Calculation power network performance prediction method and related equipment

Abstract

The invention provides a computing power network performance prediction method and related equipment. The method comprises the following steps: acquiring historical performance data of a computing power network as a training set; inputting the training set into a computing power network performance prediction model comprising a test time training layer, pre-training the computing power network performance prediction model to obtain a pre-trained computing power network performance prediction model, locking the weight of a backbone network module in the pre-trained computing power network performance prediction model, and determining the weight of the backbone network module in the pre-trained computing power network performance prediction model; training a test time training layer in the pre-trained computing power network performance prediction model, and updating the weight of the test time training layer to adapt to the change of a network performance mode; inputting the network performance data of the target computing power network into the trained computing power network performance prediction model for prediction to obtain a performance prediction result of the target computing power network; the method can adapt to the change of a computing power network performance mode while the prediction accuracy is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of computing power network performance prediction, and in particular to a computing power network performance prediction method and related equipment. Background Art

[0002] As a new type of information infrastructure, computing network integrates multi-level computing resources such as cloud platforms, edge servers, and terminal devices to provide data perception, transmission, storage, and computing services. Its core is to allocate and flexibly schedule computing resources, storage resources, and network resources between the cloud, edge, and end according to business needs, so as to improve the quality of service experience and optimize and efficiently utilize network and computing resources. The computing network has four major characteristics: resource abstraction, business guarantee, unified management and control, and elastic scheduling. These characteristics make the prediction of computing network performance complex and important.

[0003] As a link connecting all parties’ computing resources, the computing network’s network performance has a decisive impact on computing power utilization, so issues such as network latency and bandwidth limitations cannot be ignored. In this context, intelligent and automated technologies have become key means to improve computing network efficiency, and performance prediction is an important prerequisite for achieving computing network efficiency improvement. Accurate performance prediction can achieve dynamic adjustment and optimization of the computing network, thereby improving resource utilization and service quality.

[0004] However, there are still some problems in predicting computing power network performance: 1. The performance change pattern of the computing power network shows extremely strong nonlinearity. A single advanced general timing model cannot be directly used for computing power network performance prediction. It needs to be adapted or combined with multiple models to obtain good prediction results.

[0005] 2. The performance pattern of the computing power network will change significantly over time, which requires the model to continuously learn the changes in the performance pattern. Most models do not have the ability to learn online.

[0006] 3. Lack of a good online learning mechanism. Directly introducing online learning functions into the model will often cause the model parameters to jitter, and the weights to be positive or negative, which will reduce the model's predictive ability. Summary of the invention

[0007] The present invention provides a computing power network performance prediction method and related equipment, the purpose of which is to ensure the accuracy of the prediction while being able to adapt to changes in the computing power network performance model.

[0008] In order to achieve the above object, the present invention provides a computing power network performance prediction method, comprising: Step 1: Obtain historical performance data of the computing network as a training set; Step 2, constructing a computing power network performance prediction model, the computing power network performance prediction model includes an embedding module for embedding input data, a Mamba backbone network module for capturing nonlinear features of the data after the embedding operation, and an output module for converting the captured nonlinear features into prediction results, wherein the Mamba backbone network module includes a test time training layer; Step 3: input the training set into the computing power network performance prediction model, pre-train the computing power network performance prediction model, and obtain the pre-trained computing power network performance prediction model. The weight of the backbone network module in the pre-trained computing power network performance prediction model is locked, and the test time training layer in the pre-trained computing power network performance prediction model is trained. The weight of the test time training layer is updated to adapt to the change of the network performance mode, and the trained computing power network performance prediction model is obtained; Step 4: Input the network performance data of the target computing power network into the trained computing power network performance prediction model for prediction to obtain the performance prediction result of the target computing power network.

[0009] Furthermore, before the training set is input into the constructed computing network performance prediction model, it also includes: Select wavelet decomposition to decompose the training set and obtain low-frequency components and high-frequency components; Use windows of different scales to perform sliding average pooling on the training set to obtain the processed training set; The processed training set is concatenated with the low-frequency component and the high-frequency component and embedded to obtain the frequency domain reconstruction target.

[0010] Furthermore, the expression of the frequency domain reconstruction objective is: ; in, represents the frequency domain reconstruction target, represents the embedding operation, Represents a splicing operation, represents wavelet transform, represents sliding average pooling, Represents input network performance data.

