Dynamic load balancing method and system for multi-core heterogeneous ASIC (Application Specific Integrated Circuit) computing mainboard

Abstract

The invention relates to the technical field of computing task balancing, and discloses a dynamic load balancing method and system for a multi-core heterogeneous ASIC computing mainboard, which comprises the following steps: collecting performance data of each computing core on the multi-core heterogeneous mainboard, generating a computing core information set, analyzing the load adaptability of the computing cores based on the performance data, and obtaining load adaptation feature distribution; the load condition of the calculation core is monitored in real time to obtain a load change trend, the future load condition is predicted, load adaptive feature distribution and load prediction are combined, balance optimization is carried out, a load balancing scheme is generated to adjust load distribution of the calculation core, dynamic load balancing is achieved, the utilization rate of calculation resources is improved, and overload or idling of the calculation core is avoided. The system performance and the response speed are improved, different load requirements are met, load changes are dynamically adapted, task scheduling and energy efficiency management are optimized, and the problem that in the prior art, a task allocation scheme is dynamically and continuously changed in a balanced mode in real time, and consequently the execution effect of a mainboard is poor is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of computing task balancing, and in particular to a dynamic load balancing method and system for a multi-core heterogeneous ASIC computing motherboard. Background Art

[0002] In modern multi-core heterogeneous ASIC (application-specific integrated circuit) computing platforms, they are usually composed of multiple types of computing units, such as CPU, GPU, FPGA, and dedicated ASIC cores. These computing units have different processing capabilities, power consumption characteristics, and processing methods, and can process different types of tasks in parallel.

[0003] In a multi-core heterogeneous platform, the load of computing tasks is not only affected by the characteristics of the tasks (such as computing intensiveness, I / O intensiveness, etc.), but also changes over time. The main influencing items are the changes in the time of the computing tasks themselves and the changes in the actual performance status of the computing cores when they continue to execute tasks. Due to the above reasons, traditional real-time dynamic balancing methods are prone to continuous changes in task load distribution, resulting in poor task execution results. Summary of the invention

[0004] The purpose of the present invention is to provide a dynamic load balancing method and system for a multi-core heterogeneous ASIC computing motherboard, aiming to solve the problem in the prior art that the real-time dynamic balancing and continuously changing task allocation scheme leads to poor motherboard execution effect.

[0005] The present invention is implemented in this way. In a first aspect, the present invention provides a dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard, comprising: Collecting performance data of computing cores of a multi-core heterogeneous motherboard to obtain a computing core information set of the multi-core heterogeneous motherboard; wherein the computing core information set includes a plurality of computing core performance information; Performing a load adaptability analysis of the computing cores of the multi-core heterogeneous motherboard according to the computing core information set to obtain a load adaptation feature distribution of the multi-core heterogeneous motherboard; wherein the load adaptation feature distribution includes a load adaptation feature corresponding to each computing core; When the multi-core heterogeneous mainboard is in working state, monitoring the computing core load of the multi-core heterogeneous mainboard to obtain the computing core load status of the multi-core heterogeneous mainboard; Summarizing and analyzing the load trend of the computing core load conditions at each moment in the historical records to obtain the load change trend of the multi-core heterogeneous motherboard; Based on the load change trend, a prediction analysis is performed on the computing core load status at a future moment to obtain a predicted load form at a future moment; According to the load adaptation characteristic distribution, a balanced optimization analysis of the computing core load is performed on the computing core load status at the current moment and the predicted load form at the future moment, so as to obtain a corresponding load balancing solution; wherein the load balancing solution includes a prediction verification phase and an allocation execution phase; The load of the computing cores of the multi-core heterogeneous mainboard is distributed according to the load balancing solution to achieve dynamic load balancing of the multi-core heterogeneous mainboard.

[0006] In a second aspect, the present invention provides a dynamic load balancing system for a multi-core heterogeneous ASIC computing motherboard, which is used to implement a dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard as described in any one of the first aspects, including: A performance collection module, used for collecting performance data of computing cores of a multi-core heterogeneous motherboard to obtain a computing core information set of the multi-core heterogeneous motherboard; wherein the computing core information set includes a plurality of computing core performance information; An adaptation analysis module, used for performing load adaptability analysis of the computing cores of the multi-core heterogeneous motherboard according to the computing core information set, so as to obtain a load adaptation feature distribution of the multi-core heterogeneous motherboard; wherein the load adaptation feature distribution includes load adaptation features corresponding to each computing core; A load monitoring module, used for monitoring the computing core load of the multi-core heterogeneous mainboard when the multi-core heterogeneous mainboard is in working state, so as to obtain the computing core load status of the multi-core heterogeneous mainboard; A trend analysis module, used for summarizing and analyzing the load trend of the computing core load conditions at each moment in the historical records, so as to obtain the load change trend of the multi-core heterogeneous motherboard; A load prediction module, used to predict and analyze the computing core load status at a future moment based on the load change trend, so as to obtain a predicted load form at a future moment; A balancing optimization module, used to perform a balancing optimization analysis of the computing core load on the computing core load status at the current moment and the predicted load form at the future moment according to the load adaptation characteristic distribution, so as to obtain a corresponding load balancing solution; wherein the load balancing solution includes a prediction verification phase and an allocation execution phase; A balancing execution module is used to distribute the load of the computing cores of the multi-core heterogeneous mainboard according to the load balancing scheme to achieve dynamic load balancing of the multi-core heterogeneous mainboard.

[0007] The present invention provides a dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard, which has the following beneficial effects: The present invention collects performance data of each computing core on a multi-core heterogeneous motherboard, generates a computing core information set, analyzes the load adaptability of the computing core based on the performance data, obtains load adaptation feature distribution, monitors the computing core load status in real time, and performs trend analysis on historical load data, predicts future load status according to load change trends, performs balancing optimization in combination with load adaptation feature distribution and load prediction, generates a load balancing plan, adjusts the load distribution of the computing core according to the load balancing plan, realizes dynamic load balancing, improves computing resource utilization, avoids computing core overload or idleness, improves system performance and response speed, meets different load requirements, dynamically adapts to load changes, optimizes task scheduling and energy efficiency management, and solves the problem in the prior art that the real-time dynamic balancing and continuously changing task distribution plan leads to poor motherboard execution effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 It is a schematic diagram of the steps of a dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard provided by an embodiment of the present invention; Figure 2 It is a structural schematic diagram of a dynamic load balancing system for a multi-core heterogeneous ASIC computing motherboard provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0009] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0010] The implementation of the present invention is described in detail below in conjunction with specific embodiments.

[0011] Reference Figure 1 , Figure 2 As shown, a preferred embodiment of the present invention is provided.

