DDR read data bus flip signal eye pattern training method based on standard delay unit

Through the DDR data bus flip signal eye diagram training method based on standard delay unit, the delay structure is simplified and the training efficiency is improved, and the problems of heavy hardware resource occupation and large training errors in the existing technology are solved, so that the DDR system is large Efficient training in delayed scenarios.

CN118981439APending Publication Date: 2024-11-19SHANGHAI ANLOGIC INFOTECH CO LTD
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411084575.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-08
Publication Date
2024-11-19

AI Technical Summary

Technical Problem

The existing DDR data bus flip signal eye diagram training method occupies a lot of hardware logic resources in large delay scenarios, has complex delay structure, low training efficiency and large training errors.

Method used

Using the DDR reading data bus flip signal eye diagram training method based on standard delay units, the DBI function of enabling the reading direction is configured by configuring the DDR particles, the standard delay unit series of RDQ, RDQS and RDBI_n signals are set, and the write commands and reads are generated. The command performs training, record confidence intervals to judge the training process, reduces the standard delay unit stages to adjust the delay position, and realizes delay control.

Benefits of technology

The delay structure is simplified, training efficiency is improved, training errors are reduced, and DDR system needs are met in large delay scenarios, avoiding the use of a large amount of hardware logic resources.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118981439A_ABST
    Figure CN118981439A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of DDR system training, and discloses a DDR read data bus flip signal eye pattern training method based on a standard delay unit, which comprises the steps of responding to an MRS command, configuring DDR particles, enabling a DBI function in a read direction, and setting read delay parameters. Setting the stages of the standard delay units of the RDQ signal, the RDQS signal and the RDBIN signal, and enabling the initial value of the standard delay unit of the RDBIN signal to be greater than a threshold value; generating a write command, and writing data for satisfying the RDBIN signal flipping function into the DDR particles; updating and configuring the stage number of the standard delay unit of the RDBIN signal; reducing the stage number of the standard delay unit of the RDBIN signal; and sending a multi-time reading command, reading data written by the DDR particles, and judging a training process for reading a data bus flip signal. The method and the device have the effects of reducing the complexity of a delay structure in a large delay scene and improving the efficiency and the precision.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of DDR system training, and in particular to a DDR read data bus flip signal eye diagram training method based on a standard delay unit. Background Art

[0002] DDR (Double Data Rate) system, also known as dual data rate synchronous dynamic random access memory system, has the advantages of large storage capacity, low cost, mature interface, and can achieve a higher access rate when accessed in parallel burst. The external signal interface of DDR mainly includes: clock CK (including a pair of differential clocks CK_t and CK_c), data signal DQ (bidirectional data signal line), selection signal DQS (reference clock of data signal, used to sample data signal line), command / address signal CA, etc. In addition to the above basic signals, it also includes related signals supporting optional functions, among which DBI_n signal (standard external interface signal of particle) is a signal used by DDR particles (DDR SDRAM, synchronous dynamic random access memory with double data rate) to support data bus flip function. When the DBI function of the read direction is enabled, DBI_n is the output signal of DDR particles, which is transmitted to the DDR controller simultaneously with the read data RDQ signal. For DDR particles, there will be a corresponding RDBI_n signal for every 8-bit RDQ. When the DDR particles feed back RDQ, if the number of 0s in a byte channel (8-bit RDQ) is greater than 4, then the RDQ of the byte channel is flipped and the corresponding RDBI_n is pulled low. If the number of 0s in the byte channel is less than or equal to 4, then the RDQ is directly sent to the bus and the corresponding RDBI_n is pulled high. After receiving RDQ and RDBI_n, the DDR controller will restore the actual data based on the value of RDBI_n. The timing and transmission process of RDBI_n are the same as those of RDQ, and finally both need to be sampled by RDQS. Due to wiring and delay differences, the initial relative position of RDQS and RDBI_n is uncertain. In order to achieve the best sampling of RDBI_n, the DDR system needs to perform eye diagram training on RDBI_n to ensure that the RDBI_n signal is correctly sampled by RDQS (read data strobe).

[0003] In the DDR system, the gate signal training, read data eye diagram training, write equalization training, write delay training and write data eye diagram training are generally completed first. After completion, the DDR particles can complete the write operation normally, and the gate signal of RDQS and the relative relationship between RDQS and RDQ are in the appropriate position. On this basis, the relative position of RDBI_n and RDQS is adjusted by moving the delay time of RDBI_n, so as to find the appropriate position of RDBI_n and complete the training of the read data bus flip signal eye diagram. In the prior art, in order to ensure adaptation to a larger delay, the delay scale includes integer multiples of tCK and non-integer multiples of delay. For integer multiples of tCK delay, due to the need to increase the logic related to the integer multiple delay, more hardware logic resources will be occupied, especially for FPGA systems, more programmable logic units will be occupied, and the complexity of the delay structure will be increased; for non-integer multiples of delay, the common practice is to divide tCK into fixed equal parts, for example, tCK is fixed into 32 or 64 parts, but for each additional equal delay unit, the delay time will increase by 1 / 32tCK or 1 / 64tCK. Each equally divided delay unit needs to use different numbers of hardware delay resources at different application frequencies, and the hardware delay module needs to perform complex calculations and processing. In addition, the training error will be different for different frequencies. Taking DDR4-1600 as an example, each DDR clock cycle is 1.25ns. In the case of 64 equal divisions, the delay time of each equally divided delay unit is 20ps, and the training error will also reach 20ps.

[0004] With respect to the above-mentioned related technologies, the inventors found that the existing DDR read data bus flip signal eye diagram training that meets large delay requirements has the problems of occupying more hardware logic resources and programmable logic unit resources, having a complex delay structure, low training efficiency and increased training error. Summary of the invention

[0005] In order to improve the efficiency of DDR read data bus flip signal eye diagram training in large delay scenarios and reduce training errors, the present application provides a DDR read data bus flip signal eye diagram training method based on a standard delay unit.

[0006] In a first aspect, the present application provides a DDR read data bus flip signal eye diagram training method based on a standard delay unit.

