An adaptive memory testing method, system, device, product, and medium

Abstract

The present application relates to the field of storage test, and provides a kind of adaptive memory test method, system, equipment, product and medium, including according to sampling ratio sampling, obtain edge phase;Obtain two-dimensional probability eye diagram, construct full convolution inference network, the resolution of two-dimensional probability eye diagram is promoted by full convolution inference network, obtain fine two-dimensional probability eye diagram;The error of fine two-dimensional probability eye diagram is zero probability and offset standard deviation, determine eye diagram adjustment strategy, to adjust fine two-dimensional probability eye diagram, obtain adjustment two-dimensional probability eye diagram;The timing margin and eye diagram offset of adjustment two-dimensional probability eye diagram are calculated, and the compensation strategy is determined according to timing margin and eye diagram offset, and the phase compensation amount is obtained according to the compensation strategy;According to the selection of compensation strategy, determine the area failure probability, generate fault area mark by area failure probability, generate memory fault report according to fault area mark and phase compensation amount, complete memory test.

Description

Technical Field

[0001] This invention relates to the field of storage testing technology, and in particular to an adaptive memory testing method, system, device, product, and medium. Background Technology

[0002] In high-speed read / write tests of semiconductor memories, it is necessary to verify functional correctness, timing margin, and interface signal integrity simultaneously. As memory interface speeds increase, the offset of the sampling point relative to the eye diagram center, edge jitter, and voltage noise will shrink the effective sampling window, leading to bit errors, insufficient hold time, or fluctuations in test results.

[0003] Existing memory testing methods can evaluate interface signal quality using eye diagram analysis, and adjust high-speed interface timing through sampling point scanning, clock delay chains, or duty cycle calibration. Publicly available solutions cover high-speed oversampling, digital eye diagrams, safety windows, and sampling clock phase adjustment. Test systems also include solutions that use programmable timing circuits to change the phase relationship of test vectors. However, these solutions typically do not simultaneously address the following issues: the significant hardware resource and storage overhead associated with actual high-rate oversampling; the introduction of switching delays if eye diagram analysis and memory test vector execution are sequentially alternated; and the potential for over-response to random noise in phase calibration and down-frequency control without statistical gating, leading to sampling phase oscillations or premature down-frequency reduction.

[0004] Therefore, there is a need for an adaptive memory testing scheme that can obtain high equivalent resolution eye diagrams at lower actual oversampling rates and form a low-latency closed loop for eye diagram analysis, test vector execution, phase compensation, downsampling control, and fault reporting. Summary of the Invention

[0005] This invention aims to at least solve one of the technical problems existing in related technologies. To this end, this invention provides an adaptive memory testing method, system, device, product, and medium to achieve phase compensation and fault reporting of memory.

[0006] This invention provides an adaptive memory testing method, comprising: S1: Determine the data bit rate, determine the sampling rate based on the data bit rate, and perform sampling according to the sampling rate to obtain the edge phase; S2: Determine the edge magnitude, perform matrix mapping on the edge magnitude and edge phase to obtain a two-dimensional probability eye diagram, construct a fully convolutional inference network, and improve the resolution of the two-dimensional probability eye diagram through the fully convolutional inference network to obtain a refined two-dimensional probability eye diagram; S3: Calculate the zero probability and standard deviation of the error of the fine two-dimensional probability eye diagram. Determine the eye diagram adjustment strategy by using the zero probability and standard deviation of the error. Use the eye diagram adjustment strategy to adjust the fine two-dimensional probability eye diagram to obtain the adjusted two-dimensional probability eye diagram. S4: Calculate the timing margin and eye diagram offset of the adjusted two-dimensional probabilistic eye diagram, determine the compensation strategy based on the timing margin and eye diagram offset, and obtain the phase compensation amount based on the compensation strategy. S5: Determine the regional fault probability based on the selected compensation strategy, generate fault region markers based on the regional fault probability, generate a memory fault report based on the fault region markers and phase compensation amount, and complete the memory test.

[0007] According to an adaptive memory testing method provided by the present invention, in step S1, the sampling rate is 4 to 8.

