Cooperative working method and system of multi-core processor

Abstract

The invention discloses a cooperative work method and system of a multi-core processor, and relates to the technical field of computer system structures and multi-core processors, and the cooperative work method comprises the following specific steps: S100, state information exchange: designing a set of internal communication protocols, according to the multi-core processor, error detection and recovery are carried out through cooperation of the multiple cores, the stability and reliability of the multi-core processor are remarkably improved, when an abnormal condition is detected in the operation process of one core, the operation state of the multi-core processor is changed, and the reliability of the multi-core processor is improved. According to the method, an internal exception handling mechanism can be quickly triggered, an exception report can be sent to other health cores, an instant information sharing mechanism enables unaffected cores to quickly intervene, and the exception condition is jointly evaluated through a distributed exception detection algorithm, so that the speed and accuracy of fault recognition are accelerated, and the fault recognition efficiency is improved. And the healthy core can adjust the state of the healthy core and assist the affected core in fault recovery.

Description

Technical Field

[0001] The present invention relates to the technical field of computer architecture and multi-core processors, and in particular to a collaborative working method and system for multi-core processors. Background Art

[0002] With the rapid development of computer technology, multi-core processors have become the mainstream technology for improving computing performance. By integrating multiple processing cores, multi-core processors can significantly improve overall efficiency when processing tasks in parallel and meet increasingly complex and diverse computing needs. However, with the increase in the number of cores, how to effectively manage the collaborative work between these cores and ensure the efficient and stable operation of multi-core processors has become an urgent problem to be solved.

[0003] Traditional multi-core processors have obvious deficiencies in collaborative error handling. First, in terms of error detection, each core often operates independently, lacking effective information sharing and collaborative mechanisms, resulting in errors that may not be discovered in a timely manner and inaccurately located. Second, in terms of error recovery, once a core fails, it can only rely on external management systems for intervention, which not only has a slow response speed, but also easily affects the normal operation of other cores. In addition, when dealing with collaborative work between multiple cores, traditional methods lack a unified internal communication protocol and an efficient resource allocation strategy, resulting in overall inefficiency and an inability to fully utilize the performance advantages of multi-core processors.

[0004] Therefore, a collaborative working method and system for multi-core processors were developed. Through information exchange and sharing among the cores, real-time monitoring of the operating status, task execution and memory usage was achieved. On this basis, a distributed anomaly detection algorithm was introduced, which can accurately and quickly identify abnormal conditions in the cores. At the same time, when a core detects an error, it can quickly trigger the internal exception handling mechanism and send an exception report to other cores to request assistance in fault diagnosis and data recovery. Summary of the invention

[0005] The purpose of the present invention is to make up for the shortcomings of the prior art and to provide a collaborative working method and system for multi-core processors, which can achieve close collaboration among multiple cores by designing a unified internal communication protocol and an efficient collaborative mechanism. In terms of error handling, the method can automatically detect core anomalies and trigger a collaborative fault diagnosis process to ensure that when a core has a problem, other healthy cores can respond quickly to assist in fault diagnosis and data recovery. This collaborative working mode not only improves the stability and reliability of the multi-core processor, but also significantly enhances the fault tolerance and self-recovery capabilities of the multi-core processor, providing a more solid guarantee for complex computing tasks.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: On the one hand, a collaborative working method of a multi-core processor, the specific steps of the collaborative working method are:

[0007] S100, state information exchange: design an internal communication protocol to use the internal communication bus to exchange information about the running status, task execution, and memory usage between cores;

[0008] S200, distributed anomaly detection: using the exchanged status information, a distributed anomaly detection algorithm is implemented. Each core not only detects its own status, but also judges the status information of other cores received. If an anomaly is found, an anomaly handling mechanism is triggered, and an anomaly report is sent to other cores.

[0009] S300, collaborative fault diagnosis: The core that receives the abnormality report uses the preset fault diagnosis logic and the collected status information to quickly locate the cause of the abnormality using comparative analysis technology and generate a detailed fault diagnosis report;

[0010] S400, seamless task takeover: After determining the cause of the abnormality, the healthy core will quickly take over the tasks of the abnormal core according to the task allocation strategy;

[0011] S500, data recovery and synchronization: the collaborative core starts the data recovery process according to the fault diagnosis result. After the data recovery is completed, each core resynchronizes the status information.