[0007] This application is achieved through the following technical solutions:

[0008] A DDR read data bus flip signal eye diagram training method based on a standard delay unit comprises the following steps:

[0009] In response to the MRS command, the DDR particles are configured, the DBI function in the read direction is enabled, and the corresponding read delay parameters are set;

[0010] Setting the number of stages of the standard delay unit of the RDQ signal, the RDQS signal and the RDBI_n signal, and making the initial value of the standard delay unit of the RDBI_n signal greater than a preset threshold;

[0011] Generate a write command to write data used to satisfy the RDBI_n signal flip function into the DDR particles;

[0012] Update and configure the number of stages of the standard delay unit of the RDBI_n signal;

[0013] Reduce the number of standard delay units of the RDBI_n signal;

[0014] Send multiple read commands to read the data written by the DDR particles to meet the RDBI_n signal flip function, and determine the training process of the read data bus flip signal based on the read data;

[0015] When the data read back by multiple read commands at the same delay position for the first time are consistent with the preset expected data, the configuration result of the standard delay unit of the current RDQ signal, RDQS signal and RDBI_n signal is recorded as the left boundary of the confidence interval;

[0016] Continue to reduce the number of stages of the standard delay unit of the RDBI_n signal;

[0017] Send multiple read commands to read the data written by the DDR particles to meet the RDBI_n signal flip function, and determine the training process of the read data bus flip signal based on the read data;

[0018] When the data read back by any read command at the same delay position is inconsistent with the expected data, the configuration result of the standard delay unit of the RDQ signal, RDQS signal and RDBI_n signal corresponding to the last time the data read back by multiple read commands at the same delay position are consistent with the preset expected data is recorded as the right boundary of the confidence interval. At this time, it is determined that the training of the read data bus flip signal is completed.

[0019] In a preferred example, the present application can be further configured as follows: it also includes the following steps:

[0020] When the number of the standard delay units of the RDBI_n signal is reduced to the threshold value, and the data read back by the read command is not consistent with the preset expected data, it is determined that the read data bus inversion signal training has failed.

[0021] In a preferred example, the present application can be further configured as follows: it also includes the following steps:

[0022] When the number of the standard delay units of the RDBI_n signal decreases to the threshold value, and the data read back by multiple read commands at the same delay position are consistent with the preset expected data, the number of the standard delay units of the RDBI_n signal remains unchanged, and the number of the standard delay units of the RDQ signal and the RDQS signal is increased at the same time;

[0023] Send multiple read commands to read the data written by the DDR particles to meet the RDBI_n signal flip function, and determine the training process of the read data bus flip signal based on the read data;

[0024] When the data read back by any read command of the same delay position is inconsistent with the expected data, the number of stages of the standard delay unit of the RDBI_n signal is updated, and then the number of stages of the standard delay unit of the RDBI_n signal is continuously reduced;

[0025] Send multiple read commands to read the data written by the DDR particles to meet the RDBI_n signal flip function, and judge the training process of the read data bus flip signal based on the read data, until the data read back by any read command at the same delay position is inconsistent with the expected data, and it is determined that the read data bus flip signal training is completed.

[0026] In a preferred example, the present application can be further configured as follows: it also includes the following steps:

[0027] When the number of standard delay units of the RDQ signal and the RDQS signal increases to reach a preset target value, and the data read back by multiple read commands at the same delay position are consistent with the preset expected data, it is determined that the read data bus flip signal training has failed;

[0028] When the number of stages of the standard delay unit of the RDBI_n signal is reduced to 0, and the data read back by multiple read commands at the same delay position are all consistent with the preset expected data, it is determined that the read data bus flip signal training has failed.

[0029] In a preferred example, the present application can be further configured as follows: when it is determined that the read data bus flip signal training is completed, the following steps are also included:

[0030] Based on the programmable instruction set of the FPGA system, the center position of the confidence interval is calculated, and after subtracting the common part of the number of stages of the standard delay unit of the programmable logic of the RDQ signal, the RDQS signal and the RDBI_n signal, the number of stages of the standard delay unit of the current RDQ signal, the RDQS signal and the RDBI_n signal is used as the final configuration information and the configuration is updated;

[0031] By using the final configuration information, multiple stages of standard delay cells composed of combinational logic cascades and multiple stages of standard delay cells composed of analog circuits are designed to be connected in series for delay control.

[0032] In a second aspect, the present application provides a DDR read data bus flip signal eye diagram training device based on a standard delay unit.

[0033] This application is achieved through the following technical solutions:

[0034] A DDR read data bus flip signal eye diagram training device based on a standard delay unit includes:

[0035] The initialization module is used to respond to the MRS command, configure the DDR particles, enable the DBI function in the read direction, and set the corresponding read delay parameters;

[0036] A standard delay unit configuration module, used to set the number of stages of the standard delay units of the RDQ signal, the RDQS signal and the RDBI_n signal, and to make the initial value of the standard delay unit of the RDBI_n signal greater than a preset threshold;

[0037] A data writing module is used to generate a write command and write data used to satisfy the RDBI_n signal flip function into the DDR particles;

[0038] An update module, used for updating and configuring the number of stages of the standard delay unit of the RDBI_n signal;

[0039] A delay control module, used to reduce the number of stages of the standard delay unit of the RDBI_n signal;

[0040] The data recovery and judgment module is used to send multiple read commands, read the data written by the DDR particles to meet the RDBI_n signal flip function, and judge the training process of the read data bus flip signal based on the read data;

[0041] A confidence interval left boundary module is used to record the configuration results of the standard delay unit of the current RDQ signal, RDQS signal and RDBI_n signal as the left boundary of the confidence interval when the data read back by multiple read commands at the same delay position for the first time are consistent with the preset expected data;

[0042] The confidence interval right boundary module is used to record the configuration results of the standard delay unit of the corresponding RDQ signal, RDQS signal and RDBI_n signal when the data read back by any read command at the same delay position are consistent with the preset expected data when the data read back by multiple read commands at the same delay position last time are consistent with the preset expected data, as the right boundary of the confidence interval. At this time, it is determined that the training of the read data bus flip signal is completed.

[0043] In a preferred example, the present application can be further configured as follows: the data recovery and judgment module also includes:

[0044] The first determination unit is used to determine that the read data bus flip signal training fails when the number of stages of the standard delay unit of the RDBI_n signal decreases to the threshold value and the data read back by the read command is not consistent with the preset expected data.

[0045] In a preferred example, the present application can be further configured as follows: the data recovery and judgment module also includes:

[0046] A second determination unit is used to determine that the read data bus flip signal training fails when the number of stages of the standard delay unit of the RDQ signal and the RDQS signal increases to reach a preset target value and the data read back by multiple read commands at the same delay position are consistent with the preset expected data;

[0047] When the number of stages of the standard delay unit of the RDBI_n signal is reduced to 0, and the data read back by multiple read commands at the same delay position are all consistent with the preset expected data, it is determined that the read data bus flip signal training has failed.