[0008] According to the adaptive memory testing method provided by the present invention, step S2 further includes: S21: Determine the edge amplitude, normalize the edge amplitude to obtain a normalized amplitude, and normalize the edge phase to obtain a normalized phase; S22: Map the normalized amplitude and normalized phase to obtain an eye diagram matrix, determine the data window length, perform forgetting counting on the eye diagram matrix according to the data window length to obtain the phase amplitude probability value, and obtain a two-dimensional probability eye diagram through the phase amplitude probability value. S23: Construct the fully convolutional inference network including continuous convolutional layers, upsampling layers, and normalized output layers. Improve the resolution of the two-dimensional probability eye map through the fully convolutional inference network to obtain a refined two-dimensional probability eye map.

[0009] According to the adaptive memory testing method provided by the present invention, step S3 further includes: S31: Obtain the total number of edges and the total number of phase hits in the fine two-dimensional probability eye diagram within the sliding window, and calculate the probability that the error is zero based on the total number of edges and the total number of phase hits; S32: Calculate the total phase deviation distance, and calculate the offset standard deviation based on the total phase deviation distance; S33: Determine eye diagram adjustment strategies, including adjustment by standard deviation, frequency reduction standby, and extended data window, based on the probability of zero error and the standard deviation of the offset. Use the eye diagram adjustment strategies to adjust the fine two-dimensional probability eye diagram to obtain the adjusted two-dimensional probability eye diagram.

[0010] According to an adaptive memory testing method provided by the present invention, in step S4, an adjustment boundary value is determined, and a compensation strategy including compensation based on offset and compensation based on adjustment boundary value is determined. When the compensation strategy based on offset is adopted, the eye diagram offset is used as the phase compensation amount. When the compensation strategy based on adjustment boundary value is adopted, the adjustment boundary value is used as the phase compensation amount.

[0011] According to the adaptive memory testing method provided by the present invention, in step S5, the number of regional failures is obtained according to the selection of the compensation strategy, the regional failure probability is obtained through the number of regional failures, and a failure probability threshold is determined. When the regional failure probability is greater than the failure probability threshold, a failure region mark is generated for the storage region.

[0012] The present invention also provides an adaptive memory testing system, comprising: Phase measurement module: used to determine the data bit rate, determine the sampling rate based on the data bit rate, and perform sampling according to the sampling rate to obtain the edge phase; The fine-grained 2D probability eye diagram module is used to determine the edge magnitude, perform matrix mapping on the edge magnitude and edge phase to obtain the 2D probability eye diagram, construct a fully convolutional inference network, and improve the resolution of the 2D probability eye diagram through the fully convolutional inference network to obtain the fine-grained 2D probability eye diagram. Eye diagram adjustment module: used to calculate the zero probability of error and the standard deviation of offset of fine two-dimensional probability eye diagram, determine the eye diagram adjustment strategy by the zero probability of error and the standard deviation of offset, and use the eye diagram adjustment strategy to adjust the fine two-dimensional probability eye diagram to obtain the adjusted two-dimensional probability eye diagram; Phase compensation module: used to calculate the timing margin and eye diagram offset of the two-dimensional probabilistic eye diagram, determine the compensation strategy based on the timing margin and eye diagram offset, and obtain the phase compensation amount based on the compensation strategy; Fault Area Marking Module: Used to determine the fault probability of a region based on the selected compensation strategy, generate fault area markers based on the fault area probability, generate a memory fault report based on the fault area markers and phase compensation amount, and complete the memory test.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of an adaptive memory testing method as described above.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of an adaptive memory testing method as described above.

[0015] The present invention also provides a computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, which, when executed by a computer, enable the computer to perform the steps of any of the above-described adaptive memory testing methods.

[0016] The above-described one or more technical solutions in the embodiments of the present invention have at least one of the following technical effects: This invention provides an adaptive memory testing method, system, device, product, and medium. It selects a suitable adjustment strategy for a refined two-dimensional probabilistic eye diagram by using a zero-error probability and offset standard deviation, thereby adjusting the resolution of the two-dimensional probabilistic eye diagram using a fully convolutional inference network to improve its accuracy. Subsequently, a compensation strategy is determined based on timing margin and eye diagram offset to adjust the memory phase. Furthermore, fault regions are marked, effectively addressing and reporting memory faults.