[0012] Further, the S100 includes:

[0013] S101, design a unified internal communication protocol to specify the format, encoding rules, verification mechanism, and instruction set for sending and receiving the operation status, task execution status, and memory usage exchanged between cores;

[0014] S102, using the internal communication bus as a medium to achieve information exchange between the cores;

[0015] S103, regularly collect information about the core's operating status, task execution status, and memory usage, and encapsulate it according to the protocol, adding necessary header information and checksums;

[0016] S104, broadcasting the encapsulated information to all cores through the internal communication bus, the receiver receives and parses the sent information packet, extracts key status information, and updates the status information in the local cache.

[0017] Furthermore, the S200 includes:

[0018] S201, after receiving the status information, each multi-core processor core executes a distributed anomaly detection algorithm to identify an abnormal condition existing in the core;

[0019] S202, each core independently determines an exception based on the algorithm result, and triggers an internal exception handling mechanism when an exception is found;

[0020] S203, the core that detects the abnormality sends an abnormality report to other cores and sends an alarm to an external management system;

[0021] S204, after receiving the abnormality report, the unaffected core adjusts its status to assist the affected core in performing the task.

[0022] Furthermore, the S201 uses a distributed anomaly detection algorithm to identify abnormal conditions in the core. Assume that the multi-core processor has n cores, each core has m state information features, and the jth state information feature value of core i is recorded as x ij , define a time window T, collect state information data for analysis within this time window, and the weight of each feature is w j , for each feature j, calculate the probability distribution P(x j ), where x j is the value of this feature, and the formula is: Where k is the number of features, and the information entropy H(j) is calculated using the formula: For each feature j of each core i, calculate the weighted difference D between it and the mean of the feature ij , calculate the mean of feature j on all cores, the formula is: Calculate the weighted difference, the formula is: Define the anomaly score function S i The formula used to evaluate the abnormality of core i is: Determine an abnormal threshold based on historical data and experience. i When this threshold is exceeded, it is determined that core i is in an abnormal condition.

[0023] Furthermore, the S300 includes:

[0024] S301, the core that receives the abnormality report starts a collaborative fault diagnosis process and collects various related status information from the local and other cores;

[0025] S302, preliminarily screening and classifying the information according to the preset fault diagnosis logic, and using comparative analysis technology to compare with the normal and historical states to find differences;

[0026] S303, using a pattern recognition algorithm to process the comparison results to automatically identify the fault mode and root cause, and then comprehensively judge and locate the abnormal position.

[0027] S304: Automatically generate a fault diagnosis report containing detailed contents according to the positioning result.

[0028] Furthermore, the specific steps of using the comparative analysis technology to compare with the normal and historical states to find differences in S302 are:

[0029] (1) extracting real-time data related to the current abnormal state, historical normal state data, and historical fault state data from the state information;

[0030] (2) Standardize the extracted data;

[0031] (3) Establish a normal operation baseline based on historical normal status data and identify the characteristics of historical fault status data;

[0032] (4) Compare the current abnormal state data with the normal state benchmark to identify significant differences;

[0033] (5) Perform similarity matching between the current abnormal state data and the historical fault state data to find similar fault modes;

[0034] (6) Combine the results of difference point analysis and similarity pattern matching to evaluate the severity and impact of the difference point, as well as the matching degree and credibility of the similarity pattern with the current anomaly.

[0035] Furthermore, the S303 uses a pattern recognition algorithm to identify potential failure modes and abnormal root causes, and a set of data feature vectors X is provided, wherein each X i Represents the m eigenvalues ​​of a data point, calculate X i With other data points X j The Euclidean distance of (j≠i) is as follows: Calculate the average difference, the formula is: According to the historical statistics of the data, a normal data range threshold is set as [L k ,U k ], statistical data point X i The ratio of the number of features that exceed the normal range to the total number of features is taken as the abnormal frequency: Combining these two indicators, we get a potential failure index P i , the formula is: i =αD i +βF i , where α and β are weight coefficients calculated by calculating the P of each data point i value, sort the data, P i Data points with higher values ​​are more likely to represent potential failure modes and anomaly root causes.