[0011] Furthermore, the Mamba backbone network module also includes a first linear layer, a second linear layer, a third linear layer, a first convolutional layer, a first activation layer, a second activation layer, and a multiplier; The input end of the first linear layer and the input end of the second linear layer are both the input end of the Mamba backbone network module, and are connected to the output end of the embedding layer; An output of the first linear layer is connected to an input of the first convolutional layer, an output of the first convolutional layer is connected to an input of the first activation layer, an input of the first activation layer is connected to an input of the test-time training layer, and an output of the test-time training layer is connected to a first input of the multiplier; The output end of the second linear layer is connected to the input end of the second activation layer, and the output end of the second activation layer is connected to the second input end of the multiplier; The output end of the multiplier is connected to the input end of the third linear layer, and the output end of the third linear layer is connected to the input end of the output module.

[0012] Specifically, the output module is a multi-layer perceptron network consisting of 2 linear layers.

[0013] More specifically, step 3 includes: The computing power network performance prediction model is pre-trained using the training set to obtain a pre-trained computing power network performance prediction model, in which the weights of the backbone network modules are locked; A self-supervised loss function for the test time training layer in the computing power network performance prediction model is constructed in a self-supervised manner. The test time training layer is trained based on the self-supervised loss function. The weights of the test time training layer are updated to adapt to changes in the network performance pattern, and the trained computing power network performance prediction model is obtained.

[0014] Specifically, the self-supervised loss function is: ; in, represents the self-supervised loss function, represents the weights of the training layer at test time, represents the input of the test-time training layer, , , represents a low-rank matrix, represents the model of the test-time trained layers, Represents the frequency domain reconstruction target.

[0015] The present invention also provides a computing power network performance prediction device, comprising: The acquisition module is used to obtain the historical performance data of the computing network as a training set; A construction module is used to construct a computing power network performance prediction model, wherein the computing power network performance prediction model includes an embedding module for embedding input data, a Mamba backbone network module for capturing nonlinear features of the data after the embedding operation, and an output module for converting the captured nonlinear features into prediction results, wherein the Mamba backbone network module includes a test time training layer; A training module is used to input the training set into the computing power network performance prediction model, pre-train the computing power network performance prediction model, and obtain a pre-trained computing power network performance prediction model. The weight of the backbone network module in the pre-trained computing power network performance prediction model is locked, and the test time training layer in the pre-trained computing power network performance prediction model is trained. The weight of the test time training layer is updated to adapt to the change of the network performance mode, and the trained computing power network performance prediction model is obtained; The prediction module is used to input the network performance data of the target computing power network into the trained computing power network performance prediction model for prediction, and obtain the performance prediction result of the target computing power network.

[0016] The present invention also provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements a computing power network performance prediction method when executing the computer program.

[0017] The present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, a computing power network performance prediction method is implemented.

[0018] The above scheme of the present invention has the following beneficial effects: The present invention obtains historical performance data of a computing power network as a training set; inputs the training set into a computing power network performance prediction model including a test time training layer, pre-trains the computing power network performance prediction model, obtains a pre-trained computing power network performance prediction model, the weight of a backbone network module in the pre-trained computing power network performance prediction model is locked, trains the test time training layer in the pre-trained computing power network performance prediction model, updates the weight of the test time training layer to adapt to changes in the network performance mode, obtains a trained computing power network performance prediction model, and the test time training layer is set in a Mamba backbone network module; inputs network performance data of a target computing power network into the trained computing power network performance prediction model for prediction, and obtains a performance prediction result of the target computing power network; compared with the prior art, the present invention integrates the characteristics of a recurrent neural network and a convolutional neural network through a Mamba backbone network module, solves the problem of computational efficiency when processing long time series, uses a test time training layer to regard a hidden state as a model, so that it can better capture the nonlinear characteristics of the sequence, and can adapt to changes in the computing power network performance mode while ensuring prediction accuracy.

