering parameters of another passive simulation model provided in the present application;

[0031] Fig.10 This is the eye diagram of the DDR particle end when the original tap parameters are all 0;

[0032] Fig.11 This application provides an eye diagram after tap parameter optimization. DETAILED DESCRIPTION

[0033] The core of the present invention is to provide a tap parameter adjustment method, device, medium and product for decision feedback equalization. Through precise simulation and optimization adjustment, the optimal tap parameters can be determined, so that when the tap parameters are used to process the data signal, the bandwidth and quality of the signal can be improved to adapt to higher rate requirements.

[0034] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] On the one hand, if Figure 1 As shown, the present application provides a tap parameter adjustment method for decision feedback equalization, comprising:

[0036] S11: Acquire structural parameters of the printed circuit board, and determine a simulation topology model of the printed circuit board according to the structural parameters.

[0037] This step involves obtaining the structural parameters of the printed circuit board (PCB) and determining the simulation topology model based on these parameters to provide an accurate basis for subsequent simulation and optimization. Specifically, the structural parameters of the printed circuit board include the geometric layout, number of layers, material properties, trace length and width, impedance, via location and size of the PCB. These parameters are crucial to the accuracy of the simulation model because they directly affect the transmission characteristics and electromagnetic compatibility of the signal.

[0038] First, the geometric layout and number of layers determine the complexity and interference of the signal path. For example, a multi-layer PCB can reduce interference by separating the signal layer and the power layer, but it will also increase the length and complexity of the signal path. Material properties, such as dielectric constant and loss factor, affect the speed and attenuation of signal transmission. Trace length and width affect signal delay and reflection, and impedance mismatch can cause signal reflection and loss, thus affecting signal integrity. The location and size of the vias affect the integrity and transmission rate of the signal when switching between layers.

[0039] After determining these structural parameters, a simulation topology model is established. This model includes all traces, vias, pads, connectors and other components, and takes into account the signal path and interference sources in actual applications. Through this detailed modeling, the signal transmission process on the PCB can be simulated, and the signal delay, reflection, attenuation and other characteristics can be analyzed, thereby providing data support for the subsequent tap parameter adjustment.

[0040] In summary, this step ensures that the simulation topology model can truly reflect the actual signal transmission situation by obtaining and accurately modeling the structural parameters of the printed circuit board in detail, providing a solid foundation for subsequent simulation and optimization.

[0041] S12: Simulate the simulation topology model to determine simulation results, where the simulation results include eye diagram parameters.

[0042] This step simulates the simulation topology model to determine the simulation results including the eye diagram parameters, which is crucial for evaluating signal transmission performance and optimizing decision feedback equalization (DFE) parameters. Specifically, the simulation topology model is established based on the previously acquired printed circuit board (PCB) structural parameters, which can accurately reflect the signal path and transmission characteristics in the actual circuit. During the simulation process, electromagnetic simulation and time domain simulation are performed on the model to simulate the behavior of the signal at different frequencies. Eye diagram parameters, as an important indicator of the simulation results, provide an intuitive evaluation of signal integrity and transmission quality by displaying the superimposed signal waveform of multiple bit periods. Eye diagram parameters include eye height, eye width, jitter and noise tolerance, which can reveal the attenuation, distortion and interference of the signal during transmission. By analyzing the eye diagram, the bottlenecks and problem areas in the signal path can be determined, providing data support and optimization direction for the subsequent DFE tap parameter adjustment. In addition, the simulation results can also reveal the dynamic behavior of signal transmission under different frequencies and load conditions, helping engineers evaluate the robustness and reliability of the design.

[0043] In summary, this step obtains key eye diagram parameters and comprehensively evaluates signal transmission performance by performing detailed simulation and analysis on the simulation topology model, providing a scientific basis and guidance for optimizing DFE parameters and improving signal quality.

[0044] S13: adjusting the tap parameters of the decision feedback equalization according to the simulation results to determine the optimal tap parameters, and the optimal tap parameters are used to optimize the data signal input to the printed circuit board.

