Calculation network resource measurement method and device and electronic equipment

Abstract

The invention provides a computing network resource measurement method and device and electronic equipment, and the method comprises the steps: obtaining hardware information of a computing unit in computing network resources, transmission data of network transmission equipment, first power consumption data of the computing unit in different working states, and second power consumption data of the network transmission equipment in different working states, the performance evaluation indexes comprise an operation capability index, a transmission capability index, an operation energy consumption efficiency index, a transmission energy consumption efficiency index and a memory and communication efficiency index of the computing network resources, and the performance evaluation indexes comprise the operation capability index, the transmission capability index, the operation energy consumption efficiency index, the transmission energy consumption efficiency index and the memory and communication efficiency index of the computing network resources. Therefore, the computing resources and the network resources in the computing network resources are unified and effectively integrated, and the comprehensiveness and the accuracy of computing network resource measurement are improved; moreover, in the face of large-scale and heterogeneous computing network resources, the adaptability of computing network resource measurement can also be improved, and then the efficiency of computing network resource measurement is improved.

Description

Technical Field

[0001] The present invention relates to the field of computing power network technology, and in particular to a computing network resource measurement method, device and electronic equipment. Background Art

[0002] Large language models (LLMs), with their powerful language understanding and generation capabilities, are widely used in fields such as intelligent customer service, content creation, and intelligent translation. Large models like StableDiffusion, used in image generation, can also generate high-quality images from simple text descriptions. The number of parameters in these large models has grown exponentially, from millions initially to trillions today. Training and inference processes require massive amounts of computing resources, leading to an explosive surge in computing power demand. Statistics show that the computing power required for large model training has increased by more than tenfold annually over the past few years, making it difficult for traditional computing resource provisioning models to meet demand. To address the computing power challenges posed by large models and other emerging technologies, computing-network convergence has become a key solution. Computing-network convergence deeply integrates computing and network resources, connecting geographically distributed computing nodes through the network to enable coordinated resource allocation and sharing. This approach aims to provide users with an integrated, efficient service experience and meet the needs of various complex application scenarios.

[0003] In a converged computing and network system, accurately measuring computing and network resources is crucial. For service providers, accurately understanding the actual availability of computing and network resources helps them rationally plan resource allocation and optimize resource allocation, thereby improving resource utilization and reducing operating costs. For users, a clear understanding of their available computing and network resources allows them to more effectively select appropriate services and ensure stable application operation.

[0004] However, current network resource measurement faces numerous challenges. Different types of computing resources and network resources (such as bandwidth and latency) have unique characteristics and measurement methods. The lack of a unified and effective integration mechanism makes it difficult to form a comprehensive and accurate view of network resources. Furthermore, the complex and ever-changing network environment, dynamic network traffic, and sudden business demands make it even more difficult to accurately measure network resources in real time. Furthermore, existing measurement methods often suffer from inefficiency and poor adaptability when dealing with large-scale, heterogeneous network resources, making them unable to meet growing business demands. Summary of the Invention

[0005] The present invention provides a computing network resource measurement method, device and electronic equipment to solve the defects of the existing measurement methods in the face of large-scale, heterogeneous computing network resources, such as low efficiency and poor adaptability, and inability to meet the growing business needs.

[0006] The present invention provides a computing network resource measurement method, comprising the following steps: Obtaining hardware information of a computing unit in a computing network resource, transmission data of a network transmission device, first power consumption data of the computing unit in different working states, and second power consumption data of the network transmission device in different working states; Determining a performance evaluation indicator based on the hardware information, the transmission data, the first power consumption data, and the second power consumption data, and determining a performance evaluation result of the computing network resource based on the performance evaluation indicator; The performance evaluation indicators include the computing power indicator, transmission power indicator, computing energy efficiency indicator, transmission energy efficiency indicator and memory and communication efficiency indicator of the computing network resources; the computing power indicator is used to reflect the core computing power of various computing units, the transmission power indicator is used to reflect the transmission performance of data between different nodes, devices and the computing units in the computing network environment, the computing energy efficiency indicator is used to reflect the amount of computing tasks completed by the computing unit under unit energy consumption, the transmission energy efficiency indicator is used to reflect the amount of data transmission tasks completed by the computing unit under unit energy consumption, and the memory and communication efficiency indicator is used to reflect the data transmission speed between the computing unit and the memory and the communication bandwidth between different computing units.

[0007] According to a computing network resource measurement method provided by the present invention, the computing capacity index is determined based on the weight of each computing operation in the computing unit and the core computing capacity of each computing operation; The transmission capacity index is determined based on the weight of files of each transmission type in the network transmission device and the transmission capacity of files of each transmission type in the network transmission device; The computing energy efficiency index is determined based on the average power consumption of each computing unit during task execution and the computing capability index; The transmission energy efficiency index is determined based on the average power consumption and the transmission capacity index; The memory and communication efficiency indicators are determined based on data transmission bandwidth, transmission delay and communication protocol overhead.

[0008] According to a computing network resource measurement method provided by the present invention, each computing unit includes a central processing unit, a graphics processing unit, a neural processing unit and a field programmable gate array; The core computing capabilities of the central processing unit are integer and floating-point computing capabilities; The core computing capability of the graphics processing unit is parallel computing capability; The core computing capability of the neural processing unit is convolution computing capability; The core computing capability of the field programmable gate array is the logic operation capability.

[0009] According to a computing network resource measurement method provided by the present invention, determining a performance evaluation indicator based on the hardware information, the transmission data, the first power consumption data, and the second power consumption data, and determining a performance evaluation result of the computing network resource based on the performance evaluation indicator includes: The hardware information, the transmission data, the first power consumption data and the second power consumption data are input into a computing network resource measurement model, and the computing network resource measurement model determines a performance evaluation indicator based on the hardware information, the transmission data, the first power consumption data and the second power consumption data, and determines a performance score and / or performance level based on the performance evaluation indicator.