[0019] Other beneficial effects of the present invention will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of a flow chart of an embodiment of the present invention; Figure 2 It is a schematic diagram of the structure of the computing power network performance prediction model in an embodiment of the present invention; Figure 3 It is a schematic diagram of the structure of a computing power network performance prediction device in an embodiment of the present invention; Figure 4 Schematic diagram of the structure of a terminal device in an embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0023] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installation", "connection" and "connection" should be understood in a broad sense, for example, it can be a locking connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0024] In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0025] In view of the existing problems, the present invention provides a computing power network performance prediction method and related equipment.

[0026] like Figure 1 , Figure 2 As shown, an embodiment of the present invention provides a computing power network performance prediction method, comprising: Step 1: Obtain historical performance data of the computing network as a training set; Step 2, constructing a computing power network performance prediction model, the computing power network performance prediction model includes an embedding module for embedding input data, a Mamba backbone network module for capturing nonlinear features of the data after the embedding operation, and an output module for converting the captured nonlinear features into prediction results, wherein the Mamba backbone network module includes a test time training layer; Step 3: input the training set into the computing power network performance prediction model, pre-train the computing power network performance prediction model, and obtain the pre-trained computing power network performance prediction model. The weight of the backbone network module in the pre-trained computing power network performance prediction model is locked, and the test time training layer in the pre-trained computing power network performance prediction model is trained. The weight of the test time training layer is updated to adapt to the change of the network performance mode, and the trained computing power network performance prediction model is obtained; Step 4: Input the network performance data of the target computing power network into the trained computing power network performance prediction model for prediction to obtain the performance prediction result of the target computing power network.

[0027] In the embodiment of the present invention, the historical performance data includes but is not limited to network traffic, bandwidth, latency, etc.

[0028] In the embodiment of the present invention, the maximum performance value and the minimum performance value are first obtained from the historical performance data, and the historical performance data is standardized. The standardized expression is: ; in, represents the standardized historical performance data, Represents historical performance data, Indicates the maximum performance. Indicates the minimum performance value; Set the size of each historical performance data to T and the corresponding label size to P to obtain multiple samples and corresponding labels, taking the first 80% of the samples as the training set and the remaining 20% ​​of the samples as the test set; in the embodiment of the present invention, T is set to 24 and P is set to 10, that is, the historical performance data of the first 24 moments are used to predict the data of the next 10 moments.

[0029] Specifically, before the training set is input into the constructed computing network performance prediction model, it also includes: Select wavelet decomposition for training set Decompose to obtain low-frequency components and high-frequency components; Use windows of different sizes to train the Perform sliding average pooling to obtain the processed training set; The processed training set is concatenated with the low-frequency component and the high-frequency component and embedded to obtain the frequency domain reconstruction target.

[0030] In the embodiment of the present invention, the low-frequency component corresponds to a slowly changing signal, and the high-frequency component corresponds to a rapidly changing signal; since the signals are mixed together and have no characteristics, and the frequency domain information can often better show the specific changes in the time series, the performance series has strong nonlinear characteristics, indicating that the performance series is non-stationary, so it is necessary to decompose the training set to obtain more characteristic multiple low-frequency components and high-frequency components; the low-frequency component has a low scale and a large window, and the result is smoother, and the high-frequency component has a high scale and a small window, which retains more details. The processed training set is spliced ​​with the low-frequency component and the high-frequency component, and then mapped to the specified dimension through an embedding operation to obtain a frequency domain reconstruction target that highlights the details. The expression of the frequency domain reconstruction target is: ; in, represents the frequency domain reconstruction target, Represents an embedding operation that maps the input data to a specified dimension. Represents a splicing operation, represents wavelet transform, represents sliding average pooling, Represents input network performance data.

[0031] Most preferably, in the embodiment of the present invention, the embedding module is composed of a linear layer, which maps the input data to a specified dimension and converts the discrete data into a continuous vector with a fixed dimension. The calculation expression is: ; in, represents the embedding vector, Represents the input network performance data, Indicates an embedding operation; This transformation increases the dimensionality of the data and also increases the expressiveness of numerical data, allowing the model to process various types of data more efficiently.

[0032] It should be noted that the Mamba backbone network module is the Mamba backbone network, and the embodiment of the present invention replaces the state space model in the Mamba backbone network with a test time training layer.