[0045] In this step, the tap parameters of the decision feedback equalization (DFE) are adjusted according to the simulation results to determine the optimal tap parameters, thereby optimizing the data signal input to the printed circuit board. This process first relies on the simulation results obtained in the previous step (S12), especially the eye diagram parameters, which provide key performance indicators of signal transmission, such as eye height, eye width, jitter and noise tolerance. Based on these eye diagram parameters, the effectiveness and shortcomings of the current DFE tap parameter settings can be evaluated. The purpose of adjusting the DFE tap parameters is to reduce signal interference and distortion under different signal conditions using the decision feedback mechanism, thereby improving signal quality. Specifically, the DFE system eliminates inter-symbol interference (ISI) in the signal by using the data decision feedback of the first few bits. For example, during the adjustment process, the engineer will gradually change the value of each tap parameter within the preset parameter range with a certain step size, while keeping other tap parameters unchanged, and analyze the changes in the eye diagram parameters after each adjustment through simulation. Through this iterative optimization, the specific impact of each tap parameter on the signal quality is gradually determined, and the average change and weight of the eye diagram parameter are calculated. Ultimately, engineers will combine the optimization results of all tap parameters to determine the optimal DFE tap parameter configuration, which can maximize signal quality, reduce bit error rate, and ensure the stability and integrity of data signals during high-frequency transmission.

[0046] like Figure 2 As shown, Figure 2 The waveform of the original data signal is shown in FIG. Figure 3 The waveform of the data signal with ISI is shown in FIG. Figure 4 The waveform of the data signal after DFE correction is shown. Figure 4 It can be seen that DFE mainly reduces the tail and makes the level at the tail approach 0 level.

[0047] The present invention provides a method for adjusting the tap parameters of decision feedback equalization, which is used to solve the current problem of unused tap parameters or tap parameters relying on the default values ​​of chip manufacturers. The method determines and simulates the topological model of the printed circuit board by obtaining the structural parameters of the printed circuit board to obtain simulation results including eye diagram parameters. According to these simulation results, the tap parameters of the decision feedback equalization are adjusted to determine the optimal tap parameters and optimize the data signal input to the printed circuit board. It can be seen that the present application can determine the optimal tap parameters through precise simulation and optimization adjustment, so that when the data signal is processed using the tap parameters, the bandwidth and quality of the signal can be improved to meet higher rate requirements.

[0048] Based on the above embodiments:

[0049] In one embodiment, the tap parameters of the decision feedback equalizer are adjusted according to the simulation results to determine the optimal tap parameters, including: determining the influence weight of each tap parameter of the decision feedback equalizer on the eye diagram parameter; determining the corresponding preset adjustment method according to the influence weight of each tap parameter; adjusting each corresponding tap parameter according to the determined preset adjustment method to determine the optimal value of each tap parameter.

[0050] In one embodiment, the process of adjusting the tap parameters of a decision feedback equalization (DFE) according to simulation results to determine the optimal tap parameters includes several key steps. First, the effectiveness of the current DFE settings is evaluated through the simulation results obtained in the previous step (S12), especially the eye diagram parameters such as eye height, eye width, jitter and noise margin. Then, the influence weight of each tap parameter on these eye diagram parameters is determined. This process usually involves gradually adjusting each tap parameter and observing the influence of these adjustments on the eye diagram parameters. The average change of each tap parameter is calculated through multiple simulations to quantify its influence on signal quality. Then, according to the influence weight of each tap parameter, the corresponding preset adjustment method is determined. This means that different adjustment methods are set according to the importance of each parameter to the signal quality, so as to make more targeted adjustments during the optimization process. Finally, according to the determined preset adjustment method, each tap parameter is adjusted to find its optimal value. This step is usually an iterative process, with simulations performed after each adjustment, changes in eye diagram parameters observed after the adjustment, and continuous fine-tuning until the optimal setting for each tap parameter is found, thereby maximizing signal quality, reducing bit error rates, and ensuring the stability and integrity of data signals during high-frequency transmission. Through this systematic optimization method, DFE tap parameters can be accurately adjusted to significantly improve the transmission performance of high-speed signals such as DDR5, meeting higher rate and bandwidth requirements.

[0051] In one embodiment, the influence weight of each tap parameter of decision feedback equalization on the eye diagram parameter is determined, including: for each current tap parameter, the current tap parameter is adjusted within a preset range with a first preset step size, and the remaining tap parameters except the current tap parameter are controlled to remain unchanged; each time the current tap parameter is adjusted, the simulation topology model is simulated according to the adjusted current tap parameter to obtain the corresponding test eye diagram parameter, and this step is repeated until the preset range of the current tap parameter is covered; the influence weight of the current tap parameter on the eye diagram parameter is determined according to the first preset step size of the current tap parameter and the test eye diagram parameter.