[0010] According to a computing network resource measurement method provided by the present invention, the steps of training the computing network resource measurement model include: Acquire an initial computing network resource measurement model; the initial computing network resource measurement model includes an initial coding module, and an initial performance score prediction branch and an initial performance level prediction branch respectively connected to the initial coding module; Obtaining sample data and a tag performance evaluation result of the sample data; the sample data includes sample hardware information, sample transmission data, sample first power consumption data, and sample second power consumption data; the tag performance evaluation result includes a tag performance score and a tag performance level; Based on the initial encoding module, feature encoding is performed on the sample hardware information, the sample transmission data, the sample first power consumption data, and the sample second power consumption data, and feature fusion is performed on the encoded features to obtain fused features; Based on the initial performance score prediction branch, performing performance score prediction on the fusion feature to obtain a predicted performance score of the sample data; Based on the initial performance level prediction branch, performing performance level prediction on the fusion feature to obtain a predicted performance level of the sample data; Based on the difference between the predicted performance score and the label performance score, and the difference between the predicted performance level and the label performance level, a target loss is determined, and based on the target loss, the initial computing network resource measurement model is iterated to obtain the computing network resource measurement model.

[0011] According to a computing network resource measurement method provided by the present invention, the hardware information, the transmission data, the first power consumption data, and the second power consumption data are input into a computing network resource measurement model, the computing network resource measurement model determines a performance evaluation index based on the hardware information, the transmission data, the first power consumption data, and the second power consumption data, and determines a performance score and / or performance level based on the performance evaluation index, and then further includes: Storing the performance evaluation results, as well as the model structure and model parameters of the computing network resource measurement model in a result database; The model structure and the model parameters are visually displayed.

[0012] According to a computing network resource measurement method provided by the present invention, the visual display of the model structure and the model parameters includes: Visualize the number of neurons and the connection mode in each layer of the input layer, hidden layer and output layer in the model structure; Use either a heat map or a matrix map to visualize the weight matrix and bias vector in the model parameters.

[0013] According to a computing network resource measurement method provided by the present invention, the hardware information is obtained by the data acquisition module through the interaction between the hardware driver or the system management interface and the computing unit; The transmission data is obtained by the data acquisition module interacting with the network transmission device through the network management protocol and deep packet inspection technology.

[0014] The present invention also provides a computing network resource measurement device, comprising the following units: An acquisition unit, configured to acquire hardware information of a computing unit in a computing network resource, transmission data of a network transmission device, first power consumption data of the computing unit in different working states, and second power consumption data of the network transmission device in different working states; a determining unit, configured to determine a performance evaluation indicator based on the hardware information, the transmission data, the first power consumption data, and the second power consumption data, and determine a performance evaluation result of the computing network resource based on the performance evaluation indicator; The performance evaluation indicators include the computing power indicator, transmission power indicator, computing energy efficiency indicator, transmission energy efficiency indicator and memory and communication efficiency indicator of the computing network resources; the computing power indicator is used to reflect the core computing power of various computing units, the transmission power indicator is used to reflect the transmission performance of data between different nodes, devices and the computing units in the computing network environment, the computing energy efficiency indicator is used to reflect the amount of computing tasks completed by the computing unit under unit energy consumption, the transmission energy efficiency indicator is used to reflect the amount of data transmission tasks completed by the computing unit under unit energy consumption, and the memory and communication efficiency indicator is used to reflect the data transmission speed between the computing unit and the memory and the communication bandwidth between different computing units.

[0015] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any one of the above-described methods for measuring computing network resources is implemented.

[0016] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for measuring computing network resources.

[0017] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned methods for measuring computing network resources.

[0018] The computing network resource measurement method, device and electronic device provided by the present invention obtain the hardware information of the computing unit in the computing network resources, the transmission data of the network transmission equipment, the first power consumption data of the computing unit in different working states, and the second power consumption data of the network transmission equipment in different working states, and determine the performance evaluation indicators based on these data, thereby determining the performance evaluation results of the computing network resources, wherein the performance evaluation indicators include the computing capacity indicator, transmission capacity indicator, computing energy efficiency indicator, transmission energy efficiency indicator and memory and communication efficiency indicator of the computing network resources, thereby unifying and effectively integrating the computing resources and network resources in the computing network resources, improving the comprehensiveness and accuracy of the computing network resource measurement; and, when facing large-scale, heterogeneous computing network resources, it can also improve the adaptability of the computing network resource measurement, thereby improving the efficiency of the computing network resource measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 This is one of the flow charts of the computing network resource measurement method provided by the present invention.

[0021] Figure 2 This is the second flow chart of the computing network resource measurement method provided by the present invention.

[0022] Figure 3 This is a functional architecture diagram of the computing network resource measurement system provided by the present invention.

[0023] Figure 4 It is a structural diagram of the computing network resource measurement device provided by the present invention.

[0024] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0025] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0026] The terms "first," "second," and the like in the present invention are used to distinguish similar objects, and are not used to describe a particular order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein, and that the objects distinguished by "first," "second," and the like are generally of the same type.

[0027] Figure 1 This is one of the flow charts of the computing network resource measurement method provided by the present invention, such as Figure 1 As shown, the method includes step 110 and step 120.

[0028] Step 110: Obtain hardware information of the computing unit in the computing network resources, transmission data of the network transmission device, first power consumption data of the computing unit in different working states, and second power consumption data of the network transmission device in different working states.

[0029] Specifically, the hardware information of the computing unit in the computing network resources, the transmission data of the network transmission equipment, the first power consumption data of the computing unit in different working states, and the second power consumption data of the network transmission equipment in different working states can be obtained. Among them, the computing unit includes each computing unit including a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an NPU (Neural Processing Unit), and an FPGA (Field-Programmable Gate Array), etc., which is not specifically limited in the embodiment of the present invention.

[0030] Here, the hardware information of the computing unit includes basic information such as model, number of cores, frequency, etc., which is not specifically limited in the embodiment of the present invention.