[0012] In a first aspect, the present invention provides a dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard, comprising: S1: collecting performance data of computing cores of a multi-core heterogeneous motherboard to obtain a computing core information set of the multi-core heterogeneous motherboard; wherein the computing core information set includes performance information of several computing cores; S2: performing load adaptability analysis of the computing cores of the multi-core heterogeneous motherboard according to the computing core information set to obtain a load adaptation feature distribution of the multi-core heterogeneous motherboard; wherein the load adaptation feature distribution includes load adaptation features corresponding to each computing core; S3: When the multi-core heterogeneous mainboard is in a working state, monitoring the computing core load of the multi-core heterogeneous mainboard to obtain the computing core load status of the multi-core heterogeneous mainboard; S4: performing a load trend summary analysis on the computing core load conditions at each moment in the historical records to obtain a load change trend of the multi-core heterogeneous motherboard; S5: Based on the load change trend, predict and analyze the computing core load status at a future moment to obtain a predicted load form at a future moment; S6: performing a load balancing optimization analysis of the computing core load status at the current moment and the predicted load form at the future moment according to the load adaptation characteristic distribution, so as to obtain a corresponding load balancing solution; wherein the load balancing solution includes a prediction verification phase and an allocation execution phase; S7: Distribute the load of computing cores of the multi-core heterogeneous motherboard according to the load balancing solution to achieve dynamic load balancing of the multi-core heterogeneous motherboard.

[0013] Specifically, in step S1 of the embodiment provided by the present invention, performance data of the computing cores of the multi-core heterogeneous motherboard is collected to obtain a computing core information set of the multi-core heterogeneous motherboard. The multi-core heterogeneous motherboard represents that multiple computing cores are arranged on the motherboard, and the computing cores include multiple types. It can be understood that each computing core is suitable for different computing tasks, and the adaptability of the computing core to each type of computing task depends on the performance of the computing core.

[0014] More specifically, the computing core information set includes computing core performance information of each computing core, and the computing core performance information includes computing frequency, number of cores, processing power, resource utilization, functional data, and the like.

[0015] It can be understood that through real-time collection of computing core performance data, the system can ensure that it can understand the status of each computing unit in a timely manner and provide detailed basic data for subsequent analysis. The detailed performance data of each computing unit can help the system accurately capture the load and health status of each core, which helps to further optimize system performance.

[0016] Specifically, in step S2 of the embodiment provided by the present invention, the load adaptability analysis of the computing cores of the multi-core heterogeneous motherboard is performed based on the computing core information set to obtain the load adaptation feature distribution of the multi-core heterogeneous motherboard; wherein the load adaptation feature distribution includes the load adaptation features corresponding to each computing core.

[0017] More specifically, the collected performance data is used to perform load adaptability analysis of each computing core. The goal of this analysis is to evaluate the adaptability of different cores under different load conditions, that is, the responsiveness, efficiency, and stability of each core under different task loads. Based on the load adaptability of each core, tasks can be more reasonably assigned to the most suitable computing core, thereby improving the overall load adaptability of the system.

[0018] Specifically, in step S3 of the embodiment provided by the present invention, when the mainboard is in working state, the load status of each computing core is monitored in real time, including: the real-time load of each computing core (such as occupancy percentage), including the usage of computing resources, memory, bandwidth, etc., monitoring the delay and response time of the computing core to ensure that high real-time tasks are processed in a timely manner.

[0019] It is understandable that by monitoring the load status of each computing core, it is possible to promptly discover and resolve high load, overload or imbalance problems, ensure that system resources can be used efficiently, avoid performance bottlenecks caused by overloading of certain computing cores, and avoid waste of resources caused by idle cores.

[0020] It should be noted that the load condition of the computing core reflects the computing tasks undertaken by the computing core, and also provides feedback on the actual adaptability of the computing core to the tasks undertaken. The computing tasks undertaken by the computing core are determined by the system allocation, and the actual adaptability of the computing core to the tasks undertaken is determined by the actual performance status of the computing core itself. The actual performance status of the computing core itself will change with working time. At this time, the load change trend determined by the load condition of the computing core will change, and it is necessary to dynamically optimize the allocation of computing tasks according to the actual performance status of the computing core.

[0021] Specifically, in step S4 of the embodiment provided by the present invention, the load conditions at each moment in the historical records are summarized and analyzed in advance, and the trends and patterns of load changes are extracted. By analyzing the load changes of each core, its performance in different time periods is found, that is, when a multi-core heterogeneous motherboard executes different types of tasks, its overall performance load requirements are different, and therefore the computing core load conditions caused to each computing core on the multi-core heterogeneous motherboard are also different. This difference is reflected to a certain extent in the computing core load condition at a single moment, and is more obvious in the load change trend at continuous moments, that is, when a multi-core heterogeneous motherboard executes different designated motherboard tasks, the computing core load conditions and load change trends it exhibits are different.

[0022] More specifically, by identifying the load trend of the computing core load conditions at each moment, the type of task currently executed by the multi-core heterogeneous motherboard can be determined, thereby predicting the computing core load conditions of the multi-core heterogeneous motherboard in future time periods.

[0023] It should be noted that the computing core load conditions in future time periods are different for different designated motherboard tasks. That is to say, if the load distribution is optimized simply based on the real-time computing core load conditions, there is a possibility that the computing core load distribution plan will continue to change. Such frequent changes lead to unstable computing core performance, making real-time load distribution optimization a burden.

[0024] Therefore, the present invention provides a technical solution to pre-identify the tasks performed by a multi-core heterogeneous motherboard, obtain the computing core load status in future time periods, and analyze the optimal load pattern based on this to obtain a load balancing solution that can adapt to the computing core load status in future time periods without frequent changes, thereby ensuring the stability and efficiency of the computing core performance.

[0025] Specifically, in step S5 of the embodiment provided by the present invention, based on the load trend summary analysis, machine learning, time series analysis and other methods are used to predict the load at future times, analyze the load conditions of each core at future times, and consider factors such as task changes and load fluctuations. Through load prediction, potential load peaks can be identified in advance, and measures can be taken to balance the load. Based on the predicted data, the system can schedule resources in advance to ensure load balance.

[0026] Specifically, in steps S6 and S7 of the embodiment provided by the present invention, the predicted load form describes the possible computing core load conditions of the multi-core heterogeneous motherboard in the future time period, so that the multi-core heterogeneous motherboard can analyze the load distribution in advance according to the possible computing core load conditions in the future time period, so as to obtain a load distribution plan that can best adapt to the continuous load conditions of each computing core in the future time period without frequent changes.

[0027] More specifically, there are many specific forms of predicted load forms, because there are also many corresponding load distribution schemes. Based on this situation, the load balancing scheme is designed into a prediction verification stage and a distribution execution stage. In the prediction verification stage, the most common distribution scheme among various load distribution schemes is executed, and the load status of the computing core in this stage is continuously collected and feature analyzed. According to the results of the analysis, it is determined which specific form of the predicted load form corresponds to the actual load form. Then, the distribution execution stage is entered, and the load distribution scheme corresponding to the actual load form is called to distribute load tasks to each computing core of the multi-core heterogeneous motherboard.

[0028] The present invention provides a dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard, which has the following beneficial effects: The present invention collects performance data of each computing core on a multi-core heterogeneous motherboard, generates a computing core information set, analyzes the load adaptability of the computing core based on the performance data, obtains load adaptation feature distribution, monitors the computing core load status in real time, and performs trend analysis on historical load data, predicts future load status according to load change trends, performs balancing optimization in combination with load adaptation feature distribution and load prediction, generates a load balancing plan, adjusts the load distribution of the computing core according to the load balancing plan, realizes dynamic load balancing, improves computing resource utilization, avoids computing core overload or idleness, improves system performance and response speed, meets different load requirements, dynamically adapts to load changes, optimizes task scheduling and energy efficiency management, and solves the problem in the prior art that the real-time dynamic balancing and continuously changing task distribution plan leads to poor motherboard execution effect.