[0036] Furthermore, the S400 includes:

[0037] S401, after determining the cause of the abnormality, mark the affected core as "abnormal" and send a notification including the abnormality type, impact range and priority to the healthy core;

[0038] S402, after receiving the notification, the healthy core immediately evaluates its own computing, storage, and network, and prepares context information, data synchronization mechanism, and recovery strategy;

[0039] S403, according to the preset task allocation strategy, the healthy core will determine the best task takeover plan;

[0040] S404, before taking over the task, the healthy core synchronizes data with the abnormal core and updates the status record;

[0041] S405, after completing synchronization and migration, the healthy core officially takes over the task. During the task execution, the healthy core will continue to monitor.

[0042] Furthermore, the S500 includes:

[0043] S501, the collaborative core analyzes the fault diagnosis result, plans a data recovery strategy based on the fault diagnosis result, pre-allocates necessary data resources, and starts the data recovery process;

[0044] S502, after completing data recovery, the collaborative core verifies the recovered data, and at the same time, each core collects and organizes its own status information;

[0045] S503, determining the frequency, content and method according to the status information synchronization protocol, and receiving, summarizing and distributing the status information according to the protocol;

[0046] S504, after the status information synchronization is completed, each core verifies that it has received the latest status information.

[0047] On the other hand, a multi-core processor cooperative working system includes the following components:

[0048] Status information exchange: Design a status information exchange protocol so that each core can regularly exchange key information including running status, task execution status, and memory usage through the internal communication bus;

[0049] Distributed status monitoring module: Using the collected status information, the distributed anomaly detection algorithm is implemented to monitor the status changes of each core in real time. When an anomaly is found, the anomaly handling mechanism is immediately triggered and an anomaly report is sent to other cores.

[0050] Task dynamic reallocation module: Once a core failure is confirmed, the task reallocation mechanism will be immediately activated to evaluate the urgency and importance of the tasks currently being executed by the failed core. According to the load conditions of other cores, the task allocation strategy will be dynamically adjusted to transfer the tasks of the failed core to other cores that are working normally.

[0051] Data repair and consistency maintenance module: performs data repair operations for data corruption caused by the faulty core, reads redundant data before and after the failure, and uses data recovery algorithms to recover damaged data;

[0052] Optimization module: Continuously optimizes error detection algorithms, task allocation strategies, and data repair methods by analyzing historical error records and task execution data.

[0053] Compared with the prior art, the collaborative working method and system of a multi-core processor has the following beneficial effects:

[0054] 1. The present invention significantly improves the stability and reliability of multi-core processors by performing error detection and recovery through the collaboration of multiple cores. When a core detects an abnormal condition during operation, it can quickly trigger the internal exception handling mechanism and send an exception report to other healthy cores. This instant information sharing mechanism enables unaffected cores to intervene quickly and jointly evaluate the abnormal condition through a distributed anomaly detection algorithm, thereby accelerating the speed and accuracy of fault identification. In addition, the healthy core can adjust its own state to assist the affected core in fault recovery, effectively reducing the risk of overall paralysis of the multi-core processor due to a single core failure.

[0055] 2. The present invention plans a detailed data recovery strategy based on the fault diagnosis results and starts the recovery process. During this process, the cores work closely together to ensure the smooth progress of the data recovery process. After the recovery is completed, each core uses the status information synchronization protocol to ensure that all cores can obtain the latest status information, and update the local cache accordingly, thereby achieving consistency in the core status. This collaborative error handling mechanism not only improves the efficiency of data recovery, but also enhances the self-repair ability of multi-core processors when facing errors, providing a more reliable computing environment for complex computing tasks.