[0031] Here, the network transmission equipment may include switches and routers, etc., which are not specifically limited in the embodiment of the present invention.

[0032] Transmission data from network transmission devices may include network interface bandwidth usage, real-time transmission latency, and packet loss rate, among other data. This is not specifically limited in the present embodiment. Bandwidth usage reflects the ability of a network interface to transmit data within a specific timeframe, typically measured in bits per second (bps). Monitoring bandwidth usage can help identify network bottlenecks and optimize network resource allocation. Packet loss rate refers to the proportion of data packets lost during network transmission.

[0033] Here, the first power consumption data of the computing unit in different working states and the second power consumption data of the network transmission device in different working states are both obtained by connecting to a smart meter or integrating an energy consumption sensor inside the device.

[0034] Step 120: determining a performance evaluation indicator based on the hardware information, the transmission data, the first power consumption data, and the second power consumption data, and determining a performance evaluation result of the computing network resource based on the performance evaluation indicator; The performance evaluation indicators include the computing power indicator, transmission power indicator, computing energy efficiency indicator, transmission energy efficiency indicator and memory and communication efficiency indicator of the computing network resources; the computing power indicator is used to reflect the core computing power of various computing units, the transmission power indicator is used to reflect the transmission performance of data between different nodes, devices and the computing units in the computing network environment, the computing energy efficiency indicator is used to reflect the amount of computing tasks completed by the computing unit under unit energy consumption, the transmission energy efficiency indicator is used to reflect the amount of data transmission tasks completed by the computing unit under unit energy consumption, and the memory and communication efficiency indicator is used to reflect the data transmission speed between the computing unit and the memory and the communication bandwidth between different computing units.

[0035] Specifically, after obtaining hardware information, transmission data, first power consumption data and second power consumption data, performance evaluation indicators can be determined based on the hardware information transmission data, first power consumption data and second power consumption data, and the performance evaluation results of the computing network resources can be determined based on the performance evaluation indicators.

[0036] Here, the computing network resource measurement model can determine the performance evaluation index based on the hardware information transmission data, the first power consumption data and the second power consumption data, and determine the performance evaluation result of the computing network resource based on the performance evaluation index.

[0037] Performance evaluation metrics include the Compute Capability Index (CCI), Transfer Capability Index (TCI), Compute Energy Efficiency Index (CEEI), Transfer Energy Efficiency Index (TEEI), and Memory and Communication Efficiency Index (MCEI). The Compute Capability Index reflects the core computing capabilities of various computing units, taking into account the complexity and execution efficiency of the instruction set and assigning different weights to the operations of different types of computing units. The Transfer Capacity Index reflects the data transmission performance between different nodes, devices, and computing units in the computing network environment, and considers the transmission capabilities of different data types (such as data annotation files, high-definition video streams, and model parameter files).

[0038] The computing energy efficiency indicator reflects the amount of computing tasks completed by a computing unit per unit of energy consumption. By monitoring the energy consumption and computing output of computing units under different workloads, a correlation model between energy consumption and computing power is established. The transmission energy efficiency indicator reflects the amount of data transmission tasks completed by a computing unit per unit of energy consumption. By monitoring the energy consumption and duration of computing units transmitting different data types, a correlation model between energy consumption and transmission capacity is established.

[0039] Here, the memory and communication efficiency metrics reflect the data transmission speed between the computing unit and the memory, as well as the communication bandwidth between different computing units. This considers the impact of factors such as memory access latency and communication protocol overhead on data transmission efficiency.

[0040] The method provided by the embodiment of the present invention obtains the hardware information of the computing unit in the computing network resources, the transmission data of the network transmission equipment, the first power consumption data of the computing unit in different working states, and the second power consumption data of the network transmission equipment in different working states, and determines the performance evaluation indicators based on these data, thereby determining the performance evaluation results of the computing network resources, wherein the performance evaluation indicators include the computing capacity indicators, transmission capacity indicators, computing energy efficiency indicators, transmission energy efficiency indicators and memory and communication efficiency indicators of the computing network resources, thereby unifying and effectively integrating the computing resources and network resources in the computing network resources, improving the comprehensiveness and accuracy of the computing network resource measurement; and, when facing large-scale, heterogeneous computing network resources, it can also improve the adaptability of the computing network resource measurement, thereby improving the efficiency of the computing network resource measurement.

[0041] Based on the above embodiment, the computing capability index is determined based on the weight of each computing operation in the computing unit and the core computing capability of each computing operation; The transmission capacity index is determined based on the weight of files of each transmission type in the network transmission device and the transmission capacity of files of each transmission type in the network transmission device; The computing energy efficiency index is determined based on the average power consumption of each computing unit during task execution and the computing capability index; The transmission energy efficiency index is determined based on the average power consumption and the transmission capacity index; The memory and communication efficiency indicators are determined based on data transmission bandwidth, transmission delay and communication protocol overhead.

[0042] Specifically, the computing power index is determined based on the weight of each computing operation in the computing unit and the core computing power of each computing operation. The formula is as follows: in, Indicates the computing power index, Represents the weight of each operation, Indicates the core computing power of each operation.

[0043] Here, each operation includes integer operation, floating-point operation, matrix operation, convolution operation, activation function operation, logical operation, etc., which is not specifically limited in the embodiment of the present invention.

[0044] The transmission capacity index is determined based on the weight of each transmission type of file in the network transmission device and the transmission capacity of each transmission type of file in the network transmission device. The formula is as follows: in, Indicates the transmission capacity indicator, Indicates the weight of each transmission type of files in the network transmission device, Indicates the transmission capacity of files of various transmission types in network transmission devices.

[0045] Here, files of various transmission types include data annotation files, high-definition video streams, and model parameter files, etc., which are not specifically limited in the embodiment of the present invention.

[0046] The computing energy efficiency index is determined based on the average power consumption of each computing unit during the execution of the task and the computing power index. The formula is as follows: in, represents the computing energy efficiency index, Indicates the computing power index, Indicates the average power consumption of each computing unit during task execution.

