Hard disk testing device, system and method

Abstract

The present application discloses a hard disk test device, system and method, which collects and analyzes the hard disk's read and write speed value and sound signal, and introduces data processing and analysis algorithms at the back end to perform hard disk performance testing, so as to automatically determine whether the hard disk's performance meets the predetermined standard, thereby realizing intelligent, efficient and automated testing of the hard disk. In this way, the existing hard disk test device can be improved, making the test process more convenient and efficient, thereby improving the convenience, accuracy and efficiency of the test, so as to better evaluate the performance and life of the hard disk.

Description

Technical Field

[0001] The present application relates to the field of hard disk testing, and more specifically, to a hard disk testing device, system and method. Background Art

[0002] The hard disk is an important storage device in the computer system, and its performance directly affects the computer's operating efficiency and stability. The performance of the hard disk is mainly determined by factors such as read and write speed, temperature, noise, bad sectors, etc. These factors are in turn affected by factors such as the hard disk's manufacturing process, usage environment, and usage time. Therefore, comprehensive, accurate, and fast testing of the hard disk is a necessary means to ensure the quality of the hard disk and extend the life of the hard disk. However, the existing hard disk testing device has the following problems: first, the testing process is cumbersome, requiring manual intervention and switching of multiple testing software; second, the test results are not intuitive and require professional analysis and interpretation; third, the test efficiency is low, and multiple hard disks cannot be tested at the same time.

[0003] Therefore, an optimized hard disk testing device is desired. Summary of the invention

[0004] In order to solve the above technical problems, the present application is proposed. The embodiments of the present application provide a hard disk testing device, system and method, which collects and analyzes the read and write speed values ​​and sound signals of the hard disk, and introduces data processing and analysis algorithms at the back end to perform hard disk performance testing to automatically determine whether the performance of the hard disk meets the predetermined standard, thereby realizing intelligent, efficient and automated testing of the hard disk. In this way, the existing hard disk testing device can be improved, making the testing process simpler and more efficient, thereby improving the convenience, accuracy and efficiency of the test, so as to better evaluate the performance and life of the hard disk.

[0005] According to one aspect of the present application, a hard disk testing device is provided, comprising:

[0006] harddisk;

[0007] A bracket, used for mounting the hard disk;

[0008] Data and power cables;

[0009] Sound sensor;

[0010] A controller, the controller being communicatively connected to the sound sensor, the controller being used to perform performance detection on the hard disk and control a working state of the hard disk;

[0011] A display is used to display the test result of the hard disk.

[0012] According to another aspect of the present application, a hard disk testing method is provided, which includes:

[0013] Obtaining the read and write speed values ​​of the tested hard disk at multiple predetermined time points within a predetermined time period;

[0014] A sound signal acquisition module, used for acquiring the working sound signal of the tested hard disk collected by the sound sensor in the predetermined time period;

[0015] Arranging the read / write speed values ​​at the plurality of predetermined time points into a read / write speed timing input vector according to the time dimension;

[0016] Performing local timing feature analysis on the read / write speed timing input vector to obtain a sequence of read / write speed local timing feature vectors;

[0017] Performing local signal analysis on the working sound signal to obtain a sequence of waveform feature vectors of working sound signal segments;

[0018] Using a granularity-by-granularity interactive matching module to process the sequence of the read / write speed local time series feature vectors and the sequence of the working sound signal segment waveform feature vectors to obtain a sequence of the working sound segment-read / write speed local time series interactive fusion feature vectors;

[0019] Performing context-associative coding on the sequence of the working sound segment-reading and writing speed local temporal interaction fusion feature vectors to obtain a global interaction semantic feature;

[0020] Based on the global interactive semantic features, it is determined whether the performance of the hard disk under test meets a predetermined standard.

[0021] Compared with the prior art, the hard disk test device, system and method provided by the present application collect and analyze the hard disk read and write speed value and sound signal, and introduce data processing and analysis algorithms at the back end to perform hard disk performance testing, so as to automatically determine whether the hard disk performance meets the predetermined standard, thereby realizing intelligent, efficient and automated testing of the hard disk. In this way, the existing hard disk test device can be improved, making the test process more simple and efficient, thereby improving the convenience, accuracy and efficiency of the test, so as to better evaluate the performance and life of the hard disk. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0023] Figure 1is a block diagram of a hard disk testing device according to an embodiment of the present application;

[0024] Figure 2 is a system architecture diagram of a hard disk testing device according to an embodiment of the present application;

[0025] Figure 3 is a block diagram of a controller in a hard disk testing device according to an embodiment of the present application;

[0026] Figure 4 is a block diagram of a read / write speed local timing analysis module in a hard disk testing device according to an embodiment of the present application;

[0027] Figure 5 is a block diagram of a local waveform analysis module of a working sound signal in a hard disk testing device according to an embodiment of the present application;

[0028] Figure 6 Flow chart of a hard disk testing method according to an embodiment of the present application. DETAILED DESCRIPTION

[0029] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0030] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0031] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0032] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0033] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0034] The existing hard disk test device has the following problems: first, the test process is cumbersome, requiring manual intervention and switching of multiple test software; second, the test results are not intuitive, requiring professional personnel to analyze and interpret; third, the test efficiency is low, and multiple hard disks cannot be tested at the same time. Therefore, an optimized hard disk test device is desired.

