Memory fault diagnosis method, system, device, storage medium and program product

Abstract

The present invention provides a memory fault diagnosis method that can be applied to the field of fault diagnosis technology. The memory fault diagnosis method includes: extracting features from memory operation information within a preset time period based on time information within the preset time period to obtain an operation feature sequence, wherein the operation feature sequence includes at least one operation feature sorted according to the time information; performing an initial fault diagnosis on the memory using the operation feature sequence to obtain an initial diagnostic result; determining a diagnostic strategy feature sequence for the memory based on the initial diagnostic result and the operation feature sequence; and performing a fault diagnosis on the memory using the diagnostic strategy feature sequence to obtain a target diagnostic result. The present invention also provides a memory fault diagnosis system, device, storage medium, and program product.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis, and more particularly to a memory fault diagnosis method, system, device, storage medium and program product. Background Art

[0002] With the rapid development of information technology, servers are taking on important business computing and data storage tasks. As one of the core components of servers, the stability and reliability of memory modules are crucial to overall system performance and business continuity.

[0003] However, in actual operating environments, memory is prone to various types of failures due to factors such as memory manufacturing processes, operating environment conditions, load fluctuations, temperature changes, and device aging. Related technologies struggle to provide dynamic feedback based on the actual operating conditions of the memory, reducing the accuracy of memory fault diagnosis. Summary of the Invention

[0004] In view of the above problems, the present invention provides a memory fault diagnosis method, system, device, storage medium and program product.

[0005] According to a first aspect of the present invention, a memory fault diagnosis method is provided, comprising: performing feature extraction on memory operation information within a preset time period based on time information within a preset time period to obtain an operation feature sequence, the operation feature sequence including at least one operation feature sorted according to the time information; performing initial fault diagnosis on the memory using the operation feature sequence to obtain an initial diagnosis result; determining a diagnostic strategy feature sequence for the memory based on the initial diagnosis result and the operation feature sequence; and performing fault diagnosis on the memory using the diagnostic strategy feature sequence to obtain a target diagnosis result.

[0006] According to a second aspect of the present invention, a memory fault diagnosis system is provided, which is deployed on a server and includes: a collection device for collecting memory operation information within a preset time period; and a memory fault diagnosis device for implementing the steps of the above-mentioned memory fault diagnosis method during execution.

[0007] The third aspect of the present invention provides a memory fault diagnosis device, including: a feature extraction module, used to extract features of memory operation information within a preset time period based on time information within the preset time period, and obtain an operation feature sequence, the operation feature sequence including at least one operation feature sorted according to the time information; a first fault diagnosis module, used to perform initial fault diagnosis on the memory using the operation feature sequence, and obtain an initial diagnosis result; a first determination module, used to determine the diagnostic strategy feature sequence of the memory based on the initial diagnosis result and the operation feature sequence; a second fault diagnosis module, used to perform fault diagnosis on the memory using the diagnostic strategy feature sequence, and obtain a target diagnosis result.

[0008] The fourth aspect of the present invention provides an electronic device, comprising: one or more device processors; a memory for storing one or more computer programs, wherein the one or more device processors execute the one or more computer programs to implement the steps of the above-mentioned memory fault diagnosis method.

[0009] The fifth aspect of the present invention further provides a computer-readable storage medium having a computer program or instructions stored thereon, which implements the steps of the above-mentioned memory fault diagnosis method when the computer program or instructions are executed by the device processor.

[0010] The sixth aspect of the present invention further provides a computer program product, comprising a computer program or instructions, which implement the steps of the above-mentioned memory fault diagnosis method when executed by a device processor.

[0011] According to an embodiment of the present invention, based on the time information within the preset time period, feature extraction is performed on the memory operation information within the preset time period to obtain an operation feature sequence, so the operation feature sequence has time dimension information. The operation feature sequence is used to perform initial fault diagnosis on the memory to obtain an initial diagnostic result; then, based on the initial fault result and the operation feature sequence, the diagnostic strategy feature sequence of the memory is determined. Therefore, the diagnostic strategy feature sequence includes the initial fault result that is dynamically fed back based on the actual memory operation information and the memory operation information with time dimension information. The diagnostic strategy feature is used to perform fault diagnosis on the memory to obtain the target fault result, so the trend of the memory operation information evolving over time and the dynamic feedback information are used for fault diagnosis, thereby improving the accuracy of memory fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The above contents and other objects, features and advantages of the present invention will become more apparent through the following description of the embodiments of the present invention with reference to the accompanying drawings, in which:

[0013] Figure 1 A diagram showing an application scenario of a memory fault diagnosis method according to an embodiment of the present invention is shown;

[0014] Figure 2 A flow chart of a memory fault diagnosis method according to an embodiment of the present invention is shown;

[0015] Figure 3A A schematic diagram of a processor slot according to an embodiment of the present invention is shown;

[0016] Figure 3B A schematic diagram of a first dual in-line memory module slot according to an embodiment of the present invention is shown;

[0017] Figure 4A schematic diagram showing a memory fault diagnosis method according to another embodiment of the present invention is shown;

[0018] Figure 5A Shows a structural block diagram of a memory fault diagnosis system according to an embodiment of the present invention;

[0019] Figure 5B It shows a structural block diagram of a server according to an embodiment of the present invention;

[0020] Figure 5C It shows a structural block diagram of a memory fault diagnosis device according to an embodiment of the present invention;

[0021] Figure 6 A block diagram of an electronic device suitable for implementing a memory fault diagnosis method according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0022] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present invention. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of embodiments of the present invention. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concept of the present invention.

[0023] The terms used herein are only for describing specific embodiments and are not intended to limit the present invention. The terms "comprise", "include", etc. used herein indicate the presence of the features, steps, operations and / or components, but do not exclude the presence or addition of one or more other features, steps, operations or components.

[0024] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0025] When expressions such as "at least one of A, B, and C, etc." are used, they should generally be interpreted in accordance with the meaning commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0026] With the rapid development of information technology, servers are taking on important business computing and data storage tasks. As one of the core components of servers, the stability and reliability of memory modules are crucial to overall system performance and business continuity.

[0027] However, in actual operating environments, memory is prone to various types of failures due to factors such as memory manufacturing processes, operating environment conditions, load fluctuations, temperature changes, and device aging. Related technologies struggle to provide dynamic feedback based on the actual operating conditions of the memory, reducing the accuracy of memory fault diagnosis.

[0028] An embodiment of the present invention provides a memory fault diagnosis method including: performing feature extraction on memory operation information within a preset time period based on time information within a preset time period to obtain an operation feature sequence, the operation feature sequence including at least one operation feature sorted according to the time information; performing initial fault diagnosis on the memory using the operation feature sequence to obtain an initial diagnosis result; determining a diagnostic strategy feature sequence of the memory based on the initial diagnosis result and the operation feature sequence; and performing fault diagnosis on the memory using the diagnostic strategy feature sequence to obtain a target diagnosis result.

[0029] Figure 1 A diagram showing an application scenario of a memory fault diagnosis method according to an embodiment of the present invention is shown.

[0030] like Figure 1 As shown, the application scenario 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 is used as a medium for providing a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired or wireless communication links or optical fiber cables.