[0052] In this embodiment, when determining the weight of each tap parameter of the decision feedback equalization (DFE) on the eye diagram parameter, it is first necessary to set a preset range and a first preset step size for each tap parameter. These parameter ranges are usually determined based on technical specifications or previous empirical data to ensure that various possible parameter settings can be covered during the simulation process. For each current tap parameter, adjustments are made according to the preset step size while keeping other tap parameters fixed, so that the variables can be controlled and the independent impact of each parameter on the eye diagram parameter can be accurately measured.

[0053] Whenever the current tap parameters are adjusted, the simulation topology model must be re-simulated. During the simulation process, the transmission of the data signal in the designed PCB structure is simulated to obtain new eye diagram parameters. Eye diagram parameters include eye height, eye width, jitter, etc. These parameters directly reflect the quality and stability of the data signal.

[0054] Repeat the above steps until the entire preset range of the current tap parameter is covered, and collect and record the eye diagram parameter data obtained from each simulation. Then, according to the first preset step size of each current tap parameter and the change of the test eye diagram parameter, calculate and determine the influence weight of each tap parameter on the eye diagram parameter. This process usually includes statistical analysis of the average change of the eye diagram parameter after each parameter is adjusted, or weighted calculation based on its contribution to the change of the eye diagram parameter to determine the final influence weight.

[0055] For example, the value range is between -1 and 1. The tap parameter tap mainly includes tap1, tap2, tap3 and tap4. In order to determine the weights of the four taps, the following case simulations are established to determine the impact of each tap on the DDR5 signal eye diagram.

[0056] Case 1: The values ​​of tap2, tap3, and tap4 are fixed to 0, and the value of tap1 will be traversed: -1~1, with a step size of 0.4 (that is, the first preset step size).

[0057] Case 2: Other parameter settings are consistent with case 1. Only DEF tap is modified. The values ​​of tap1, tap3, and tap4 are fixed to 0. The value of tap2 is traversed from -1 to 1 with a step size of 0.4.

[0058] Case 3: Other parameter settings are consistent with case 1. Only DEF tap is modified. The values ​​of tap1, tap2, and tap4 are fixed to 0. The value of tap3 is traversed from -1 to 1 with a step size of 0.4.

[0059] Case 4: Other parameter settings are consistent with case 1, only DEF tap is modified, tap1, tap2, and tap3 are fixed to 0, and tap4 is traversed from -1 to 1 with a step size of 0.4.

[0060] In the above experiment, only the DFE tap is changed, and other conditions remain the same, ensuring the rigor of the experiment. Through the above experiment, the impact of each tap change on the DDR5 eye diagram is obtained, and the weight is established to facilitate subsequent precise optimization.

[0061] Through the above detailed steps and precise simulation analysis, engineers can fully understand the specific impact of each DFE tap parameter on signal quality under different settings, so as to accurately adjust and optimize these parameters to ensure the best eye diagram performance and signal stability in high-speed data transmission.

[0062] In one embodiment, the weight of the influence of the current tap parameter on the eye diagram parameter is determined based on the first preset step size of the current tap parameter and the test eye diagram parameter, including: calculating the change between the test eye diagram parameter of the current adjustment of the current tap parameter and the test eye diagram parameter of the last adjustment of the current tap parameter; determining the weight of the influence of the current tap parameter on the eye diagram parameter according to the change amount and the first preset step size. In one embodiment, the weight of the influence of the current tap parameter on the eye diagram parameter is determined based on the change amount and the first preset step size, including: summarizing the change amount of the test eye diagram parameter corresponding to each adjustment of the current tap parameter within the entire preset range; calculating the average change amount corresponding to the current tap parameter based on the change amount; determining the weight of the influence of the current tap parameter on the eye diagram parameter based on the average change amount.

[0063] In an embodiment, the process of determining the influence weight of the current tap parameter on the eye diagram parameter according to the first preset step size of the current tap parameter and the test eye diagram parameter can be divided into several key steps. First, for each current tap parameter, a preset step size is set, that is, the size of the parameter is changed in each adjustment. This step size is determined according to the accuracy requirements of the simulation and the expected parameter variation range.

[0064] Secondly, electromagnetic simulation is performed on the designed printed circuit board to obtain the test eye diagram parameters after each adjustment of the current tap parameters. The eye diagram parameters reflect the transmission quality of the data signal, including indicators such as eye height, eye width, and jitter.