[0035] In the technical solution of the present application, a hard disk testing device is proposed. Figure 1 FIG. 1 is a block diagram of a hard disk testing device according to an embodiment of the present application. Figure 1 As shown, a hard disk testing device 300 according to an embodiment of the present application includes: a hard disk 310; a bracket 320 for installing the hard disk; a data cable and a power cable 330; a sound sensor 340; a controller 350, which can be communicatively connected to the sound sensor, and is used to perform performance detection on the hard disk and control the working status of the hard disk; a display 360, which is used to display the test results of the hard disk.

[0036] In particular, the hard disk 310. It is worth mentioning that a hard disk is a computer storage device for storing and reading data. It is usually used to save operating systems, applications, files and other data. When a computer needs to read data on a hard disk, the read / write head will locate the corresponding track (circular track) according to the instruction, and then read or write data on the track. The reading and writing of data is achieved through the magnetic change of magnetic materials. The read / write head will interpret the data or write the data to the disk according to the magnetic change.

[0037] In particular, the bracket 320 is used to mount the hard disk. It should be understood that when installing a hard disk, a hard disk bracket or hard disk tray is usually required to fix the hard disk inside a computer or other device. The hard disk bracket provides a stable mounting position and ensures that the hard disk is well connected to other components.

[0038] In particular, the data line and the power line 330. It is worth mentioning that the data line and the power line are two common connection lines in computers and other electronic devices, which transmit different types of signals and energy between devices.

[0039] In particular, the sound sensor 340. It is worth mentioning that the sound sensor is a device used to detect and measure the sound in the environment.

[0040] In particular, the controller 350 can be communicatively connected to the sound sensor, and the controller is used to perform performance detection on the hard disk and control the working state of the hard disk. In particular, in a specific example of the present application, Figure 2 and Figure 3 As shown, the controller 350 includes: a read / write speed acquisition module 351, which is used to obtain the read / write speed values ​​of the tested hard disk at multiple predetermined time points within a predetermined time period; a sound signal acquisition module 352, which is used to obtain the working sound signal of the tested hard disk in the predetermined time period collected by the sound sensor; a read / write speed timing arrangement module 353, which is used to arrange the read / write speed values ​​of the multiple predetermined time points into a read / write speed timing input vector according to the time dimension; a read / write speed local timing analysis module 354, which is used to perform local timing feature analysis on the read / write speed timing input vector to obtain a sequence of read / write speed local timing feature vectors; a working sound signal local waveform analysis module 355, which is used to perform signal analysis on the working sound signal. Perform local signal analysis to obtain a sequence of working sound signal segment waveform feature vectors; a sound segment-reading and writing speed local timing feature interaction module 356, used to use a granular interaction matching module to process the sequence of the reading and writing speed local timing feature vectors and the sequence of the working sound signal segment waveform feature vectors to obtain a sequence of the working sound segment-reading and writing speed local timing interaction fusion feature vectors; a global semantic interaction association coding module 357, used to context-associately encode the sequence of the working sound segment-reading and writing speed local timing interaction fusion feature vectors to obtain a global interaction semantic feature; a hard disk performance detection module 358, used to determine whether the performance of the tested hard disk meets the predetermined standard based on the global interaction semantic feature.

[0041] Specifically, the read / write speed acquisition module 351 is used to obtain the read / write speed values ​​of the tested hard disk at multiple predetermined time points within a predetermined time period. In one example, the read / write speed sensor can be used to obtain the read / write speed values ​​of the tested hard disk at multiple predetermined time points within a predetermined time period.

[0042] Specifically, the sound signal acquisition module 352 is used to obtain the working sound signal of the tested hard disk collected by the sound sensor during the predetermined time period. It is worth mentioning that the sound sensor is a device used to detect and measure the sound in the environment. It is also called a sound sensor, a sound detector or a weak sound sensor. The sound sensor can convert the sound signal into an electrical signal so that a computer or other electronic device can further process and analyze it.