[0047] Here, the transmission energy efficiency index is determined based on the average power consumption and the transmission capacity index, and the formula is as follows: in, represents the transmission energy efficiency index, Indicates the transmission capacity indicator, Indicates the average power consumption of each computing unit during task execution.

[0048] Here, the memory and communication efficiency indicators are determined based on data transmission bandwidth, transmission delay, and communication protocol overhead, and the formula is as follows: in, Represents memory and communication efficiency indicators, Indicates the data transmission bandwidth, Indicates the transmission delay, Indicates the communication protocol overhead.

[0049] The method provided by the embodiments of this invention, on the one hand, considers the weights of different computing operations and transmission types in its computing capacity and transmission capacity indicators, making the evaluation results more accurate and more realistically reflecting the actual performance of computing units and network transmission equipment. On the other hand, the computing energy efficiency indicators and transmission energy efficiency indicators can be used to evaluate the amount of computing and transmission tasks per unit energy consumption, helping to optimize resource utilization. This solution can adapt to different types of computing resources (such as CPUs and GPUs) and network resources (such as bandwidth and latency), providing a unified evaluation framework that helps to form a comprehensive and accurate view of computing and network resources.

[0050] Based on the above embodiment, each computing unit includes a central processing unit, a graphics processing unit, a neural processing unit and a field programmable gate array; The core computing capabilities of the central processing unit are integer and floating-point computing capabilities; The core computing capability of the graphics processing unit is parallel computing capability; The core computing capability of the neural processing unit is convolution computing capability; The core computing capability of the field programmable gate array is the logic operation capability.

[0051] Specifically, the computing units include central processing units (CPUs), graphics processing units (GPUs), neural processing units (NPUs), and field-programmable gate arrays (FPGAs). The computing power indicator is used to measure the core computing capabilities of each type of computing unit. The core computing capabilities of the CPU are integer and floating-point computing capabilities, the GPU is parallel computing capability, the NPU is convolution computing capability, and the FPGA is logical computing capability.

[0052] Parallel computing capability refers to the ability of a graphics processing unit (GPU) to perform multiple computing tasks simultaneously. Convolution computing capability refers to the ability of a neural processing unit (NPU) to perform convolution operations. Logic computing capability refers to the ability of a field-programmable gate array (FPGA) to perform logic operations.

[0053] This segmented evaluation method for the core computing capabilities of different computing units can more comprehensively understand and optimize system performance, improve resource utilization efficiency, reduce costs, and support a wider range of application scenarios.

[0054] Based on the above embodiment, step 120 includes: Step 121: Input the hardware information, the transmission data, the first power consumption data, and the second power consumption data into a computing network resource measurement model; the computing network resource measurement model determines a performance evaluation indicator based on the hardware information, the transmission data, the first power consumption data, and the second power consumption data; and determines a performance score and / or performance level based on the performance evaluation indicator.

[0055] Specifically, the hardware information, transmission data, first power consumption data and second power consumption data are input into the computing network resource measurement model, and the computing network resource measurement model determines the performance evaluation indicators based on the hardware information, transmission data, first power consumption data and second power consumption data, and determines the performance score and / or performance level based on the performance evaluation indicators.

[0056] Here, the computing network resource measurement model can be a deep learning model or a machine learning model. The computing network resource measurement model can include a performance score prediction branch and a performance level prediction branch. Accordingly, the performance evaluation results can include performance scores, performance levels, performance scores and performance levels, etc. The embodiment of the present invention does not make specific limitations on this.

[0057] Based on the above embodiment, the steps of training the computing network resource measurement model include: Step 210: Obtain an initial computing network resource measurement model; the initial computing network resource measurement model includes an initial encoding module, and an initial performance score prediction branch and an initial performance level prediction branch respectively connected to the initial encoding module; Step 220: Obtain sample data and a tag performance evaluation result of the sample data; the sample data includes sample hardware information, sample transmission data, sample first power consumption data, and sample second power consumption data; the tag performance evaluation result includes a tag performance score and a tag performance level; Step 230: Based on the initial encoding module, feature encoding is performed on the sample hardware information, the sample transmission data, the sample first power consumption data, and the sample second power consumption data, and feature fusion is performed on the encoded features to obtain fused features. Step 240: Based on the initial performance score prediction branch, predict the performance score of the fusion feature to obtain a predicted performance score of the sample data; Step 250: Based on the initial performance level prediction branch, perform performance level prediction on the fusion feature to obtain a predicted performance level of the sample data; Step 260, based on the difference between the predicted performance score and the label performance score, and the difference between the predicted performance level and the label performance level, determine the target loss, and based on the target loss, perform parameter iteration on the initial computing network resource measurement model to obtain the computing network resource measurement model.

[0058] Specifically, an initial computing network resource measurement model is obtained, wherein the initial computing network resource measurement model may include an initial encoding module, and an initial performance score prediction branch and an initial performance level prediction branch respectively connected to the initial encoding module.

[0059] Here, the parameters of the initial computing network resource measurement model may be pre-set or randomly generated, and this embodiment of the present invention does not impose any specific limitation on this.

[0060] Then obtain sample data and label performance evaluation results of the sample data, wherein the sample data includes sample hardware information, sample transmission data, sample first power consumption data, and sample second power consumption data, and the label performance evaluation results include label performance score and label performance level.

[0061] Furthermore, based on the initial coding module, feature coding is performed on the sample hardware information, the sample transmission data, the sample first power consumption data and the sample second power consumption data, and feature fusion is performed on the coding features after feature coding to obtain fused features.

[0062] Here, the initial encoding module may include a multi-layer convolutional neural network (CNN) with a cascade structure, a deep neural network (DNN), or a combination of CNN and DNN, etc., which is not specifically limited in the embodiment of the present invention.