[0017] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating an adaptive memory testing method provided by the present invention.

[0020] Figure 2 This is a schematic diagram of the structure of an adaptive memory testing system provided by the present invention.

[0021] Figure 3 This is a schematic diagram of the structure of an adaptive memory testing device provided by the present invention.

[0022] Figure label: 100. Phase measurement module; 200. Fine two-dimensional probabilistic eye diagram module; 300. Eye diagram adjustment module; 400. Phase compensation module; 500. Fault area marking module; 810. Processor; 820. Communication interface; 830. Memory; 840. Communication bus. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention. The following embodiments are used to illustrate this invention but cannot be used to limit the scope of this invention.

[0024] In the description of the embodiments of the present invention, it should be noted that the terms "first", "second" and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0025] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0026] The following is combined Figures 1 to 3 Specific embodiments of the present invention are described below. Figure 1 A flowchart illustrating an adaptive memory testing method provided by the present invention includes: S1: Determine the data bit rate, determine the sampling rate based on the data bit rate, and perform sampling according to the sampling rate to obtain the edge phase; Furthermore, the purpose of this stage is to sample and obtain the edge phase. Specifically, in step S1, the sampling rate is 4 to 8.

[0027] The specific implementation method for the above steps in this embodiment is as follows: Here, during the testing of the memory under test, the oversampling unit first needs to determine the data bit rate. Based on the data bit rate and experience, a sampling multiplier of 4 to 8 can be determined. Then, the read data or interface return data of a certain region of the memory is sampled according to the product of the data bit rate and the sampling multiplier, resulting in a sampled data set. Subsequently, when a sign flip occurs between two adjacent sampling points in the sampled data set, these two sampling points are taken as edge sampling points, allowing the calculation of the edge phase of the edge sampling points. : in, This represents the phase of the earlier edge sampling point among the two edge sampling points. The phase difference between the two edge sampling points. The decision level is 0 in this embodiment because the standard for determining the edge sampling point is the occurrence of sign flipping. This represents the level of the first edge sampling point out of the two edge sampling points. It represents the level of the latter edge sampling point among the two edge sampling points.

[0028] S2: Determine the edge magnitude, perform matrix mapping on the edge magnitude and edge phase to obtain a two-dimensional probability eye diagram, construct a fully convolutional inference network, and improve the resolution of the two-dimensional probability eye diagram through the fully convolutional inference network to obtain a refined two-dimensional probability eye diagram; Furthermore, the objective of this stage is to obtain a two-dimensional probability eye diagram, and then to improve the resolution of the two-dimensional probability eye diagram to obtain a refined two-dimensional probability eye diagram. Specifically, step S2 further includes: S21: Determine the edge amplitude, normalize the edge amplitude to obtain a normalized amplitude, and normalize the edge phase to obtain a normalized phase; S22: Map the normalized amplitude and normalized phase to obtain an eye diagram matrix, determine the data window length, perform forgetting counting on the eye diagram matrix according to the data window length to obtain the phase amplitude probability value, and obtain a two-dimensional probability eye diagram through the phase amplitude probability value. S23: Construct the fully convolutional inference network including continuous convolutional layers, upsampling layers, and normalized output layers. Improve the resolution of the two-dimensional probability eye map through the fully convolutional inference network to obtain a refined two-dimensional probability eye map.

[0029] The specific implementation method for the above steps in this embodiment is as follows: First, the edge amplitude needs to be determined, which is the amplitude of the decision level. Then, the edge amplitude is normalized, mapping it to a value range of 0-1 to obtain the normalized amplitude. When normalizing the phase, the edge phase needs to be normalized, mapping it to a value range of 0-1 to obtain the normalized phase, which is then used as the phase of the normalized amplitude. Next, the normalized amplitude and normalized phase are mapped: the normalized amplitude is mapped to a pre-constructed amplitude matrix, and the normalized phase is mapped to a pre-constructed phase matrix. The phase matrix and amplitude matrix form the eye diagram matrix. Then, the time length for acquiring the normalized amplitude and normalized phase is determined, i.e., the data window length. After accumulating a large amount of data within the data window length, a probability value can be obtained by updating the eye diagram matrix using a forgetting count method. This probability value is the probability that the amplitude of the normalized level corresponding to a normalized phase value is the edge amplitude; this is called the phase amplitude probability value. A two-dimensional probability eye diagram can be obtained by plotting the normalized phase value on the horizontal axis and the phase amplitude probability value on the vertical axis.