[0043] Specifically, the read / write speed timing arrangement module 353 is used to arrange the read / write speed values ​​at the plurality of predetermined time points into a read / write speed timing input vector according to the time dimension. Considering that the read / write speed values ​​of the tested hard disk have a dynamic change rule of timing in the time dimension, in order to capture the read / write performance change of the hard disk in the time dimension, it is necessary to arrange the read / write speed values ​​at the plurality of predetermined time points into a read / write speed timing input vector according to the time dimension to integrate the distribution information of the read / write speed values ​​in the timing.

[0044] Specifically, the read / write speed local timing analysis module 354 is used to perform local timing feature analysis on the read / write speed timing input vector to obtain a sequence of read / write speed local timing feature vectors. In particular, in a specific example of the present application, Figure 4 As shown, the read-write speed local timing analysis module 354 includes: a vector segmentation unit 3541, used to perform vector segmentation on the read-write speed timing input vector to obtain a sequence of read-write speed local timing input vectors; a read-write speed local timing feature extraction unit 3542, used to pass the sequence of read-write speed local timing input vectors through a timing feature extractor based on a one-dimensional convolutional layer to obtain a sequence of read-write speed local timing feature vectors.

[0045] More specifically, the vector segmentation unit 3541 is used to perform vector segmentation on the read-write speed timing input vector to obtain a sequence of read-write speed local timing input vectors. Considering that the read-write performance of the hard disk may have different characteristics and patterns in different time periods. For example, in certain time periods, the read-write speed of the hard disk may show obvious fluctuations or peaks, while in other time periods it may be relatively stable. Therefore, the read-write speed timing input vector is further vector segmented to obtain a sequence of read-write speed local timing input vectors. By segmenting the timing input vector of the read-write speed into local timing input vectors, the local timing characteristics of the read-write speed in the time dimension can be better captured, so as to more fully and accurately understand the read-write performance of the hard disk.

[0046] More specifically, the read / write speed local time series feature extraction unit 3542 is used to pass the sequence of the read / write speed local time series input vectors through a time series feature extractor based on a one-dimensional convolution layer to obtain the sequence of the read / write speed local time series feature vectors. That is, in the technical solution of the present application, the sequence of the read / write speed local time series input vectors is subjected to feature mining in a time series feature extractor based on a one-dimensional convolution layer to extract the local time series dynamic feature information of the read / write speed value of the tested hard disk in each local time segment, thereby obtaining the sequence of the read / write speed local time series feature vectors. Specifically, the sequence of local timing input vectors of read-write speeds is passed through a timing feature extractor based on a one-dimensional convolutional layer to obtain a sequence of local timing feature vectors of read-write speeds, including: using each layer of the timing feature extractor based on a one-dimensional convolutional layer to perform the following on the input data in the forward pass of the layer: performing convolution processing on the input data to obtain a convolution feature map; performing pooling based on a feature matrix on the convolution feature map to obtain a pooling feature map; and performing nonlinear activation on the pooling feature map to obtain an activation feature map; wherein the output of the last layer of the timing feature extractor based on the one-dimensional convolutional layer is the sequence of local timing feature vectors of read-write speeds, and the input of the first layer of the timing feature extractor based on the one-dimensional convolutional layer is the sequence of local timing input vectors of read-write speeds.

[0047] It is worth noting that the 1D convolutional layer is a type of layer in a convolutional neural network (CNN) that processes 1D sequence data. It extracts local features of sequence data by applying a 1D convolution operation. The 1D convolution operation performs a sliding window convolution operation on the input sequence using a filter (also called a convolution kernel or feature detector). A filter is a small weight matrix that calculates the output value by element-wise multiplication with different parts of the input sequence and summing them. By performing convolution operations on the input sequence at different locations, the 1D convolutional layer can capture local patterns and features at different locations. A 1D convolutional layer usually contains multiple filters, each of which can learn different features. The weight matrix of each filter is a learnable parameter that is optimized through the back-propagation algorithm and training data. The output of the 1D convolutional layer is a series of feature maps, where each feature map corresponds to a filter. The feature map is the result of performing a convolution operation on the input sequence. The number and dimensionality of the output feature maps can be controlled by choosing different numbers and sizes of filters. After the 1D convolutional layer, an activation function (such as ReLU) is usually applied to introduce nonlinearity. In addition, pooling layers can be used to reduce the spatial size of feature maps and reduce the number of parameters to prevent overfitting. One-dimensional convolutional layers are widely used in processing one-dimensional data such as text, audio, and time series. They can extract useful local features from sequences and provide better input representations for subsequent tasks such as classification and regression.