[0029] Preferably, the step of performing load adaptability analysis of computing cores on the multi-core heterogeneous mainboard according to the computing core information set to obtain a load adaptation feature distribution of the multi-core heterogeneous mainboard includes: S21: performing performance simulation processing on each computing core on the multi-core heterogeneous motherboard according to the computing core performance information corresponding to each computing core in the computing core information set, so as to obtain performance simulation characteristics for digital feedback of the computing performance of each computing core; S22: Acquire several preset target load task types of the multi-core heterogeneous motherboard, and perform reverse analysis of load adaptability of various preset target load task types according to the performance simulation characteristics, so as to obtain load adaptability indexes of various target load task types corresponding to the computing cores corresponding to the performance simulation characteristics; S23: Combining the load adaptability indexes of the computing cores to obtain the load adaptation characteristics of the computing cores, and combining the load adaptation characteristics of the computing cores to obtain the load adaptation characteristic distribution of the multi-core heterogeneous motherboard.

[0030] Specifically, based on the previously collected computing core performance data (such as frequency, number of cores, load, temperature, power consumption, etc.) as computing core performance information, the performance of each computing core is simulated using this information, and the simulation characteristics (such as theoretical execution speed, memory bandwidth requirements, energy efficiency ratio, etc.) are calculated. This step usually requires the use of simulation tools, models or computing frameworks (for example, simulation tools based on CPU and GPU architecture) to virtually simulate the processing capabilities of each computing core.

[0031] More specifically, according to different application scenarios, several typical load task types are set (such as computing-intensive tasks, memory-intensive tasks, IO-intensive tasks, etc.). Each task type has different resource requirements and computing modes. The computing requirements, memory requirements, degree of parallelism, etc. of these task types are modeled to facilitate subsequent load adaptability analysis.

[0032] More specifically, based on the performance simulation characteristics of each computing core, reverse analysis is performed to evaluate the adaptability of the computing core under different task types. This process usually includes: analyzing the response ability of each core to different load tasks (such as task completion time, response delay, throughput, etc.) based on the resource consumption model of each core, and based on the reverse analysis results, deriving the load adaptability index of each computing core corresponding to each target load task type. These indexes reflect the efficiency and effectiveness of the computing core in processing specific tasks. Common adaptability indexes include task execution efficiency, delayed response, energy consumption efficiency, etc.

[0033] More specifically, the load adaptability index of each computing core is combined to generate the load adaptability feature of the core, that is, the comprehensive performance of the core under different load conditions. Further, the load adaptability features of all computing cores are combined to form the load adaptability feature distribution of the entire multi-core heterogeneous motherboard. This reflects the adaptability of the entire system under different load tasks.

[0034] More specifically, the load adaptation characteristics of each computing core are combined to obtain the overall load adaptation characteristic distribution. These distributions usually show the performance differences of different cores under various load tasks, providing a basis for subsequent load balancing and scheduling strategies.

[0035] It is understandable that the load adaptability index can quantify the adaptability of different computing cores to different task types, thereby achieving optimal load distribution at runtime and improving system efficiency. Load adaptability analysis provides system administrators with a detailed view of computing core adaptability, helping them optimize task allocation and resource scheduling of computing cores, improving overall computing performance, and avoiding unnecessary waste of resources. Analysis based on load adaptability characteristics can achieve dynamic scheduling during task execution, so that tasks can run on the most suitable computing core, thereby reducing execution time and latency. An in-depth understanding of the distribution of load adaptability characteristics of multi-core heterogeneous systems enables the system to intelligently adjust task allocation according to load conditions to achieve load balancing, thereby maximizing resource utilization and improving overall performance. Comprehensive analysis of power consumption and computing power can help reduce unnecessary energy consumption and achieve better energy efficiency.

[0036] Preferably, the step of monitoring the computing core load of the multi-core heterogeneous mainboard to obtain the computing core load status of the multi-core heterogeneous mainboard includes: S31: Collecting data on computing resource scheduling for each computing core of the multi-core heterogeneous mainboard to obtain resource scheduling characteristics of each computing core; S32: collecting temperature and power consumption data of each computing core of the multi-core heterogeneous mainboard through a preset sensor group to obtain operating pressure characteristics of each computing core; S33: Performing a weighted comprehensive analysis on the resource scheduling characteristics and the operating pressure characteristics of the computing core according to a preset weight evaluation standard to obtain a real-time load index of the computing core; wherein the real-time load index is used to describe the numerical ratio of the performance load of the computing core at the current moment to the maximum executable performance load of the computing core; S34: performing a theoretical deviation analysis of the feature correlation between the resource scheduling feature and the operation pressure feature of the computing core according to the load adaptation feature of the computing core, and performing a mapping conversion of the load performance quality of the computing core according to the theoretical deviation result obtained by the analysis, so as to obtain a performance quality index of the computing core; wherein the performance quality index is used to describe the difference between the computing core performance exhibited by the computing core corresponding to the current load task and the theoretical computing core performance; S35: Combine the real-time load index and the load quality index of the computing core to obtain the unit load status of the computing core, and combine the unit load status of each computing core on the multi-core heterogeneous motherboard to obtain the computing core load status of the multi-core heterogeneous motherboard.

[0037] Specifically, the resource scheduler of the system collects the resource allocation information of each computing core, including task queues, the amount of computing tasks assigned, processing time, etc. These data help understand the load distribution and resource usage of each computing core.

[0038] More specifically, through the temperature sensors and power consumption monitoring sensors installed on the computing cores, the temperature and power consumption data of each computing core are collected in real time. These data reflect the operating pressure of the computing core and are crucial for evaluating the load status.

[0039] More specifically, the collected resource scheduling characteristics and operating pressure characteristics (such as temperature and power consumption) are weighted and comprehensively analyzed according to preset weights, and the real-time load index of each computing core is calculated. This index represents the proportion of the current load of the computing core to its maximum executable performance load, and is usually used to reflect the load density and resource pressure of the computing core.

[0040] More specifically, by performing theoretical deviation analysis on the correlation between the resource scheduling characteristics and the operating pressure characteristics of the computing core, the difference between the performance of the computing core under the current load conditions and the theoretical optimal performance is evaluated. This step uses mathematical modeling and data mining methods to identify the deviation between the computing core load condition and its theoretical performance. Based on the results of the theoretical deviation analysis, a mapping transformation of the load performance of each computing core is performed to obtain a performance index, which measures the gap between the actual performance of the computing core under the current load and the theoretical performance, and reflects whether there is a bottleneck in the performance of the computing core.

[0041] More specifically, the real-time load index and the performance index are comprehensively processed to obtain the unit load status of each computing core. The unit load status indicates the comprehensive performance of the computing core's resource load, operating pressure and performance deviation. By combining and analyzing the unit load status of all computing cores on a multi-core heterogeneous motherboard, the computing core load status of the entire system is obtained, which provides comprehensive data support for performance monitoring, load balancing and dynamic scheduling of multi-core systems.

