Heterogeneous system performance bottleneck monitoring and tuning method and device based on AI, equipment and medium

Abstract

The invention discloses an AI (artificial intelligence)-based heterogeneous system performance bottleneck monitoring and tuning method, device, equipment and medium, belongs to the technical field of performance testing and tuning, and aims to solve the technical problem of how to analyze and predict the performance bottleneck of a heterogeneous system in real time through artificial intelligence so as to propose targeted optimization suggestions. According to the technical scheme, the method comprises the steps that various performance data in the heterogeneous system are collected, and the collected various performance data in the heterogeneous system are analyzed in real time through the AI technology; the performance bottleneck in the heterogeneous system is detected, and adjustment and optimization work of load balancing, resource redistribution and algorithm optimization is automatically executed; the method specifically comprises the following steps: data acquisition and preprocessing; training an AI model; performance bottleneck monitoring and analysis; a tuning strategy is generated and automatically implemented; and feeding back the effect and updating the AI model.

Description

Technical Field

[0001] The present invention relates to the technical field of performance testing and tuning, and specifically to an AI-based heterogeneous system performance bottleneck monitoring and tuning method, device, equipment and medium. Background Art

[0002] In modern computing environments, heterogeneous systems are becoming increasingly mainstream. These systems integrate different types of processing units, but since each processing unit has different architectures and performance characteristics, system performance optimization becomes complex and challenging. Traditional performance monitoring and tuning methods often cannot effectively adapt to this diverse computing environment, resulting in uneven resource allocation and failure to identify performance bottlenecks in a timely manner.

[0003] With the advancement of artificial intelligence technology, all walks of life are using artificial intelligence (AI) technology to solve industry problems.

[0004] Therefore, how to use artificial intelligence to achieve real-time analysis and prediction of the performance bottlenecks of heterogeneous systems, so as to put forward targeted optimization suggestions and improve the overall efficiency and stability of heterogeneous systems is a technical problem that needs to be solved urgently. Summary of the invention

[0005] The technical task of the present invention is to provide an AI-based heterogeneous system performance bottleneck monitoring and tuning method, device, equipment and medium to solve the problem of how to use artificial intelligence to realize real-time analysis and prediction of the performance bottleneck of heterogeneous systems, thereby proposing targeted optimization suggestions and improving the overall efficiency and stability of heterogeneous systems.

[0006] The technical task of the present invention is achieved in the following way: a method for monitoring and optimizing the performance bottleneck of a heterogeneous system based on AI, which collects various performance data in the heterogeneous system, analyzes the collected various performance data in the heterogeneous system in real time through AI technology, detects the performance bottleneck in the heterogeneous system and automatically performs load balancing, resource reallocation and algorithm optimization optimization; the details are as follows:

[0007] Data collection and preprocessing: Deploy monitoring tools to collect and preprocess data to ensure data accuracy and consistency, so that AI can conduct effective analysis.

[0008] Training AI models: Use machine learning or deep learning algorithms to train models, identify patterns and regularities of performance bottlenecks in heterogeneous systems, and build AI models that accurately predict and identify performance bottlenecks;

[0009] Performance bottleneck monitoring and analysis: Use trained AI models to monitor the performance of heterogeneous systems in real time and identify potential bottlenecks. Perform detailed analysis on the identified bottlenecks to determine the causes and impact of the heterogeneous system performance bottlenecks.

[0010] Generate tuning strategies and automatically implement them: Based on the performance bottleneck analysis results, the AI ​​model generates corresponding tuning strategies and implements the generated tuning strategies to adjust heterogeneous system configurations through automated tools.

[0011] Effect feedback and AI model update: Evaluate the effect of tuning strategies, monitor changes in heterogeneous system performance, ensure that problems are resolved, and continuously update and optimize AI models based on actual effects and feedback data to improve the accuracy of detection and tuning, thereby continuously optimizing heterogeneous system performance and ensuring that tuning strategies are always effective and adapt to changes in heterogeneous systems.