[0048] In a preferred example, the present application can be further configured as follows: further comprising:

[0049] An adjustment module is used to keep the number of standard delay units of the RDBI_n signal unchanged and increase the number of standard delay units of the RDQ signal and the RDQS signal at the same time when the number of standard delay units of the RDBI_n signal is reduced to the threshold value and the data read back by multiple read commands at the same delay position are consistent with the preset expected data; when the data read back by any read command at the same delay position is inconsistent with the expected data, update the number of standard delay units of the RDBI_n signal, and then continue to reduce the number of standard delay units of the RDBI_n signal until the data read back by any read command at the same delay position is inconsistent with the expected data, and determine that the read data bus flip signal training is completed.

[0050] In a third aspect, the present application provides a computer device.

[0051] This application is achieved through the following technical solutions:

[0052] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of any of the above-mentioned methods for training eye diagrams of DDR read data bus flip signals based on standard delay units are implemented.

[0053] In a fourth aspect, the present application provides a computer-readable storage medium.

[0054] This application is achieved through the following technical solutions:

[0055] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the computer program implements the steps of any of the above-mentioned DDR read data bus flip signal eye diagram training methods based on a standard delay unit.

[0056] In a fifth aspect, the present application provides a computer program product.

[0057] This application is achieved through the following technical solutions:

[0058] A computer program product comprises a computer program, wherein when the computer program is executed by a processor, the computer program implements the steps of any of the above-mentioned methods for training eye diagrams of DDR read data bus flip signals based on standard delay units.

[0059] In summary, compared with the prior art, the technical solution provided by this application has at least the following beneficial effects:

[0060] Delay control is performed based on standard delay units. In view of the phenomenon that the routing delays of RDQS, RDQ and RDBI_n in the read direction are similar and the SKEW is not much different, the RDQS signal, RDQ signal and RDBI_n signal are delayed at the standard delay unit level. No tCK integer multiple delay related logic is added, and there is no need to occupy more hardware logic resources and programmable logic unit resources. The delay structure is relatively simple, and the delay time can meet the SKEW requirements of almost all DDR systems. In the eye diagram training process, a write command is generated to write fixed data into the DDR, and then a read command is generated to generate a GATE gating signal according to the read command, which is used to select the appropriate RDQS signal, read the sampled RDQ signal and RDBI_n signal from the FIFO, perform read data recovery and judge the progress of the DDR read data bus flip signal eye diagram training. At the same time, the standard delay unit is used for delay control, and the training accuracy can reach a high level within the applicable frequency range. The delay structure is relatively simple, which improves the efficiency of the DDR read data bus flip signal eye diagram training in the large delay scenario and reduces the training error. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 A complete flowchart of a DDR read data bus flip signal eye diagram training method based on a standard delay unit is provided for an exemplary embodiment of the present application.

[0062] Figure 2A relative position relationship diagram of the RDBI_n signal and the RDQS signal / RDQ signal when the read data eye diagram training method of the DDR read data bus flip signal eye diagram based on the standard delay unit is completed and not adjusted according to an exemplary embodiment of the present application.

[0063] Figure 3 A diagram showing the relative position relationship of the RDQS signal, the RDQ signal and the RDBI_n signal after the initial setting of the DDR read data bus flip signal eye diagram training method based on the standard delay unit provided by an exemplary embodiment of the present application is completed.

[0064] Figure 4 A DDR read data bus flip signal eye diagram training method based on a standard delay unit provided in an exemplary embodiment of the present application is a schematic diagram of a delay adjustment process of a RDBI_n signal starting from an initial position.

[0065] Figure 5 A structural block diagram of a DDR training platform based on an FPGA system provided as an exemplary embodiment of the present application.

[0066] Figure 6 A position relationship diagram of the RDQS_T signal after gated logic processing in a DDR read data bus flip signal eye diagram training method based on a standard delay unit provided by an exemplary embodiment of the present application.

[0067] Figure 7 A schematic diagram of the positional relationship between RDQS_T0 / T1 signals, RDQ signals, etc. in a DDR read data bus flip signal eye diagram training method based on a standard delay unit provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0068] This specific embodiment is merely an explanation of the present application and is not a limitation of the present application. After reading this specification, those skilled in the art may make modifications to the present embodiment without any creative contribution as needed, but such modifications are protected by the patent law as long as they are within the scope of the claims of the present application.

[0069] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0070] In addition, the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article, unless otherwise specified, generally means that the associated objects before and after are in an "or" relationship.

[0071] An embodiment of the present application provides a DDR read data bus flip signal eye diagram training method based on a standard delay unit, and the main steps of the method are described as follows.

[0072] In response to the MRS command, the DDR particles are configured, the DBI function in the read direction is enabled, and the corresponding read delay parameters are set;

[0073] Setting the number of stages of the standard delay unit of the RDQ signal, the RDQS signal and the RDBI_n signal, and making the initial value of the standard delay unit of the RDBI_n signal greater than a preset threshold;

[0074] Generate a write command to write data used to satisfy the RDBI_n signal flip function into the DDR particles;

[0075] Update and configure the number of stages of the standard delay unit of the RDBI_n signal;

[0076] Reduce the number of standard delay units of the RDBI_n signal;

[0077] Send multiple read commands to read the data written by the DDR particles to meet the RDBI_n signal flip function, and determine the training process of the read data bus flip signal based on the read data;

[0078] When the data read back by multiple read commands at the same delay position for the first time are consistent with the preset expected data, the configuration result of the standard delay unit of the current RDQ signal, RDQS signal and RDBI_n signal is recorded as the left boundary of the confidence interval;

[0079] Continue to reduce the number of stages of the standard delay unit of the RDBI_n signal;

[0080] Send multiple read commands to read the data written by the DDR particles to meet the RDBI_n signal flip function, and determine the training process of the read data bus flip signal based on the read data;

[0081] When the data read back by any read command at the same delay position is inconsistent with the expected data, the configuration result of the standard delay unit of the RDQ signal, RDQS signal and RDBI_n signal corresponding to the last time the data read back by multiple read commands at the same delay position are consistent with the preset expected data is recorded as the right boundary of the confidence interval. At this time, it is determined that the training of the read data bus flip signal is completed.

[0082] This embodiment can greatly increase the use scenarios of the system by delaying RDQS, RDQ, and RDBI_n, and meet the DDR system requirements when RDQ / RDBI_n and RDQS are large in SKEW.

[0083] The RDQS signal includes a pair of differential signals RDQS_t and RDQS_c.