[0033] Specifically, the Mamba backbone network module also includes a first linear layer, a second linear layer, a third linear layer, a first convolutional layer, a first activation layer, a second activation layer, and a multiplier; The input end of the first linear layer and the input end of the second linear layer are both the input end of the Mamba backbone network module, and are connected to the output end of the embedding layer; An output of the first linear layer is connected to an input of the first convolutional layer, an output of the first convolutional layer is connected to an input of the first activation layer, an input of the first activation layer is connected to an input of the test-time training layer, and an output of the test-time training layer is connected to a first input of the multiplier; The output end of the second linear layer is connected to the input end of the second activation layer, and the output end of the second activation layer is connected to the second input end of the multiplier; The output end of the multiplier is connected to the input end of the third linear layer, and the output end of the third linear layer is connected to the input end of the output module.

[0034] In an embodiment of the present invention, after the historical performance data is processed by the embedding module, the obtained embedding vector is input into the Mamba backbone network module and divided into two branches. The first branch first passes through the first linear layer, then passes through the first convolutional layer, and then is activated by the first activation layer and input into the test time training layer for prediction to obtain the first feature vector; the second branch first passes through the second linear layer, and then is activated by the second activation layer to obtain the second feature vector, and then the first feature vector and the second feature vector output by the two branches are multiplied by the multiplier and processed by the third linear layer as the output input and output module of the Mamba backbone network module.

[0035] In the embodiment of the present invention, the first activation layer and the second activation layer are both silu functions.

[0036] In the embodiment of the present invention, the key to the test time training layer is not to compress the context into a hidden state of a fixed size. , but the hidden state As a model, hidden state Equivalent to the weight of a model, which can be a linear model, a small neural network model, or any other form of model, the output rule of the training layer at test time can be expressed as: ; in, represents the input of the test-time training layer, represents the model of the test-time trained layers, represents the output of the training layer at test time, i.e., Updated weights For input predictions made; In an embodiment of the present invention, in order to enable the model to capture the mutation of performance data, the test time training layer is improved so that the improved test time training layer can enhance the nonlinear ability of the model and be more sensitive to performance changes. That is, the reconstruction goal of the test time training layer should further highlight the detail features and trend features.

[0037] Most preferably, the output module is a multi-layer perceptron network consisting of 2 linear layers.

[0038] The embodiment of the present invention converts the output results of the backbone network into performance prediction results for a period of time in the future through a multi-layer perception network.

[0039] Specifically, step 3 includes: The computing power network performance prediction model is pre-trained using the training set to obtain a pre-trained computing power network performance prediction model, in which the weights of the backbone network modules are locked; A self-supervised loss function for the test time training layer in the computing power network performance prediction model is constructed in a self-supervised manner. The test time training layer is trained based on the self-supervised loss function. The weights of the test time training layer are updated to adapt to changes in the network performance pattern, and the trained computing power network performance prediction model is obtained.

[0040] Specifically, the embodiment of the present invention uses a training set to pre-train the computing power network performance prediction model, wherein the parameters of the Mamba backbone network and the test time training layer are trained in a supervised and self-supervised manner respectively.

[0041] The Mamba backbone network performs supervised training and calculates the loss after the output module. The loss function is: ; in, For real performance results, is the prediction result of the entire model.

[0042] The specific training process is: (1): Set the batch size and divide the training set into several batches; (2): Each batch of data is fed into the computing network performance prediction model to calculate the LOSS loss function; (3): Perform iterative optimization based on the Adam algorithm until the set number of iterations.

[0043] (4): After the iteration, the trained model parameters are saved.

[0044] In order to adapt to the unknown changes in network performance patterns, the weights of the test time training layer are always updated independently. The self-supervised loss function of the test time training layer in the computing power network performance prediction model is constructed in a self-supervised manner. The test time training layer is trained based on the self-supervised loss function to obtain the weight of the test time training layer.

[0045] In this embodiment of the present invention, the update rule of the test-time training layer is to update the training layer under a self-supervised loss Further gradient descent is performed on , and the calculation expression is: ; in, represents the learning rate; Loss Function The choice is to reconstruct itself. In order to make the learning problem non-trivial, the embodiment of the present invention first inputs Treated as a corrupted input , and then optimize the loss function, the expression is: ; By refactoring the input This process, model Able to discover The correlation between the dimensions in order to obtain the information from the destroyed Reconstruct the original information . The input sequence Mapping to a sequence The process of is programmed into the forward propagation of the test-time training layer, using the hidden state, update rule, and output rule described above. Even at test time, the test-time training layer still trains a different sequence of weights for each input sequence , that is, online learning is realized.