[0031] A user may use a first terminal device 101, a second terminal device 102, or a third terminal device 103 to interact with a server 105 via a network 104 to receive or send messages, etc. Various communication client applications may be installed on the first terminal device 101, the second terminal device 102, or the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (for example only).

[0032] The first terminal device 101 , the second terminal device 102 , and the third terminal device 103 may be various electronic devices having display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

[0033] Server 105 can be a server that provides various services, such as a background management server (for example only) that supports memory fault diagnosis requests issued by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103. The background management server can process the received user memory fault diagnosis request and obtain memory-related information to be diagnosed, including a preset time period and memory operation information. Based on the time information within the preset time period, the server can extract features from the memory operation information within the preset time period to obtain an operation feature sequence. The server can use the operation feature sequence to perform an initial fault diagnosis on the memory to obtain an initial diagnosis result. Based on the initial diagnosis result and the operation feature sequence, the server can determine a diagnostic strategy feature sequence for the memory. The server can use the diagnostic strategy feature sequence to perform fault diagnosis on the memory to obtain a target diagnostic result. The server can then feed the target diagnostic result (e.g., a web page, information, or data obtained or generated based on the user request) back to the terminal device.

[0034] It should be noted that the memory fault diagnosis method provided in the embodiment of the present invention can generally be executed by the server 105. Accordingly, the memory fault diagnosis device provided in the embodiment of the present invention can generally be set in the server 105. The memory fault diagnosis method provided in the embodiment of the present invention can also be executed by a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Accordingly, the memory fault diagnosis device provided in the embodiment of the present invention can also be set in a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105.

[0035] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only . According to the implementation requirements, there can be any number of terminal devices, networks and servers.

[0036] Figure 2 A flow chart of a memory fault diagnosis method according to an embodiment of the present invention is shown.

[0037] like Figure 2 As shown, the memory fault diagnosis method of this embodiment includes operations S210 to S240, and the memory fault diagnosis method can be executed by a server.

[0038] In operation S210 , feature extraction is performed on memory operation information within the preset time period according to time information within the preset time period to obtain an operation feature sequence.

[0039] According to an embodiment of the present invention, the preset time period may be a time window for extracting memory operation information, for example, the time window may be 5 minutes, 15 minutes, 30 minutes, 1 hour, 6 hours, 24 hours, 48 ​​hours, 72 hours, 7 days, etc.

[0040] According to an embodiment of the present invention, the time information may include the start time to the end time of the time window. For example, the preset time period may be 5 minutes, the start time is 08:00, and the end time is 08:05. The time information may include any time between 08:00 and 08:05.

[0041] According to embodiments of the present invention, memory operation information can represent the operating status of the memory. For example, memory operation information can include memory log information, memory read information, memory usage, memory environment, etc. After obtaining memory operation information, data cleaning and preprocessing can be performed to improve data quality.

[0042] According to an embodiment of the present invention, an operation feature sequence includes at least one operation feature sorted by time. For example, the operation feature sequence Ht may be {H1, H2, H3, ..., Hn}. H1, H2, H3, ..., Hn may be sorted from earliest to latest time. H1, H2, H3, ..., Hn may respectively include error behavior, memory usage, memory environment, and historical memory failure information.

[0043] For example, by calculating the average value of memory operation information within a preset time period based on time information within the preset time period, long-term trends in memory operation can be captured. For example, the average value of memory usage within the preset time period can be used. Operation characteristics can be determined based on the average value of memory usage within the preset time period.

[0044] For example, the maximum and minimum memory operation values ​​can be obtained from the memory operation information within a preset time period. Based on the maximum and minimum memory operation values, the memory operation variation within the preset time period can be obtained, and the stability characteristics of the memory operation can be extracted.

[0045] For example, the preset time period can be 24 hours. A time series decomposition method can be used to decompose the memory operation information within the preset time period into a trend component, a seasonal component, and a residual component. The trend component can reflect whether the memory operation information has a long-term growth or decline trend. The seasonal component can reflect whether there are regular fluctuations in the memory operation information. The residual component can indicate that the memory operation information has no obvious regularity or pattern. The operation characteristics can be determined based on the trend, seasonality, and residual components.

[0046] In operation S220 , an initial fault diagnosis is performed on the memory using the running feature sequence to obtain an initial diagnosis result.

[0047] For example, the operating characteristics can be derived based on an average value of memory operating information within a preset time period. The memory operating information at each time period can be compared with the average value to obtain a target number of operating characteristics greater than an average threshold. If the target number is greater than the threshold, the initial fault result can be determined as a memory failure within the preset time period. If the target number is less than the threshold, the initial fault result can be determined as no memory failure within the preset time period.

[0048] For example, the operational characteristics may be stability characteristics. A machine learning algorithm (such as a random forest or deep learning model) is used to classify the stability characteristics to determine the memory fault type corresponding to the stability characteristics. An initial diagnostic result is determined based on the memory fault type.

[0049] For example, the trend, seasonality, and residual components can be used as running features and fed into a classification algorithm (such as a decision tree, random forest, or support vector machine) to obtain initial memory diagnostic results. If the residual component exhibits significant abnormal fluctuations, anomaly detection algorithms can be used to further detect faults. Isolation Forest is a tree-based algorithm that isolates outliers by randomly selecting features and partitioning the data. Anomaly detection algorithms can include Isolation Forest and One-Class Support Vector Machine (SVM). One-Class SVM is based on a support vector machine algorithm that considers residual components that exceed the normal sample boundary as anomalies.

[0050] In operation S230 , a stored diagnostic strategy feature sequence is determined based on the initial diagnostic result and the operating feature sequence.

[0051] According to an embodiment of the present invention, each fault type corresponds to a different quantization value. For example, the fault type may be an error correcting code (ECC) check error in the memory, and the quantization value may be 2; the fault type may be an uncorrectable error (UE) in the memory, and the quantization value may be 1; the fault type may be a correctable error (CE) in the memory, and the quantization value may be 0.

[0052] For example, the initial diagnosis result and the operation feature sequence can be spliced ​​together to obtain a diagnosis strategy feature sequence.

[0053] For example, the memory is repaired according to the initial diagnosis result to obtain a repair result, and the initial diagnosis result, the repair result and the operation feature sequence are spliced ​​together to obtain a diagnosis strategy feature sequence.

[0054] For example, model parameters in a machine learning algorithm related to an initial diagnosis result are determined, and the initial diagnosis result, the model parameters, and the running feature sequence are concatenated to obtain a diagnostic strategy feature sequence.

[0055] In operation S240 , a fault diagnosis is performed on the memory using the diagnostic strategy feature sequence to obtain a target diagnostic result.

[0056] According to an embodiment of the present invention, a diagnosis strategy feature sequence may be used to determine a fault diagnosis strategy.

[0057] For example, the diagnostic strategy feature sequence is matched with reference diagnostic strategy feature sequences of multiple fault diagnostic strategies in a database to obtain a target fault diagnostic strategy that matches the diagnostic strategy feature sequence. The target fault diagnostic strategy is used to diagnose the memory fault and obtain a target diagnostic result.