[0084] After receiving the read command, the DDR chip sends out a read data signal RDQ and a read data select signal RDQS at a fixed interval of read delay RL time.

[0085] In one embodiment, the DDR read data bus flip signal eye diagram training method based on the standard delay unit further includes the following steps:

[0086] When the number of the standard delay units of the RDBI_n signal is reduced to the threshold value, and the data read back by the read command is not consistent with the preset expected data, it is determined that the read data bus inversion signal training has failed.

[0087] In one embodiment, the DDR read data bus flip signal eye diagram training method based on the standard delay unit further includes the following steps:

[0088] When the number of the standard delay units of the RDBI_n signal decreases to the threshold value, and the data read back by multiple read commands at the same delay position are consistent with the preset expected data, the number of the standard delay units of the RDBI_n signal remains unchanged, and the number of the standard delay units of the RDQ signal and the RDQS signal is increased at the same time;

[0089] Send multiple read commands to read the data written by the DDR particles to meet the RDBI_n signal flip function, and determine the training process of the read data bus flip signal based on the read data;

[0090] When the data read back by any read command of the same delay position is inconsistent with the expected data, the number of stages of the standard delay unit of the RDBI_n signal is updated, and then the number of stages of the standard delay unit of the RDBI_n signal is continuously reduced;

[0091] Send multiple read commands to read the data written by the DDR particles to meet the RDBI_n signal flip function, and judge the training process of the read data bus flip signal based on the read data, until the data read back by any read command at the same delay position is inconsistent with the expected data, and it is determined that the read data bus flip signal training is completed.

[0092] In one embodiment, the DDR read data bus flip signal eye diagram training method based on the standard delay unit further includes the following steps:

[0093] When the number of standard delay units of the RDQ signal and the RDQS signal increases to reach a preset target value, and the data read back by multiple read commands at the same delay position are consistent with the preset expected data, it is determined that the read data bus flip signal training has failed;

[0094] When the number of stages of the standard delay unit of the RDBI_n signal is reduced to 0, and the data read back by multiple read commands at the same delay position are all consistent with the preset expected data, it is determined that the read data bus flip signal training has failed.

[0095] In one embodiment, when determining that the read data bus flip signal training is completed, the following steps are also included:

[0096] Based on the programmable instruction set of the FPGA system, the center position of the confidence interval is calculated, and after subtracting the common part of the number of stages of the standard delay unit of the programmable logic of the RDQ signal, the RDQS signal and the RDBI_n signal, the number of stages of the standard delay unit of the current RDQ signal, the RDQS signal and the RDBI_n signal is used as the final configuration information and the configuration is updated;

[0097] By using the final configuration information, multiple stages of standard delay cells composed of combinational logic cascades and multiple stages of standard delay cells composed of analog circuits are designed to be connected in series for delay control.

[0098] Taking the eye diagram training of the read data bus flip signal RDBI_n as an example, the embodiments of the present application are further described in detail below in conjunction with the drawings in the specification.

[0099] Reference Figure 1 , configure the DDR particles through the MRS command, enable the DBI (data bus inversion) function in the read direction, and configure the corresponding read delay parameters. After the DBI function is turned on, the DDR particles will decide whether to invert the read data according to the number of 0s on each byte. The DBI function is divided into the read direction and the write direction. This embodiment mainly describes the functions related to the read direction and the corresponding DBI_n signal eye diagram training method. After enabling the read direction, the DBI_n signal is the output signal of the particle, which is used to indicate whether the particle has flipped the data.

[0100] The number of standard delay units of the RDQ signal, RDQS signal and RDBI_n signal is set to determine the appropriate RDQSDL64 and RDQSSDL96. It is worth noting that the RDBI_n signal and the DBI_n signal in this embodiment have the same meaning, both referring to the DBI_n interface signal of the DDR particle to the outside when the read data bus flip function is enabled.

[0101] The standard delay cells include standard delay cells SDL (Standard Cell Delay Line) with fixed delay function composed of combinational logic cascades and user-defined standard delay cells CDL (Customer delay line) composed of analog circuits.

[0102] Figure 2 This is the possible relative position relationship between the RDBI_n signal and the RDQS signal / RDQ signal when the data eye diagram training is completed and no adjustment is made. Only RDQS_T0 is selected for illustration. When RDQSSDL96 changes, RDQS_T0 and RDQS_T1 will move at the same time.

[0103] In order to ensure that the initial training position of RDBI_n is at the fail position, the initial value of the standard delay unit of the RDBI_n signal is set to be greater than the preset threshold value A0, so as to set a larger initial value for the SDL64 / 32 / CDL4 of RDBI_n. In addition, due to SKEW (i.e., different signals may have signal offset at the receiving end due to wiring and other issues, so that the signals cannot be completely aligned), RDBI_n may have different initial positions such as the above positions 1, 2, and 3. It is necessary to ensure that in all the above cases, RDBI_n has found the left boundary position from fail to pass before it is reduced to the preset threshold value A. Therefore, it is necessary to add the same delay setting to the configuration of RDQSSDL96 and RDQ SDL64 based on the reading data eye diagram training results. The setting of threshold A is mainly used for the subsequent search for the right boundary, and the value is set to be small.

[0104] Figure 3 To complete the initial setting of the relative position relationship of the RDQS signal, RDQ signal and RDBI_n signal, position 3 is taken as an example in the figure.

[0105] Reference Figure 1 Then, a write command is generated to write a fixed data pattern into the DDR particles. The data pattern needs to ensure that when the address data is read back, RDBI_n in phase 0 / 2 / 4 / 6 and phase 1 / 3 / 5 / 7 has sufficient flips.

[0106] Update and configure SDL64 / 32 / CDL4 for RDBI_n.

[0107] At this point, the eye diagram training for writing and reading data has been completed, so by sending a read command, it is possible to indirectly determine whether the sampling of RDBI_n is correct by checking whether the read-back data is consistent with the expected data.

[0108] Based on the read data, the training progress is judged and the settings of SDL64 / 32 / CDL4 of RDBI_n are recorded.

[0109] Reference Figure 4 , starting from the initial position, the delay setting of RDBI_n is continuously reduced. The minimum step size can be 1CDL each time, and the size of each step can be adjusted according to the accuracy requirements.

[0110] Send a read command to read back the data written to the DDR particles.