[0046] The most important part of training a layer at test time is the self-supervision task, as it determines the weights The types of features learned from the test set. In the design of this task, the original test-time training layer takes a more end-to-end approach, directly optimizing the self-supervised task to achieve the ultimate goal of next token prediction. Starting from a simple reconstruction task, some outer loop parameters are added to make this task learnable, so the self-supervised loss is optimized to: ; The output becomes: ; , , is a series of low-rank matrices that can To carry out destruction, , , are the outer loop parameters, learned in the Mamba backbone network module, and the weights of the test-time training layer It is an inner loop parameter and does not participate in the outer loop update. After the computing network performance prediction model training is completed, the outer loop parameters are no longer updated, but the new performance data can update the inner loop parameters, so that the model can learn the subsequent performance change pattern, but it will cause large jitter in the overall model.

[0047] The improved TTT layer calculates the gradient through the improved self-supervised loss function, and further updates the weight W and the bias term b. During the subsequent use of the model, the backbone network parameters will not change, but when data is input into the TTT layer, the TTT layer parameters will be updated, which can avoid large jitters in the entire model and adapt the model to new data.

[0048] Specifically, the self-supervised loss function is: ; in, represents the self-supervised loss function, represents the weights of the training layer at test time, represents the input of the test-time training layer, , , represents a low-rank matrix, represents the model of the test-time trained layers, represents the frequency domain reconstruction target; Since there are a lot of redundant features in the high-frequency and low-frequency components, it is possible to use the low-rank matrix to filter.

[0049] In the embodiment of the present invention, the model used in the test time training layer is It is a linear model and can be expressed as: ; in, For Model Input, is the weight of the model, is the bias term of the function.

[0050] In order to test and evaluate the prediction accuracy of the model, the embodiment of the present invention inputs the test set into the trained computing power network performance prediction model for prediction, obtains the single-step prediction result, and evaluates the model with two indicators: root mean square error and absolute percentage mean error. The calculation expression is: ; ; in, represents the root mean square error, represents the absolute percentage mean error, Represents the prediction result of the model, represents the labels in the training set, Indicates the number of prediction results or labels.

[0051] The embodiment of the present invention collects network performance data of a certain computing power node in the target computing power network, with an interval of one hour, 24 groups per day, a total of 204 days of network performance data, the data of the first 194 days is used as a training set, the data of the last 10 days is used as a test set, the review window size is set to 24, and samples are continuously generated in the form of a sliding window. When the prediction window size is set to 10, a multi-step prediction is performed to predict the performance values ​​of the next 10 hours. The relationship between the prediction error and the number of layers of the Mamba backbone network module is shown in Table 1: Table 1 Error size corresponding to the number of layers of different Mamba backbone network modules

[0052] It can be seen from Table 1 above that when the number of layers of the Mamba backbone network module is set to 3, the model performs best.

[0053] The number of layers of the Mamba backbone network module is set to 3, and single-step prediction and multi-step prediction are performed respectively. The results are shown in Table 2: Table 2 Single-step prediction and multi-step prediction results

[0054] The prediction step size is set to 5 and compared with other time series prediction models. The results are shown in Table 3: Table 3 Comparative experimental results when the prediction step length is 5

[0055] It can be seen from Table 3 above that the method provided by the present invention exceeds some benchmark timing models and advanced timing models.

[0056] The embodiment of the present invention compares the original test time training layer with the improved test time training layer, and the comparison results are shown in Table 4 below: Table 4 Comparison results of the original test time training layer and the improved test time training layer

[0057] It can be seen from Table 4 above that the result error of the improved test time training layer is lower and has stronger nonlinear ability, which proves the effectiveness of the embodiment of the present invention.