[0058] For example, the diagnostic strategy feature sequence can be input into a strategy classification algorithm (such as a decision tree, random forest, or support vector machine) to obtain a target fault diagnosis strategy. The target fault diagnosis strategy can be achieved by adjusting model parameters based on the initial diagnostic results, or by selecting a target feature extraction method from a variety of optional feature extraction methods (such as average values, stability features, etc.). While the specific form of the target fault diagnosis strategy is not limited, it should be noted that the target fault diagnosis strategy is used to adjust the fault diagnosis process and improve the accuracy of fault diagnosis based on feedback from operations S210 to S230.

[0059] According to an embodiment of the present invention, based on the time information within the preset time period, feature extraction is performed on the memory operation information within the preset time period to obtain an operation feature sequence, so the operation feature sequence has time dimension information. The operation feature sequence is used to perform initial fault diagnosis on the memory to obtain an initial diagnostic result; then, based on the initial fault result and the operation feature sequence, the diagnostic strategy feature sequence of the memory is determined. Therefore, the diagnostic strategy feature sequence includes the initial fault result that is dynamically fed back based on the actual memory operation information and the memory operation information with time dimension information. The diagnostic strategy feature is used to perform fault diagnosis on the memory to obtain the target fault result, thereby utilizing the trend of the memory operation information evolving over time and the dynamic feedback information for fault diagnosis, thereby improving the accuracy of memory fault diagnosis.

[0060] According to an embodiment of the present invention, the memory operation information includes at least one of the following: historical fault diagnosis information of the memory, electrical environment information, operating system access information, operating system operation load information, and memory fault events recorded in the operating system log.

[0061] According to an embodiment of the present invention, operating system access information, operating system running load information, memory failure events recorded in the operating system log and historical fault diagnosis information can be obtained from the operating system through periodic scanning, real-time monitoring of data interfaces, etc.

[0062] Memory failure events occurring within a preset time period are extracted from the operating system log. Memory failure events can include the failure type (e.g., ECC, UE, CE), the memory address where the failure occurred, and the time of occurrence. Multiple memory failure events can be sorted by occurrence time to record a time series of failure evolution. For example, memory failure events can include the number of UE or CE failures, the growth rate, the hotspot row number, the error entropy, the propagation speed, and the error density.

[0063] Operational load information can include the server's service processor utilization, memory usage, the time the service processor is waiting for input and output operations to complete, I / O (input / output) activity related to switching, task queue length, etc.

[0064] Operating system access information can include current hot rows, page hit rates, read / write bandwidth, cache miss rates, access mode transitions, etc. It can also construct memory row heat maps and behavior graphs to assist in spatial anomaly detection.

[0065] Because faults often have an evolutionary path, memory operation information can include historical fault diagnosis information and memory fault events recorded in logs. Based on the time information within a preset time period, feature extraction can be performed on the memory operation information within the preset time period to obtain an operation feature sequence. This operation feature sequence captures the fault's evolutionary trend over time.

[0066] Historical fault diagnosis information can include historical fault diagnosis process information, historical diagnosis results, historical repair results, etc. Historical diagnosis results can include whether they are consistent with the actual fault (such as the fault location and fault type); historical repair results include whether actual repair operations (such as page migration) were triggered and the repair time.

[0067] According to an embodiment of the present invention, power environment information can be obtained by a baseboard management controller. The power environment information may include the temperature, voltage, current, power consumption slope, etc. of the DIMM (Dual Inline Memory Module) at the address where the memory failure event occurred.

[0068] According to an embodiment of the present invention, memory operation information can have multi-dimensional information, such as historical fault diagnosis information of the memory, electrical environment information, operating system access information, operating system operation load information, and memory fault events recorded in the operating system log. Therefore, the operation feature sequence can have the trend of evolution of faults, electrical environment, operating system access, operating load, etc. over time. Fault diagnosis is performed using the operation feature sequence, which improves the reliability of memory fault diagnosis.

[0069] According to an embodiment of the present invention, the above method also includes: obtaining memory failure events within a preset time period from the log; locating the memory address where the memory failure event occurs according to the preset access path of the operating system, and the memory operation information also includes the memory address; obtaining the electrical environment information at the memory address.

[0070] According to an embodiment of the present invention, a memory address may be a preset access path set according to a processor socket (CPU Socket), a memory channel (Channel), a dual in-line memory module slot, a memory array level (Bank), a chip (Chip), and a row / column (Row / Column).

[0071] According to an embodiment of the present invention, the minimum unit of the memory address where the memory fault event occurs can be located step by step according to a preset access path.

[0072] Figure 3A A schematic diagram of a processor slot according to an embodiment of the present invention is shown.

[0073] like Figure 3A As shown, the processor slot is the interface on the motherboard for inserting the service processor. The server supports multiple processor slots, each of which houses a multi-core service processor. Each processor slot integrates one or more integrated memory controllers (IMCs). It interacts with system memory through the first and second memory channels. In a non-uniform memory access (NUMA) architecture, the memory and I / O resources directly controlled by each processor slot constitute a NUMA node. In actual deployments, a NUMA node consists of a processor slot, the first and second DIMM slots on the first memory channel, and the third and fourth DIMM slots on the second memory channel.

[0074] Figure 3B A schematic diagram of a first dual in-line memory module slot according to an embodiment of the present invention is shown.

[0075] like Figure 3B As shown, each memory array level of the first dual in-line memory module slot has multiple chips, such as chip 1, chip 2, chip 3, chip 4, and chip 5. The memory array level is typically organized in a two-dimensional array, such as chip 5 being divided into rows and columns. Row errors, column errors, byte errors, and the like may occur on chip 5.

[0076] According to an embodiment of the present invention, historical fault diagnosis information at the memory address, operating system access information related to the memory address, and operating system operation load information can be obtained, and the memory operation data obtained is related to the memory fault event, which improves the data quality and reduces the data volume of the memory operation information, thereby improving the computing performance of the server.

[0077] According to an embodiment of the present invention, memory failure events within a preset time period are obtained from a log; the memory address where the memory failure event occurs is located according to a preset access path of the operating system; and then the electrical environment information at the memory address is obtained. The memory operation information thus obtained is information related to the memory failure event, which can reduce the data volume of the memory operation information, improve the computing performance of the server, and improve the response timeliness of fault diagnosis.

[0078] According to an embodiment of the present invention, based on the time information within the preset time period, feature extraction is performed on the memory operation information within the preset time period to obtain an operation feature sequence, including: determining attribute statistical features based on the time information and multiple attribute data in the memory operation information; vectorizing the memory operation information based on the time information to obtain an operation status feature sequence; and splicing the attribute statistical features and the operation status feature sequence to obtain an operation feature sequence.

[0079] According to an embodiment of the present invention, attribute statistical features can represent changes in attribute data within a preset time period. For example, attribute statistical features can include the maximum value, minimum value, extreme value, variance, average value, cumulative value, etc. of the attribute data within the preset time period.

[0080] According to an embodiment of the present invention, the vectorized memory operation information is sorted according to time information to obtain an operation status feature sequence. The operation status feature sequence includes at least one operation status feature, which can be obtained by splicing multiple attribute dimensions. The operation status feature can represent the instantaneous operation status of the memory.