[0030] Finally, a fully convolutional inference network is constructed, consisting of continuous convolutional layers, upsampling layers, and normalized output layers. This network employs fixed-point quantized weights, line-buffered convolutions, and a pipelined multiply-accumulate structure to reduce on-chip storage and multiplier usage. The two-dimensional probability eye diagram is input into the fully convolutional inference network, which then performs resolution enhancement based on the data characteristics of the eye diagram, increasing the resolution by 4-8 times to obtain a refined two-dimensional probability eye diagram. In the refined two-dimensional probability eye diagram, the phase amplitude probability curve is one where the phase amplitude probability value is highest at a certain phase, and gradually decreases as the phase moves to either side.

[0031] S3: Calculate the zero probability and standard deviation of the error of the fine two-dimensional probability eye diagram. Determine the eye diagram adjustment strategy by using the zero probability and standard deviation of the error. Use the eye diagram adjustment strategy to adjust the fine two-dimensional probability eye diagram to obtain the adjusted two-dimensional probability eye diagram. Furthermore, the objective of this stage is to calculate the probability of zero error and the standard deviation of the offset, thereby using an eye diagram adjustment strategy to adjust the refined two-dimensional probability eye diagram, resulting in an adjusted two-dimensional probability eye diagram. Specifically, step S3 further includes: S31: Obtain the total number of edges and the total number of phase hits in the fine two-dimensional probability eye diagram within the sliding window, and calculate the probability that the error is zero based on the total number of edges and the total number of phase hits; S32: Calculate the total phase deviation distance, and calculate the offset standard deviation based on the total phase deviation distance; S33: Determine eye diagram adjustment strategies, including adjustment by standard deviation, frequency reduction standby, and extended data window, based on the probability of zero error and the standard deviation of the offset. Use the eye diagram adjustment strategies to adjust the fine two-dimensional probability eye diagram to obtain the adjusted two-dimensional probability eye diagram.

[0032] The specific implementation method for the above steps in this embodiment is as follows: First, a variable-length sliding window needs to be determined based on the amount of data in the eye diagram matrix; here, the sliding window is a time length. Then, the total number of edges and the total number of phase hits within the sliding window are obtained for the refined two-dimensional probabilistic eye diagram. The total number of edges represents the total number of sign flips between two adjacent sampling points within the sliding window. The total number of phase hits represents the total number of times the normalized phase, when the normalized level is equal to the normalized decision level, has a phase amplitude probability value higher than a preset minimum probability threshold within the sliding window. Dividing the total number of phase hits by the total number of edges yields the probability of zero error.

[0033] Next, the total phase deviation distance is calculated. Here, we first need to calculate the phase deviation distance of the fine two-dimensional probability eye diagram when the sign flips between two sampling points. : in, In a refined two-dimensional probability eye diagram, the phase amplitude probability value to the left of the highest point of the phase amplitude probability curve is equal to the preset minimum probability threshold. In a refined two-dimensional probability eye diagram, the phase amplitude probability value to the right of the highest point of the phase amplitude probability curve is equal to the preset minimum probability threshold. The total phase deviation distance is obtained by summing the phase deviation distances when sign flips occur between any two adjacent sampling points in the sliding window.

[0034] Then, the average phase deviation distance is obtained by calculating the total phase deviation distance over each sign flip between two adjacent sampling points. This allows us to calculate the standard deviation of the offset. : in, This represents the number of sign flips between two adjacent sampling points within the sliding window. ( ) is used to sum the contents within the parentheses.