[0063] After obtaining the fused features, the performance score of the fused features can be predicted based on the initial performance score prediction branch to obtain the predicted performance score of the sample data. Furthermore, the performance level of the fused features can be predicted based on the initial performance level prediction branch to obtain the predicted performance level of the sample data. The predicted performance level can be high, medium, or low.

[0064] Finally, based on the difference between the predicted performance score and the label performance score, the first loss is determined. Based on the difference between the predicted performance level and the label performance level, the second loss is determined. Based on the first loss and the second loss, the target loss is determined. Based on the target loss, the parameters of the initial computing network resource measurement model are iterated, and the initial computing network resource measurement model after the parameter iteration is completed is used as the computing network resource measurement model.

[0065] It can be understood that the smaller the difference between the predicted performance score and the label performance score, the smaller the first loss; the larger the difference between the predicted performance score and the label performance score, the larger the first loss.

[0066] It can be understood that the smaller the difference between the predicted performance level and the label performance level, the smaller the second loss; and the larger the difference between the predicted performance level and the label performance level, the larger the second loss.

[0067] It should be noted that these sample data are used to train the computing network resource measurement model. By adjusting the parameters of the computing network resource measurement model, the model's prediction results are made as close as possible to the label performance evaluation results. In this process, cross-validation, regularization and other technologies can be used to prevent model overfitting and improve the generalization ability of the computing network resource measurement model.

[0068] Based on the above embodiment, step 121 further includes: Step 1211: storing the performance evaluation results, as well as the model structure and model parameters of the computing network resource measurement model in a result database; Step 1212: Visually display the model structure and the model parameters.

[0069] Specifically, the performance evaluation results, as well as the model structure and parameters of the computing network resource measurement model, are stored in a result database. This allows the model structure and parameters to be read from the result database and visualized. For example, the number of neurons and the connection method in each layer of the input layer, hidden layer, and output layer of the model structure can be visualized.

[0070] Here, either a heat map or a matrix map can be used to visualize the weight matrix and bias vector in the model parameters.

[0071] As you can see, the information about the computing network resource measurement model is read from the result database. The model structure visualization section presents each component of the computing network resource measurement model in a clear hierarchical structure, such as the input layer, hidden layer, and output layer of the neural network. For complex models, a layered layout is automatically created, displaying the number of neurons in each layer and their connection structure, allowing users to intuitively understand the model architecture.

[0072] For the parameters of each layer of the model, such as the weight matrix and bias vector, heat maps, matrix diagrams and other methods are used in the model parameter visualization part to visualize the distribution and changes of the parameters, provide users with intuitive data display, and help users understand the model performance.

[0073] Based on the above embodiment, the hardware information is obtained by the data acquisition module through interaction with the computing unit via a hardware driver or a system management interface; The transmission data is obtained by the data acquisition module interacting with the network transmission device through the network management protocol and deep packet inspection technology.

[0074] Specifically, hardware information is obtained by the data acquisition module through interaction with the computing unit via the hardware driver or system management interface, and transmission data is obtained by the data acquisition module through interaction with the network transmission device via the Simple Network Management Protocol (SNMP) and Deep Packet Inspection (DPI) technology.

[0075] Here, the Network Management Protocol (NMP) is a standard protocol used to manage and monitor network devices (such as routers, switches, and servers). It allows network administrators to centrally collect device status information, configure device parameters, and monitor network performance and faults. Deep packet inspection (DPI) is a network monitoring technology that examines not only the header information of network packets but also the payload (i.e., the packet's contents). This technology can identify and classify network traffic, including applications, services, and protocols.

[0076] Based on any of the above embodiments, Figure 2 This is the second flow chart of the computing network resource measurement method provided by the present invention. Figure 2 As shown, the method includes: First, in the data collection stage, hardware information of the computing unit, transmission data of the network transmission device, and first power consumption data and second power consumption data are obtained.

[0077] Then, in the data modeling stage, the hardware information of the computing unit, the transmission data of the network transmission equipment, and the first power consumption data and the second power consumption data are input into the computing network resource measurement model to obtain the comprehensive performance evaluation result output by the computing network resource measurement model, and the comprehensive performance evaluation result is stored in the structural database.

[0078] In the model visualization phase, the model information is read from the result database, and the model structure and parameter visualization content in the model information are displayed.

[0079] Based on any of the above embodiments, Figure 3 This is the functional architecture diagram of the computing network resource measurement system provided by the present invention, such as Figure 3As shown in the figure, (1) Data acquisition module: After the system is started, the data acquisition module starts working. For computing units (CPU, GPU, NPU, FPGA, etc.), by interacting with the hardware driver or system management interface, it collects hardware parameters in real time, such as model, number of cores, frequency and other basic information. At the same time, it monitors the operating status data, such as current load, temperature, etc., and temporarily stores this data in the local cache.

[0080] For network transmission equipment (switches, routers, etc.), network management protocols are used to collect data such as network interface bandwidth usage, real-time transmission delay, and packet loss rate. Deep packet inspection technology and application-layer protocol parsing algorithms are used to classify, count, and record the transmission of different data types (data annotation files, high-definition video streams, model parameter files, etc.), and the data is also stored in a local cache.

[0081] By connecting to smart meters or energy consumption sensors integrated inside devices, real-time power consumption data of the computing unit and transmission unit under different working conditions is obtained, and association is established with previously collected data, stored in the local cache, and waits for transmission to subsequent modules.

[0082] (2) Data Modeling Module: After data collection is completed, the collected parameters and the calculation formulas for each indicator are input into the computing network resource measurement model. Prior to this, the model training and optimization phase has used a large amount of performance data (including measured data and simulated data) of different types of computing units under various workloads to train the neural network model.

[0083] During the training process, model parameters are adjusted and cross-validation and regularization techniques are used to prevent overfitting, ensuring that the model's predictions are as close as possible to actual performance data, thereby improving the model's generalization capabilities. After receiving input data, the trained model performs computational processing and outputs a comprehensive performance evaluation of the computing unit, such as a performance level (high, medium, or low) or a specific performance score. The results are then stored in a results database.