[0042] It is understandable that by collecting and analyzing data such as resource scheduling, temperature and power consumption of computing cores, it is possible to monitor the load of each core in real time, promptly discover resource bottlenecks and operating pressures, and support rapid load adjustment. The calculation of real-time load index and performance index enables the load status and performance deviation of each computing core to be quantified. Theoretical deviation analysis helps identify performance bottlenecks and unbalanced loads, and provides optimization space for system scheduling. Through the performance index, task allocation can be dynamically adjusted to avoid overloading or excessive idleness of computing cores, thereby improving overall computing efficiency. Through the analysis of the load status of computing cores, the system can implement intelligent load balancing strategies, reasonably allocate tasks, avoid overloading of certain computing cores, and idleness of other core resources, thereby improving overall system performance and resource utilization.

[0043] Preferably, the step of summarizing and analyzing the load trend of the computing core load conditions at each moment in the historical records to obtain the load change trend of the multi-core heterogeneous motherboard includes: S41: Arranging the computing core load conditions at each moment of the historical records according to the record timestamps corresponding to the computing core load conditions to obtain a historical load condition sequence; S42: extracting the load change trend of the historical load condition sequence according to a pre-trained load trend recognition algorithm to obtain the load change trend of the historical load condition sequence; The load trend identification algorithm is a pre-trained machine learning algorithm, and the pre-training steps of the load trend identification algorithm include: S421: Build a machine learning algorithm model; S422: Collecting a plurality of training data groups for model training of the machine learning algorithm model; wherein the training data groups include a computing core load state sequence corresponding to a specific type of task, and a load change trend corresponding to the computing core load state sequence; S423: Substituting each of the training data groups into the machine learning algorithm model, and allowing the machine learning algorithm model to perform model training according to each of the training data groups, so that the machine learning algorithm model obtains a computing core load state sequence corresponding to a specific type of task, and an identification relationship between load change trends corresponding to the computing core load state sequence; S424: Use the identification relationship as the algorithm body, and perform algorithm conversion on the algorithm body to obtain an identification unit, generate an execution unit and a result unit with associated operation logic based on the identification unit, so as to obtain a load trend identification algorithm composed of the identification unit, the execution unit and the result unit; wherein the execution unit is used to input a historical load condition sequence to be identified, the identification unit is used to identify the corresponding task type of the historical load condition sequence received by the execution unit according to the identification relationship, so as to obtain the corresponding load change trend, and the result unit is used to output the load change trend identified by the identification unit.

[0044] Specifically, the load status of each computing core is arranged in chronological order according to the timestamp of the load status record of each computing core. This step ensures the temporal order of historical records so that subsequent analysis can reflect the load change trend over time. By sorting the load status of each computing core in time, the load status sequence of each computing core at different times is obtained.

[0045] More specifically, a pre-trained load trend recognition algorithm is used to analyze the historical load condition sequence and extract the trend of load changes. The algorithm uses a machine learning model to identify the load change pattern (such as increase, decrease, stability, etc.) in different time periods, extract the load change trend from the historical load sequence, and apply it to predict future load patterns, helping to determine possible load peaks or troughs in the computing core in the future.

[0046] More specifically, a suitable machine learning algorithm model is constructed, such as regression analysis, time series analysis, neural network, etc. These models can extract the trend of load changes based on the load condition sequence, and collect multiple training data sets for training machine learning models. Each training data set contains: The core load condition sequence and the corresponding load change trend (e.g., load increase, load decrease, stable load, etc.) are calculated.

[0047] More specifically, these training data sets are input into the machine learning algorithm model for model training. By learning the relationship between the load condition sequence and the load change trend under different types of tasks (such as high-load tasks and low-load tasks), the model gradually builds up the ability to recognize load patterns. Through training, the model can capture the recognition relationship between the load condition sequence and the load change trend, that is, the law of load change under specific tasks.

[0048] More specifically, the recognition relationship obtained from model training is used as the basis of the recognition unit. The recognition unit is used to analyze the input historical load condition sequence, identify the load change trend of the corresponding task type, and generate an execution unit and a result unit with associated operation logic in combination with the recognition unit. The execution unit is responsible for receiving the historical load condition sequence to be identified and passing it to the recognition unit. The recognition unit identifies the load change trend in the load condition sequence according to the pre-trained recognition relationship, and the result unit outputs the identified load change trend according to the output of the recognition unit.

[0049] More specifically, the execution unit passes the input historical load condition sequence to the identification unit. The identification unit analyzes the load change trend based on the identification relationship and outputs the identified load change trend through the result unit. The output load change trend can be an upward trend, a downward trend, a stable trend, or a more complex trend type (such as periodic fluctuations, etc.).

[0050] It is understandable that by training historical load sequences with machine learning algorithms, the pattern of load changes can be automatically identified. Based on load trend identification, the system can predict the changing trend of the computing core load. By identifying the load change trend, the system can achieve more intelligent task scheduling. The identification of load change trends can help the system avoid problems such as overload and overheating, optimize the resource allocation of multi-core heterogeneous motherboards, and improve the stability and operating efficiency of the system. Through machine learning training, the recognition algorithm can adapt to load changes under different types of tasks, whether it is a high-load task or a low-load task, it can accurately identify and predict load trends.

[0051] Preferably, the step of predicting and analyzing the computing core load condition at a future moment based on the load change trend to obtain the predicted load form at a future moment includes: S51: Directly predicting the computing core load status at a future moment based on the load change trend to obtain a first prediction form at the future moment; wherein the first prediction form includes a plurality of predicted load statuses at the future moments arranged in chronological order; S52: performing similarity analysis on the load change trend of the multi-core heterogeneous mainboard according to the load change trends corresponding to several specific types of tasks pre-stored in the database to obtain similarity parameters, and evaluating the similarity parameters according to a preset threshold to obtain load change trends corresponding to several specific types of tasks that meet the preset threshold, and directly predicting the computing core load status at future moments according to the load change trends corresponding to several specific types of tasks that meet the preset threshold to obtain several second prediction forms at future moments; wherein the second prediction form includes several predicted load conditions at future moments arranged in chronological order; S53: The first prediction form and a plurality of the second prediction forms together constitute a predicted load form at a future moment.

[0052] Specifically, based on the existing load change trend data (historical load patterns extracted by machine learning models), the computing core load conditions at future moments are predicted. This prediction is based on the identified load change patterns (for example, rising, falling, stable, etc. trends) and uses time series methods (such as regression analysis, deep learning, etc.) to directly predict the load conditions at future moments.

[0053] More specifically, the load conditions at future moments are obtained through direct prediction. These load conditions are arranged in chronological order. The first prediction form includes the predicted load data for multiple future moments. For example, at the first moment, the second moment, and the nth moment in the future, what will the system predict the load to be like? This step is based on historical data and load change trend models, and can provide preliminary prediction results for future loads, helping the system to schedule and optimize loads in advance.

[0054] More specifically, the database stores the load change trend data of different specific task types. Each task type will have a load change trend associated with it. For example, the computing demand of the task may be stable, increase sharply or decrease suddenly over time. A similarity analysis is performed based on the load change trend of the current multi-core heterogeneous motherboard (trend information obtained in the first step) and the load change trends of different task types stored in the database. This usually uses a similarity measurement method (such as cosine similarity, Euclidean distance, etc.) to evaluate the similarity between the current load trend and the stored task load trend. Through similarity analysis, the load change trends of several task types that are most similar to the current load trend are obtained. The similarity parameter is used to quantify the degree of match between the current load and different task loads. According to the preset threshold, the similarity parameter is evaluated to screen out the load trends of those task types that meet the threshold requirements. That is, only those task load trends that are highly similar to the current load change trend are selected.