[0035] Finally, eye diagram adjustment strategies, including adjustment by standard deviation, frequency reduction and standby, and extended data window, are determined based on the probability of zero error and the standard deviation offset. Here, when the probability of zero error is higher than a preset threshold and the standard deviation offset is lower than a preset threshold, the eye diagram adjustment strategy based on standard deviation is adopted. This means that the phase corresponding to the highest point of the phase amplitude probability curve in the refined two-dimensional probabilistic eye diagram is changed to the original phase plus the standard deviation offset, thereby shifting the phase amplitude probability curve as a whole. When the probability of zero error is lower than a preset threshold and the standard deviation offset is higher than a preset threshold, the eye diagram adjustment strategy of frequency reduction and standby is adopted. This means that the error of the refined two-dimensional probabilistic eye diagram is considered large at this time, requiring manual intervention to acquire new data to obtain a new refined two-dimensional probabilistic eye diagram, and a new eye diagram adjustment strategy is selected for the new refined two-dimensional probabilistic eye diagram. When the probability of zero error and the standard deviation of deviation do not fall into either of the above two categories, an eye diagram adjustment strategy of extending the data window is adopted. This involves increasing the amount of data in the eye diagram matrix to obtain a new, refined two-dimensional probability eye diagram. The eye diagram adjustment strategy is then re-selected for this new refined two-dimensional probability eye diagram. By adjusting the refined two-dimensional probability eye diagram using the above adjustment strategy, the adjusted two-dimensional probability eye diagram is obtained.

[0036] S4: Calculate the timing margin and eye diagram offset of the adjusted two-dimensional probabilistic eye diagram, determine the compensation strategy based on the timing margin and eye diagram offset, and obtain the phase compensation amount based on the compensation strategy. Furthermore, the objective of this stage is to determine a compensation strategy based on timing margin and eye diagram offset, and to obtain the phase compensation amount based on the compensation strategy. Specifically, in step S4, an adjustment boundary value is determined, and a compensation strategy including compensation based on offset and compensation based on adjustment boundary value is determined. When the compensation strategy based on offset is adopted, the eye diagram offset is used as the phase compensation amount; when the compensation strategy based on adjustment boundary value is adopted, the adjustment boundary value is used as the phase compensation amount.

[0037] The specific implementation method for the above steps in this embodiment is as follows: This requires calculating the timing margin and eye diagram offset for adjusting the 2D probabilistic eye diagram. Calculating the timing margin involves identifying the phase corresponding to the highest point of the phase amplitude probability curve in the adjusted 2D probabilistic eye diagram and the phase at which the phase amplitude probability value equals a preset minimum probability threshold. The phase difference between these two phases is then calculated, and the minimum of these phase differences is taken as the timing margin. Calculating the eye diagram offset involves calculating the phase difference between the phase corresponding to the highest point of the phase amplitude probability curve in the adjusted 2D probabilistic eye diagram and the phase corresponding to the highest point of the phase amplitude probability curve in the refined 2D probabilistic eye diagram, thus obtaining the eye diagram offset.

[0038] Subsequently, a compensation strategy is determined, and the phase compensation amount is obtained based on the compensation strategy. Specifically, the compensation strategy includes compensation based on offset and compensation based on adjustment boundary values. When the timing margin is greater than a preset timing margin threshold and the eye diagram offset is less than a preset eye diagram offset threshold, the compensation strategy based on offset is adopted, that is, the eye diagram offset is used as the phase compensation amount for that region of the memory; otherwise, the compensation strategy based on adjustment boundary values ​​is adopted, that is, the adjustment boundary value is predetermined and used as the phase compensation amount for that region of the memory.

[0039] S5: Determine the regional fault probability based on the selected compensation strategy, generate fault region markers based on the regional fault probability, generate a memory fault report based on the fault region markers and phase compensation amount, and complete the memory test.

[0040] Furthermore, the objective of this stage is to generate fault area markers, generate memory fault reports based on the fault area markers and phase compensation amounts, and complete the memory test. Specifically, in step S5, the number of area faults is obtained based on the selected compensation strategy, the area fault probability is obtained from the number of area faults, and a fault probability threshold is determined. When the area fault probability is greater than the fault probability threshold, a fault area marker is generated for the memory area.