[0065] Next, calculate the change between the test eye diagram parameters after each adjustment of the current tap parameters and the eye diagram parameters at the last adjustment. These changes reflect the specific impact of different settings of the current tap parameters on the eye diagram parameters. Summarize these changes to obtain the overall change of the test eye diagram parameters corresponding to each adjustment of the current tap parameters within the entire preset range.

[0066] Then, by counting and analyzing these changes, the average change of the current tap parameter is calculated. The average change is the weighted average of the changes in the eye diagram parameters during all adjustment processes, which can quantify the overall impact of the current tap parameter on the signal quality.

[0067] Finally, the weight of the impact of the current tap parameter on the eye diagram parameter is determined based on the average change. This weight reflects the relative importance of the current tap parameter to the eye diagram parameter, taking into account the parameter changes within the entire preset range. Through this precise analysis method, engineers can adjust the DFE tap parameters in a targeted manner, optimize signal quality, and ensure optimal performance and stability in high-speed data transmission.

[0068] In one embodiment, after determining the influence weight of the current tap parameter on the eye diagram parameter according to the first preset step size of the current tap parameter and the test eye diagram parameter, it also includes: sorting the influence weight according to the influence weight of each tap parameter on the eye diagram parameter; determining the corresponding preset adjustment method according to the influence weight of each tap parameter, including: determining the adjustment method of the first N tap parameters after the influence weights are sorted from large to small as the first adjustment method.

[0069] The adjustment method of the last M tap parameters after the influence weights are sorted from large to small is determined as the second adjustment method, wherein N and M are both integers, and the sum of N and M is the total number of tap parameters.

[0070] In an embodiment, after determining the influence weight of the current tap parameter on the eye diagram parameter, a further step includes sorting these influence weights. This sorting process is to determine which tap parameters should be given priority when adjusting and which parameters can be processed later. First, the influence weights of the eye diagram parameters are sorted according to each tap parameter. The tap parameters with larger influence weights are ranked in front. These parameters have a greater impact on the eye diagram parameters and need to be adjusted first to achieve the best signal quality. Secondly, according to the sorting result, the adjustment method of the first N tap parameters after the influence weights are sorted from large to small is determined as the first adjustment method. This means that on these important tap parameters, engineers will use more sophisticated and frequent adjustments to ensure the optimal performance of the eye diagram parameters under different settings. At the same time, the adjustment method of the last M tap parameters after the influence weights are sorted from large to small is determined as the second adjustment method. These parameters have relatively small effects on the eye diagram parameters, so the adjustment can be more flexible or performed over a longer time interval to reduce the workload or improve the stability of the system.

[0071] N and M are determined by: selecting the values ​​of N and M based on past experience with similar projects; or by performing a series of simulation experiments to evaluate the impact of different combinations of N and M on performance indicators (such as eye diagram parameters), and selecting the best N and M values ​​based on simulation results and performance evaluation; or by building a statistical model or machine learning model to predict the impact of different combinations of N and M on performance indicators. The model can be trained based on historical data and simulation results to provide recommended N and M values.

[0072] In summary, through this method of sorting and hierarchical adjustment, engineers can formulate reasonable adjustment strategies based on the actual impact of each tap parameter on the eye diagram parameters, thereby effectively optimizing the transmission quality and stability of the data signal.

[0073] In one embodiment, the first adjustment method is: for each current tap parameter, the current tap parameter is adjusted within a preset range with a second preset step size, and the remaining tap parameters except the current tap parameter are controlled to remain unchanged; whenever the current tap parameter is adjusted, the simulation topology model is simulated according to the adjusted current tap parameter to obtain the corresponding first optimal eye diagram parameter; the optimal range of the current tap parameter is determined according to each first optimal eye diagram parameter; for each current tap parameter, the current tap parameter is adjusted within the optimal range with a third preset step size, and the remaining tap parameters except the current tap parameter are controlled to remain unchanged, and the third preset step size is smaller than the second preset step size; whenever the current tap parameter is adjusted, the simulation topology model is simulated according to the adjusted current tap parameter to obtain the corresponding second optimal eye diagram parameter; the optimal value of the current tap parameter is determined according to each second optimal eye diagram parameter.