[0048] It is worth mentioning that in other specific examples of the present application, the read-write speed timing input vector can also be subjected to local timing feature analysis in other ways to obtain a sequence of local timing feature vectors of the read-write speed, for example: the read-write speed timing data is represented as an input vector sequence, each vector contains the read-write speed information of a time step; select a suitable window size and step size; start from the starting position of the input vector sequence, and slide the window according to the step size. Each time the window is slid, the read-write speed vector in the window is taken as the local timing feature; the read-write speed vector in each window is subjected to feature extraction. Common feature extraction methods include statistical features (such as mean, variance, maximum value, minimum value), frequency domain features (such as Fourier transform, power spectral density) and time domain features (such as autocorrelation, difference), etc.; the features extracted from each window are combined into a feature vector. In this way, each window corresponds to a local timing feature vector; the local timing feature vectors obtained from all windows are sequentially combined into a feature vector sequence.

[0049] Specifically, the working sound signal local waveform analysis module 355 is used to perform local signal analysis on the working sound signal to obtain a sequence of working sound signal segment waveform feature vectors. In particular, in a specific example of the present application, Figure 5 As shown, the working sound signal local waveform analysis module 355 includes: a working sound signal segmentation unit 3551, used to perform signal segmentation on the working sound signal to obtain a sequence of working sound signal segments; a sound waveform local time series feature extraction unit 3552, used to pass the sequence of working sound signal segments through a sound waveform feature extractor based on a convolutional neural network model to obtain a sequence of waveform feature vectors of the working sound signal segments.

[0050] More specifically, the working sound signal segmentation unit 3551 is used to segment the working sound signal to obtain a sequence of working sound signal segments. It should be understood that the hard disk will produce certain sounds when working, and these sound signals contain information about the working state and operation of the hard disk. At the same time, there are noise signals in these sound signals. In order to better analyze the working state and noise level of the hard disk, it is necessary to extract the local time series characteristics of the working sound of the hard disk. Therefore, in the technical solution of the present application, the working sound signal is further segmented to obtain a sequence of working sound signal segments. The segmented working sound signal segments can be used for subsequent feature extraction and analysis. For example, the frequency component, energy distribution, peak value and other characteristics of the sound signal can be detected to evaluate the working state and noise level of the hard disk.

[0051] More specifically, the sound waveform local time series feature extraction unit 3552 is used to pass the sequence of the working sound signal segments through a sound waveform feature extractor based on a convolutional neural network model to obtain a sequence of waveform feature vectors of the working sound signal segments. That is, in the technical solution of the present application, the sequence of the working sound signal segments is subjected to feature mining in a sound waveform feature extractor based on a convolutional neural network model to extract the local time series feature distribution information in the working sound signal segments, thereby obtaining a sequence of waveform feature vectors of the working sound signal segments. By segmenting and extracting features from the working sound signal, the working status and noise characteristics of the hard disk can be understood in more detail. This helps to detect whether the hard disk has abnormal sounds or the noise level exceeds a threshold, thereby discovering potential problems in advance and taking corresponding measures.

[0052] It is worth mentioning that Convolutional Neural Network (CNN) is a deep learning model that is widely used in computer vision tasks such as image classification, object detection, and image segmentation. The design of CNN is inspired by the working principle of biological vision systems. The core idea of ​​CNN is to extract the features of input data through components such as convolutional layers, pooling layers, and fully connected layers, and use these features for classification or other tasks. Below is the general structure and main components of CNN: Convolutional layer: The convolutional layer is one of the core components of CNN. It uses a set of learnable filters (also called convolution kernels) to perform convolution operations on the input data to extract local features of the input data. The convolution operation performs product and sum operations on the local area of ​​the input data through a sliding window to generate a feature map. Activation function: After the convolution layer, an activation function is usually applied to introduce nonlinearity. Common activation functions include ReLU, Sigmoid, and Tanh. Activation functions can increase the expressiveness of the network and help the network learn nonlinear relationships; Pooling layer: The pooling layer is used to reduce the spatial size of the feature map and reduce the number of parameters. Common pooling operations include maximum pooling and average pooling. The pooling operation reduces the size of the feature map by selecting the maximum or average value in the local area of ​​the input data while retaining important feature information; Fully connected layer: The fully connected layer flattens the feature map output by the pooling layer into a vector and connects it to one or more fully connected layers. The neurons in the fully connected layer are connected to all the neurons in the previous layer, and the features are combined and classified by learning weights; Output layer: The output layer is usually a fully connected layer, which is used to map the feature map of the previous layer to the final output category or predicted value. The activation function of the output layer depends on the requirements of the task, such as the Softmax function for multi-category classification problems and the linear activation function for regression problems. The training process of CNN usually uses the backpropagation algorithm and the gradient descent optimization method to update the weight parameters in the network to minimize the loss function. During the training process, CNN gradually improves the feature extraction and representation capabilities of the input data by repeatedly iteratively adjusting the weights.