[0084] (3) Model Visualization Module: This module reads information about the computing network resource measurement model from the result database and presents the various components of the model in a clear hierarchical structure in the model structure visualization section, such as the input layer, hidden layer, and output layer of the neural network. For complex models, a layered layout is automatically performed, displaying the number of neurons in each layer and the connection method, allowing users to intuitively understand the model architecture.

[0085] For the parameters of each layer of the model, such as the weight matrix and bias vector, heat maps, matrix diagrams and other methods are used in the model parameter visualization part to visualize the distribution and changes of the parameters, provide users with intuitive data display, and help users understand the model performance.

[0086] In summary, the embodiments of the present invention obtain the hardware parameters, operating status and energy consumption data of the computing unit and network equipment in real time, providing comprehensive support for resource evaluation. The data modeling module uses deep learning technology, combined with a large amount of measured and simulated data to train the optimization model, accurately evaluate the performance of the computing unit, and output a reliability performance level or score. The model visualization module displays the model architecture and parameter changes in the form of intuitive hierarchical structures and heat maps, which makes it easy for users to quickly understand the model performance and resource status. Overall, the system can achieve accurate measurement and efficient management of resources, help optimize resource allocation, improve system operation efficiency and energy consumption management level, and is suitable for a variety of complex computing network environments.

[0087] The method provided by the embodiments of the present invention addresses the current problems of a lack of integrated mechanisms in computing network resource measurement, difficulty in achieving accurate real-time measurement, low efficiency, and poor adaptability, thereby forming a comprehensive and accurate view of computing network resources. This method helps service providers rationally plan resource layout and optimize configuration, thereby improving resource utilization and reducing costs. It also helps users clearly understand available computing network resources, accurately select appropriate services, and ensure stable application operation. This, in turn, promotes the in-depth development of computing network integration and enhances the overall level of information technology services.

[0088] The computing network resource measurement device provided by the present invention is described below. The computing network resource measurement device described below and the computing network resource measurement method described above can be referenced to each other.

[0089] Based on any of the above embodiments, the present invention provides a computing network resource measurement device, Figure 4 This is a schematic diagram of the structure of the computing network resource measurement device provided by the present invention. Figure 4 As shown, the device includes: An acquisition unit 410 is configured to acquire hardware information of a computing unit in a computing network resource, transmission data of a network transmission device, first power consumption data of the computing unit in different working states, and second power consumption data of the network transmission device in different working states; a determining unit 420 configured to determine a performance evaluation indicator based on the hardware information, the transmission data, the first power consumption data, and the second power consumption data, and determine a performance evaluation result of the computing network resource based on the performance evaluation indicator; The performance evaluation indicators include the computing power indicator, transmission power indicator, computing energy efficiency indicator, transmission energy efficiency indicator and memory and communication efficiency indicator of the computing network resources; the computing power indicator is used to reflect the core computing power of various computing units, the transmission power indicator is used to reflect the transmission performance of data between different nodes, devices and the computing units in the computing network environment, the computing energy efficiency indicator is used to reflect the amount of computing tasks completed by the computing unit under unit energy consumption, the transmission energy efficiency indicator is used to reflect the amount of data transmission tasks completed by the computing unit under unit energy consumption, and the memory and communication efficiency indicator is used to reflect the data transmission speed between the computing unit and the memory and the communication bandwidth between different computing units.

[0090] The device provided by the embodiment of the present invention obtains the hardware information of the computing unit in the computing network resources, the transmission data of the network transmission equipment, the first power consumption data of the computing unit in different working states, and the second power consumption data of the network transmission equipment in different working states, and determines the performance evaluation indicators based on these data, thereby determining the performance evaluation results of the computing network resources, wherein the performance evaluation indicators include the computing capacity indicators, transmission capacity indicators, computing energy efficiency indicators, transmission energy efficiency indicators and memory and communication efficiency indicators of the computing network resources, thereby unifying and effectively integrating the computing resources and network resources in the computing network resources, improving the comprehensiveness and accuracy of the computing network resource measurement; and, when facing large-scale, heterogeneous computing network resources, it can also improve the adaptability of the computing network resource measurement, thereby improving the efficiency of the computing network resource measurement.

[0091] Based on any of the above embodiments, the computing capacity indicator is determined based on the weight of each computing operation in the computing unit and the core computing capacity of each computing operation; The transmission capacity index is determined based on the weight of files of each transmission type in the network transmission device and the transmission capacity of files of each transmission type in the network transmission device; The computing energy efficiency index is determined based on the average power consumption of each computing unit during task execution and the computing capability index; The transmission energy efficiency index is determined based on the average power consumption and the transmission capacity index; The memory and communication efficiency indicators are determined based on data transmission bandwidth, transmission delay and communication protocol overhead.

[0092] Based on any of the above embodiments, each computing unit includes a central processing unit, a graphics processing unit, a neural processing unit and a field programmable gate array; The core computing capabilities of the central processing unit are integer and floating-point computing capabilities; The core computing capability of the graphics processing unit is parallel computing capability; The core computing capability of the neural processing unit is convolution computing capability; The core computing capability of the field programmable gate array is the logic operation capability.

[0093] Based on any of the foregoing embodiments, the determining unit 420 is specifically configured to: The hardware information, the transmission data, the first power consumption data and the second power consumption data are input into a computing network resource measurement model, and the computing network resource measurement model determines a performance evaluation indicator based on the hardware information, the transmission data, the first power consumption data and the second power consumption data, and determines a performance score and / or performance level based on the performance evaluation indicator.