[0041] The specific implementation method for the above steps in this embodiment is as follows: Here, a fault counter with an initial count of 0 needs to be constructed first. Then, the methods recorded in steps S1 to S4 are used to obtain multiple phase compensation values ​​for the same area of ​​the memory at different times, and the corresponding compensation strategies for these phase compensation values ​​are determined. Thus, when the compensation strategy corresponding to a phase compensation value is compensation based on the adjustment boundary value, the value of the fault counter is incremented by 1. When the compensation strategy corresponding to a phase compensation value is compensation based on the offset, the phase compensation value is considered to be caused by normal random error, and the fault counter does not activate. In this way, the reading of the fault counter can be used as the number of area faults.

[0042] The region failure probability is then obtained by dividing the number of region failures by the number of phase compensation values ​​for the same region at different times. A failure probability threshold is then determined. This threshold is based on the memory's importance, storage reliability requirements, and data corruption tolerance. Higher reliability requirements or lower data corruption tolerance result in a lower failure probability threshold. When the region failure probability exceeds the threshold, a failure region marker is generated for the storage region corresponding to the phase compensation value. Finally, the failure region markers and phase compensation values ​​for each region of the memory are obtained using the above method. A memory failure report is then generated based on these markers and values. In the report, the failure region marker indicates that a failure has occurred in the storage region corresponding to the phase compensation value. The average of multiple phase compensation values ​​is used to compensate for the phase deviation of the storage level in the corresponding region. This completes the memory testing.

[0043] The following describes an adaptive memory testing apparatus provided by the present invention. The adaptive memory testing apparatus described below and the adaptive memory testing method described above can be referred to in correspondence.

[0044] Figure 2 An example is a schematic diagram of the structure of an adaptive memory testing system, such as... Figure 2 As shown, an adaptive memory testing method for performing the above-described method includes: Phase measurement module 100: used to determine the data bit rate, determine the sampling rate based on the data bit rate, and perform sampling based on the sampling rate to obtain the edge phase; The fine two-dimensional probabilistic eye diagram module 200 is used to determine the edge magnitude, perform matrix mapping on the edge magnitude and edge phase to obtain the two-dimensional probabilistic eye diagram, construct a fully convolutional inference network, and improve the resolution of the two-dimensional probabilistic eye diagram through the fully convolutional inference network to obtain the fine two-dimensional probabilistic eye diagram. Eye diagram adjustment module 300: used to calculate the zero probability of error and the standard deviation of offset of the fine two-dimensional probability eye diagram, determine the eye diagram adjustment strategy by the zero probability of error and the standard deviation of offset, and use the eye diagram adjustment strategy to adjust the fine two-dimensional probability eye diagram to obtain the adjusted two-dimensional probability eye diagram; Phase compensation module 400: used to calculate the timing margin and eye diagram offset of the two-dimensional probabilistic eye diagram, determine the compensation strategy based on the timing margin and eye diagram offset, and obtain the phase compensation amount based on the compensation strategy. Fault Area Marking Module 500: Used to determine the fault probability of a region based on the selection of the compensation strategy, generate fault area markers based on the fault area markers and the phase compensation amount, and generate a memory fault report to complete the memory test.

[0045] Figure 3 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 3 As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 can call a computer program in the memory 830 to execute an adaptive memory testing method, which includes: S1: Determine the data bit rate, determine the sampling rate based on the data bit rate, and perform sampling according to the sampling rate to obtain the edge phase; S2: Determine the edge magnitude, perform matrix mapping on the edge magnitude and edge phase to obtain a two-dimensional probability eye diagram, construct a fully convolutional inference network, and improve the resolution of the two-dimensional probability eye diagram through the fully convolutional inference network to obtain a refined two-dimensional probability eye diagram; S3: Calculate the zero probability and standard deviation of the error of the fine two-dimensional probability eye diagram. Determine the eye diagram adjustment strategy by using the zero probability and standard deviation of the error. Use the eye diagram adjustment strategy to adjust the fine two-dimensional probability eye diagram to obtain the adjusted two-dimensional probability eye diagram. S4: Calculate the timing margin and eye diagram offset of the adjusted two-dimensional probabilistic eye diagram, determine the compensation strategy based on the timing margin and eye diagram offset, and obtain the phase compensation amount based on the compensation strategy. S5: Determine the regional fault probability based on the selected compensation strategy, generate fault region markers based on the regional fault probability, generate a memory fault report based on the fault region markers and phase compensation amount, and complete the memory test.