[0012] As a preferred option, the deployment monitoring tools are as follows:

[0013] Deploy performance monitoring tools in heterogeneous systems to collect various indicator data during the operation of heterogeneous systems; among them, the indicator data during the operation of heterogeneous systems include CPU utilization, memory usage, GPU load, and network latency.

[0014] As a preference, the collected data is preprocessed as follows:

[0015] Data cleaning and normalization: Clean the collected data, remove noise and outliers to obtain cleaned data, and normalize the cleaned data to obtain normalized data;

[0016] Feature selection: Extract key features from the normalized data; the key features include load changes of the processing unit, task execution time, and resource usage.

[0017] Preferably, the machine learning or deep learning algorithms include regression analysis, decision trees and neural networks.

[0018] Preferably, the causes of performance bottlenecks in heterogeneous systems include computationally intensive tasks and insufficient memory.

[0019] Preferably, the tuning strategy includes adjusting load balancing, optimizing task scheduling, and adjusting resource allocation.

[0020] An AI-based heterogeneous system performance bottleneck monitoring and tuning device, the device comprising:

[0021] Data collection and preprocessing module, which is used to deploy monitoring tools, collect data and preprocess the collected data to ensure the accuracy and consistency of the data, so as to facilitate effective analysis by AI;

[0022] AI model training module, which is used to train models using machine learning or deep learning algorithms, identify patterns and regularities of performance bottlenecks in heterogeneous systems, and build AI models that accurately predict and identify performance bottlenecks;

[0023] The performance bottleneck monitoring and analysis module is used to use the trained AI model to monitor the performance of heterogeneous systems in real time and identify potential bottlenecks. It also conducts detailed analysis of the identified bottlenecks to determine the causes and impact of the heterogeneous system performance bottlenecks.

[0024] The tuning strategy generation module is used to generate corresponding tuning strategies based on the performance bottleneck analysis results and AI models; and implement the generated tuning strategies to adjust the heterogeneous system configurations through automated tools;

[0025] The effect feedback and AI model update module is used to evaluate the effect of the tuning strategy, monitor the performance changes of heterogeneous systems, ensure that the problems are resolved, and continuously update and optimize the AI ​​model based on the actual effects and feedback data to improve the accuracy of detection and tuning, thereby continuously optimizing the performance of heterogeneous systems and ensuring that the tuning strategy is always effective and adapts to changes in heterogeneous systems.

[0026] Preferably, the data acquisition and preprocessing module includes:

[0027] The performance monitoring tool deployment submodule is used to deploy performance monitoring tools in heterogeneous systems and collect various indicator data during the operation of heterogeneous systems. The indicator data during the operation of heterogeneous systems include CPU utilization, memory usage, GPU load, and network latency.

[0028] The data cleaning and normalization submodule is used to clean the collected data, remove noise and outliers to obtain cleaned data, and normalize the cleaned data to obtain normalized data;

[0029] The feature selection submodule is used to extract key features from the normalized data; the key features include the load changes of the processing unit, the task execution time and the resource usage.

[0030] An electronic device comprising: a memory and at least one processor;

[0031] Wherein, the memory stores a computer program;

[0032] The at least one processor executes the computer program stored in the memory, so that the at least one processor executes the above-mentioned AI-based heterogeneous system performance bottleneck monitoring and tuning method.

[0033] A computer-readable storage medium having a computer program stored therein, wherein the computer program can be executed by a processor to implement the above-mentioned AI-based heterogeneous system performance bottleneck monitoring and tuning method.