[0058] The embodiment of the present invention obtains historical performance data of a computing power network as a training set; inputs the training set into a computing power network performance prediction model including a test time training layer, pre-trains the computing power network performance prediction model, obtains a pre-trained computing power network performance prediction model, the weight of a backbone network module in the pre-trained computing power network performance prediction model is locked, trains the test time training layer in the pre-trained computing power network performance prediction model, updates the weight of the test time training layer to adapt to changes in the network performance mode, obtains a trained computing power network performance prediction model, and the test time training layer is set in a Mamba backbone network module; inputs the network performance data of a target computing power network into the trained computing power network performance prediction model for prediction, and obtains a performance prediction result of the target computing power network; compared with the prior art, the embodiment of the present invention solves the problem of computational efficiency when processing long time series by integrating the characteristics of a recurrent neural network and a convolutional neural network through a Mamba backbone network module, and uses the test time training layer to regard a hidden state as a model, so that it can better capture the nonlinear characteristics of the sequence, and can adapt to changes in the computing power network performance mode while ensuring prediction accuracy.

[0059] Corresponding to the computing power network performance prediction method described in the above embodiment, Figure 3 As shown, the embodiment of the present invention further provides a computing power network performance prediction device 100, and the computing power network performance prediction device 100 includes: An acquisition module 101 is used to acquire historical performance data of a computing network as a training set; A construction module 102 is used to construct a computing power network performance prediction model, which includes an embedding module for embedding input data, a Mamba backbone network module for capturing nonlinear features of the data after the embedding operation, and an output module for converting the captured nonlinear features into prediction results, wherein the Mamba backbone network module includes a test time training layer; The training module 103 is used to input the training set into the computing power network performance prediction model, pre-train the computing power network performance prediction model, and obtain the pre-trained computing power network performance prediction model. The weight of the backbone network module in the pre-trained computing power network performance prediction model is locked, and the test time training layer in the pre-trained computing power network performance prediction model is trained. The weight of the test time training layer is updated to adapt to the change of the network performance mode, so as to obtain the trained computing power network performance prediction model; The prediction module 104 is used to input the network performance data of the target computing power network into the trained computing power network performance prediction model for prediction to obtain the performance prediction result of the target computing power network.

[0060] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0061] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0062] The embodiment of the present invention also provides a terminal device, such as Figure 4 As shown, the terminal device D10 of this embodiment includes: at least one processor D100 ( Figure 4 Only one processor is shown in the figure), a memory D101, and a computer program D102 stored in the memory D101 and executable on the at least one processor D100. When the processor D100 executes the computer program D102, the above-mentioned computing power network performance prediction method is implemented.

[0063] The terminal device D10 may be a computing device such as a desktop computer, a notebook, a PDA, a server, a server cluster, a cloud server, etc. The terminal device may include, but is not limited to, a processor D100 and a memory D101. Those skilled in the art will appreciate that Figure 4 This is only an example of the terminal device D10 and does not constitute a limitation on the terminal device D10. The terminal device D10 may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, it may also include input and output devices, network access devices, etc.

[0064] The processor D100 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. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0065] In some embodiments, the memory D101 may be an internal storage unit of the terminal device D10, such as a hard disk or memory of the terminal device D10. In other embodiments, the memory D101 may also be an external storage device of the terminal device D10, such as a plug-in hard disk, a smart memory card (SMC, SmartMedia Card), a secure digital (SD, Secure Digital) card, a flash card (Flash Card), etc. equipped on the terminal device D10. Further, the memory D101 may also include both an internal storage unit of the terminal device D10 and an external storage device. The memory D101 is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as the program code of the computer program, etc. The memory D101 may also be used to temporarily store data that has been output or is to be output.

[0066] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0067] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0068] An embodiment of the present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, a computing power network performance prediction method is implemented.

[0069] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device that can carry the computer program code to the construction device / terminal device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, RandomAccess Memory), electric carrier signal, telecommunication signal and software distribution medium. For example, a USB flash drive, a mobile hard disk, a disk or an optical disk.

[0070] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A computing power network performance prediction method, characterized in that: include: Step 1: Obtain historical performance data of the computing network as a training set; Step 2, constructing a computing power network performance prediction model, the computing power network performance prediction model includes an embedding module for embedding input data, a Mamba backbone network module for capturing nonlinear features of the data after the embedding operation, and an output module for converting the captured nonlinear features into prediction results, wherein the Mamba backbone network module includes a test time training layer; Step 3, inputting the training set into the computing power network performance prediction model, pre-training the computing power network performance prediction model, obtaining a pre-trained computing power network performance prediction model, wherein the weight of the backbone network module in the pre-trained computing power network performance prediction model is locked, and training the test time training layer in the pre-trained computing power network performance prediction model, and updating the weight of the test time training layer to adapt to changes in the network performance mode, thereby obtaining a trained computing power network performance prediction model; Step 4: Input the network performance data of the target computing power network into the trained computing power network performance prediction model for prediction to obtain the performance prediction result of the target computing power network.