[0056] Other advantages, objectives and features of the present invention will be set forth in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0058] Figure 1 A process framework diagram of a collaborative working method of a multi-core processor;

[0059] Figure 2 The present invention is a flowchart of a cooperative working system of multi-core processors. DETAILED DESCRIPTION

[0060] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] Embodiment 1:

[0062] A collaborative working method for a multi-core processor comprises the following steps: state information exchange, distributed anomaly detection, collaborative fault diagnosis, seamless task takeover, and data recovery and synchronization.

[0063] First, in the state information exchange stage (S100), a unified internal communication protocol is designed to specify in detail the format, encoding rules, verification mechanism and sending and receiving instruction sets of the operating status, task execution status and memory usage exchanged between the cores. The protocol ensures that the cores can exchange key information accurately and use the internal communication bus as the medium for information exchange to achieve efficient communication between the cores. The operating status, task execution status and memory usage information of each core are regularly collected. According to the protocol specifications, the collected information is encapsulated into a data packet, and necessary header information and verification codes are added to ensure the integrity and accuracy of the data. The encapsulated information packet is broadcast to all cores through the internal communication bus. After the receiving core receives the information packet, it parses it, extracts the key state information, and updates the state information in the local cache to maintain the consistency of the state information between the cores.

[0064] Next, the distributed anomaly detection stage (S200) is entered. After receiving the status information, each multi-core processor core executes a distributed anomaly detection algorithm. The algorithm identifies the abnormal conditions in the core based on the collected status information data. The algorithm considers multiple status information features and evaluates the degree of abnormality of the core by calculating the probability distribution, information entropy, and weighted difference index of the features. Each core independently judges the abnormality based on the algorithm results. Once an abnormality is found, the internal exception handling mechanism is immediately triggered, such as restarting some functions of the abnormal core and adjusting the task priority. The core that detects the abnormality will send an abnormality report to other cores. At the same time, it will also send an alarm to the external management system so that the administrator can understand and handle the abnormality in time. After receiving the abnormality report, the unaffected healthy core will adjust its own status to assist the affected abnormal core in task and data backup.

[0065] Subsequently, the collaborative fault diagnosis stage (S300) is entered. The core that receives the abnormality report starts the collaborative fault diagnosis process, collects various relevant status information from the local and other cores, and preliminarily screens and classifies the collected information according to the preset fault diagnosis logic. The current abnormal status data is compared with the normal status benchmark and the historical fault status data using comparative analysis technology to find out the significant differences and similar fault modes. The comparison results are processed with the help of pattern recognition algorithms. The potential fault index of the data points is evaluated by calculating the Euclidean distance and average difference between the data points, and then the abnormal position is comprehensively judged and located. According to the positioning results, a fault diagnosis report containing detailed content is automatically generated. The report content includes the abnormality type, location, cause, impact range and recommended recovery measures.

[0066] Then, the seamless task takeover phase (S400) is entered. After the abnormal core is determined, the core is marked as "abnormal" and a notification containing the abnormality type, impact range and priority is sent to all healthy cores. After receiving the abnormality notification, the healthy core will immediately begin to evaluate its own computing power, storage resources and network status to ensure that they have sufficient resources to take over the affected tasks. The healthy core will prepare context information, data synchronization mechanism and recovery strategy. The healthy core determines the best task takeover plan based on the preset task allocation strategy and its own evaluation results. The plan considers multiple factors, such as load balancing between cores, the urgency of tasks and the communication cost between cores. Once the task takeover plan is determined, the healthy core will prepare to execute the takeover operation. Before officially taking over the task, the healthy core needs to synchronize data with the abnormal core. After the data synchronization is completed, the healthy core will update its own status record. After completing data synchronization and status update, the healthy core officially takes over the task originally performed by the abnormal core and continues to execute the task until completion. During the task execution process, the healthy core will continue to monitor to ensure the smooth execution of the task and the stable operation of the multi-core processor.