[0074] In one embodiment, the second adjustment method is: for each current tap parameter, the current tap parameter is adjusted within a preset range with a second preset step size, and the remaining tap parameters except the current tap parameter are controlled to remain unchanged; whenever the current tap parameter is adjusted, the simulation topology model is simulated according to the adjusted current tap parameter to obtain the corresponding first optimal eye diagram parameter; the optimal value of the current tap parameter is determined according to each first optimal eye diagram parameter.

[0075] In both embodiments, the first adjustment method and the second adjustment method are intended to determine the influence of each tap parameter of decision feedback equalization (DFE) on the eye diagram parameters and further optimize these parameters to improve the quality and stability of the data signal.

[0076] The first adjustment method: First, for each current tap parameter, fine-tune it within the preset range with a second preset step size, while keeping other tap parameters unchanged. In this way, the influence range of each parameter can be covered and explored to determine the performance of the eye diagram parameters under different settings. After each adjustment, the current tap parameter is simulated using the simulation topology model, and the corresponding first optimal eye diagram parameter is obtained. Repeat this process until all possibilities within the preset range are covered. Then, based on each obtained first optimal eye diagram parameter, determine the optimal range of each tap parameter. Next, within the optimal range, use a finer third preset step size to further adjust the current tap parameter. Similarly, other tap parameters remain unchanged to explore the impact of each parameter in more detail. After simulation, the second optimal eye diagram parameter is obtained, and this process is repeated again until the optimal value of each tap parameter is determined.

[0077] Second adjustment method: The second adjustment method is similar to the first adjustment method in basic logic, but different in details. Also, each current tap parameter is adjusted within the preset range with the second preset step size, keeping other parameters unchanged. By simulating the adjusted tap parameters, the corresponding first optimal eye diagram parameters are obtained. The difference is that the second adjustment method directly determines the optimal value of the current tap parameter based on each first optimal eye diagram parameter, without the need for further refinement within the range.

[0078] The first adjustment method is suitable for scenarios that require a more precise and comprehensive understanding of the impact of each tap parameter on the eye diagram parameters. Through two stages of adjustment (i.e., coarse adjustment and fine adjustment), parameters can be systematically explored and optimized to ensure the best performance under all possible settings. This approach is very effective for complex signal optimization tasks or applications that require a deep understanding of the impact of each parameter.

[0079] The second adjustment method corresponds to a tap parameter with a smaller weight, so in order to improve efficiency, a coarse adjustment is performed directly, and the optimal value can be determined in a short time.

[0080] For example, the influence weight ranking obtained above is tap1>tap2>tap3>tap4.

[0081] Tune tap1: The weight of tap1 is relatively large, so first make a rough adjustment: fix the parameters of tap2, tap3, and tap4 to 0, and traverse tap1: -1~1, with a step size of 0.2. After finding the optimal parameters, narrow the range and make fine adjustments: the step size is 0.02.

[0082] After finding the optimal parameters of tap1, optimize the parameters of tap2: tap2 also has a relatively high weight, and the optimization method is the same as tap1, as follows: fix the tap1 parameters to the optimal parameters found above, fix tap3 and tap4 to 0, and make a rough adjustment for tap2: -1~1, with a step size of 0.2. After finding the optimal parameters, narrow the range and make fine adjustments: the step size is 0.02.

[0083] Next, we optimize tap3. Fix tap1 and tap2 as the optimal parameters found, fix tap4 to 0, and optimize tap3. The tap3 parameter weight is small, the coarse adjustment parameter can be -1~1, and the step size is 0.4.

[0084] The optimization method for tap4 is the same as tap3. Fix tap1, tap2, and tap3 as the optimal parameters found, and optimize tap4. The tap4 parameter weight is small, and the coarse adjustment parameter can be -1~1 with a step size of 0.4.

[0085] Finally, after determining the small range of parameters for each tap, adaptive optimization is performed.

[0086] In one embodiment, obtaining the structural parameters of the printed circuit board includes: obtaining the scattering parameter of the longest line and the scattering parameter of the shortest line in the printed circuit board.

[0087] In an embodiment, the process of obtaining the structural parameters of a printed circuit board involves two key steps: obtaining the scattering parameters of the longest routing and obtaining the scattering parameters of the shortest routing. First, for the routing of the printed circuit board, the design team will select the longest and shortest routings as key investigation objects. These routings are usually regarded as extreme cases, and their scattering parameters have a significant impact on the signal transmission performance of the entire circuit board. Long routings usually introduce greater signal attenuation and delay, while short routings may cause signal reflection and interference problems. Secondly, the process of obtaining scattering parameters involves measuring or simulating these selected routings using test equipment. Measuring scattering parameters includes complex transmission line characteristics, such as impedance, reflection coefficient, transmission loss, etc. of the transmission line. These parameters depend not only on the physical properties of the routing, such as width, thickness, and material, but are also closely related to factors such as the surrounding environment and grounding method.