[0111] Whether the sampling of RDBI_n is correct is indirectly judged based on whether the read-back data is consistent with the expected data. When the first stable pass position is found, the current RDQ / RDQS / RDBI_n configuration result is recorded as the left boundary of the pass interval, and the delay setting of RDBI_n is continued to be reduced; when the first data fail position is found, the last stable pass position is used as the right boundary of the pass interval.

[0112] During the above training process, each delay position will perform multiple read operations of the read command. A stable pass means that the read data of multiple read operations at the same position are consistent with the expected data, and a fail means that the read data is inconsistent with the expected data in one of the multiple read operations.

[0113] Reference Figure 1 If a complete confidence interval pass is found, the read data bus flip signal eye diagram training is determined to be successful (completed) and marked as pass.

[0114] If a complete confidence interval pass is not found and the SDL64 / 32 / CDL4 of RDBI_n is not reduced to the threshold A, the delay setting of RDBI_n continues to be reduced, and the SDL64 / 32 / CDL4 of RDBI_n is updated and configured, and the read data continues to be read to detect the training process.

[0115] If a complete confidence interval pass is not found and SDL64 / 32 / CDL4 of RDBI_n is reduced to threshold A, that is, the right boundary is still not found when RDBI_n is reduced to the set threshold A, then start to increase the delay settings of SDL96 of RDQS and SDL64 of RDQ at the same time, update SDL64 of RDQ and SDL96 of RDQS, and continue to read the read data to detect the training process and judge the settings of RDQ and RDQS.

[0116] The SDL32 / CDL4 on RDQS is used to compensate for the deviation of RDQS duty, and the SDL32 / CDL4 on RDQ is used to compensate for the deviation of RDQ SKEW. Therefore, during the eye diagram training of RDBI_n, the configuration of the above SDL32 / CDL4 cannot be changed. Therefore, the CDL4 of RDQS and RDQ cannot be adjusted.

[0117] Reference Figure 1 ,To ensure that the accuracy of the eye diagram training of RDBI_n can reach 5ps, when the SDL96 of RDQS and the SDL64 of RDQ are increased to the first data fail position, the delay of RDBI_n is increased by 1step, and then the delay setting of RDBI_n is reduced again, the SDL64 / 32 / CDL4 of RDBI_n is updated and configured, and the read data is continued to be read to detect the training process, and the setting of SDL64 / 32 / CDL4 of RDBI_n is judged so as to find the right boundary using the minimum CDL4step.

[0118] The adjustment of SDL64 / 32 / CDL4 of RDBI_n starts from A+1step, where step is the delay size increased by a single time when increasing the delay settings of SDL96 of RDQS and SDL64 of RDQ, to ensure that the setting of RDBI_n at this time is in the pass interval, and continue to reduce the setting of SDL64 / 32 / CDL4 of RDBI_n until a fail position appears, that is, a complete confidence interval pass is found, the training is judged to be successful, marked as pass, and the current configuration result of RDQ / RDQS / RDBI_n is recorded as the right boundary.

[0119] If the SDL96 of RDQS and the SDL64 of RDQ increase to reach the set maximum threshold B, and the fail position is still not found, the training is determined to have failed, marked as fail, and the training is terminated.

[0120] If the fail position is not found (the complete confidence interval pass is not found) and the SDL64 / 32 / CDL4 of RDBI_n is not reduced to 0, continue to reduce the delay setting of RDBI_n, update and configure the SDL64 / 32 / CDL4 of RDBI_n, and continue to read the read data to detect the training process.

[0121] If the fail position is not found (the complete confidence interval pass is not found) and the SDL64 / 32 / CDL4 of RDBI_n decreases to 0, and the right boundary is still not found, the training is judged to have failed, marked as fail, and the training is terminated.

[0122] Figure 1 In the embodiment, SDL96 is a series connection of 96 levels of SDL units. The overall delay time of RDQS can be adjusted by adjusting the number of SDL units passed during the RDQS_T transmission process through configuration information.

[0123] SDL64 / SDL32 are 64-level and 32-level SDL units connected in series, respectively, to adjust the delay time of RDQ and RDBI_n during training. Through the coordination of RDQSSDL96, RDQ SDL64, and RDBI_n SDL64, the training method can be adapted to the DDR system with a larger SKEW. SDL32 on RDQ is mainly used to compensate for the SKEW between different RDQs.

[0124] CDL4 is a series connection of 4 levels of CDL. You can adjust the overall delay time by adjusting the number of CDLs passed during the RDQS / RDQ / RDBI_n transmission process through configuration information. The delay time of each level of CDL is about 5ps. The delay time of the 4-level CDL series connection is approximately equal to the delay time of the 1-level SDL. When the delay time of the used CDL reaches 4 levels, one level of SDL will be added and the value of CDL will be cleared.

[0125] Finally, the center positions of the left and right boundaries are calculated by the instruction as the difference to subtract the common delay part of SDL64 / 32 of RDQSSDL96, RDQ SDL64 and RDBI_n to obtain the final configuration information.

[0126] After the traditional method is trained, the training results will be directly configured on the corresponding delay line, which may cause the delay configured on the delay line to be relatively large, which is not an integer multiple of tCK (DDR clock cycle). When PVT changes, the delay time of this part of the delay will usually change with PVT (short for process, voltage and temperature), causing RDQS, RDQ and RDBI_n to drift, and the stability of the overall system will decrease.

[0127] This scheme is based on the delay structure of SDL and CDL. At the end of training, the final configuration information is used to design multiple levels of standard delay units composed of combinational logic cascades and multiple levels of standard delay units composed of analog circuits in series for delay control, thereby reducing the absolute value configured on the standard delay unit and improving the stability of the system.

[0128] In summary, the DDR read data bus flip signal eye diagram training method based on the standard delay unit is based on the delay structure of the standard delay unit SDL and CDL for delay control. Aiming at the phenomenon that the routing delays of RDQS, RDQ and RDBI_n in the read direction are similar and the SKEW is not much different, the RDQS signal, RDQ signal and RDBI_n signal are delayed at the standard delay unit level, without adding tCK integer multiple delay related logic, without occupying more hardware logic resources and programmable logic unit resources, the delay structure is relatively simple, and the delay time can meet the SKEW requirements of almost all DDR systems; in the eye diagram training process, a write command is generated to write fixed data into the DDR, and then a read command is generated to generate RDQS according to the read command.

[0129] GATE (gating signal of read enable signal) is used to control the passage time of read enable signal. By selecting the appropriate RDQS signal, the sampled RDQ signal and RDBI_n signal are read out from FIFO to recover the read data and judge the progress of DDR read data bus flip signal eye diagram training. At the same time, standard delay unit is used for delay control, and the training accuracy can reach a high level within the applicable frequency range. The delay structure is relatively simple, which improves the efficiency of DDR read data bus flip signal eye diagram training in large delay scenarios.