[0067] Finally, the data recovery and synchronization stage (S500) is entered. When the tasks of the abnormal core have been taken over by the healthy core, the collaborative core begins to parse the generated detailed diagnostic report. Based on the diagnostic results, the collaborative core formulates a data recovery strategy. The collaborative core pre-allocates the necessary computing resources, storage resources and network resources, starts the data recovery process, and monitors the entire recovery process to ensure the smooth progress of the recovery operation. After the data recovery is completed, the collaborative core verifies the recovered data. At the same time, all cores collect and organize their own status information, including operating status, task execution status and memory usage. After collecting its own status information, each core follows the pre-defined status information synchronization protocol to determine the synchronization frequency, specific content to be synchronized and the synchronization method. Each core receives the status information sent by other cores through the internal communication bus in accordance with the protocol requirements, summarizes and distributes it to ensure that all cores can obtain the latest status information. After all status information synchronization is completed, each core verifies that it has received the latest status information.

[0068] By implementing the steps of the above embodiment, the multi-core processor can effectively handle core exceptions, recover data, and synchronize status information of each core, thereby maintaining the stability and reliability of the system.

[0069] Embodiment 2:

[0070] A collaborative working method and system for a multi-core processor in this embodiment particularly emphasizes that by designing a unified internal communication protocol, real-time exchange and synchronization of the operating status, task execution status and memory usage between cores are achieved. When an abnormal situation is detected, the exception handling mechanism can be quickly triggered, and the root cause of the fault can be accurately located through a collaborative fault diagnosis process.

[0071] After receiving the status information broadcasted by other cores through the internal communication bus, each core in the multi-core processor will execute the distributed anomaly detection algorithm to identify the abnormal conditions that may exist in each core. Suppose the multi-core processor has n cores, each core has m types of status information features, and the jth status information feature value of core i is recorded as x ij , define a time window T, collect state information data in this time window for subsequent analysis, and set the weight of each feature to w j , which is used to reflect the importance of the feature in anomaly detection. All state information in all cores and time windows is traversed to count the number of occurrences of each feature value. Then, the information entropy H(j) is calculated. The formula is: For each feature j of each core i, calculate the weighted difference D between it and the mean of the feature ij , calculate the mean of feature j on all cores, the formula is: Calculate the weighted difference, the formula is: Define the anomaly score function S i The formula used to evaluate the abnormality of core i is: Determine an abnormal threshold based on historical data and experience. i When this threshold is exceeded, it is determined that core i is in an abnormal condition.

[0072] Each core independently determines whether it is abnormal based on the algorithm results. If an abnormality is found, the internal exception handling mechanism is immediately triggered. This mechanism may include resource isolation, error logging, and fault isolation to prevent the spread of abnormalities and affect the stability of multi-core processors. When an abnormal core is detected, it will send an exception report to other cores, detailing the type of abnormality, time of occurrence, and scope of impact. At the same time, an alarm will be sent to the external management system so that the system administrator can intervene in time. After receiving the exception report, the unaffected healthy core will adjust its status and assist the affected abnormal core in fault recovery based on the content of the exception report and its own capabilities.

[0073] The core that receives the exception report will start the collaborative fault diagnosis process and collect various related status information from the local and other cores. This status information includes but is not limited to the core's operating status, task execution status, and memory usage.

[0074] According to the preset fault diagnosis logic, the collected information is preliminarily screened and classified, and then the comparative analysis technology is used to compare this information with the normal state and historical state to find out the differences. The specific steps are: extract the real-time data, historical normal state data and historical fault state data related to the current abnormal state from the state information, standardize the extracted data to ensure the comparability of data between different data sources, establish a normal operation baseline based on the historical normal state data, and identify the characteristics of the historical fault state data, compare the current abnormal state data with the normal state baseline, and identify significant differences. At the same time, the current abnormal state data and the historical fault state data are matched for similarity to find similar fault modes. Combined with the results of difference point analysis and similarity pattern matching, the severity and impact range of the difference points are evaluated to provide a basis for subsequent positioning and repair.

[0075] According to the preset fault diagnosis logic, the information is initially screened and classified, and the pattern recognition algorithm is used to identify potential fault modes and abnormal root causes. There is a set of data feature vectors X = {X1, X2, ..., X n}, where each X i =(x i1 ,x i2 ,…,x im ) represents the m eigenvalues ​​of a data point, and calculates X i With other data points X j The Euclidean distance of (j≠i) is as follows: Calculate the average difference, the formula is: According to the historical statistics of the data, a normal data range threshold is set as [L k ,U k ], statistical data point X i The ratio of the number of features that exceed the normal range to the total number of features is taken as the abnormal frequency: Combining these two indicators, we get a potential failure index P i , the formula is: i =αD i +βF i , where α and β are weight coefficients calculated by calculating the P of each data point i value, sort the data, P i Data points with higher values ​​are more likely to represent potential failure modes and anomaly root causes.