[0088] By comprehensively utilizing the scattering parameters of the longest and shortest traces, the design team can comprehensively evaluate the signal integrity and transmission performance of the circuit board, which helps optimize the design and ensure stable signal transmission. The acquisition of these scattering parameters is an important step in the design process, providing basic data and reference for subsequent simulation analysis and circuit board optimization.

[0089] like Figure 5As shown, in one embodiment, the structural parameters of the printed circuit board are obtained, and the simulation topology model of the printed circuit board is determined according to the structural parameters, including: obtaining the first scattering parameter of the printed circuit board; obtaining the second scattering parameter of other components associated with the printed circuit board, and the other components include the central processing unit package, the central processing unit socket, the memory slot, and the memory module package; determining the third scattering parameter of the overall memory link according to the first scattering parameter and the second scattering parameter; the overall memory link includes the printed circuit board and other components; and determining the passive simulation model of the overall memory link according to the third scattering parameter of the overall memory link. In one embodiment, after determining the passive simulation schematic diagram of the overall memory link according to the third scattering parameter of the overall memory link, it also includes: obtaining the simulation model of the central processing unit and the simulation model of the dynamic random access memory particle; determining the active simulation model according to the simulation model of the central processing unit, the simulation model of the dynamic random access memory particle, and the passive simulation model of the overall memory link.

[0090] In this embodiment, the structural parameters of the printed circuit board and the related scattering parameters are obtained in order to establish a complete simulation model of the overall memory link. The specific steps are as follows. First, the structural parameters of the printed circuit board are obtained, which include key physical properties such as the length, width, and thickness of the traces on the printed circuit board; these parameters have a direct impact on the transmission characteristics of the signal, and therefore are the basis for establishing the simulation model. Secondly, the scattering parameters of other components directly associated with the printed circuit board are obtained, including the central processing unit package, the central processing unit socket, the memory slot, and the memory module package. The scattering parameters of each component reflect its impact on signal transmission, such as reflection, attenuation, interference, etc. These parameters are obtained in order to comprehensively consider the impact of each component on the signal integrity in the entire system. Next, based on the first scattering parameter of the printed circuit board and the second scattering parameter of other components, the third scattering parameter of the overall memory link is determined. This step will consider the comprehensive impact of all components to form the comprehensive scattering parameter of the overall memory link, that is, the signal transmission characteristics of the entire system. Subsequently, the passive simulation model of the overall memory link is determined. The passive simulation model is mainly used to analyze the transmission of signals between different components, including basic characteristics such as impedance matching of the transmission line and transmission loss of the signal. As Figure 6 , Figure 7 , Figure 8 , Fig. 9 As shown in the figure, the insertion loss diagram and return loss diagram corresponding to the scattering parameters under the passive simulation model are given. Specifically, Figure 6 In the figure, at m4 and frequency freq=2.100GHz, the dB of the scattering parameter S(1,13) is -7.922, and the dB of the scattering parameter S(2,14) is -8.266. Figure 7In the figure, at m3 and frequency freq=2.100GHz, the dB of the scattering parameter S(1,25) is -8.057, and the dB of the scattering parameter S(2,26) is -8.405. Figure 8 In the figure, at m2 and frequency freq=2.100GHz, the dB of the scattering parameter S(1,1) is -10.039, and the dB of the scattering parameter S(2,2) is -9.481. Fig.10 In the figure, at m1, frequency freq=2.100GHz, the dB of the scattering parameter S(13,14) is -8.753, and the dB of the scattering parameter S(25,25) is -8.206. Finally, based on the simulation models of the CPU and DRAM particles and the established passive simulation model of the overall memory link, the active simulation model is determined. The active simulation model takes into account the influence of actual circuits such as the driving circuit and feedback circuit, and is used to more accurately evaluate the actual transmission performance of the signal and optimize the design.

[0091] Through the above steps, all physical parameters and electrical characteristics of the printed circuit board and its related components can be comprehensively considered to establish a complete simulation model, so as to perform system-level signal integrity analysis and optimization design. These tasks are crucial to ensure the stable operation and performance optimization of high-speed digital systems such as memory.