[0130] Since DDR4 is applicable to a wide frequency range (800MTS~3200MTS), using the traditional delay method, each equally divided delay unit needs to use different numbers of hardware delay resources at different frequencies, and the hardware delay module needs to perform complex calculations and processing. In addition, for different frequencies, the training error will also be different. Taking DDR4-1600 as an example, each DDR clock cycle is 1.25ns. In the case of 64 equal divisions, the delay time of each equally divided delay unit is 20ps, and the training error will also reach 20ps. This method uses standard delay units SDL and CDL for delay control, where SDL is a combinational logic cascade with fixed The standard delay unit with delay function has a simple structure and a fixed delay time of about 20ps. CDL is a standard delay unit with fixed delay function composed of analog circuits and a fixed delay time of about 5ps. By using a delay structure composed of multiple levels of SDL and CDL in series, the accuracy of the training results can be as high as 5ps at all frequencies. The training accuracy is high and can meet the DDR system in which the initial position of RDBI_n is located before and after RDQS. After the training, the overall configuration of the standard delay unit is small, which can greatly reduce the number of standard delay units configured on RDQS / RDQ / RDBI_n, and the overall system is less affected by PVT, and the stability is improved.

[0131] The DDR read data bus flip signal eye diagram training method based on standard delay unit designs a training module based on instruction set control based on FPGA. Through instruction configuration, flexible adjustment and configuration of delay information can be achieved. The training process is realized by a combination of training algorithms and instructions, and no additional external software participation is required.

[0132] In the DDR read data bus flip signal eye diagram training method based on the standard delay unit, the DDR control logic has the ability to adjust the RDQS, RDQ and RDBI_n delays and can adapt to larger delays.

[0133] The DDR read data bus inversion signal eye diagram training method based on the standard delay unit can be applied to all DDR particle types and systems that support the RDBI (read data bus inversion) function, such as DDR5 / LPDDR, etc. In addition to being used on the FPGA platform, it can be applied to various integrated circuit systems and SOCs that use DDR particles. The bit width is expanded according to demand and is not limited to x8 particles. Various DDR particle types such as x16 are also applicable. In addition, it is applicable to other similar systems such as LPDDR.

[0134] The delay structure of the DDR read data bus flip signal eye diagram training method based on the standard delay unit can also be adjusted according to actual needs, so that the training process can be further simplified, which will not be repeated here.

[0135] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0136] The embodiment of the present application also provides a DDR read data bus flip signal eye diagram training device based on a standard delay unit, and the DDR read data bus flip signal eye diagram training device based on a standard delay unit corresponds to the DDR read data bus flip signal eye diagram training method based on a standard delay unit in the above embodiment. The DDR read data bus flip signal eye diagram training device based on a standard delay unit includes:

[0137] The initialization module is used to respond to the MRS command, configure the DDR particles, enable the DBI function in the read direction, and set the corresponding read delay parameters;

[0138] A standard delay unit configuration module, used to set the number of stages of the standard delay units of the RDQ signal, the RDQS signal and the RDBI_n signal, and to make the initial value of the standard delay unit of the RDBI_n signal greater than a preset threshold;

[0139] A data writing module is used to generate a write command and write data used to satisfy the RDBI_n signal flip function into the DDR particles;

[0140] An update module, used for updating and configuring the number of stages of the standard delay unit of the RDBI_n signal;

[0141] A delay control module, used to reduce the number of stages of the standard delay unit of the RDBI_n signal;

[0142] The data recovery and judgment module is used to send multiple read commands, read the data written by the DDR particles to meet the RDBI_n signal flip function, and judge the training process of the read data bus flip signal based on the read data;

[0143] A confidence interval left boundary module is used to record the configuration results of the standard delay unit of the current RDQ signal, RDQS signal and RDBI_n signal as the left boundary of the confidence interval when the data read back by multiple read commands at the same delay position for the first time are consistent with the preset expected data;

[0144] The confidence interval right boundary module is used to record the configuration results of the standard delay unit of the corresponding RDQ signal, RDQS signal and RDBI_n signal when the data read back by any read command at the same delay position are consistent with the preset expected data when the data read back by multiple read commands at the same delay position last time are consistent with the preset expected data, as the right boundary of the confidence interval. At this time, it is determined that the training of the read data bus flip signal is completed.

[0145] In one embodiment, the data recovery and judgment module further includes:

[0146] The first determination unit is used to determine that the read data bus flip signal training fails when the number of stages of the standard delay unit of the RDBI_n signal decreases to the threshold value and the data read back by the read command is not consistent with the preset expected data.

[0147] In one embodiment, the data recovery and judgment module further includes:

[0148] A second determination unit is used to determine that the read data bus flip signal training fails when the number of stages of the standard delay unit of the RDQ signal and the RDQS signal increases to reach a preset target value and the data read back by multiple read commands at the same delay position are consistent with the preset expected data;

[0149] When the number of the standard delay units of the RDBI_n signal is reduced to 0, and the data read back by multiple read commands at the same delay position are consistent with the preset expected data, it is determined that the read data bus flip signal training has failed.

[0150] In one embodiment, the DDR read data bus flip signal eye diagram training device based on the standard delay unit further includes:

[0151] An adjustment module is used to keep the number of standard delay units of the RDBI_n signal unchanged and increase the number of standard delay units of the RDQ signal and the RDQS signal at the same time when the number of standard delay units of the RDBI_n signal is reduced to the threshold value and the data read back by multiple read commands at the same delay position are consistent with the preset expected data; when the data read back by any read command at the same delay position is inconsistent with the expected data, update the number of standard delay units of the RDBI_n signal, and then continue to reduce the number of standard delay units of the RDBI_n signal until the data read back by any read command at the same delay position is inconsistent with the expected data, and determine that the read data bus flip signal training is completed.

[0152] Reference Figure 5Based on the FPGA (Field Programmable Gate Array, that is, an integrated circuit with programmable characteristics pre-designed and implemented on a silicon chip) system, a DDR training platform based on a programmable instruction set was built using its internal resources. By combining and calling FPGA resources, the training of the DDR read data bus flip signal eye diagram can be achieved.