[0046] Furthermore, when the computer program in the aforementioned memory 830 can be implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0047] On the other hand, the present invention also provides a computer program product, the computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, wherein when the program instructions are executed by a computer, the computer is able to execute an adaptive memory testing method provided by the above methods, the method comprising: S1: Determine the data bit rate, determine the sampling rate based on the data bit rate, and perform sampling according to the sampling rate to obtain the edge phase; S2: Determine the edge magnitude, perform matrix mapping on the edge magnitude and edge phase to obtain a two-dimensional probability eye diagram, construct a fully convolutional inference network, and improve the resolution of the two-dimensional probability eye diagram through the fully convolutional inference network to obtain a refined two-dimensional probability eye diagram; S3: Calculate the zero probability and standard deviation of the error of the fine two-dimensional probability eye diagram. Determine the eye diagram adjustment strategy by using the zero probability and standard deviation of the error. Use the eye diagram adjustment strategy to adjust the fine two-dimensional probability eye diagram to obtain the adjusted two-dimensional probability eye diagram. S4: Calculate the timing margin and eye diagram offset of the adjusted two-dimensional probabilistic eye diagram, determine the compensation strategy based on the timing margin and eye diagram offset, and obtain the phase compensation amount based on the compensation strategy. S5: Determine the regional fault probability based on the selected compensation strategy, generate fault region markers based on the regional fault probability, generate a memory fault report based on the fault region markers and phase compensation amount, and complete the memory test.

[0048] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the aforementioned adaptive memory testing methods, the method comprising: S1: Determine the data bit rate, determine the sampling rate based on the data bit rate, and perform sampling according to the sampling rate to obtain the edge phase; S2: Determine the edge magnitude, perform matrix mapping on the edge magnitude and edge phase to obtain a two-dimensional probability eye diagram, construct a fully convolutional inference network, and improve the resolution of the two-dimensional probability eye diagram through the fully convolutional inference network to obtain a refined two-dimensional probability eye diagram; S3: Calculate the zero probability and standard deviation of the error of the fine two-dimensional probability eye diagram. Determine the eye diagram adjustment strategy by using the zero probability and standard deviation of the error. Use the eye diagram adjustment strategy to adjust the fine two-dimensional probability eye diagram to obtain the adjusted two-dimensional probability eye diagram. S4: Calculate the timing margin and eye diagram offset of the adjusted two-dimensional probabilistic eye diagram, determine the compensation strategy based on the timing margin and eye diagram offset, and obtain the phase compensation amount based on the compensation strategy. S5: Determine the regional fault probability based on the selected compensation strategy, generate fault region markers based on the regional fault probability, generate a memory fault report based on the fault region markers and phase compensation amount, and complete the memory test.

[0049] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0050] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0051] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An adaptive memory testing method, characterized in that, include: S1: Determine the data bit rate, determine the sampling rate based on the data bit rate, and perform sampling according to the sampling rate to obtain the edge phase; S2: Determine the edge magnitude, perform matrix mapping on the edge magnitude and edge phase to obtain a two-dimensional probability eye diagram, construct a fully convolutional inference network, and improve the resolution of the two-dimensional probability eye diagram through the fully convolutional inference network to obtain a refined two-dimensional probability eye diagram; S3: Calculate the zero probability and standard deviation of the error of the fine two-dimensional probability eye diagram. Determine the eye diagram adjustment strategy by using the zero probability and standard deviation of the error. Use the eye diagram adjustment strategy to adjust the fine two-dimensional probability eye diagram to obtain the adjusted two-dimensional probability eye diagram. S4: Calculate the timing margin and eye diagram offset of the adjusted two-dimensional probabilistic eye diagram, determine the compensation strategy based on the timing margin and eye diagram offset, and obtain the phase compensation amount based on the compensation strategy. S5: Determine the regional fault probability based on the selected compensation strategy, generate fault region markers based on the regional fault probability, generate a memory fault report based on the fault region markers and phase compensation amount, and complete the memory test.