[0081] In order to enhance the model's ability to perceive the fault evolution trend, the memory operation information is vectorized according to the time information to obtain the operation status feature sequence. For example, the preset time period includes a total of w moments, which can be t-w+1, t-w+2 to t. The memory operation information can include n attribute data. The memory operation information of w moments is spliced ​​together to obtain the memory operation information sequence within the preset time period. .

[0082] (1);

[0083] w is the number of moments in the preset time period, and n is the number of attribute data. The memory operation information is vectorized according to the time information to obtain a vector feature sequence. The vector feature sequence is expanded into a one-dimensional feature vector to obtain the operation status feature sequence , R wn Represents a w×n dimensional vector space. The running state feature sequence includes at least one running state feature, for example, the running state feature sequence Including running status characteristics vec(S t-w+1 )、vec(S t-w+2 )……vec(S t ).

[0084] (2);

[0085] For example, there are two operating status features within a preset time period, and the memory operation information may include two attribute data. The vector feature sequence is shown in formula (3):

[0086] (3);

[0087] The vector feature sequence Expanded into a one-dimensional feature vector is the running state feature sequence vec( ), vec( )=[10 20 12 22].

[0088] According to an embodiment of the present invention, the attribute statistical features are concatenated with the operation status feature sequence expanded into a one-dimensional feature vector to obtain an operation feature sequence.

[0089] According to an embodiment of the present invention, attribute statistical features are determined based on multiple attribute data in time information and memory operation information; the attribute statistical features and the operation status feature sequence are spliced ​​to obtain an operation feature sequence, so that the operation feature sequence has a trend of attribute data changing over time, and thus the operation feature sequence is used to perform initial fault diagnosis, thereby improving the accuracy of the initial diagnosis results.

[0090] According to an embodiment of the present invention, the attribute statistical features include at least one of the following: a gradient feature that characterizes the operating change trend of the memory within a preset time period, a fluctuation feature that characterizes the fluctuation of the attribute data within a preset time period, and an adjacent state feature that characterizes the change trend of the attribute data at adjacent moments.

[0091] According to an embodiment of the present invention, the gradient feature k t The formula is as follows:

[0092] (4);

[0093] Where i represents the i-th moment in the preset time period, , .

[0094] According to an embodiment of the present invention, the fluctuation characteristics The formula is as follows:

[0095] (5);

[0096] According to an embodiment of the present invention, the proximity state feature The formula is as follows:

[0097] (6);

[0098] ε can be 0.0001 to avoid the denominator being 0.

[0099] For example, attribute statistics It can include gradient features, fluctuation features and proximity state features. Attribute statistical features are one-dimensional feature vectors. and operating state feature sequence Splice and get the running feature sequence X t ,Right now .

[0100] According to an embodiment of the present invention, the three attribute statistical features, namely, the gradient feature that characterizes the operating change trend of the memory within a preset time period, the fluctuation feature that characterizes the fluctuation of the attribute data within the preset time period, and the proximity feature that characterizes the changing trend of the attribute data at adjacent moments, can reflect the changes in the memory operation within the preset time period and improve the reliability of the fault diagnosis results.

[0101] According to an embodiment of the present invention, an initial fault diagnosis is performed on the memory using an operation feature sequence to obtain an initial diagnosis result, including: time encoding the operation features according to the order of the operation features in the operation feature sequence to obtain a time feature; embedding the time feature into a preset position of the operation feature to obtain a time perception feature; processing the time perception feature using a multi-head self-attention mechanism to obtain a target attention feature; and performing an initial fault diagnosis on the memory using a target attention feature sequence to obtain an initial diagnosis result, wherein the target attention feature sequence includes at least one target attention feature.

[0102] According to an embodiment of the present invention, when the sequence number of the running feature in the running feature sequence is an even number, that is, i is an even number, the even time feature is as follows:

[0103] (7);

[0104] When the sequence number of the running feature in the running feature sequence is odd, that is, i is an odd number, the odd time feature is as follows:

[0105] (8);

[0106] In formula (7) and formula (8), j is any positive integer, d h =Z / h, where Z is the dimension of the running feature sequence and h is the number of attention heads in the multi-head attention mechanism. The temporal features PE include even-numbered temporal features and odd-numbered temporal features.

[0107] According to an embodiment of the present invention, the preset position can be the head or tail of the running state feature in the running feature. The time feature PE is embedded in the preset position of the running feature to obtain a time-aware feature sequence. as follows:

[0108] (9);

[0109] According to an embodiment of the present invention, the operating features are temporally encoded based on their order in the operating feature sequence to obtain a time feature; the time feature is then embedded into a preset position within the operating feature to obtain a time-aware feature. The time-aware feature is processed using a multi-head self-attention mechanism to obtain a target attention feature. This allows the multi-head self-attention mechanism to extract global dependency features between moments, meaning that the target attention feature represents the global characteristics of memory operation within a preset time period. Using the target attention feature sequence to perform initial fault diagnosis on the memory and obtain an initial diagnostic result can reduce the influence of local information on the diagnosis and improve the accuracy of the initial diagnostic result.

[0110] According to an embodiment of the present invention, a multi-head self-attention mechanism is used to process time perception features to obtain target attention features, including: using a multi-head self-attention mechanism to process time perception features to obtain initial self-attention features; normalizing the time perception features and the initial self-attention features to obtain intermediate attention features; and inputting the intermediate attention features into a position feedforward network to obtain target attention features.

[0111] According to an embodiment of the present invention, a multi-head attention mechanism is used to map time-aware features into query (Q), key (K), and value (V). The formula is as follows:

[0112] (10);

[0113] W Q 、W K 、W V are the weight matrices for query, key, and value, respectively. The weight matrices are trainable parameters.

[0114] The output of the i-th attention mechanism is calculated based on Q, K, and V. The formula is as follows:

[0115] (11);

[0116] QK T is the dot product between the query and the key, The output of the single-head attention mechanism is obtained by the activation function (such as the softmax function). i .

[0117] The initial self-attention feature MultiHead(X) is as follows:

[0118] (12);

[0119] W O is the output weight matrix, which maintains the stability of the multi-head attention mechanism during training. The output of the first attention mechanism head1 to the output of the h-th attention mechanism headh ​​are concatenated to obtain the initial self-attention feature.

[0120] LayerNorm is a normalization function, time-aware feature sequence After concatenating with the initial self-attention feature MultiHead(X), the feature sequence Z1 to be normalized is obtained.

[0121] (13);

[0122] The feature sequence to be normalized is input into the position feedforward network (FFN) to obtain the feedback feature sequence FFN (Z1).

[0123] (14);

[0124] W1 and W2 are the network weight matrices of the position feedforward network, b1 and b2 are the two biases of the position feedforward network, and ReLU is the activation function.

[0125] After concatenating the feedback feature sequence FFN (Z1) and the feature sequence to be normalized Z1, normalization is performed to obtain the target attention feature sequence, which includes at least one target attention feature. The target attention feature sequence H t The formula is as follows:

[0126] (15);

[0127] The above steps constitute a block. For memory fault diagnosis, it can include three blocks.

[0128] According to an embodiment of the present invention, the target attention feature sequence is input into the fault prediction model to obtain an initial diagnosis result. For example, the fault prediction model can be a decision tree, and the output of the decision tree is:

[0129] (16);

[0130] is the mth decision tree, represents the probability of memory failure within the next △t time, f GBDT Represents a decision tree function.