[0053] It is worth mentioning that in other specific examples of the present application, the working sound signal can also be subjected to local signal analysis in other ways to obtain a sequence of waveform feature vectors of the working sound signal segments, for example: preprocessing the working sound signal; dividing the working sound signal into continuous frames. This can be achieved by dividing the signal into time windows of fixed length, and a window function (such as a Hamming window) is usually used to reduce boundary effects. Each frame usually has overlapping parts to preserve continuity in the time domain; feature extraction is performed on each frame to obtain a waveform feature vector of the frame. Common feature extraction methods include short-time Fourier transform (STFT), Mel-spectral coefficients (MFCC), linear predictive coding (LPC), etc.; the waveform feature vectors of each frame are combined into a sequence; and the feature vector sequence is further analyzed and processed.

[0054] Specifically, the sound segment-reading and writing speed local timing feature interaction module 356 is used to use a granular interactive matching module to process the sequence of the reading and writing speed local timing feature vectors and the sequence of the working sound signal segment waveform feature vectors to obtain the sequence of the working sound segment-reading and writing speed local timing interactive fusion feature vectors. Considering that the sequence of the reading and writing speed local timing feature vectors reflects the changes in the reading and writing performance of the hard disk in different time periods, and the sequence of the working sound signal segment waveform feature vectors provides information on the working state and noise characteristics of the hard disk in different time periods. Therefore, in order to be able to fuse the interactive information between the hard disk's reading and writing performance and sound features to obtain a more comprehensive feature representation, in the technical solution of the present application, a granular interactive matching module is further used to process the sequence of the reading and writing speed local timing feature vectors and the sequence of the working sound signal segment waveform feature vectors to obtain a sequence of the working sound segment-reading and writing speed local timing interactive fusion feature vectors. It should be understood that the granular interactive matching module can perform timing feature interaction and matching of read / write performance and sound characteristics at different time granularities, so as to better capture the correlation and mutual influence between read / write performance and working sound. Through granular interactive matching, the correlation information between read / write speed and working sound can be fused into a feature vector sequence. Such a feature vector sequence can more comprehensively describe the performance and working status of the hard disk, and provide richer information for performance evaluation, fault detection or other related tasks. More specifically, the granular interactive matching module is used to process the sequence of the read / write speed local timing feature vectors and the sequence of the working sound signal segment waveform feature vectors to obtain the sequence of the working sound segment-read / write speed local timing interactive fusion feature vectors, including: calculating the correlation between each read / write speed local timing feature vector in the sequence of the read / write speed local timing feature vector and each working sound signal segment waveform feature vector in the sequence of the working sound signal segment waveform feature vector using the following formula, wherein the formula is: S ij =h i T r j , where S ij represents the correlation between the i-th local time series feature vector of the read / write speed in the sequence of the local time series feature vectors of the read / write speed and the j-th working sound signal segment waveform feature vector in the sequence of the working sound signal segment waveform feature vector, h i represents the i-th local time series feature vector of the read / write speed in the sequence of the local time series feature vectors of the read / write speed, and r jRepresenting the jth working sound signal segment waveform feature vector in the sequence of working sound signal segment waveform feature vectors; based on the correlation between each reading and writing speed local timing feature vector in the sequence of reading and writing speed local timing feature vectors and all working sound signal segment waveform feature vectors in the sequence of working sound signal segment waveform feature vectors and all working sound signal segment waveform feature vectors in the sequence of working sound signal segment waveform feature vectors, interactively updating each reading and writing speed local timing feature vector in the sequence of reading and writing speed local timing feature vectors to obtain a sequence of updated reading and writing speed local timing feature vectors; based on the correlation between each working sound signal segment waveform feature vector in the sequence of working sound signal segment waveform feature vectors and all reading and writing speed local timing feature vectors in the sequence of reading and writing speed local timing feature vectors and the reading and writing speed local timing feature all the local timing feature vectors of reading and writing speeds in the sequence of vectors, interactively update each working sound signal segment waveform feature vector in the sequence of working sound signal segment waveform feature vectors to obtain a sequence of updated working sound signal segment waveform feature vectors; fuse the sequence of the local timing feature vectors of reading and writing speeds and the sequence of the updated local timing feature vectors of reading and writing speeds to obtain a sequence of interactively fused local timing feature vectors of reading and writing speeds; fuse the sequence of the working sound signal segment waveform feature vectors and the sequence of the updated working sound signal segment waveform feature vectors to obtain a sequence of interactively fused working sound signal segment waveform feature vectors; and, splice the sequence of interactively fused local timing feature vectors of reading and writing speeds and the sequence of interactively fused working sound signal segment waveform feature vectors to obtain a sequence of working sound segment-reading and writing speed local timing interactively fused feature vectors.