[0034] The AI-based heterogeneous system performance bottleneck monitoring and tuning method, device, equipment and medium of the present invention have the following advantages:

[0035] (1) The present invention can analyze and predict the performance bottleneck of the system in real time, thereby proposing targeted optimization suggestions, and ultimately improving the overall efficiency and stability of the heterogeneous system. It can be applied to various fields that require high-performance computing and resource optimization, such as data centers, high-performance computing, cloud computing environments, and intelligent manufacturing;

[0036] (ii) The present invention can automatically analyze and integrate performance data from different processing units, identify potential bottlenecks, and thus simplify the diagnosis process of complex performance problems in heterogeneous systems;

[0037] (iii) The present invention can monitor system load and resource usage in real time, dynamically adjust configuration and optimization strategies according to actual needs, and improve the system's adaptive ability and overall performance;

[0038] (iv) The heterogeneous system of the present invention can generate and implement targeted optimization strategies for different processing units, thereby achieving the best performance of each processing unit, rather than relying on a unified optimization solution;

[0039] (V) The present invention can significantly improve performance: AI technology can identify and optimize performance bottlenecks in real time, significantly improving overall system efficiency and response speed;

[0040] (vi) The present invention reduces manual intervention: automated monitoring and tuning reduce the need for manual analysis and adjustment, and reduce maintenance costs;

[0041] (VII) The present invention achieves dynamic adaptation: it can dynamically adjust the configuration according to real-time data, adapt to load changes, and optimize resource usage. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The present invention is further described below in conjunction with the accompanying drawings.

[0043] Attached Figure 1 It is a flowchart of the performance bottleneck monitoring and tuning method of heterogeneous systems based on AI;

[0044] Attached Figure 2 A flowchart of the working process of the AI-based heterogeneous system performance bottleneck monitoring and tuning device DETAILED DESCRIPTION

[0045] The AI-based heterogeneous system performance bottleneck monitoring and tuning method, device, equipment and medium of the present invention are described in detail below with reference to the drawings and specific embodiments of the specification.

[0046] Embodiment 1:

[0047] As attached Figure 1 As shown, this embodiment provides an AI-based heterogeneous system performance bottleneck monitoring and tuning method, which collects various performance data in the heterogeneous system, analyzes the collected various performance data in the heterogeneous system in real time through AI technology, detects the performance bottleneck in the heterogeneous system and automatically performs load balancing, resource reallocation and algorithm optimization tuning; the details are as follows:

[0048] S1. Data collection and preprocessing: Deploy monitoring tools, collect data and preprocess the collected data to ensure the accuracy and consistency of the data, so as to facilitate effective analysis by AI;

[0049] S2. Training AI models: Use machine learning or deep learning algorithms to train models, identify patterns and regularities of performance bottlenecks in heterogeneous systems, and build AI models that accurately predict and identify performance bottlenecks;

[0050] S3. Performance bottleneck monitoring and analysis: Use the trained AI model to monitor the performance of heterogeneous systems in real time and identify potential bottlenecks. Perform detailed analysis on the identified bottlenecks to determine the causes and impact of the heterogeneous system performance bottlenecks.

[0051] S4. Generate tuning strategies and automatically implement them: Based on the performance bottleneck analysis results, the AI ​​model generates corresponding tuning strategies and implements the generated tuning strategies to adjust the heterogeneous system configurations through automated tools.

[0052] S5. Effect feedback and AI model update: Evaluate the effect of the tuning strategy, monitor the performance changes of heterogeneous systems, ensure that the problems are solved, and continuously update and optimize the AI ​​model based on the actual results and feedback data to improve the accuracy of detection and tuning, thereby continuously optimizing the performance of heterogeneous systems and ensuring that the tuning strategy is always effective and adapts to changes in heterogeneous systems.

[0053] The deployment monitoring tool in step S1 of this embodiment is specifically as follows:

[0054] Deploy performance monitoring tools in heterogeneous systems to collect various indicator data during the operation of heterogeneous systems; among them, the indicator data during the operation of heterogeneous systems include CPU utilization, memory usage, GPU load, and network latency.

[0055] The preprocessing of the collected data in step S1 of this embodiment is specifically as follows:

[0056] ① Data cleaning and normalization: Clean the collected data, remove noise and outliers to obtain cleaned data, and normalize the cleaned data to obtain normalized data;

[0057] ② Feature selection: Extract key features from the normalized data; the key features include the load changes of the processing unit, task execution time, and resource usage.