[0131] The training objective of the decision tree can be the total loss Reached minimum value.

[0132] (17);

[0133] represents the binary cross entropy loss, is the first regularization coefficient, is the second regularization coefficient, Represents a sample pair (u, v) in the set β, which is used to calculate the difference loss between the sample pairs. are the squares of the Euclidean norms of the network weight matrices W1 and W2, respectively. is the output obtained by inputting sample u into the mth decision tree, is the output of inputting sample v into the mth decision tree. Sample u and sample v correspond to the same memory address. For example, sample u and sample v correspond to the same row or column.

[0134] According to an embodiment of the present invention, a multi-head self-attention mechanism is used to process time-aware features to generate initial self-attention features. This captures the diverse information of time-aware features and enhances the expressiveness of fault prediction models used for fault diagnosis. The multi-attention mechanism utilizes parallel computation, improving computational efficiency. The time-aware features and initial self-attention features are normalized to generate intermediate attention features. These intermediate attention features are then fed into a position feedforward network to generate target attention features. This allows the intermediate attention features to undergo independent nonlinear transformations at each position, improving the stability of the fault prediction model.

[0135] According to an embodiment of the present invention, the memory operation information includes historical fault diagnosis information of the memory, and the historical fault diagnosis information includes at least one of the following: historical repair results corresponding to historical target prediction strategies and historical risk thresholds.

[0136] For example, the start time of the preset time period is t, and the historical risk threshold can be the risk threshold set by the fault prediction model in the fault diagnosis at time t-1. The historical repair result corresponding to the historical target prediction strategy can be the repair result of the fault at time t-1, where the repair result can be 1 for successful repair and 0 for failed repair.

[0137] Exemplarily, the memory operation information may further include a change rate of the number of memory fault events within a preset time period, a change rate of electrical environment information within a preset time period, and a change rate of memory access frequency within a preset time period.

[0138] According to an embodiment of the present invention, determining a diagnostic strategy feature sequence in memory according to an initial diagnostic result and an operation feature sequence includes: determining a diagnostic strategy feature sequence according to a time perception feature sequence and an initial diagnostic result.

[0139] According to an embodiment of the present invention, the time perception feature is concatenated with the initial diagnosis result to obtain a diagnosis strategy feature sequence.

[0140] For example, the diagnostic strategy feature sequence G t The formula is as follows:

[0141] (18);

[0142] According to an embodiment of the present invention, the diagnostic strategy feature sequence may further include a risk threshold for initial fault diagnosis, a repair method for initial fault diagnosis, the number of erroneous behaviors earlier than a preset time period, a conversion rate of the electrical environment, and the like.

[0143] The risk threshold of initial fault diagnosis and the repair method of initial fault diagnosis can provide real feedback on the process of initial fault diagnosis so as to determine the appropriate target prediction strategy, thereby enabling the fault detection model to quickly adapt to changes in scenarios.

[0144] The number of erroneous behaviors earlier than the preset time period, the conversion rate of the electrical environment, etc., can provide a more comprehensive analysis of the operating status information, and can determine the reliability of the memory operating data at the beginning of the preset time period, thereby improving the data quality of the diagnostic strategy characteristics and improving the computing performance.

[0145] According to an embodiment of the present invention, a memory diagnostic strategy feature sequence is determined based on a time-aware feature sequence and initial diagnostic results. Specifically, the diagnostic strategy feature sequence dynamically feeds back initial fault results obtained based on actual memory operation information, providing both temporal evolution trend information and feedback information about memory faults.

[0146] According to an embodiment of the present invention, a diagnostic strategy feature sequence is used to perform fault diagnosis on a memory to obtain a target diagnostic result, including: inputting the diagnostic strategy feature sequence into a strategy prediction model to obtain a target prediction strategy, the strategy prediction model is trained based on the reward function values ​​of multiple prediction strategies of a fault prediction model corresponding to a historical diagnostic strategy feature sequence, the fault prediction model is used to perform initial fault diagnosis on the memory using the operating feature sequence to obtain an initial diagnostic result; and using the fault prediction model to perform fault diagnosis on the memory according to the target prediction strategy to obtain a target diagnostic result.

[0147] According to an embodiment of the present invention, the strategy prediction model may be a multi-layer perceptron network.

[0148] For example, the strategy prediction model may adopt a three-layer multi-layer perceptron.

[0149] (19);

[0150] Diagnostic strategy feature sequence G t Input function of 3-layer multilayer perceptron (MLP) network , the dimension of the input layer is consistent with the state space, the hidden layer can be 128-dimensional, and the output layer is the corresponding prediction strategy A t The size of the space, the dimension of the output vector can be determined according to the number of prediction strategies. The probability of multiple prediction strategies being selected is output through the softmax function. .

[0151] According to an embodiment of the present invention, a fault prediction model and a strategy prediction model are used in combination. An initial diagnostic result is obtained by inputting an operating feature sequence into the fault prediction model. The initial diagnostic result is concatenated with the operating feature sequence to obtain a diagnostic strategy feature. The diagnostic strategy feature is then input into the strategy prediction model to obtain a target prediction strategy. This provides a predictive feedback mechanism within the fault diagnosis process. Therefore, in dynamic scenarios (such as unstable loads or sudden increases in memory temperature), the parameters of the fault prediction model can be adaptively adjusted to accommodate a variety of scenario changes, thereby improving the accuracy of fault diagnosis in different scenarios.

[0152] Figure 4 A schematic diagram of a memory fault diagnosis method according to an embodiment of the present invention is shown.

[0153] like Figure 4 As shown, the target attention feature sequence is input into the fault prediction model 410 to obtain an initial diagnosis result. The initial diagnosis result is concatenated with the time perception feature sequence to obtain a diagnosis strategy feature sequence 420. Diagnosis strategy feature sequence 420 is input into the strategy prediction model 430 to obtain a target prediction strategy. Using the fault prediction model 410, memory faults are diagnosed according to the target prediction strategy, resulting in a target diagnosis result 440.

[0154] According to an embodiment of the present invention, the prediction strategy includes one of the following: using a fault prediction model to process a target attention feature sequence to obtain a target diagnosis result; determining a target operation feature subsequence corresponding to preset memory operation information from the operation feature sequence; updating the risk threshold of the fault prediction model based on the target operation feature subsequence, and using the updated fault prediction model to perform fault diagnosis on the memory, the preset memory operation information includes the electrical environment information and operating system access information of the operating system, and the risk threshold is used to compare with the output value of the fault prediction model so as to determine the diagnosis result of the memory based on the comparison result; according to a preset time length, determining the target memory operation information for fault diagnosis from the memory operation information within a preset time period; using a repair method for scheduling the storage space of the memory; temporarily not performing any operation.

[0155] According to an embodiment of the present invention, the prediction strategy may involve the data dimension of the input fault prediction model, whether the fault modification module needs to be linked, memory operation information used for feature extraction, etc.

[0156] The prediction strategy is shown in Table 1.

[0157] Table 1

[0158]

[0159] According to an embodiment of the present invention, the preset time length may be 10 minutes, and the time length within the preset time period may be 15 minutes. For example, target memory operation information within 10 minutes for fault diagnosis is determined from the memory operation information within the preset time period.