[0055] More specifically, based on the load change trends of the selected task types that are similar to the current load change trends, a second prediction form is generated. These load trends will be arranged in chronological order to predict the computing core load conditions at future moments. Through the similarity analysis of the load trends of task types, more targeted load forecasts can be provided for the system under different task load modes, reducing the errors that may occur in single model predictions.

[0056] More specifically, the load prediction data obtained through direct prediction (first prediction form) is integrated with the load prediction data obtained through task type similarity analysis (second prediction form) to finally form the predicted load form at the future moment. The first prediction form is a basic prediction that quickly provides the load change trend at the future moment; the second prediction form combines the specific load trend of the task type to make the prediction result more accurate and multi-dimensional, making up for the limitations of a single prediction method. Through the combination of the two, the system can more accurately capture future load changes, especially in multi-task and complex load situations.

[0057] Preferably, the steps of performing a balanced optimization analysis of the computing core load on the computing core load status at the current moment and the predicted load form at the future moment according to the load adaptation characteristic distribution to obtain a corresponding load balancing solution include: S61: combining and processing the computing core load conditions at the current moment according to the first prediction form in the prediction load form to obtain a first load prediction sequence; S62: combining and processing the computing core load conditions at the current moment according to each second prediction form in the prediction load form to obtain a plurality of second load prediction sequences; S63: performing load distribution simulation on the first load prediction sequence according to the load adaptation characteristic distribution to obtain a first load distribution mode corresponding to the first load prediction sequence; S64: performing load distribution simulation on each of the second load prediction sequences according to the load adaptation characteristic distribution to obtain a second load distribution mode corresponding to each of the second load prediction sequences; S65: Perform sequence verification and identification scheme analysis on the first load prediction sequence and each of the second load prediction sequences to obtain a prediction verification scheme; wherein the prediction verification scheme is used to continuously collect and identify features of the computing core load status within a period of time in the future, so as to determine a load prediction sequence that is consistent with the actual load status of the multi-core heterogeneous motherboard in the first load prediction sequence and each of the second load prediction sequences as a verification load sequence; S66: extracting common features from the first load distribution mode and each of the second load distribution modes to obtain a verification phase distribution mode; wherein the verification phase distribution mode is used to distribute load tasks to each computing core of the multi-core heterogeneous motherboard when executing the prediction verification scheme; S67: The prediction verification scheme and the verification phase allocation mode together constitute the prediction verification phase of the load balancing scheme; S68: constructing a distribution execution phase based on the verification load sequence obtained by the prediction verification scheme and the first load distribution mode and each of the second load distribution modes; S69: Combine the prediction verification phase and the allocation execution phase to obtain a load balancing solution.

[0058] Specifically, based on the first prediction form (i.e., direct load prediction for future moments), the computing core load conditions at the current moment are combined and processed, and the load at the current moment is combined with the future load prediction data using historical load data and prediction trends to generate a first load prediction sequence. This sequence gives load predictions for future moments in chronological order. This step provides a preliminary load prediction sequence for subsequent load balancing, helping the system understand the load trends at the current moment and in the future, and preparing for optimizing load distribution.

[0059] More specifically, the computing core load conditions at the current moment are combined and processed according to the second prediction form (i.e., the load prediction obtained based on the similarity analysis of task types), and multiple load prediction models based on task type similarity are used to perform combined analysis on the load trend of each task type to generate several second load prediction sequences. These sequences represent the load prediction trends under different task types, and multiple load prediction sequences are generated by integrating the load change trends of multiple tasks, thereby providing a reference basis for multiple load distribution schemes for multi-core heterogeneous motherboards.

[0060] More specifically, based on the first load prediction sequence, load adaptation feature distribution (such as information such as the load capacity and task processing capacity of each computing core) is used to perform load distribution simulation. The computing cores are simulated to distribute load according to the load requirements of the first load prediction sequence to generate a first load prediction sequence and a corresponding first load distribution mode. The simulated load distribution scheme can provide a basis for subsequent load balancing strategies, ensuring the reasonable distribution of load in a multi-core computing system, thereby avoiding excessive or low load on some cores.

[0061] More specifically, based on multiple second load prediction sequences, load distribution simulation is performed on each prediction sequence through load adaptation characteristic distribution, and each second load prediction sequence is simulated separately to obtain multiple second load prediction sequences and corresponding second load distribution modes. This multi-angle simulation analysis provides multiple possible load distribution schemes, which helps the system select the optimal scheme from multiple possibilities and improves the flexibility and accuracy of load balancing.

[0062] More specifically, the first load prediction sequence and multiple second load prediction sequences are input, sequence verification and identification are performed on these prediction sequences, the matching degree with the actual load conditions is analyzed, and the prediction verification scheme is combined to continuously collect and feature identify future load conditions. Through continuous verification and identification, the load prediction can be gradually adjusted to make it closer to the actual load conditions, thereby improving the accuracy and reliability of the prediction. Through this dynamic verification, the problem of load imbalance caused by prediction deviation can be avoided.

[0063] More specifically, common features are extracted from all load distribution modes to generate a verification phase distribution mode, that is, how to reasonably distribute the load during the verification phase. This step integrates the common features between different load distribution modes and ensures that the load distribution performed during the verification phase can achieve the optimal effect under different load predictions.

[0064] More specifically, through the verification load sequence and the corresponding load distribution mode obtained by the prediction verification scheme, in the prediction verification stage, the load balancing strategy is verified and adjusted according to the verification load sequence and the first and second load distribution modes. Through continuous monitoring and adjustment in the prediction verification stage, the actual load can be matched more accurately, further improving the accuracy and stability of the load balancing scheme.

[0065] More specifically, based on the optimization of the verification phase, combined with the prediction verification plan, verification load sequence and load distribution mode, the distribution execution phase is constructed. In the distribution execution phase, the system performs load task allocation and scheduling according to the verification results to ensure that the load distribution of the computing core complies with the predetermined optimization plan. Through the optimization of the execution phase, load balancing operations can be actually performed to ensure that the load of the multi-core computing platform is efficient and balanced, thereby improving the utilization of system resources.

[0066] More specifically, the optimization results of the prediction and verification phase and the implementation plan of the allocation and execution phase combine the prediction and verification phase with the allocation and execution phase to ultimately form a complete load balancing solution. This solution ultimately implements a dynamic and precise load balancing mechanism that can adjust load distribution based on prediction data and perform real-time optimization based on actual load conditions, thereby ensuring the optimal allocation of computing resources and improving the overall system performance and stability.

[0067] It can be understood that by combining the first load prediction form and the second load prediction form, the computing requirements under different load modes can be fully considered, providing diverse perspectives and bases for load balancing. The prediction verification scheme and the verification stage allocation mode enable the system to dynamically adjust the load distribution scheme when the actual load occurs, avoiding the staticness and lack of adaptability of traditional load balancing methods. Through different load distribution simulations (the first load distribution mode and the second load distribution mode), the system can simulate load distribution from multiple angles and ultimately select the optimal solution. Through the verification and adjustment of the load prediction in the prediction verification stage, the system can adapt to various load changes and update the load balancing scheme in real time. This scheme can more efficiently utilize computing resources, reduce resource waste, and ensure the stable operation of the system through precise load prediction and distribution optimization. Through the processing and optimization of different load predictions, it can flexibly respond to different computing tasks and load modes, and improve the adaptability and response speed of the overall system.