[0058] The machine learning or deep learning algorithm in step S2 of this embodiment includes regression analysis, decision tree and neural network.

[0059] The causes of the heterogeneous system performance bottleneck in step S3 of this embodiment include computationally intensive tasks and insufficient memory.

[0060] The optimization strategy in step S4 of this embodiment includes adjusting load balancing, optimizing task scheduling, and adjusting resource allocation.

[0061] Embodiment 2:

[0062] This embodiment provides an AI-based heterogeneous system performance bottleneck monitoring and tuning device, which includes:

[0063] Data collection and preprocessing module, which is used to deploy monitoring tools, collect data and preprocess the collected data to ensure the accuracy and consistency of the data, so as to facilitate effective analysis by AI;

[0064] AI model training module, which is used to train models using machine learning or deep learning algorithms, identify patterns and regularities of performance bottlenecks in heterogeneous systems, and build AI models that accurately predict and identify performance bottlenecks;

[0065] The performance bottleneck monitoring and analysis module is used to use the trained AI model to monitor the performance of heterogeneous systems in real time and identify potential bottlenecks. It also conducts detailed analysis of the identified bottlenecks to determine the causes and impact of the heterogeneous system performance bottlenecks.

[0066] The tuning strategy generation module is used to generate corresponding tuning strategies based on the performance bottleneck analysis results and AI models; and implement the generated tuning strategies to adjust the heterogeneous system configurations through automated tools;

[0067] The effect feedback and AI model update module is used to evaluate the effect of the tuning strategy, monitor the performance changes of heterogeneous systems, ensure that the problems are resolved, and continuously update and optimize the AI ​​model based on the actual effects and feedback data to improve the accuracy of detection and tuning, thereby continuously optimizing the performance of heterogeneous systems and ensuring that the tuning strategy is always effective and adapts to changes in heterogeneous systems.

[0068] The data acquisition and preprocessing module in this embodiment includes:

[0069] The performance monitoring tool deployment submodule is used to deploy performance monitoring tools in heterogeneous systems and collect various indicator data during the operation of heterogeneous systems. The indicator data during the operation of heterogeneous systems include CPU utilization, memory usage, GPU load, and network latency.

[0070] The data cleaning and normalization submodule is used to clean the collected data, remove noise and outliers to obtain cleaned data, and normalize the cleaned data to obtain normalized data;

[0071] The feature selection submodule is used to extract key features from the normalized data; the key features include the load changes of the processing unit, the task execution time and the resource usage.

[0072] As attached Figure 2 As shown, the working process of the device is as follows:

[0073] (1) Deploy AI monitoring tools: Deploy performance monitoring tools in heterogeneous systems to collect various indicator data during system operation, such as CPU utilization, memory usage, GPU load, network latency, etc.

[0074] (2) Data cleaning and normalization: Clean the collected data, remove noise and outliers, and normalize the data for subsequent analysis;

[0075] (3) Feature selection: extract key features from the raw data, such as load changes of processing units, task execution time, resource usage, etc.

[0076] (4) Model training: Use machine learning or deep learning algorithms (such as regression analysis, decision trees, neural networks, etc.) to train models and identify patterns and regularities of performance bottlenecks;

[0077] (5) Real-time monitoring: Deploy trained AI models to monitor system performance in real time and identify potential bottlenecks;

[0078] (6) Bottleneck analysis: Analyze the identified bottlenecks in detail to determine their causes (such as computationally intensive tasks, insufficient memory, etc.) and the scope of their impact;

[0079] (7) Generate tuning strategies and implement them: Based on the bottleneck analysis results, the AI ​​system generates corresponding tuning strategies, such as adjusting load balancing, optimizing task scheduling, adjusting resource allocation, etc.; implement the generated tuning strategies and adjust system configurations through automated tools;

[0080] (8) Feedback and iteration: Evaluate the effectiveness of the tuning strategy, monitor system performance changes, and ensure that problems are resolved. Based on actual results and feedback data, continuously update and optimize the AI ​​model to improve the accuracy of detection and tuning;

[0081] (9) Output report: Provide users with an easy-to-use interface or command to generate performance monitoring results and tuning suggestions to help users understand system status and improvement measures.