[0094] Based on any of the above embodiments, the further comprising a training unit, wherein the training unit is specifically configured to: Acquire an initial computing network resource measurement model; the initial computing network resource measurement model includes an initial coding module, and an initial performance score prediction branch and an initial performance level prediction branch respectively connected to the initial coding module; Obtaining sample data and a tag performance evaluation result of the sample data; the sample data includes sample hardware information, sample transmission data, sample first power consumption data, and sample second power consumption data; the tag performance evaluation result includes a tag performance score and a tag performance level; Based on the initial encoding module, feature encoding is performed on the sample hardware information, the sample transmission data, the sample first power consumption data, and the sample second power consumption data, and feature fusion is performed on the encoded features to obtain fused features; Based on the initial performance score prediction branch, performing performance score prediction on the fusion feature to obtain a predicted performance score of the sample data; Based on the initial performance level prediction branch, performing performance level prediction on the fusion feature to obtain a predicted performance level of the sample data; Based on the difference between the predicted performance score and the label performance score, and the difference between the predicted performance level and the label performance level, a target loss is determined, and based on the target loss, the initial computing network resource measurement model is iterated to obtain the computing network resource measurement model.

[0095] Based on any of the above embodiments, it further includes a visualization unit, and the visualization unit specifically includes: A storage unit, configured to store the performance evaluation results, and the model structure and model parameters of the computing network resource measurement model in a result database; A visualization display unit is used to visualize the model structure and the model parameters.

[0096] Based on any of the above embodiments, the visual display unit is specifically used to: Visualize the number of neurons and the connection mode in each layer of the input layer, hidden layer and output layer in the model structure; Use either a heat map or a matrix map to visualize the weight matrix and bias vector in the model parameters.

[0097] Based on any of the above embodiments, the hardware information is obtained by the data acquisition module interacting with the computing unit through a hardware driver or a system management interface; The transmission data is obtained by the data acquisition module interacting with the network transmission device through the network management protocol and deep packet inspection technology.

[0098] Figure 5 Schematic diagram of the structure of the electronic device provided by the present invention, such as Figure 5 As shown, the electronic device may include: a processor (processor) 510, a communication interface (Communications Interface) 520, a memory (memory) 530 and a communication bus 540, wherein the processor 510, the communication interface 520, and the memory 530 communicate with each other through the communication bus 540. The processor 510 can call the logic instructions in the memory 530 to execute the computing network resource measurement method, which includes: obtaining hardware information of the computing unit in the computing network resource, transmission data of the network transmission device, first power consumption data of the computing unit in different working states, and second power consumption data of the network transmission device in different working states; determining performance evaluation indicators based on the hardware information, the transmission data, the first power consumption data and the second power consumption data, and determining the performance evaluation result of the computing network resource based on the performance evaluation indicators; the performance evaluation indicators include the computing power indicator, transmission power indicator, computing energy efficiency indicator, transmission energy efficiency indicator and memory and communication efficiency indicator of the computing network resource; the computing power indicator is used to reflect the core computing power of various computing units, the transmission power indicator is used to reflect the transmission performance of data between different nodes, devices and the computing units in the computing network environment, the computing energy efficiency indicator is used to reflect the amount of computing tasks completed by the computing unit under unit energy consumption, the transmission energy efficiency indicator is used to reflect the amount of data transmission tasks completed by the computing unit under unit energy consumption, and the memory and communication efficiency indicator is used to reflect the data transmission speed between the computing unit and the memory and the communication bandwidth between different computing units.

[0099] Furthermore, the logic instructions in the aforementioned memory 530 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0100] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the computing network resource measurement method provided by the above methods. The method includes: obtaining hardware information of the computing unit in the computing network resource, transmission data of the network transmission device, first power consumption data of the computing unit in different working states, and second power consumption data of the network transmission device in different working states; determining a performance evaluation index based on the hardware information, the transmission data, the first power consumption data and the second power consumption data, and determining the performance evaluation index of the computing network resource based on the performance evaluation index. Evaluation results; the performance evaluation indicators include the computing power indicator, transmission power indicator, computing energy efficiency indicator, transmission energy efficiency indicator and memory and communication efficiency indicator of the computing network resources; the computing power indicator is used to reflect the core computing power of various computing units, the transmission power indicator is used to reflect the transmission performance of data between different nodes, devices and the computing units in the computing network environment, the computing energy efficiency indicator is used to reflect the amount of computing tasks completed by the computing unit under unit energy consumption, the transmission energy efficiency indicator is used to reflect the amount of data transmission tasks completed by the computing unit under unit energy consumption, and the memory and communication efficiency indicator is used to reflect the data transmission speed between the computing unit and the memory and the communication bandwidth between different computing units.

[0101] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the computing network resource measurement method provided by the above-mentioned methods, the method comprising: obtaining hardware information of a computing unit in the computing network resource, transmission data of a network transmission device, first power consumption data of the computing unit in different working states, and second power consumption data of the network transmission device in different working states; determining a performance evaluation index based on the hardware information, the transmission data, the first power consumption data and the second power consumption data, and determining a performance evaluation result of the computing network resource based on the performance evaluation index; the performance evaluation index comprises The computing power indicators, transmission power indicators, computing energy efficiency indicators, transmission energy efficiency indicators and memory and communication efficiency indicators of the computing network resources; the computing power indicators are used to reflect the core computing power of various computing units, the transmission power indicators are used to reflect the transmission performance of data between different nodes, devices and the computing units in the computing network environment, the computing energy efficiency indicators are used to reflect the amount of computing tasks completed by the computing unit under unit energy consumption, the transmission energy efficiency indicators are used to reflect the amount of data transmission tasks completed by the computing unit under unit energy consumption, and the memory and communication efficiency indicators are used to reflect the data transmission speed between the computing unit and the memory and the communication bandwidth between different computing units.