[0092] In one embodiment, the method further includes: setting the drive resistance of the central processing unit and the terminal resistance parameters of the dynamic random access memory particles to fixed values.

[0093] In this embodiment, in addition to establishing a complete memory overall link simulation model, it also includes setting the drive resistance of the central processing unit and the terminal resistance parameters of the dynamic random access memory particles to fixed values. The purpose of this step is to regard these resistance parameters as constant during the simulation process, so that when analyzing and optimizing the simulation results, it is possible to focus on the adjustment and optimization of other influencing factors.

[0094] Specifically, the driving resistance of the CPU refers to the resistance value at the output of the CPU when sending signals to the printed circuit board and related components. This value affects the transmission rate, reflection and interference of the signal. Setting it to a fixed value means that this resistance value will not be changed during the simulation process, thus maintaining consistency. The terminal resistance parameter of the dynamic random access memory particle refers to the terminal resistance setting of the memory particle when receiving the signal. This parameter affects the transmission quality and stability of the signal in the memory particle. Similarly, setting it to a fixed value can simplify the simulation process, making it easier to control variables during analysis.

[0095] By setting these resistor parameters to fixed values, simulation engineers can focus more on adjusting and optimizing other important design parameters, such as the tap parameters of decision feedback equalization and the layout of signal lines, to ensure the signal integrity and performance of the entire system are optimal.

[0096] like Fig.10 and Fig.11 As shown, Fig.10 This is the eye diagram at the DDR particle end when the original tap parameters are all 0. Fig.11 This application provides the eye diagram after tap parameter optimization. The eye diagrams are all the results under bit error BER=1E-16. From the figure, it can be seen that the DFE tap with the optimal parameters can optimize the eye diagram, increase the eye height EH by 60mV, and increase the eye width by 10.6ps.

[0097] On the other hand, the present application provides a tap parameter adjustment device for decision feedback equalization, comprising: a memory for storing a computer program; a processor for implementing the steps of the tap parameter adjustment method for decision feedback equalization when executing the computer program. For other introductions to the tap parameter adjustment device for decision feedback equalization, please refer to the above embodiments, which will not be elaborated in this application.

[0098] On the other hand, the present application provides a non-volatile storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the tap parameter adjustment method of the decision feedback equalization are implemented. For other introductions to the non-volatile storage medium, please refer to the above embodiments, and the present application will not repeat them here.

[0099] On the other hand, the present application provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the tap parameter adjustment method for decision feedback equalization. For other introductions to the computer program product, please refer to the above embodiments, which will not be repeated in this application.

[0100] It should also be noted that, in this specification, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises", "comprising" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the statement "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0101] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A tap parameter adjustment method for decision feedback equalization, characterized in that: include: Acquiring structural parameters of a printed circuit board, and determining a simulation topology model of the printed circuit board according to the structural parameters; Simulating the simulation topology model to determine simulation results, wherein the simulation results include eye diagram parameters; Adjusting the tap parameters of the decision feedback equalization according to the simulation results to determine the optimal tap parameters, wherein the optimal tap parameters are used to optimize the data signal input to the printed circuit board; Adjusting the tap parameters of the decision feedback equalization according to the simulation results to determine the optimal tap parameters includes: For each current tap parameter, adjusting the current tap parameter with a first preset step size within a preset range, and controlling the remaining tap parameters except the current tap parameter to remain unchanged; Whenever the current tap parameters are adjusted, the simulation topology model is simulated according to the adjusted current tap parameters to obtain corresponding test eye diagram parameters; Calculating the change between the test eye diagram parameter of the current tap parameter being adjusted and the test eye diagram parameter of the last time the current tap parameter was adjusted; Determine the influence weight of the current tap parameter on the eye diagram parameter according to the variation and the first preset step size; Determine a corresponding preset adjustment method according to the influence weight of each of the tap parameters; Each corresponding tap parameter is adjusted according to the determined preset adjustment method to determine an optimal value of each tap parameter.

2. The tap parameter adjustment method for decision feedback equalization according to claim 1, characterized in that: Determining the influence weight of the current tap parameter on the eye diagram parameter according to the variation and the first preset step size includes: Summarize the change amount of the test eye diagram parameter corresponding to each adjustment of the current tap parameter within the entire preset range; Calculate the average change amount corresponding to the current tap parameter according to the change amount; The influence weight of the current tap parameter on the eye diagram parameter is determined according to the average change amount.