[0055] Specifically, the global semantic interaction association encoding module 357 is used to context-associate the sequence of the working sound segment-reading and writing speed local time series interaction fusion feature vectors to obtain global interaction semantic features. It should be understood that the sequence of the working sound segment-reading and writing speed local time series interaction fusion feature vectors fuses the interaction feature information between the hard disk's read and write performance and the working sound. However, these feature vectors are still of local time series granularity and can only provide limited local time series feature representation. Therefore, in order to obtain more global semantic features, a converter-based context encoder can be used. Based on this, in the technical solution of the present application, the sequence of the working sound segment-reading and writing speed local time series interaction fusion feature vectors is further encoded through a converter-based context encoder to capture the semantic association and context information of the entire sequence, that is, the context association feature information based on the time series globality between the semantic interaction features of the sound segment and the reading and writing speed in each local time period, thereby obtaining a global interaction semantic feature vector. More specifically, the sequence of the working sound segment-reading and writing speed local temporal interaction fusion feature vectors is passed through a converter-based context encoder to obtain a global interactive semantic feature vector, including: arranging the sequence of the working sound segment-reading and writing speed local temporal interaction fusion feature vectors in one dimension to obtain a global working sound segment-reading and writing speed temporal interaction fusion feature vector; calculating the product of the global working sound segment-reading and writing speed temporal interaction fusion feature vector and the transposed vector of the working sound segment-reading and writing speed local temporal interaction fusion feature vector in the sequence of the working sound segment-reading and writing speed local temporal interaction fusion feature vector to obtain a plurality of self-attention association matrices; and performing a one-dimensional transformation on each self-attention segment in the plurality of self-attention association matrices. The force association matrix is ​​standardized to obtain multiple standardized self-attention association matrices; each standardized self-attention association matrix in the multiple standardized self-attention association matrices is passed through a Softmax classification function to obtain multiple probability values; each probability value in the multiple probability values ​​is used as a weight to weight the working sound segment-reading and writing speed local timing interaction fusion feature vector in the sequence of the working sound segment-reading and writing speed local timing interaction fusion feature vector to obtain the multiple context semantic working sound segment-reading and writing speed local timing interaction fusion feature vectors; and the multiple context semantic working sound segment-reading and writing speed local timing interaction fusion feature vectors are cascaded to obtain the global interaction semantic feature vector.

[0056] Specifically, the hard disk performance detection module 358 is used to determine whether the performance of the tested hard disk meets the predetermined standard based on the global interactive semantic features. That is, in a specific example of the application, the global interactive semantic feature vector is passed through a classifier to obtain a classification result, and the classification result is used to indicate whether the performance of the tested hard disk meets the predetermined standard. That is to say, the local temporal interactive features of the sound clips and read / write speed of the tested hard disk are used to perform classification processing based on the global context-related feature information, so as to perform hard disk performance testing to automatically determine whether the performance of the hard disk meets the predetermined standard, thereby realizing intelligent, efficient and visual testing of the hard disk. More specifically, the global interactive semantic feature vector is passed through a classifier to obtain a classification result, and the classification result is used to indicate whether the performance of the tested hard disk meets the predetermined standard, including: using multiple fully connected layers of the classifier to fully connect the global interactive semantic feature vector to obtain a coded classification feature vector; and passing the coded classification feature vector through the Softmax classification function of the classifier to obtain the classification result.

[0057] In the technical solution of the present application, the sequence of local timing feature vectors of reading and writing speeds and the sequence of waveform feature vectors of working sound signal segments respectively express the timing correlation characteristics of the reading and writing speed values ​​and the working sound signals in the local time domain determined from the global time domain via signal segmentation. In this way, after using a granular interaction matching module to process the sequence of local timing feature vectors of reading and writing speeds and the sequence of waveform feature vectors of working sound signal segments, the global interaction semantic feature vector includes, in addition to the timing correlation characteristics of the reading and writing speed values ​​and the working sound signal segments in the local time domain, timing interaction features based on the distribution of local time domain sequences. That is, the global interaction semantic feature vector has a multi-dimensional timing feature representation with multiple time domain scales.

[0058] However, considering the distribution differences of temporal feature representations under multiple time domain scales and multi-dimensional features, the local feature distribution will be sparser relative to the overall feature representation of the global interaction semantic feature vector, that is, the out-of-distribution sparse sub-manifold relative to the overall high-dimensional feature manifold. This will result in poor constraints on the global interaction semantic feature vector to the predetermined class probability category representation in the probability space when the global interaction semantic feature vector is mapped to class probability through a classifier, affecting the accuracy of the classification results.