[0082] Embodiment 3:

[0083] This embodiment also provides an electronic device, including: a memory and a processor;

[0084] Wherein, the memory stores computer-executable instructions;

[0085] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the AI-based heterogeneous system performance bottleneck monitoring and tuning method in any embodiment of the present invention.

[0086] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor may be a microprocessor or any conventional processor, etc.

[0087] The memory can be used to store computer programs and / or modules. The processor realizes various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the terminal, etc. In addition, the memory can also include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SMC), a secure digital (SD) card, a flash memory card, at least one disk storage period, a flash memory device, or other volatile solid-state storage devices.

[0088] Embodiment 4:

[0089] This embodiment also provides a computer-readable storage medium, which stores a plurality of instructions, which are loaded by a processor to enable the processor to execute the AI-based heterogeneous system performance bottleneck monitoring and tuning method in any embodiment of the present invention. Specifically, a system or device equipped with a storage medium can be provided, on which a software program code that implements the functions of any of the above embodiments is stored, and a computer (or CPU or MPU) of the system or device reads and executes the program code stored in the storage medium.

[0090] In this case, the program code itself read from the storage medium can realize the function of any one of the above-mentioned embodiments, and thus the program code and the storage medium storing the program code constitute a part of the present invention.

[0091] The storage medium embodiments for providing the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RYM, DVD-RW, DVD+RW), a magnetic tape, a non-volatile memory card, and a ROM. Alternatively, the program code can be downloaded from a server computer via a communication network.

[0092] In addition, it should be clear that the functions of any of the above embodiments can be implemented not only by executing the program code read by the computer, but also by enabling an operating system operating on the computer to complete part or all of the actual operations based on instructions from the program code.

[0093] In addition, it can be understood that the program code read from the storage medium is written to a memory provided in an expansion board inserted into the computer or written to a memory provided in an expansion unit connected to the computer, and then based on the instructions of the program code, a CPU installed on the expansion board or the expansion unit is enabled to perform part or all of the actual operations, thereby realizing the functions of any of the above-mentioned embodiments.

[0094] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring and optimizing performance bottlenecks of heterogeneous systems based on AI, characterized in that: This method collects various performance data in heterogeneous systems, analyzes the collected performance data in heterogeneous systems in real time through AI technology, detects performance bottlenecks in heterogeneous systems, and automatically performs load balancing, resource reallocation, and algorithm optimization tuning; the details are as follows: Data collection and preprocessing: Deploy monitoring tools to collect and preprocess data to ensure data accuracy and consistency, so that AI can conduct effective analysis. Training AI models: Use machine learning or deep learning algorithms to train models, identify patterns and regularities of performance bottlenecks in heterogeneous systems, and build AI models that accurately predict and identify performance bottlenecks; Performance bottleneck monitoring and analysis: Use trained AI models to monitor the performance of heterogeneous systems in real time and identify potential bottlenecks. Perform detailed analysis on the identified bottlenecks to determine the causes and impact of the heterogeneous system performance bottlenecks. Generate tuning strategies and automatically implement them: Based on the performance bottleneck analysis results, the AI ​​model generates corresponding tuning strategies and implements the generated tuning strategies to adjust heterogeneous system configurations through automated tools. Effect feedback and AI model update: Evaluate the effect of tuning strategies, monitor changes in heterogeneous system performance, ensure that problems are resolved, and continuously update and optimize AI models based on actual effects and feedback data to improve the accuracy of detection and tuning, thereby continuously optimizing heterogeneous system performance and ensuring that tuning strategies are always effective and adapt to changes in heterogeneous systems.