[0076] A fault diagnosis report containing detailed information is automatically generated based on the positioning results. The report will list the location, cause, impact range and recommended repair measures of the fault in detail, providing guidance for subsequent fault repair work.

[0077] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A collaborative working method of a multi-core processor, characterized in that: The specific steps of this collaborative working method are: S100, state information exchange: design an internal communication protocol to use the internal communication bus to exchange information about the running status, task execution, and memory usage between cores; S200, distributed anomaly detection: using the exchanged status information, a distributed anomaly detection algorithm is implemented. Each core not only detects its own status, but also judges the status information of other cores received. If an anomaly is found, an anomaly handling mechanism is triggered, and an anomaly report is sent to other cores. S300, collaborative fault diagnosis: The core that receives the abnormality report uses the preset fault diagnosis logic and the collected status information to quickly locate the cause of the abnormality using comparative analysis technology and generate a detailed fault diagnosis report; S400, seamless task takeover: After determining the cause of the abnormality, the healthy core will quickly take over the tasks of the abnormal core according to the task allocation strategy; S500, data recovery and synchronization: the collaborative core starts the data recovery process according to the fault diagnosis result. After the data recovery is completed, each core resynchronizes the status information.

2. The method for collaborative working of a multi-core processor according to claim 1, characterized in that: The S100 includes: S101, design a unified internal communication protocol to specify the format, encoding rules, verification mechanism, and instruction set for sending and receiving the operation status, task execution status, and memory usage exchanged between cores; S102, using the internal communication bus as a medium to achieve information exchange between the cores; S103, regularly collect information about the core's operating status, task execution status, and memory usage, and encapsulate it according to the protocol, adding necessary header information and checksums; S104, broadcasting the encapsulated information to all cores through the internal communication bus, the receiver receives and parses the sent information packet, extracts key status information, and updates the status information in the local cache.

3. The collaborative working method of a multi-core processor according to claim 1, characterized in that: The S200 includes: S201, after receiving the status information, each multi-core processor core executes a distributed anomaly detection algorithm to identify an abnormal condition existing in the core; S202, each core independently determines an exception based on the algorithm result, and triggers an internal exception handling mechanism when an exception is found; S203, the core that detects the abnormality sends an abnormality report to other cores and sends an alarm to an external management system; S204, after receiving the abnormality report, the unaffected core adjusts its status to assist the affected core in performing the task.

4. The method for collaborative working of a multi-core processor according to claim 3, characterized in that: S201 uses a distributed anomaly detection algorithm to identify abnormal conditions in the core. Assume that the multi-core processor has n cores, each core has m state information features, and the jth state information feature value of core i is recorded as x ij , define a time window T, collect state information data for analysis within this time window, and the weight of each feature is w j , for each feature j, calculate the probability distribution P(x j ), where x j is the value of this feature, and the formula is: Where k is the number of features, and the information entropy H(j) is calculated using the formula: For each feature j of each core i, calculate the weighted difference D between it and the mean of the feature ij , calculate the mean of feature j on all cores, the formula is: Calculate the weighted difference, the formula is: Define the anomaly score function S i The formula used to evaluate the abnormality of core i is: Determine an abnormal threshold based on historical data and experience. i When this threshold is exceeded, it is determined that core i is in an abnormal condition.

5. The method for collaborative working of a multi-core processor according to claim 1, characterized in that: The S300 includes: S301, the core that receives the abnormality report starts a collaborative fault diagnosis process and collects various related status information from the local and other cores; S302, preliminarily screening and classifying the information according to the preset fault diagnosis logic, and using comparative analysis technology to compare with the normal and historical states to find differences; S303, using a pattern recognition algorithm to process the comparison results to automatically identify the fault mode and root cause, and then comprehensively judge and locate the abnormal position. S304: Automatically generate a fault diagnosis report containing detailed contents according to the positioning result.