2. The adaptive memory testing method according to claim 1, characterized in that, In step S1, the sampling factor is 4 to 8.

3. The adaptive memory testing method according to claim 1, characterized in that, Step S2 further includes: S21: Determine the edge amplitude, normalize the edge amplitude to obtain a normalized amplitude, and normalize the edge phase to obtain a normalized phase; S22: Map the normalized amplitude and normalized phase to obtain an eye diagram matrix, determine the data window length, perform forgetting counting on the eye diagram matrix according to the data window length to obtain the phase amplitude probability value, and obtain a two-dimensional probability eye diagram through the phase amplitude probability value. S23: Construct the fully convolutional inference network including continuous convolutional layers, upsampling layers, and normalized output layers. Improve the resolution of the two-dimensional probability eye map through the fully convolutional inference network to obtain a refined two-dimensional probability eye map.

4. The adaptive memory testing method according to claim 1, characterized in that, Step S3 further includes: S31: Obtain the total number of edges and the total number of phase hits in the fine two-dimensional probability eye diagram within the sliding window, and calculate the probability that the error is zero based on the total number of edges and the total number of phase hits; S32: Calculate the total phase deviation distance, and calculate the offset standard deviation based on the total phase deviation distance; S33: Determine eye diagram adjustment strategies, including adjustment by standard deviation, frequency reduction standby, and extended data window, based on the probability of zero error and the standard deviation of the offset. Use the eye diagram adjustment strategies to adjust the fine two-dimensional probability eye diagram to obtain the adjusted two-dimensional probability eye diagram.

5. The adaptive memory testing method according to claim 1, characterized in that, In step S4, the adjustment boundary value is determined, and a compensation strategy including compensation based on offset and compensation based on adjustment boundary value is determined. When the compensation strategy based on offset is adopted, the eye diagram offset is used as the phase compensation amount. When the compensation strategy based on adjustment boundary value is adopted, the adjustment boundary value is used as the phase compensation amount.

6. The adaptive memory testing method according to claim 1, characterized in that, In step S5, the number of regional failures is obtained based on the selection of the compensation strategy, the regional failure probability is obtained through the number of regional failures, and the failure probability threshold is determined. When the regional failure probability is greater than the failure probability threshold, a failure region marker is generated for the storage region.

7. An adaptive memory testing system for performing an adaptive memory testing method as described in any one of claims 1 to 6, characterized in that, include: Phase measurement module: used to determine the data bit rate, determine the sampling rate based on the data bit rate, and perform sampling according to the sampling rate to obtain the edge phase; The fine-grained 2D probability eye diagram module is used to determine the edge magnitude, perform matrix mapping on the edge magnitude and edge phase to obtain the 2D probability eye diagram, construct a fully convolutional inference network, and improve the resolution of the 2D probability eye diagram through the fully convolutional inference network to obtain the fine-grained 2D probability eye diagram. Eye diagram adjustment module: used to calculate the zero probability of error and the standard deviation of offset of fine two-dimensional probability eye diagram, determine the eye diagram adjustment strategy by the zero probability of error and the standard deviation of offset, and use the eye diagram adjustment strategy to adjust the fine two-dimensional probability eye diagram to obtain the adjusted two-dimensional probability eye diagram; Phase compensation module: used to calculate the timing margin and eye diagram offset of the two-dimensional probabilistic eye diagram, determine the compensation strategy based on the timing margin and eye diagram offset, and obtain the phase compensation amount based on the compensation strategy; Fault Area Marking Module: Used to determine the fault probability of a region based on the selected compensation strategy, generate fault area markers based on the fault area probability, generate a memory fault report based on the fault area markers and phase compensation amount, and complete the memory test.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements an adaptive memory testing method as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements an adaptive memory testing method as described in any one of claims 1 to 6.

10. A computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, characterized in that, When the program instructions are executed by the computer, the computer is able to perform an adaptive memory testing method as described in any one of claims 1 to 6.

Images