[0160] A repair method of the storage space of the scheduling memory is adopted. For example, the repair method of the storage space of the scheduling memory may be to execute memory page migration or adjust the storage capacity of the memory in a linked manner.

[0161] Temporarily not performing any operation means not triggering any behavior, that is, observing state.

[0162] According to embodiments of the present invention, prediction strategies can include factors such as the data dimensions input to the fault prediction model, whether a fault correction module needs to be linked, and memory operation information used for feature extraction. These strategies can be adjustments to the initial fault diagnosis process. Utilizing a fault prediction model and following a targeted prediction strategy for memory fault diagnosis results in more accurate results, improving the robustness of the fault prediction model.

[0163] According to an embodiment of the present invention, the strategy prediction model is trained based on the following method: using the fault prediction model to process the historical diagnosis strategy features according to multiple prediction strategies to obtain the reward function values ​​of multiple prediction strategies; determining the target reward value based on the multiple reward function values; updating the parameters of the pre-trained strategy prediction model based on the target reward value until the target reward value reaches the preset reward value to obtain the strategy prediction model.

[0164] According to an embodiment of the present invention, prediction strategy A t The reward function value R t+1 The formula is as follows:

[0165] (20);

[0166] Indicates whether the diagnosis result obtained according to the prediction strategy is the same as the actual fault result, 1 if they are the same, and 0 if they are not the same. Indicates whether the diagnosis result obtained according to the prediction strategy has an error (memory address, fault type, etc.). If so, it is 1, otherwise it is 0. Indicates the resource cost consumed in executing the prediction strategy. Indicates the repair result according to the diagnosis result obtained by the prediction strategy. If the repair is successful, it is 1, otherwise it is 0. α1, α2, α3 and α4 are 、 、 、 The preset reward weights.

[0167] The preset reward value can be within the time step, the prediction strategy At The target reward value reaches the maximum value. The formula of the target reward value J(θ) is as follows:

[0168] (twenty one);

[0169] γ is the discount factor (γ can be 0.9), T is the time step of the strategy prediction model, is the expected function, π is the ratio of circumference to circumference, and the parameter of the policy prediction model is θ. By continuously updating the parameter θ so that J(θ) reaches its maximum value, the policy prediction model is obtained.

[0170] According to an embodiment of the present invention, a fault prediction model is used to process historical diagnostic strategy features according to multiple prediction strategies to obtain reward function values ​​of multiple prediction strategies; a target reward value is determined based on the multiple reward function values; the parameters of the pre-trained strategy prediction model are updated based on the target reward value until the target reward value reaches the preset reward value, thereby obtaining a strategy prediction model, and realizing the training of the strategy prediction model using the output results of the fault prediction model, so that the strategy prediction model is more in line with the scenario of the fault prediction model, thereby improving the accuracy of the memory diagnosis results.

[0171] Figure 5A Shown is a structural block diagram of a memory fault diagnosis system according to an embodiment of the present invention.

[0172] The memory fault diagnosis system 500 can be deployed on the server 105 and can include a collection device 510 and a memory fault diagnosis device 520. The collection device 510 is used to collect memory operation information within a preset time period. The memory fault diagnosis device 520 is used to implement the steps of the memory fault diagnosis method when executed.

[0173] Figure 5B A structural block diagram of a server according to an embodiment of the present invention is shown.

[0174] According to an embodiment of the present invention, the server 105 may include a baseboard management controller 1051, a memory 1052, an operating system 1053 and a service processor 1054. The baseboard management controller 1051 is connected to the memory controller to monitor the electrical environment information of the memory 1052; the memory 1052 is connected to the operating system 1053, and the operating system 1053 is used to schedule the storage space of the memory 1052.

[0175] According to an embodiment of the present invention, the memory 1052 is connected to the service processor 1054 of the server 105 via a memory channel. The server 105 may include multiple service processors 1054.

[0176] According to an embodiment of the present invention, the collection device 510 is used to: collect electrical environment information of the memory 1052 from the baseboard management controller 1051; collect operating system access information, operating system 1053 operation load information, and memory fault events recorded in the log of the operating system 1053 from the operating system 1053; collect historical fault diagnosis information of the memory 1052 from the memory fault diagnosis device 520, and the memory operation information includes at least one of the following: historical fault diagnosis information, electrical environment information, operating system access information, operation load information, and memory fault events.

[0177] For example, you can use performance analysis tools to obtain operating system access information. This information can include current hot rows, page hit rates, read and write bandwidth utilization, etc. Performance analysis tools can use perf (PerformanceEvents) event sampling.

[0178] For example, the collection device 510 can collect memory failure events using tools for monitoring and logging hardware errors. Tools for monitoring and logging hardware errors can include mcelog (Machine Check Exception Logging), the EDAC (Error Detection and Correction) framework, and RasDaemon (Reliability, Availability, and Serviceability Daemon). Memory failure events can be scanned periodically or monitored in real time through kernel interfaces.

[0179] For example, the acquisition device 510 can access the temperature, voltage, current, and power consumption data of the dual in-line memory modules under load through the IPMI (Intelligent Platform Management Interface). The acquisition device 510 can also access the serial presence detect (SPD) element of each dual in-line memory module through the system management bus to obtain the module temperature and calibrated electrical parameters.

[0180] For example, the collection device 510 can obtain the model parameters corresponding to the historical target prediction strategy in the historical fault diagnosis information through the internal log of the memory fault diagnosis device 520, obtain the historical risk threshold through the early warning linkage system, and obtain the historical repair results corresponding to the historical target prediction strategy through the basic input and output system.

[0181] According to an embodiment of the present invention, the acquisition device 510 may include a plurality of acquisition probes, which are respectively deployed in the baseboard management controller 1051 , the operating system 1053 , and the memory fault diagnosis device 520 .

[0182] Figure 5C A structural block diagram of a memory fault diagnosis device according to an embodiment of the present invention is shown.

[0183] like Figure 5C As shown, the memory fault diagnosis device 520 of this embodiment includes a feature extraction module 521 , a first fault diagnosis module 522 , a first determination module 523 and a second fault diagnosis module 524 .

[0184] Feature extraction module 521 is configured to extract features from memory operation information within a preset time period based on the time information within the preset time period, thereby obtaining an operation feature sequence. The operation feature sequence includes at least one operation feature sorted by the time information. In one embodiment, feature extraction module 521 can be configured to perform operation S210 described above, and will not be further described here.

[0185] The first fault diagnosis module 522 is used to perform initial fault diagnosis on the memory using the running feature sequence to obtain an initial diagnosis result. In one embodiment, the first fault diagnosis module 522 can be used to perform the operation S220 described above, which will not be repeated here.

[0186] The first determining module 523 is used to determine the stored diagnostic strategy feature sequence according to the initial diagnostic result and the operation feature sequence. In one embodiment, the first determining module 523 can be used to perform the operation S230 described above, which will not be repeated here.

[0187] The second fault diagnosis module 524 is used to perform fault diagnosis on the memory using the diagnostic strategy feature sequence to obtain a target diagnostic result. In one embodiment, the second fault diagnosis module 524 can be used to perform the operation S240 described above, which will not be repeated here.