6. The method for collaborative working of a multi-core processor according to claim 5, characterized in that: The specific steps of using the comparative analysis technology in S302 to compare with the normal and historical states to find differences are as follows: (1) extracting real-time data related to the current abnormal state, historical normal state data, and historical fault state data from the state information; (2) Standardize the extracted data; (3) Establish a normal operation baseline based on historical normal status data and identify the characteristics of historical fault status data; (4) Compare the current abnormal state data with the normal state benchmark to identify significant differences; (5) Perform similarity matching between the current abnormal state data and the historical fault state data to find similar fault modes; (6) Combining the results of difference point analysis and similarity pattern matching, evaluate the severity and impact scope of the difference point, as well as the matching degree and credibility of the similarity pattern with the current anomaly.

7. The method for collaborative working of a multi-core processor according to claim 5, characterized in that: S303 uses a pattern recognition algorithm to identify potential failure modes and abnormal root causes, and a set of data feature vector sets X is provided, wherein each X i Represents the m eigenvalues ​​of a data point, calculate X i With other data points X j The Euclidean distance of (j≠i) is as follows: Calculate the average difference, the formula is: According to the historical statistics of the data, a normal data range threshold is set as [L k ,U k ], statistical data point X i The ratio of the number of features that exceed the normal range to the total number of features is taken as the abnormal frequency: Combining these two indicators, we get a potential failure index P i , the formula is: i =αD i +βF i , where α and β are weight coefficients calculated by calculating the P of each data point i value, sort the data, P i Data points with higher values ​​are more likely to represent potential failure modes and anomaly root causes.

8. The method for collaborative working of a multi-core processor according to claim 1, characterized in that: The S400 includes: S401, after determining the cause of the abnormality, mark the affected core as "abnormal" and send a notification containing the abnormality type, impact range and priority to the healthy core; S402, after receiving the notification, the healthy core immediately evaluates its own computing, storage, and network, and prepares context information, data synchronization mechanism, and recovery strategy; S403, according to the preset task allocation strategy, the healthy core will determine the best task takeover plan; S404, before taking over the task, the healthy core synchronizes data with the abnormal core and updates the status record; S405, after completing synchronization and migration, the healthy core officially takes over the task. During the task execution, the healthy core will continue to monitor.

9. The method for collaborative working of a multi-core processor according to claim 1, characterized in that: The S500 includes: S501, the collaborative core analyzes the fault diagnosis result, plans a data recovery strategy based on the fault diagnosis result, pre-allocates necessary data resources, and starts the data recovery process; S502, after completing data recovery, the collaborative core verifies the recovered data, and at the same time, each core collects and organizes its own status information; S503, determining the frequency, content and method according to the status information synchronization protocol, and receiving, summarizing and distributing the status information according to the protocol; S504, after the status information synchronization is completed, each core verifies that it has received the latest status information.

10. A multi-core processor collaborative working system, characterized in that: The system consists of the following components: Status information exchange: Design a status information exchange protocol so that each core can regularly exchange key information including running status, task execution status, and memory usage through the internal communication bus; Distributed status monitoring module: Using the collected status information, the distributed anomaly detection algorithm is implemented to monitor the status changes of each core in real time. When an anomaly is found, the anomaly handling mechanism is immediately triggered and an anomaly report is sent to other cores. Task dynamic reallocation module: Once a core failure is confirmed, the task reallocation mechanism will be immediately activated to evaluate the urgency and importance of the tasks currently being executed by the failed core. According to the load conditions of other cores, the task allocation strategy will be dynamically adjusted to transfer the tasks of the failed core to other cores that are working normally. Data repair and consistency maintenance module: performs data repair operations for data corruption caused by the faulty core, reads redundant data before and after the failure, and uses data recovery algorithms to recover damaged data; Optimization module: Continuously optimizes error detection algorithms, task allocation strategies, and data repair methods by analyzing historical error records and task execution data.

Images