[0153] Specifically, the general IO of the FPGA system is used as the interface with the DDR particles, relying on the rich interface resources of the FPGA and the rich delay control logic on the interface to perform delay control, including:

[0154] The DDR training control module is used for the DDR training control logic based on the instruction set. It sends instructions through the APB access interface to control the internal logic. The DDR training control module can complete the basic read and write access to the DDR, control the delay unit, process the read-back data, and judge the training results.

[0155] The command generation module is used to generate read / write commands. During the training process of the read data bus flip signal eye diagram, a write command is first generated to write a fixed data pattern into the DDR particles, and then a read command is generated. A gating signal is generated according to the read command to select the appropriate RDQS signal. At the same time, according to the read command, a read valid (RD_VALID) signal is generated at the set time, and the sampled RDQ and RDBI_n are read out from the FIFO for subsequent read data judgment.

[0156] The read data recovery and judgment module is used to recover the read data and make read data judgment. After receiving the valid RDQ and RDBI_n, the required read data is recovered according to the value of RDBI_n, and it is judged whether the read data is consistent with the expected data.

[0157] The delay control module is used to adjust the delay time according to the judgment result of the read data. The configuration and adjustment of all standard delay units during the training process are completed through the instruction control delay control module.

[0158] Finally, the DDR training control module is used to judge the training process based on the comparison results of the read data.

[0159] The gate control logic of the FPGA system is used to delay and process the GATE signal, and the RDQS-related logic is selected through the GATE. After the RDQS returned by the particle passes through the gate control logic, only four complete cycle signals are retained to facilitate the sampling of RDQ and RDBI_n. The gate control logic does not include the delay of RDQS, and the gate control logic only processes the RDQS_T returned by the particle, and does not process the RDQS_C signal.

[0160] Figure 5 In the example, SDL32_0 / 1 is a series connection of 32 levels of SDL units, which are standard delay units on the RDQS_T0 / T1 path. Because the duty cycle of RDQS_T may deviate from the ideal value of 50%, and the delay time from RDQS_T0 / T1 to the same DQ sampling position also has SKEW, when RDQS_T0 / T1 samples the same RDQ phase0 / 2 / 4 / 6 and phase1 / 3 / 5 / 7, the center position of the eye diagram may have a certain deviation. This deviation value is compensated by the setting of SDL32_0 / 1 during the reading data eye diagram training process.

[0161] RD FIFO is used to temporarily store the RDQ and RDBI_n data sampled by RDQS, and when RD_VALID is valid, the data is taken out and sent to the read data recovery and judgment module for comparison of read data and judgment of the training process.

[0162] IOB stands for input / output buffer, which is the external interface resource of FPGA. In this embodiment, IOB is the interface between DDR particles and FPGA.

[0163] This example diagram mainly describes the data transmission direction and delay structure of RDQS, RDQ, and RDBI_n during the read process. During the write operation, the transmission direction of signals such as DQ / DQS is opposite to that in the example diagram. To facilitate understanding of the RDBI training process and delay structure, no example of the write direction is added here.

[0164] After the gating logic processing, the RDQS_T signal is as follows Figure 6 shown.

[0165] After completing the read data eye diagram training, the rising edge of RDQS_T0 / T1 is located at the center of RDQ phase0 / 2 / 4 / 6 and phase1 / 3 / 5 / 7 respectively.

[0166] Reference Figure 7During the training process, the rising edge of RDQS_T0 / T1 samples RDQ phase0 / 2 / 4 / 6 and phase1 / 3 / 5 / 7 respectively. After the duty cycle of RDQS_T is compensated by SDL32_0 / 1, during the eye diagram training of RDBI_n, only the SDL64 / 32 / CDL4 of RDBI_n needs to be moved separately to ensure that the rising edge of RDQS_T0 / T1 is located at the center of RDBI_n phase0 / 2 / 4 / 6 and phase1 / 3 / 5 / 7 respectively. In addition, if the delay of RDQS and RDQ needs to be adjusted during the RDBI_n eye diagram training, then the SDL96 of RDQS and the SDL64 of RDQ need to be moved at the same time to ensure that the relative position relationship between RDQS and RDQ remains unchanged.

[0167] The specific limitations on the DDR read data bus flip signal eye diagram training device based on the standard delay unit can also be found in the above limitations on the DDR read data bus flip signal eye diagram training method based on the standard delay unit, which will not be repeated here.

[0168] Each module in the above-mentioned DDR read data bus flip signal eye diagram training device based on standard delay unit can be fully or partially implemented by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.

[0169] In one embodiment, a computer device is provided, which may be a server. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, any of the above-mentioned DDR read data bus flip signal eye diagram training methods based on a standard delay unit is implemented.

[0170] In one embodiment, a computer-readable storage medium is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, any of the above-mentioned DDR read data bus flip signal eye diagram training methods based on a standard delay unit is implemented.

[0171] In one embodiment, a computer program product is provided. The computer program product includes a computer program. When the computer program is executed by a processor, the computer program implements any of the above-mentioned DDR read data bus flip signal eye diagram training methods based on a standard delay unit.

[0172] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. When the computer program is executed, it may include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in the present application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0173] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

Claims

1. DDR read data bus flip signal eye diagram training method based on standard delay unit, It is characterized in that The following steps are included: In response to the MRS command, the DDR particles are configured, the DBI function in the read direction is enabled, and the corresponding read delay parameters are set; Setting the number of stages of the standard delay unit of the RDQ signal, the RDQS signal and the RDBI_n signal, and making the initial value of the standard delay unit of the RDBI_n signal greater than a preset threshold; Generate a write command to write data used to satisfy the RDBI_n signal flip function into the DDR particles; Update and configure the number of stages of the standard delay unit of the RDBI_n signal; Reduce the number of standard delay units of the RDBI_n signal; Send multiple read commands to read the data written by the DDR particles to meet the RDBI_n signal flip function, and determine the training process of the read data bus flip signal based on the read data; When the data read back by multiple read commands at the same delay position for the first time are consistent with the preset expected data, the configuration result of the standard delay unit of the current RDQ signal, RDQS signal and RDBI_n signal is recorded as the left boundary of the confidence interval; Continue to reduce the number of stages of the standard delay unit of the RDBI_n signal; Send multiple read commands to read the data written by the DDR particles to meet the RDBI_n signal flip function, and determine the training process of the read data bus flip signal based on the read data; When the data read back by any read command at the same delay position is inconsistent with the expected data, the configuration result of the standard delay unit of the RDQ signal, RDQS signal and RDBI_n signal corresponding to the last time the data read back by multiple read commands at the same delay position are consistent with the preset expected data is recorded as the right boundary of the confidence interval. At this time, it is determined that the training of the read data bus flip signal is completed.