[0188] According to an embodiment of the present invention, the first fault diagnosis module 522 includes a time encoding submodule, an embedding submodule, a processing submodule, and a first fault diagnosis submodule. The time encoding submodule is configured to time-encode the operation features according to their order in the operation feature sequence to obtain time features; the embedding submodule is configured to embed time features at preset positions in the operation features to obtain time-aware features; the processing submodule is configured to process the time-aware features using a multi-head self-attention mechanism to obtain target attention features; and the first fault diagnosis submodule is configured to perform initial fault diagnosis on the memory using the target attention feature sequence to obtain an initial diagnosis result, wherein the target attention feature sequence includes at least one target attention feature.

[0189] According to an embodiment of the present invention, the processing submodule includes a processing unit, a normalization unit, and an input unit. The processing unit is used to process the time perception features using a multi-head self-attention mechanism to obtain initial self-attention features; the normalization unit is used to normalize the time perception features and the initial self-attention features to obtain intermediate attention features; and the input unit is used to input the intermediate attention features into the position feedforward network to obtain the target attention features.

[0190] According to an embodiment of the present invention, the first determining module 523 includes a first determining submodule. The first determining submodule is configured to determine a diagnosis strategy feature sequence according to the time perception feature sequence and the initial diagnosis result.

[0191] According to an embodiment of the present invention, the apparatus further includes a first acquisition module, a positioning module, and a second acquisition module. The first acquisition module is configured to acquire memory failure events within a preset time period from a log; the positioning module is configured to locate the memory address where the memory failure event occurred based on a preset access path of the operating system, wherein the memory operation information also includes the memory address; and the second acquisition module is configured to acquire electrical environment information at the memory address.

[0192] According to an embodiment of the present invention, the feature extraction module 521 includes a second determination submodule, a vector processing submodule, and a splicing submodule. The second determination submodule is configured to determine attribute statistical features based on time information and multiple attribute data in the memory operation information; the vector processing submodule is configured to vectorize the memory operation information based on the time information to obtain an operation status feature sequence; and the splicing submodule is configured to splice the attribute statistical features with the operation status feature sequence to obtain an operation feature sequence.

[0193] According to an embodiment of the present invention, the second fault diagnosis module 524 includes an input submodule and a second fault diagnosis submodule. The input submodule is used to input the diagnostic strategy feature sequence into the strategy prediction model to obtain a target prediction strategy. The strategy prediction model is trained based on the reward function values ​​of multiple prediction strategies of the fault prediction model corresponding to the historical diagnostic strategy feature sequence. The fault prediction model is used to perform initial fault diagnosis on the memory using the operating feature sequence to obtain an initial diagnosis result. The second fault diagnosis submodule is used to use the fault prediction model to perform fault diagnosis on the memory according to the target prediction strategy to obtain a target diagnosis result.

[0194] According to an embodiment of the present invention, any multiple modules among the feature extraction module 521, the first fault diagnosis module 522, the first determination module 523, and the second fault diagnosis module 524 may be combined into a single module, or any one of these modules may be split into multiple modules. Alternatively, at least part of the functionality of one or more of these modules may be combined with at least part of the functionality of other modules and implemented in a single module. According to an embodiment of the present invention, at least one of the feature extraction module 521, the first fault diagnosis module 522, the first determination module 523, and the second fault diagnosis module 524 may be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application-specific integrated circuit (ASIC), or may be implemented in hardware or firmware through any other reasonable means of circuit integration or packaging, or implemented in any one of software, hardware, and firmware, or any suitable combination of these. Alternatively, at least one of the feature extraction module 521 , the first fault diagnosis module 522 , the first determination module 523 and the second fault diagnosis module 524 may be at least partially implemented as a computer program module, which may perform corresponding functions when executed.

[0195] Figure 6 A block diagram of an electronic device suitable for implementing a memory fault diagnosis method according to an embodiment of the present invention is shown.

[0196] like Figure 6 As shown, an electronic device 600 according to an embodiment of the present invention includes a device processor 601, which can perform various appropriate actions and processes based on programs stored in a read-only memory (ROM) 602 or programs loaded from a storage unit 608 into a random access memory (RAM) 603. The device processor 601 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or related chipsets and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The device processor 601 may also include onboard memory for caching purposes. The device processor 601 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0197] RAM 603 stores various programs and data required for the operation of electronic device 600. Device processor 601, ROM 602, and RAM 603 are interconnected via bus 604. Device processor 601 executes the programs stored in ROM 602 and / or RAM 603 to perform various operations according to the method flow of the embodiment of the present invention. It should be noted that the programs may also be stored in one or more memories other than ROM 602 and RAM 603. Device processor 601 may also execute the programs stored in the one or more memories to perform various operations according to the method flow of the embodiment of the present invention.

[0198] According to an embodiment of the present invention, electronic device 600 may further include an input / output (I / O) interface 605, which is also connected to bus 604. Electronic device 600 may also include one or more of the following components connected to I / O interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including devices such as a cathode ray tube (CRT), liquid crystal display (LCD), and speakers; a storage section 608 including a hard disk; and a communication section 609 including a network interface card such as a LAN card or modem. Communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to I / O interface 605 as needed. Removable media 611, such as a magnetic disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed in drive 610 as needed, so that computer programs read from the removable media can be installed into storage section 608 as needed.

[0199] The present invention also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments, or may exist independently and not incorporated into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of the present invention.

[0200] According to an embodiment of the present invention, a computer-readable storage medium may be a non-volatile computer-readable storage medium, and may include, for example, but not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to an embodiment of the present invention, a computer-readable storage medium may include the ROM 602 and / or RAM 603 described above, and / or one or more memories other than ROM 602 and RAM 603.

[0201] Embodiments of the present invention also include a computer program product, which includes a computer program containing program code for executing the method shown in the flowchart. When the computer program product is executed in a computer system, the program code is used to enable the computer system to implement the memory fault diagnosis method provided by the embodiment of the present invention.

[0202] The computer program executes the above functions defined in the system / apparatus of the embodiment of the present invention when executed by the device processor 601. According to the embodiment of the present invention, the above-described system, apparatus, modules, units, etc. can be implemented by computer program modules.

[0203] In one embodiment, the computer program may be stored on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may be transmitted and distributed in the form of a signal on a network medium, downloaded and installed via the communication portion 609, and / or installed from a removable medium 611. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0204] In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 609 and / or installed from a removable medium 611. When the computer program is executed by the device processor 601, the aforementioned functions defined in the system of the embodiment of the present invention are performed. According to the embodiment of the present invention, the systems, devices, apparatuses, modules, units, etc. described above can be implemented by computer program modules.

[0205] According to an embodiment of the present invention, the program code for executing the computer program provided by the embodiment of the present invention can be written in any combination of one or more programming languages. Specifically, these computer programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C" or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect via the Internet).

[0206] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the above-mentioned module, program segment, or a part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0207] It will be understood by those skilled in the art that the features described in the various embodiments of the present invention may be combined and / or coupled in various ways, even if such combinations or couplings are not explicitly described in the present invention. In particular, the features described in the various embodiments of the present invention may be combined and / or coupled in various ways without departing from the spirit and teachings of the present invention. All such combinations and / or couplings fall within the scope of the present invention.