[0102] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0103] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0104] 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 make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A computing network resource measurement method, characterized in that: include: Obtaining hardware information of a computing unit in a computing network resource, transmission data of a network transmission device, first power consumption data of the computing unit in different working states, and second power consumption data of the network transmission device in different working states; Determining a performance evaluation indicator based on the hardware information, the transmission data, the first power consumption data, and the second power consumption data, and determining a performance evaluation result of the computing network resource based on the performance evaluation indicator; The performance evaluation indicators include the computing power indicator, transmission power indicator, computing energy efficiency indicator, transmission energy efficiency indicator and memory and communication efficiency indicator of the computing network resources; the computing power indicator is used to reflect the core computing power of various computing units, the transmission power indicator is used to reflect the transmission performance of data between different nodes, devices and the computing units in the computing network environment, the computing energy efficiency indicator is used to reflect the amount of computing tasks completed by the computing unit under unit energy consumption, the transmission energy efficiency indicator is used to reflect the amount of data transmission tasks completed by the computing unit under unit energy consumption, and the memory and communication efficiency indicator is used to reflect the data transmission speed between the computing unit and the memory and the communication bandwidth between different computing units.

2. The computing network resource measurement method according to claim 1, characterized in that: The computing capability indicator is determined based on the weight of each computing operation in the computing unit and the core computing capability of each computing operation; The transmission capacity index is determined based on the weight of files of each transmission type in the network transmission device and the transmission capacity of files of each transmission type in the network transmission device; The computing energy efficiency index is determined based on the average power consumption of each computing unit during task execution and the computing capability index; The transmission energy efficiency index is determined based on the average power consumption and the transmission capacity index; The memory and communication efficiency indicators are determined based on data transmission bandwidth, transmission delay and communication protocol overhead.

3. The computing network resource measurement method according to claim 2, characterized in that: Each computing unit includes a central processing unit, a graphics processing unit, a neural processing unit and a field programmable gate array; The core computing capabilities of the central processing unit are integer and floating-point computing capabilities; The core computing capability of the graphics processing unit is parallel computing capability; The core computing capability of the neural processing unit is convolution computing capability; The core computing capability of the field programmable gate array is the logic operation capability.

4. The computing network resource measurement method according to any one of claims 1 to 3, characterized in that: The determining a performance evaluation indicator based on the hardware information, the transmission data, the first power consumption data, and the second power consumption data, and determining a performance evaluation result of the computing network resource based on the performance evaluation indicator includes: The hardware information, the transmission data, the first power consumption data and the second power consumption data are input into a computing network resource measurement model, and the computing network resource measurement model determines a performance evaluation indicator based on the hardware information, the transmission data, the first power consumption data and the second power consumption data, and determines a performance score and / or performance level based on the performance evaluation indicator.

5. The method for measuring computing network resources according to claim 4, characterized in that: The training steps of the computing network resource measurement model include: Acquire an initial computing network resource measurement model; the initial computing network resource measurement model includes an initial coding module, and an initial performance score prediction branch and an initial performance level prediction branch respectively connected to the initial coding module; Obtaining sample data and a tag performance evaluation result of the sample data; the sample data includes sample hardware information, sample transmission data, sample first power consumption data, and sample second power consumption data; the tag performance evaluation result includes a tag performance score and a tag performance level; Based on the initial encoding module, feature encoding is performed on the sample hardware information, the sample transmission data, the sample first power consumption data, and the sample second power consumption data, and feature fusion is performed on the encoded features to obtain fused features; Based on the initial performance score prediction branch, performing performance score prediction on the fusion feature to obtain a predicted performance score of the sample data; Based on the initial performance level prediction branch, performing performance level prediction on the fusion feature to obtain a predicted performance level of the sample data; Based on the difference between the predicted performance score and the label performance score, and the difference between the predicted performance level and the label performance level, a target loss is determined, and based on the target loss, the initial computing network resource measurement model is iterated to obtain the computing network resource measurement model.

6. The method for measuring computing network resources according to claim 4, characterized in that: The step of inputting the hardware information, the transmission data, the first power consumption data, and the second power consumption data into a computing network resource measurement model, determining a performance evaluation indicator based on the hardware information, the transmission data, the first power consumption data, and the second power consumption data by the computing network resource measurement model, and determining a performance score and / or performance level based on the performance evaluation indicator, further comprising: Storing the performance evaluation results, as well as the model structure and model parameters of the computing network resource measurement model in a result database; The model structure and the model parameters are visually displayed.

7. The method for measuring computing network resources according to claim 6, characterized in that: The visual display of the model structure and the model parameters includes: Visualize the number of neurons and the connection mode in each layer of the input layer, hidden layer and output layer in the model structure; Use either a heat map or a matrix map to visualize the weight matrix and bias vector in the model parameters.

8. The computing network resource measurement method according to any one of claims 1 to 3, characterized in that: The hardware information is obtained by the data acquisition module through the interaction between the hardware driver or the system management interface and the computing unit; The transmission data is obtained by the data acquisition module interacting with the network transmission device through the network management protocol and deep packet inspection technology.

9. A computing network resource measurement device, characterized in that: include: An acquisition unit, configured to acquire hardware information of a computing unit in a computing network resource, transmission data of a network transmission device, first power consumption data of the computing unit in different working states, and second power consumption data of the network transmission device in different working states; a determining unit, configured to determine a performance evaluation indicator based on the hardware information, the transmission data, the first power consumption data, and the second power consumption data, and determine a performance evaluation result of the computing network resource based on the performance evaluation indicator; The performance evaluation indicators include the computing power indicator, transmission power indicator, computing energy efficiency indicator, transmission energy efficiency indicator and memory and communication efficiency indicator of the computing network resources; the computing power indicator is used to reflect the core computing power of various computing units, the transmission power indicator is used to reflect the transmission performance of data between different nodes, devices and the computing units in the computing network environment, the computing energy efficiency indicator is used to reflect the amount of computing tasks completed by the computing unit under unit energy consumption, the transmission energy efficiency indicator is used to reflect the amount of data transmission tasks completed by the computing unit under unit energy consumption, and the memory and communication efficiency indicator is used to reflect the data transmission speed between the computing unit and the memory and the communication bandwidth between different computing units.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the computing network resource measurement method according to any one of claims 1 to 8 is implemented.

Images