2. The computing power network performance prediction method according to claim 1 is characterized in that: Before inputting the training set into the constructed computing power network performance prediction model, it also includes: Select wavelet decomposition to decompose the training set to obtain low-frequency components and high-frequency components; Using windows of multiple different scales to perform sliding average pooling processing on the training set to obtain a processed training set; The processed training set is concatenated with the low-frequency component and the high-frequency component and an embedding operation is performed to obtain a frequency domain reconstruction target.

3. The computing power network performance prediction method according to claim 2 is characterized in that: The expression of the frequency domain reconstruction objective is: in, represents the frequency domain reconstruction target, represents the embedding operation, Represents a splicing operation, represents wavelet transform, represents sliding average pooling, Represents input network performance data.

4. The computing power network performance prediction method according to claim 3 is characterized in that: The Mamba backbone network module also includes a first linear layer, a second linear layer, a third linear layer, a first convolutional layer, a first activation layer, a second activation layer, and a multiplier; The input end of the first linear layer and the input end of the second linear layer are both input ends of the Mamba backbone network module, and are connected to the output end of the embedding layer; An output of the first linear layer is connected to an input of the first convolutional layer, an output of the first convolutional layer is connected to an input of the first activation layer, an input of the first activation layer is connected to an input of the test-time training layer, and an output of the test-time training layer is connected to a first input of the multiplier; The output end of the second linear layer is connected to the input end of the second activation layer, and the output end of the second activation layer is connected to the second input end of the multiplier; The output end of the multiplier is connected to the input end of the third linear layer, and the output end of the third linear layer is connected to the input end of the output module.

5. The computing power network performance prediction method according to claim 4 is characterized in that: The output module is a multi-layer perception network composed of 2 linear layers.

6. The computing power network performance prediction method according to claim 5 is characterized in that: The step 3 comprises: Pre-training the computing power network performance prediction model using the training set to obtain a pre-trained computing power network performance prediction model, wherein the weight of the backbone network module in the pre-trained computing power network performance prediction model is locked; A self-supervised loss function of the test time training layer in the computing power network performance prediction model is constructed in a self-supervised manner, the test time training layer is trained based on the self-supervised loss function, and the weight of the test time training layer is updated to adapt to changes in the network performance pattern, so as to obtain a trained computing power network performance prediction model.

7. The computing power network performance prediction method according to claim 6, characterized in that: The self-supervised loss function is: in, represents the self-supervised loss function, represents the weights of the training layer at test time, represents the input of the test-time training layer, , , represents a low-rank matrix, represents the model of the test-time trained layers, Represents the frequency domain reconstruction target.

8. A computing power network performance prediction device, characterized in that: include: The acquisition module is used to obtain the historical performance data of the computing network as a training set; A construction module is used to construct a computing power network performance prediction model, wherein the computing power network performance prediction model includes an embedding module for embedding input data, a Mamba backbone network module for capturing nonlinear features of the data after the embedding operation, and an output module for converting the captured nonlinear features into prediction results, wherein the Mamba backbone network module includes a test time training layer; A training module, used for inputting the training set into the computing power network performance prediction model, pre-training the computing power network performance prediction model, obtaining a pre-trained computing power network performance prediction model, wherein the weight of the backbone network module in the pre-trained computing power network performance prediction model is locked, and training the test time training layer in the pre-trained computing power network performance prediction model, and updating the weight of the test time training layer to adapt to changes in the network performance mode, thereby obtaining a trained computing power network performance prediction model; The prediction module is used to input the network performance data of the target computing power network into the trained computing power network performance prediction model for prediction to obtain the performance prediction result of the target computing power network.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the computing power network performance prediction method as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the computing power network performance prediction method as described in any one of claims 1 to 7 is implemented.

Images