[0068] Preferably, the step of performing sequence verification and identification scheme analysis on the first load prediction sequence and each of the second load prediction sequences to obtain a prediction verification scheme includes: S651: planning a certain range of time intervals towards the future time based on the current time, so as to obtain an initial range of the verification phase; S652: Perform difference analysis on the first load prediction sequence and each of the second load prediction sequences based on the initial range of the verification phase to obtain sequence difference characteristics of the first load prediction sequence and each of the second load prediction sequences in the initial range of the verification phase; S653: Using the difference amplitude corresponding to the sequence difference feature as a constraint condition, extending or shortening the initial range of the verification phase to obtain a verification phase duration range that meets a preset standard; S654: performing corresponding sequence content interception and content feature extraction on the first load prediction sequence and each of the second load prediction sequences according to the verification phase duration range, so as to obtain a first sequence identification feature corresponding to the first load prediction sequence and a second sequence identification feature corresponding to each of the second load prediction sequences; S655: Analyze the key identification features of the first sequence identification features and each of the second sequence identification features to obtain a prediction verification method; S656: The verification phase duration range and the prediction verification method together constitute a prediction verification plan.

[0069] Specifically, based on the current moment, a certain range of time intervals is selected to plan the initial range of the verification phase, and a reasonable time window is determined within which the load forecast sequence is verified and analyzed. The selection of this time interval is the basis for subsequent analysis, providing a preliminary time frame for subsequent verification and analysis, ensuring that subsequent steps can perform load forecasting and analysis within a suitable time period.

[0070] More specifically, within the initial time interval, a difference analysis is performed on the first load prediction sequence and the second load prediction sequence to analyze the difference characteristics between the first load prediction sequence and each second load prediction sequence, especially within a preset time window. These differences may include deviations between predicted values ​​and actual values, trend changes, etc. Through difference analysis, the accuracy of each load prediction sequence can be quantified, and their changing trends and deviations in the verification stage can be found out to help adjust the prediction model later.

[0071] More specifically, based on the sequence difference characteristics obtained in the difference analysis, the difference amplitude is determined, and the initial range of the verification phase is adjusted according to the difference amplitude, which may be to extend or shorten the time window to ensure that the verification phase contains sufficient load change information and meets the preset standards. By adjusting the duration of the verification phase, the verification effect can be optimized to avoid insufficient prediction accuracy due to too short a verification cycle, or overfitting due to too long a cycle.

[0072] More specifically, according to the adjusted verification phase duration range, the first load prediction sequence and the second load prediction sequence are intercepted and feature extracted. Within the determined verification phase, the corresponding content is intercepted from the load prediction sequence, and sequence identification features are extracted, such as key indicators such as trends, periodicity, and fluctuations. Through feature extraction, the key patterns of each load prediction sequence can be accurately described, providing representative data support for subsequent verification schemes.

[0073] More specifically, the extracted first sequence identification features and each second sequence identification feature are obtained, and key analysis is performed on these features to find the most representative identification features, and how these features affect the accuracy of load prediction is analyzed to form an effective prediction verification method. The analysis of key identification features helps to identify the most critical change patterns in load forecasting and enhance the accuracy and adaptability of the prediction verification scheme.

[0074] More specifically, the verification phase duration range and prediction verification method obtained through the above steps are combined to form a complete prediction verification scheme for subsequent load prediction verification and adjustment. This scheme provides a dynamic prediction verification method based on time interval adjustment and feature extraction, which can continuously adapt to the verification requirements under different load scenarios.

[0075] It is understandable that through difference analysis and sequence feature extraction, the duration of the verification phase can be flexibly adjusted to ensure that the verification process is more in line with actual load changes. By extracting key identification features, multi-dimensional analysis and optimization of load forecasts are performed, thereby enhancing the ability to predict future load conditions. By intercepting and analyzing the characteristics of sequence content, the key change trends in load forecasts can be accurately captured, making the prediction verification process more accurate and effective. With the dynamic adjustment of the time interval, the prediction verification plan can be automatically optimized according to changes in actual load conditions, adapt to different load scenarios, and improve the overall load balancing capability of the system.

[0076] Preferably, the step of distributing the load of the computing cores of the multi-core heterogeneous mainboard according to the load balancing scheme to achieve dynamic load balancing of the multi-core heterogeneous mainboard includes: S71: performing load task allocation corresponding to the verification phase allocation mode on the multi-core heterogeneous mainboard according to the prediction verification phase of the load balancing solution, so that each computing core of the multi-core heterogeneous mainboard performs the load task corresponding to the verification phase allocation mode in the prediction verification phase; S72: performing load sequence verification corresponding to the prediction verification scheme on the multi-core heterogeneous mainboard according to the prediction verification phase of the load balancing scheme to obtain the verification load sequence; S73: Select a corresponding mode for the verification load sequence according to the distribution execution phase of the load balancing scheme, so as to use the first load distribution mode or the second load distribution mode corresponding to the verification load sequence as the distribution execution mode, and distribute load tasks to the multi-core heterogeneous motherboard according to the distribution execution mode.

[0077] Specifically, according to the prediction verification phase of the load balancing solution, load tasks are allocated to the computing cores of the multi-core heterogeneous motherboard. The allocation mode will be determined based on the initial load prediction results of the verification phase. In the prediction verification phase, corresponding load tasks are allocated to each computing core according to the requirements of the load balancing solution. The allocation of these load tasks should ensure that the load of the computing cores is as balanced as possible during the verification phase to avoid overloading some computing cores while other computing cores are idle. Through the allocation of load tasks in the prediction verification phase, each computing core can be load balanced in the initial stage, avoiding performance bottlenecks caused by excessive load concentration on certain cores, and enhancing the predictability of load balancing.

[0078] More specifically, according to the prediction verification phase of the load balancing solution, load sequence verification is performed to serialize and verify the load tasks assigned to each computing core. The accuracy of load balancing is verified by calculating the execution effect of the load tasks. The load sequence verification process includes detecting whether the load task sequence executed by each computing core in a given time period is consistent with the expected load pattern. The load sequence verification ensures that the load of each computing core is consistent with the requirements of the prediction verification phase, thereby verifying the effectiveness of the load distribution solution. This process can timely discover potential problems in load distribution and correct them.

[0079] More specifically, according to the allocation execution phase of the load balancing solution, a distribution mode corresponding to the verification load sequence is selected, and based on the verification load sequence, the corresponding load distribution mode is determined. The first load mode and the second load mode correspond to different types of tasks, respectively. Based on the prediction verification phase, the type of task performed by the mainboard can be determined, thereby calling the first load mode or the second load mode, thereby more accurately adapting to load changes and improving system stability and computing efficiency.

[0080] More specifically, the selected distribution execution mode (the first load distribution mode or the second load distribution mode) performs actual load task distribution on each computing core of the multi-core heterogeneous motherboard according to the selected load distribution mode. Task distribution will take into account the performance characteristics and current load conditions of the computing core to ensure the continuity of load balancing. Through precise task distribution, the loads of different computing cores can be kept balanced under the distribution mode, avoiding excessive load concentration or waste of resources, thereby improving the overall computing efficiency and system response speed.