[0208] The above describes embodiments of the present invention. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of the present invention. Although each embodiment has been described separately above, this does not mean that the measures in each embodiment cannot be advantageously used in combination. Without departing from the scope of the present invention, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present invention.

Claims

1. A memory fault diagnosis method, characterized in that: The memory fault diagnosis method comprises: Extracting features of the memory operation information within the preset time period according to time information within the preset time period to obtain an operation feature sequence, wherein the operation feature sequence includes at least one operation feature sorted according to the time information; Performing initial fault diagnosis on the memory using the operation characteristic sequence to obtain an initial diagnosis result; Determining a diagnostic strategy feature sequence for the memory based on the initial diagnostic result and the operating feature sequence, the diagnostic strategy feature sequence also including a risk threshold and a repair method for the initial fault diagnosis, the risk threshold being used for comparison with an output value of a fault prediction model to determine a diagnostic result for the memory based on the comparison result, the fault prediction model being used to perform an initial fault diagnosis on the memory using the operating feature sequence to obtain an initial diagnostic result; Performing fault diagnosis on the memory using the diagnostic strategy feature sequence to obtain a target diagnostic result includes: Inputting the diagnostic strategy feature sequence into a strategy prediction model to obtain a target prediction strategy, wherein the strategy prediction model is trained based on reward function values ​​of multiple prediction strategies of the fault prediction model corresponding to the historical diagnostic strategy feature sequence; The fault prediction model is used to perform fault diagnosis on the memory according to the target prediction strategy to obtain the target diagnosis result.

2. The memory fault diagnosis method according to claim 1, wherein: The performing initial fault diagnosis on the memory using the operation characteristic sequence to obtain an initial diagnosis result includes: Time-encoding the operation features according to the order of the operation features in the operation feature sequence to obtain time features; Embedding the time feature into a preset position of the operation feature to obtain a time perception feature; Using a multi-head self-attention mechanism to process the time perception features to obtain target attention features; An initial fault diagnosis is performed on the memory using a target attention feature sequence to obtain the initial diagnosis result, wherein the target attention feature sequence includes at least one target attention feature.

3. The memory fault diagnosis method according to claim 2, wherein: The multi-head self-attention mechanism is used to process the time perception feature to obtain the target attention feature, including: Processing the time perception feature using the multi-head self-attention mechanism to obtain an initial self-attention feature; Normalizing the time perception feature and the initial self-attention feature to obtain an intermediate attention feature; The intermediate attention feature is input into the position feedforward network to obtain the target attention feature.

4. The memory fault diagnosis method according to claim 2, wherein: Determining the diagnostic strategy feature sequence stored in memory according to the initial diagnostic result and the operating feature sequence includes: The diagnosis strategy feature sequence is determined according to the time perception feature sequence and the initial diagnosis result.

5. The memory fault diagnosis method according to any one of claims 1 to 4, characterized in that: The memory operation information includes at least one of the following: historical fault diagnosis information of the memory, electrical environment information, operating system access information, operating system operation load information, and memory fault events recorded in the operating system log.

6. The memory fault diagnosis method according to claim 5, characterized in that: The memory fault diagnosis method further includes: Obtaining memory failure events within the preset time period from the log of the operating system; Locating the memory address where the memory failure event occurs according to a preset access path of the operating system, wherein the memory operation information further includes the memory address; The electrical environment information at the memory address is obtained.

7. The memory fault diagnosis method according to claim 1, wherein: The step of extracting features of the memory operation information within the preset time period based on the time information within the preset time period to obtain an operation feature sequence includes: Determining attribute statistical features based on the time information and multiple attribute data in the memory operation information; Performing vectorization processing on the memory operation information according to the time information to obtain an operation state feature sequence; The attribute statistical features and the operation status feature sequence are spliced ​​together to obtain the operation feature sequence.

8. The memory fault diagnosis method according to claim 7, characterized in that: The attribute statistical features include at least one of the following: a gradient feature characterizing the running change trend of the memory within the preset time period, a fluctuation feature characterizing the fluctuation of the attribute data within the preset time period, and an adjacent state feature characterizing the change trend of the attribute data at adjacent moments.

9. The memory fault diagnosis method according to claim 2, wherein: The prediction strategy includes one of the following: Processing the target attention feature sequence using the fault prediction model to obtain the target diagnosis result; Determining a target operation feature subsequence corresponding to preset memory operation information from the operation feature sequence; Updating the risk threshold of the fault prediction model based on the target operation feature subsequence, and performing memory fault diagnosis using the updated fault prediction model, wherein the preset memory operation information includes operating system electrical environment information and operating system access information; Determining target memory operation information for fault diagnosis from the memory operation information within the preset time period according to a preset time period; A repair method for scheduling storage space of the memory is adopted; Do nothing at this time.

10. The memory fault diagnosis method according to claim 1, wherein: The memory operation information includes historical fault diagnosis information of the memory, and the historical fault diagnosis information includes at least one of the following: historical repair results corresponding to historical target prediction strategies and historical risk thresholds.

11. The memory fault diagnosis method according to claim 1, wherein: The strategy prediction model is trained based on the following method: Processing the historical diagnosis strategy features according to the plurality of prediction strategies using the fault prediction model to obtain reward function values ​​of the plurality of prediction strategies; Determining a target reward value based on the multiple reward function values; The parameters of the pre-trained strategy prediction model are updated based on the target reward value until the target reward value reaches a preset reward value, thereby obtaining the strategy prediction model.

12. A memory fault diagnosis system, deployed on a server, characterized in that: include: A collection device for collecting memory operation information within a preset time period; A memory fault diagnosis device, configured to implement the steps of the memory fault diagnosis method according to any one of claims 1 to 11 when executed.

13. The memory fault diagnosis system according to claim 12, characterized in that: The server includes a baseboard management controller, a memory and an operating system, wherein the baseboard management controller is connected to the memory controller so as to monitor the electrical environment information of the memory; The memory is connected to the operating system, and the operating system is used to schedule the storage space of the memory.

14. The memory fault diagnosis system according to claim 13, characterized in that: The collecting device is used for: Collecting electrical environment information of the memory from the baseboard management controller; Collecting operating system access information, operating system load information, and memory failure events recorded in a log of the operating system from the operating system; The historical fault diagnosis information of the memory is collected from the memory fault diagnosis device, and the memory operation information includes at least one of the following: the historical fault diagnosis information, the electrical environment information, the operating system access information, the operation load information and the memory fault event.

15. The memory fault diagnosis system according to claim 13, wherein: The memory is connected to the service processor of the server via a memory channel.

16. The memory fault diagnosis system according to claim 13, wherein: The acquisition device includes a plurality of acquisition probes, and the plurality of acquisition probes are respectively deployed on the baseboard management controller, the operating system and the memory fault diagnosis device.

17. An electronic device comprising: one or more device processors; a memory for storing one or more computer programs, It is characterized in that the one or more device processors execute the one or more computer programs to implement the steps of the memory fault diagnosis method according to any one of claims 1 to 11.

18. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: When the computer program or instruction is executed by a device processor, the steps of the memory fault diagnosis method according to any one of claims 1 to 11 are implemented.

19. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a device processor, the steps of the memory fault diagnosis method according to any one of claims 1 to 11 are implemented.

Images