[0059] In a preferred example, the global interaction semantic feature vector is passed through a classifier to obtain a classification result, including the following steps: calculating the point sum of the global interaction semantic feature vector, the square root of the length of the global interaction semantic feature vector and the inverse of the square root of the two-norm of the global interaction semantic feature vector to obtain a global interaction semantic offset feature vector; calculating the exponential function of the global interaction semantic offset feature vector with a natural constant as the base to obtain a global interaction semantic offset prediction feature vector; calculating the point product of the global interaction semantic feature vector and the one-norm and a weight hyperparameter of the global interaction semantic feature vector to obtain a global interaction semantic constraint feature vector; calculating the point sum of the global interaction semantic offset prediction feature vector and the global interaction semantic constraint feature vector to obtain an optimized global interaction semantic feature vector; and, passing the optimized global interaction semantic feature vector through the classifier to obtain the classification result.

[0060] That is, the global interaction semantic feature vector is optimized by the following optimization formula to obtain an optimized global interaction semantic feature vector; wherein the optimization formula is:

[0061]

[0062] Wherein, V represents the global interaction semantic feature vector, ||V||1 represents the first norm of the global interaction semantic feature vector, ||V||2 represents the second norm of the global interaction semantic feature vector, L represents the length of the global interaction semantic feature vector, β represents the weight hyperparameter, ⊕ represents dot addition, Denotes the dot product, and V' denotes the optimized global interaction semantic feature vector.

[0063] Specifically, considering the multi-time-domain scale and multi-dimensional feature distribution of the global interaction semantic feature vector for the overall distribution regression constraint of the classifier, in the above preferred example, the structured norm of the global interaction semantic feature vector is used as the local canonical coordinate for each eigenvalue of the global interaction semantic feature vector to determine the class rotation offset of the vector overall distribution representation of the global interaction semantic feature vector relative to the eigenvalue for the class offset prediction direction of each eigenvalue of the global interaction semantic feature vector as the center, and the bounding box of the vector eigenvalue distribution of the global interaction semantic feature vector is used for eigenvalue constraint to improve the constraint of the global interaction semantic feature vector under the overall regression distribution, thereby improving the training speed of the model and the accuracy of the classification result obtained by the global interaction semantic feature vector through the classifier. In this way, the automatic test of the hard disk can be realized based on the read and write speed and sound of the hard disk, thereby improving the convenience, accuracy and efficiency of the test, so as to better evaluate the performance and life of the hard disk.

[0064] In particular, the display 360 is used to display the test result of the hard disk. It is worth mentioning that the display is an output device in a computer system for displaying computer-generated images and contents to a user.

[0065] As described above, the hard disk test device 300 according to the embodiment of the present application can be implemented in various wireless terminals, such as a server with a hard disk test algorithm. In a possible implementation, the hard disk test device 300 according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the hard disk test device 300 can be a software module in the operating system of the wireless terminal, or can be an application developed for the wireless terminal; of course, the hard disk test device 300 can also be one of the many hardware modules of the wireless terminal.

[0066] Alternatively, in another example, the hard disk testing device 300 and the wireless terminal may also be separate devices, and the hard disk testing device 300 may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0067] Furthermore, a hard disk testing method is also provided.

[0068] Figure 6 FIG. 1 is a flow chart of a hard disk testing method according to an embodiment of the present application. Figure 6 As shown, according to the hard disk testing method of the embodiment of the present application, the steps include: S1, obtaining the read and write speed values ​​of the tested hard disk at multiple predetermined time points within a predetermined time period; S2, obtaining the working sound signal of the tested hard disk collected by the sound sensor in the predetermined time period; S3, arranging the read and write speed values ​​of the multiple predetermined time points into a read and write speed timing input vector according to the time dimension; S4, performing local timing feature analysis on the read and write speed timing input vector to obtain a sequence of read and write speed local timing feature vectors; S5, performing local signal analysis on the working sound signal to obtain a sequence of working sound signal segment waveform feature vectors; S6, using a granularity-by-granularity interaction matching module to process the sequence of the read and write speed local timing feature vectors and the sequence of the working sound signal segment waveform feature vectors to obtain a sequence of the working sound segment-read and write speed local timing interaction fusion feature vectors; S7, context-association encoding the sequence of the working sound segment-read and write speed local timing interaction fusion feature vectors to obtain a global interaction semantic feature; S8, based on the global interaction semantic feature, determining whether the performance of the tested hard disk meets the predetermined standard.

[0069] In summary, the hard disk testing method according to the embodiment of the present application is explained, which collects and analyzes the read and write speed values ​​and sound signals of the hard disk, and introduces data processing and analysis algorithms at the back end to perform hard disk performance testing to automatically determine whether the performance of the hard disk meets the predetermined standard, thereby realizing intelligent, efficient and automated testing of the hard disk. In this way, the existing hard disk testing device can be improved, making the testing process simpler and more efficient, thereby improving the convenience, accuracy and efficiency of the test, so as to better evaluate the performance and life of the hard disk.