[0081] It is understandable that using the information from the prediction and verification phase to perform preliminary load task allocation avoids problems caused by unbalanced load distribution and load fluctuations, improves the system's prediction capabilities and load balancing effects, and accurately verifies load tasks through load sequence verification to ensure that the load tasks of each computing core can be effectively executed, reducing the execution imbalance problem caused by load prediction errors. On a multi-core heterogeneous motherboard, the precise allocation of load tasks ensures the maximum performance of each computing core, avoids resource waste and performance bottlenecks, and enhances the computing power and response speed of the entire computing platform. By continuously adjusting and verifying the load distribution plan, the system can operate stably under different load environments, ensure the efficient execution of computing tasks, and improve the system's adaptability to changing loads through dynamic scheduling.

[0082] Reference Figure 2 As shown, in a second aspect, the present invention provides a dynamic load balancing system for a multi-core heterogeneous ASIC computing motherboard, which is used to implement a dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard as described in any one of the first aspects, including: A performance collection module, used for collecting performance data of computing cores of a multi-core heterogeneous motherboard to obtain a computing core information set of the multi-core heterogeneous motherboard; wherein the computing core information set includes a plurality of computing core performance information; An adaptation analysis module, used for performing load adaptability analysis of the computing cores of the multi-core heterogeneous motherboard according to the computing core information set, so as to obtain a load adaptation feature distribution of the multi-core heterogeneous motherboard; wherein the load adaptation feature distribution includes load adaptation features corresponding to each computing core; A load monitoring module, used for monitoring the computing core load of the multi-core heterogeneous mainboard when the multi-core heterogeneous mainboard is in working state, so as to obtain the computing core load status of the multi-core heterogeneous mainboard; A trend analysis module, used for summarizing and analyzing the load trend of the computing core load conditions at each moment in the historical records, so as to obtain the load change trend of the multi-core heterogeneous motherboard; A load prediction module, used to predict and analyze the computing core load status at a future moment based on the load change trend, so as to obtain a predicted load form at a future moment; A balancing optimization module, used to perform a balancing optimization analysis of the computing core load on the computing core load status at the current moment and the predicted load form at the future moment according to the load adaptation characteristic distribution, so as to obtain a corresponding load balancing solution; wherein the load balancing solution includes a prediction verification phase and an allocation execution phase; A balancing execution module is used to distribute the load of the computing cores of the multi-core heterogeneous mainboard according to the load balancing scheme to achieve dynamic load balancing of the multi-core heterogeneous mainboard.

[0083] In this embodiment, for the specific implementation of each module in the above system embodiment, please refer to the above method embodiment, which will not be repeated here.

[0084] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard, characterized in that: include: Collecting performance data of computing cores of a multi-core heterogeneous motherboard to obtain a computing core information set of the multi-core heterogeneous motherboard; wherein the computing core information set includes a plurality of computing core performance information; Performing a load adaptability analysis of the computing cores of the multi-core heterogeneous motherboard according to the computing core information set to obtain a load adaptation feature distribution of the multi-core heterogeneous motherboard; wherein the load adaptation feature distribution includes a load adaptation feature corresponding to each computing core; When the multi-core heterogeneous mainboard is in working state, monitoring the computing core load of the multi-core heterogeneous mainboard to obtain the computing core load status of the multi-core heterogeneous mainboard; Summarizing and analyzing the load trend of the computing core load conditions at each moment in the historical records to obtain the load change trend of the multi-core heterogeneous motherboard; Based on the load change trend, a prediction analysis is performed on the computing core load status at a future moment to obtain a predicted load form at a future moment; According to the load adaptation characteristic distribution, a balanced optimization analysis of the computing core load is performed on the computing core load status at the current moment and the predicted load form at the future moment, so as to obtain a corresponding load balancing solution; wherein the load balancing solution includes a prediction verification phase and an allocation execution phase; The load of the computing cores of the multi-core heterogeneous mainboard is distributed according to the load balancing solution to achieve dynamic load balancing of the multi-core heterogeneous mainboard.

2. The dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard according to claim 1, characterized in that: The step of performing load adaptability analysis of the computing cores of the multi-core heterogeneous mainboard according to the computing core information set to obtain a load adaptation feature distribution of the multi-core heterogeneous mainboard includes: According to the computing core performance information corresponding to each computing core in the computing core information set, performance simulation processing is performed on each computing core on the multi-core heterogeneous motherboard to obtain performance simulation characteristics for digital feedback of the computing performance of each computing core; Acquire several preset target load task types of the multi-core heterogeneous motherboard, and perform reverse analysis of load adaptability of various preset target load task types according to the performance simulation characteristics, so as to obtain load adaptability indexes of various target load task types corresponding to the computing cores corresponding to the performance simulation characteristics; The load adaptability indexes of the computing cores are combined to obtain the load adaptation characteristics of the computing cores, and the load adaptation characteristics of the computing cores are combined to obtain the load adaptation characteristic distribution of the multi-core heterogeneous motherboard.

3. The dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard according to claim 1, characterized in that: The step of monitoring the computing core load of the multi-core heterogeneous mainboard to obtain the computing core load status of the multi-core heterogeneous mainboard includes: Performing data collection on computing resource scheduling for each computing core of the multi-core heterogeneous mainboard to obtain resource scheduling characteristics of each computing core; The temperature and power consumption data of each computing core of the multi-core heterogeneous mainboard are collected through a preset sensor group to obtain the operating pressure characteristics of each computing core; A weighted comprehensive analysis is performed on the resource scheduling characteristics and the operating pressure characteristics of the computing core according to a preset weight evaluation standard to obtain a real-time load index of the computing core; wherein the real-time load index is used to describe the numerical ratio of the performance load of the computing core at the current moment to the maximum executable performance load of the computing core; A theoretical deviation analysis of the feature correlation is performed on the resource scheduling features and the operation pressure features of the computing core according to the load adaptation features of the computing core, and a mapping conversion of the load performance quality of the computing core is performed according to the theoretical deviation results obtained by the analysis to obtain a performance quality index of the computing core; wherein the performance quality index is used to describe the difference between the computing core performance exhibited by the computing core corresponding to the current load task and the theoretical computing core performance; The real-time load index and load quality index of the computing core are combined to obtain the unit load status of the computing core, and the unit load status of each computing core on the multi-core heterogeneous motherboard are combined to obtain the computing core load status of the multi-core heterogeneous motherboard.

4. The dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard according to claim 1, characterized in that: The step of summarizing and analyzing the load trend of the computing core load conditions at each moment in the historical records to obtain the load change trend of the multi-core heterogeneous motherboard includes: Arrange the computing core load conditions at each moment of the historical records according to the record timestamps corresponding to the computing core load conditions to obtain a historical load condition sequence; Extracting the load change trend of the historical load condition sequence according to a pre-trained load trend recognition algorithm to obtain the load change trend of the historical load condition sequence; The load trend identification algorithm is a pre-trained machine learning algorithm, and the pre-training steps of the load trend identification algorithm include: Build machine learning algorithm models; Collecting a plurality of training data sets for model training of the machine learning algorithm model; wherein the training data sets include a computing core load state sequence corresponding to a specific type of task, and a load change trend corresponding to the computing core load state sequence; Substituting each of the training data groups into the machine learning algorithm model, allowing the machine learning algorithm model to perform model training according to each of the training data groups, so that the machine learning algorithm model obtains a computing core load status sequence corresponding to a specific type of task, and an identification relationship between load change trends corresponding to the computing core load status sequence; The identification relationship is used as the algorithm body, and the algorithm body is converted by algorithm to obtain an identification unit, and an execution unit and a result unit with associated operation logic are generated based on the identification unit to obtain a load trend identification algorithm composed of the identification unit, the execution unit and the result unit; wherein the execution unit is used to input a historical load condition sequence to be identified, the identification unit is used to identify the corresponding task type of the historical load condition sequence received by the execution unit according to the identification relationship to obtain the corresponding load change trend, and the result unit is used to output the load change trend identified by the identification unit.