2. The AI-based heterogeneous system performance bottleneck monitoring and tuning method according to claim 1 is characterized in that: The deployment monitoring tools are as follows: Deploy performance monitoring tools in heterogeneous systems to collect various indicator data during the operation of heterogeneous systems; among them, the indicator data during the operation of heterogeneous systems include CPU utilization, memory usage, GPU load, and network latency.

3. The AI-based heterogeneous system performance bottleneck monitoring and tuning method according to claim 1 is characterized in that: The preprocessing of the collected data is as follows: Data cleaning and normalization: Clean the collected data, remove noise and outliers to obtain cleaned data, and normalize the cleaned data to obtain normalized data; Feature selection: Extract key features from the normalized data; the key features include load changes of the processing unit, task execution time, and resource usage.

4. The AI-based heterogeneous system performance bottleneck monitoring and tuning method according to claim 1 is characterized in that: Machine learning or deep learning algorithms include regression analysis, decision trees, and neural networks.

5. The AI-based heterogeneous system performance bottleneck monitoring and tuning method according to claim 1 is characterized in that: Causes of performance bottlenecks in heterogeneous systems include compute-intensive tasks and insufficient memory.

6. The AI-based heterogeneous system performance bottleneck monitoring and tuning method according to any one of claims 1 to 5, characterized in that: Tuning strategies include adjusting load balancing, optimizing task scheduling, and adjusting resource allocation.

7. An AI-based heterogeneous system performance bottleneck monitoring and tuning device, characterized in that: The device includes: Data collection and preprocessing module, which is used to deploy monitoring tools, collect data and preprocess the collected data to ensure the accuracy and consistency of the data, so as to facilitate effective analysis by AI; AI model training module, which is used to train models using machine learning or deep learning algorithms, identify patterns and regularities of performance bottlenecks in heterogeneous systems, and build AI models that accurately predict and identify performance bottlenecks; The performance bottleneck monitoring and analysis module is used to use the trained AI model to monitor the performance of heterogeneous systems in real time and identify potential bottlenecks. It also conducts detailed analysis of the identified bottlenecks to determine the causes and impact of the heterogeneous system performance bottlenecks. The tuning strategy generation module is used to generate corresponding tuning strategies based on the performance bottleneck analysis results and AI models; and implement the generated tuning strategies to adjust the heterogeneous system configurations through automated tools; The effect feedback and AI model update module is used to evaluate the effect of the tuning strategy, monitor the performance changes of heterogeneous systems, ensure that the problems are resolved, and continuously update and optimize the AI ​​model based on the actual effects and feedback data to improve the accuracy of detection and tuning, thereby continuously optimizing the performance of heterogeneous systems and ensuring that the tuning strategy is always effective and adapts to changes in heterogeneous systems.

8. The AI-based heterogeneous system performance bottleneck monitoring and tuning device according to claim 7, characterized in that: The data acquisition and preprocessing module includes: The performance monitoring tool deployment submodule is used to deploy performance monitoring tools in heterogeneous systems and collect various indicator data during the operation of heterogeneous systems. The indicator data during the operation of heterogeneous systems include CPU utilization, memory usage, GPU load, and network latency. The data cleaning and normalization submodule is used to clean the collected data, remove noise and outliers to obtain cleaned data, and normalize the cleaned data to obtain normalized data; The feature selection submodule is used to extract key features from the normalized data; the key features include the load changes of the processing unit, the task execution time and the resource usage.

9. An electronic device, characterized in that: include: memory and at least one processor; Wherein, the memory stores a computer program; The at least one processor executes the computer program stored in the memory, so that the at least one processor executes the AI-based heterogeneous system performance bottleneck monitoring and tuning method as described in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which can be executed by a processor to implement the AI-based heterogeneous system performance bottleneck monitoring and tuning method as described in any one of claims 1 to 5.

Images