3. The tap parameter adjustment method for decision feedback equalization according to claim 1, characterized in that: After determining the influence weight of the current tap parameter on the eye diagram parameter according to the first preset step size of the current tap parameter and the test eye diagram parameter, the method further includes: Sorting the influence weights according to the influence weight of each tap parameter on the eye diagram parameter; Determining a corresponding preset adjustment method according to the influence weight of each of the tap parameters includes: Determine the adjustment modes of the first N tap parameters after the influence weights are sorted from large to small as the first adjustment mode; The adjustment method of the last M tap parameters after sorting the influence weights from large to small is determined as the second adjustment method, wherein N and M are both integers, and the sum of N and M is the total number of the tap parameters.

4. The tap parameter adjustment method for decision feedback equalization according to claim 3, characterized in that: The first adjustment method is: For each current tap parameter, adjusting the current tap parameter within a preset range with a second preset step size, and controlling the remaining tap parameters except the current tap parameter to remain unchanged; Whenever the current tap parameter is adjusted, the simulation topology model is simulated according to the adjusted current tap parameter to obtain the corresponding first optimal eye diagram parameter; Determining the optimal range of the current tap parameter according to each of the first optimal eye diagram parameters; For each of the current tap parameters, adjusting the current tap parameter with a third preset step size within the optimal range, controlling the remaining tap parameters except the current tap parameter to remain unchanged, and the third preset step size is smaller than the second preset step size; Whenever the current tap parameter is adjusted, the simulation topology model is simulated according to the adjusted current tap parameter to obtain a corresponding second optimal eye diagram parameter; The optimal value of the current tap parameter is determined according to each of the second optimal eye diagram parameters.

5. The tap parameter adjustment method for decision feedback equalization according to claim 3, characterized in that: The second adjustment method is: For each current tap parameter, adjusting the current tap parameter within a preset range with a second preset step size, and controlling the remaining tap parameters except the current tap parameter to remain unchanged; Whenever the current tap parameter is adjusted, the simulation topology model is simulated according to the adjusted current tap parameter to obtain the corresponding first optimal eye diagram parameter; The optimal value of the current tap parameter is determined according to each of the first optimal eye diagram parameters.

6. The tap parameter adjustment method for decision feedback equalization according to claim 1, characterized in that: Get the structural parameters of the printed circuit board, including: A scattering parameter of the longest line and a scattering parameter of the shortest line in the printed circuit board are obtained.

7. The tap parameter adjustment method for decision feedback equalization according to any one of claims 1 to 6, characterized in that: Acquiring structural parameters of a printed circuit board, and determining a simulation topology model of the printed circuit board according to the structural parameters, including: Acquiring a first scattering parameter of the printed circuit board; Acquire second scattering parameters of other components associated with the printed circuit board, the other components including a central processing unit package, a central processing unit socket, a memory slot, and a memory module package; Determine a third scattering parameter of a memory overall link according to the first scattering parameter and the second scattering parameter; the memory overall link includes the printed circuit board and the other components; A passive simulation model of the memory overall link is determined according to a third scattering parameter of the memory overall link.

8. The tap parameter adjustment method for decision feedback equalization according to claim 7, characterized in that: After determining the passive simulation model of the memory overall link according to the third scattering parameter of the memory overall link, the method further includes: Obtaining a simulation model of a central processing unit and a simulation model of a dynamic random access memory particle; The active simulation model is determined according to the simulation model of the central processing unit, the simulation model of the dynamic random access memory particle, and the passive simulation model of the overall memory link.

9. The tap parameter adjustment method for decision feedback equalization according to claim 8, characterized in that: Also includes: The driving resistance of the central processing unit and the terminal resistance parameters of the dynamic random access memory particles are set to fixed values.

10. A tap parameter adjustment device for decision feedback equalization, characterized in that: include: Memory for storing computer programs; A processor, used to implement the steps of the tap parameter adjustment method for decision feedback equalization as described in any one of claims 1 to 9 when executing a computer program.

11. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a computer program, and when the computer program is executed by the processor, the steps of the tap parameter adjustment method for decision feedback equalization according to any one of claims 1 to 9 are implemented.

12. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the steps of the tap parameter adjustment method for decision feedback equalization according to any one of claims 1 to 9 are implemented.

Images