5. The dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard according to claim 1, characterized in that: The steps of predicting and analyzing the computing core load status at a future moment based on the load change trend to obtain the predicted load form at a future moment include: Directly predicting the computing core load status at a future moment based on the load change trend to obtain a first prediction form at the future moment; wherein the first prediction form includes a plurality of predicted load statuses at the future moments arranged in chronological order; According to the load change trends corresponding to several specific types of tasks pre-stored in the database, similarity analysis is performed on the load change trends of the multi-core heterogeneous motherboard to obtain similarity parameters, and the similarity parameters are evaluated according to a preset threshold to obtain load change trends corresponding to several specific types of tasks that meet the preset threshold, and the computing core load conditions at future moments are directly predicted according to the load change trends corresponding to several specific types of tasks that meet the preset threshold to obtain several second prediction forms at future moments; wherein the second prediction form includes several predicted load conditions at future moments arranged in chronological order; The first prediction form and a plurality of the second prediction forms together constitute a predicted load form at a future moment.

6. The dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard according to claim 5, characterized in that: The steps of performing a balanced optimization analysis of the computing core load on the computing core load status at the current moment and the predicted load form at the future moment according to the load adaptation characteristic distribution to obtain a corresponding load balancing solution include: Combining and processing the computing core load conditions at the current moment according to the first prediction form in the prediction load form to obtain a first load prediction sequence; Combining and processing the computing core load status at the current moment according to each second prediction form in the prediction load form to obtain a plurality of second load prediction sequences; Performing a load distribution simulation on the first load prediction sequence according to the load adaptation characteristic distribution to obtain a first load distribution mode corresponding to the first load prediction sequence; Performing load distribution simulation on each of the second load prediction sequences according to the load adaptation characteristic distribution to obtain a second load distribution mode corresponding to each of the second load prediction sequences; Performing sequence verification and identification scheme analysis on the first load prediction sequence and each of the second load prediction sequences to obtain a prediction verification scheme; wherein the prediction verification scheme is used to continuously collect and identify features of the computing core load conditions within a period of time in the future, so as to determine a load prediction sequence that is consistent with the actual load condition of the multi-core heterogeneous motherboard in the first load prediction sequence and each of the second load prediction sequences as a verification load sequence; Extracting common features from the first load distribution mode and each of the second load distribution modes to obtain a verification phase distribution mode; wherein the verification phase distribution mode is used to distribute load tasks to each computing core of the multi-core heterogeneous motherboard when executing the prediction verification scheme; The prediction verification scheme and the verification phase allocation mode together constitute the prediction verification phase of the load balancing scheme; Constructing a distribution execution phase based on the verification load sequence obtained by the prediction verification scheme and the first load distribution mode and each of the second load distribution modes; The prediction verification phase is combined with the allocation execution phase to obtain a load balancing solution.

7. The dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard according to claim 6, characterized in that: The step of performing sequence verification and identification scheme analysis on the first load prediction sequence and each of the second load prediction sequences to obtain a prediction verification scheme includes: Plan a certain time interval into the future based on the current moment to obtain the initial range of the verification phase; Performing a difference analysis on the first load prediction sequence and each of the second load prediction sequences based on the initial range of the verification phase to obtain sequence difference characteristics of the first load prediction sequence and each of the second load prediction sequences in the initial range of the verification phase; Taking the difference amplitude corresponding to the sequence difference feature as a constraint condition, the initial range of the verification phase is extended or shortened to obtain a verification phase duration range that meets the preset standard; According to the verification phase duration range, corresponding sequence content interception and content feature extraction are performed on the first load prediction sequence and each of the second load prediction sequences to obtain a first sequence identification feature corresponding to the first load prediction sequence and a second sequence identification feature corresponding to each of the second load prediction sequences; Analyzing the key identification features of the first sequence identification features and each of the second sequence identification features to obtain a prediction verification method; The verification phase duration range and the prediction verification method together constitute a prediction verification plan.

8. The dynamic load balancing method for a multi-core heterogeneous ASIC computing motherboard according to claim 6, characterized in that: The steps of distributing the load of the computing cores of the multi-core heterogeneous mainboard according to the load balancing scheme to achieve dynamic load balancing of the multi-core heterogeneous mainboard include: According to the prediction verification phase of the load balancing scheme, load tasks corresponding to the verification phase allocation mode are allocated to the multi-core heterogeneous mainboard, so that each computing core of the multi-core heterogeneous mainboard performs the load tasks corresponding to the verification phase allocation mode in the prediction verification phase; According to the prediction verification phase of the load balancing solution, the multi-core heterogeneous mainboard is subjected to load sequence verification corresponding to the prediction verification solution to obtain the verification load sequence; According to the distribution execution phase of the load balancing scheme, a corresponding mode is selected for the verification load sequence, so that the first load distribution mode or the second load distribution mode corresponding to the verification load sequence is used as the distribution execution mode, and load tasks are distributed to the multi-core heterogeneous mainboard according to the distribution execution mode.

9. A dynamic load balancing system for a multi-core heterogeneous ASIC computing motherboard, characterized in that: A method for implementing a dynamic load balancing method of a multi-core heterogeneous ASIC computing motherboard according to any one of claims 1 to 8, comprising: A performance collection module, used for collecting performance data of computing cores of a multi-core heterogeneous motherboard to obtain a computing core information set of the multi-core heterogeneous motherboard; wherein the computing core information set includes a plurality of computing core performance information; An adaptation analysis module, used for performing load adaptability analysis of the computing cores of the multi-core heterogeneous motherboard according to the computing core information set, so as to obtain a load adaptation feature distribution of the multi-core heterogeneous motherboard; wherein the load adaptation feature distribution includes load adaptation features corresponding to each computing core; A load monitoring module, used for monitoring the computing core load of the multi-core heterogeneous mainboard when the multi-core heterogeneous mainboard is in working state, so as to obtain the computing core load status of the multi-core heterogeneous mainboard; A trend analysis module, used for summarizing and analyzing the load trend of the computing core load conditions at each moment in the historical records, so as to obtain the load change trend of the multi-core heterogeneous motherboard; A load prediction module, used to predict and analyze the computing core load status at a future moment based on the load change trend, so as to obtain a predicted load form at a future moment; A balancing optimization module, used to perform a balancing optimization analysis of the computing core load on the computing core load status at the current moment and the predicted load form at the future moment according to the load adaptation characteristic distribution, so as to obtain a corresponding load balancing solution; wherein the load balancing solution includes a prediction verification phase and an allocation execution phase; A balancing execution module is used to distribute the load of the computing cores of the multi-core heterogeneous mainboard according to the load balancing scheme to achieve dynamic load balancing of the multi-core heterogeneous mainboard.

Images