2. The DDR read data bus flip signal eye diagram training method based on the standard delay unit according to claim 1, It is characterized in that The following steps are also included: When the number of the standard delay units of the RDBI_n signal is reduced to the threshold value, and the data read back by the read command is not consistent with the preset expected data, it is determined that the read data bus inversion signal training has failed.

3. The DDR read data bus flip signal eye diagram training method based on the standard delay unit according to claim 1, It is characterized in that The following steps are also included: When the number of the standard delay units of the RDBI_n signal decreases to the threshold value, and the data read back by multiple read commands at the same delay position are consistent with the preset expected data, the number of the standard delay units of the RDBI_n signal remains unchanged, and the number of the standard delay units of the RDQ signal and the RDQS signal is increased at the same time; Send multiple read commands to read the data written by the DDR particles to meet the RDBI_n signal flip function, and determine the training process of the read data bus flip signal based on the read data; When the data read back by any read command of the same delay position is inconsistent with the expected data, the number of stages of the standard delay unit of the RDBI_n signal is updated, and then the number of stages of the standard delay unit of the RDBI_n signal is continuously reduced; Send multiple read commands to read the data written by the DDR particles to satisfy the RDBI_n signal flip function, and judge the training process of the read data bus flip signal based on the read data, until the data read back by any read command at the same delay position is inconsistent with the expected data, and determine that the read data bus flip signal training is completed, and record the configuration results of the standard delay units of the corresponding RDQ signal, RDQS signal and RDBI_n signal when the data read back by multiple read commands at the same delay position last time are consistent with the preset expected data.

4. The DDR read data bus flip signal eye diagram training method based on the standard delay unit according to claim 3, It is characterized in that The following steps are also included: When the number of standard delay units of the RDQ signal and the RDQS signal increases to reach a preset target value, and the data read back by multiple read commands at the same delay position are consistent with the preset expected data, it is determined that the read data bus flip signal training has failed; When the number of stages of the standard delay unit of the RDBI_n signal is reduced to 0, and the data read back by multiple read commands at the same delay position are all consistent with the preset expected data, it is determined that the read data bus flip signal training has failed.

5. The DDR read data bus flip signal eye diagram training method based on the standard delay unit according to any one of claims 1 to 4, It is characterized in that When the read data bus flip signal training is completed, The following steps are included: Based on the programmable instruction set of the FPGA system, the center position of the confidence interval is calculated, and after subtracting the common part of the number of stages of the standard delay unit of the programmable logic of the RDQ signal, the RDQS signal and the RDBI_n signal, the number of stages of the standard delay unit of the current RDQ signal, the RDQS signal and the RDBI_n signal is used as the final configuration information and the configuration is updated; By using the final configuration information, multiple stages of standard delay cells composed of combinational logic cascades and multiple stages of standard delay cells composed of analog circuits are designed to be connected in series for delay control.

6. DDR read data bus flip signal eye diagram training device based on standard delay unit, It is characterized in that include, The initialization module is used to respond to the MRS command, configure the DDR particles, enable the DBI function in the read direction, and set the corresponding read delay parameters; A standard delay unit configuration module, used to set the number of stages of the standard delay units of the RDQ signal, the RDQS signal and the RDBI_n signal, and to make the initial value of the standard delay unit of the RDBI_n signal greater than a preset threshold; A data writing module is used to generate a write command and write data used to satisfy the RDBI_n signal flip function into the DDR particles; An update module, used for updating and configuring the number of stages of the standard delay unit of the RDBI_n signal; A delay control module, used to reduce the number of stages of the standard delay unit of the RDBI_n signal; The data recovery and judgment module is used to send multiple read commands, read the data written by the DDR particles to meet the RDBI_n signal flip function, and judge the training process of the read data bus flip signal based on the read data; A confidence interval left boundary module is used to record the configuration results of the standard delay unit of the current RDQ signal, RDQS signal and RDBI_n signal as the left boundary of the confidence interval when the data read back by multiple read commands at the same delay position for the first time are consistent with the preset expected data; The confidence interval right boundary module is used to record the configuration results of the standard delay unit of the corresponding RDQ signal, RDQS signal and RDBI_n signal when the data read back by any read command at the same delay position are consistent with the preset expected data when the data read back by multiple read commands at the same delay position last time are consistent with the preset expected data, as the right boundary of the confidence interval. At this time, it is determined that the training of the read data bus flip signal is completed.

7. The DDR read data bus flip signal eye diagram training device based on the standard delay unit according to claim 6, It is characterized in that The data recovery and judgment module also includes: The first determination unit is used to determine that the read data bus flip signal training fails when the number of stages of the standard delay unit of the RDBI_n signal decreases to the threshold value and the data read back by the read command is not consistent with the preset expected data.

8. The DDR read data bus flip signal eye diagram training device based on the standard delay unit according to claim 6, It is characterized in that The data recovery and judgment module also includes: The second determination unit is used to determine that the training of the read data bus flip signal has failed when the number of stages of the standard delay unit of the RDQ signal and the RDQS signal increases to reach a preset target value and the data read back by multiple read commands at the same delay position are all consistent with the preset expected data; and to determine that the training of the read data bus flip signal has failed when the number of stages of the standard delay unit of the RDBI_n signal decreases to 0 and the data read back by multiple read commands at the same delay position are all consistent with the preset expected data.

9. The DDR read data bus flip signal eye diagram training device based on the standard delay unit according to claim 6, It is characterized in that Also includes, An adjustment module is used to keep the number of standard delay units of the RDBI_n signal unchanged and increase the number of standard delay units of the RDQ signal and the RDQS signal at the same time when the number of standard delay units of the RDBI_n signal is reduced to the threshold value and the data read back by multiple read commands at the same delay position are consistent with the preset expected data; when the data read back by any read command at the same delay position is inconsistent with the expected data, update the number of standard delay units of the RDBI_n signal, and then continue to reduce the number of standard delay units of the RDBI_n signal until the data read back by any read command at the same delay position is inconsistent with the expected data, and determine that the read data bus flip signal training is completed.

10. A computer device, It is characterized in that The invention comprises a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method according to any one of claims 1 to 5.

11. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

12. A computer program product, It is characterized in that The invention comprises a computer program, which implements the steps of the method according to any one of claims 1 to 5 when being executed by a processor.

Citation Information

Cited By

  • Training system and method for high-speed parallel port IP

    CN119250001A

  • DDR-based training debugging system and method

    CN121072421A