[0070] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A hard disk testing device, characterized in that: include: harddisk; A bracket, used for mounting the hard disk; Data and power cables; Sound sensor; A controller, the controller being communicatively connected to the sound sensor, the controller being used to perform performance detection on the hard disk and control the working state of the hard disk; A display, used to display the test result of the hard disk; Wherein, the controller comprises: The read / write speed acquisition module is used to obtain the read / write speed values ​​of the tested hard disk at multiple predetermined time points within a predetermined time period; A sound signal acquisition module, used for acquiring the working sound signal of the tested hard disk collected by the sound sensor in the predetermined time period; A read / write speed timing arrangement module, used to arrange the read / write speed values ​​at the plurality of predetermined time points into a read / write speed timing input vector according to a time dimension; A read / write speed local timing analysis module, used for performing local timing feature analysis on the read / write speed timing input vector to obtain a sequence of read / write speed local timing feature vectors; A working sound signal local waveform analysis module, used for performing local signal analysis on the working sound signal to obtain a sequence of working sound signal segment waveform feature vectors; A sound segment-reading and writing speed local time series feature interaction module, used to use a granularity-by-granularity interaction matching module to process the sequence of the reading and writing speed local time series feature vectors and the sequence of the working sound signal segment waveform feature vectors to obtain a sequence of working sound segment-reading and writing speed local time series interaction fusion feature vectors; A global semantic interaction association coding module, used for performing context association coding on the sequence of the working sound segment-reading and writing speed local temporal interaction fusion feature vectors to obtain a global interaction semantic feature; The hard disk performance detection module is used to determine whether the performance of the tested hard disk meets the predetermined standard based on the global interactive semantic features.

2. The hard disk testing device according to claim 1, characterized in that: The read / write speed local timing analysis module comprises: A vector segmentation unit, used for performing vector segmentation on the read / write speed timing input vector to obtain a sequence of read / write speed local timing input vectors; The read / write speed local time series feature extraction unit is used to pass the sequence of the read / write speed local time series input vectors through a time series feature extractor based on a one-dimensional convolutional layer to obtain the sequence of the read / write speed local time series feature vectors.

3. The hard disk testing device according to claim 2, characterized in that: The working sound signal local waveform analysis module comprises: A working sound signal segmentation unit, used for performing signal segmentation on the working sound signal to obtain a sequence of working sound signal segments; The sound waveform local time series feature extraction unit is used to pass the sequence of the working sound signal segments through a sound waveform feature extractor based on a convolutional neural network model to obtain a sequence of waveform feature vectors of the working sound signal segments.

4. The hard disk testing device according to claim 3, characterized in that: The sound segment-reading and writing speed local time series feature interaction module is used to: The correlation between each read / write speed local time series feature vector in the sequence of the read / write speed local time series feature vector and each working sound signal segment waveform feature vector in the sequence of the working sound signal segment waveform feature vector is calculated by the following formula, wherein the formula is: ， in, represents the correlation between the i-th local time series feature vector of the read / write speed in the sequence of the local time series feature vectors of the read / write speed and the j-th working sound signal segment waveform feature vector in the sequence of the working sound signal segment waveform feature vector, represents the i-th local time series feature vector of the read / write speed in the sequence of local time series feature vectors of the read / write speed, and The j-th working sound signal segment waveform feature vector in the sequence representing the working sound signal segment waveform feature vector; Based on the correlation between each read / write speed local timing feature vector in the sequence of the read / write speed local timing feature vectors and all working sound signal segment waveform feature vectors in the sequence of the working sound signal segment waveform feature vectors and all working sound signal segment waveform feature vectors in the sequence of the working sound signal segment waveform feature vectors, interactively updating each read / write speed local timing feature vector in the sequence of the read / write speed local timing feature vectors to obtain a sequence of updated read / write speed local timing feature vectors; Based on the correlation between each working sound signal segment waveform feature vector in the sequence of the working sound signal segment waveform feature vector and all the reading and writing speed local timing feature vectors in the sequence of the reading and writing speed local timing feature vectors and all the reading and writing speed local timing feature vectors in the sequence of the reading and writing speed local timing feature vectors, interactively updating each working sound signal segment waveform feature vector in the sequence of the working sound signal segment waveform feature vector to obtain a sequence of updated working sound signal segment waveform feature vectors; Fusion of the sequence of local time series feature vectors of read / write speeds and the sequence of updated local time series feature vectors of read / write speeds to obtain a sequence of interactively fused local time series feature vectors of read / write speeds; fusing the sequence of the working sound signal segment waveform feature vectors and the sequence of the updated working sound signal segment waveform feature vectors to obtain a sequence of interactively fused working sound signal segment waveform feature vectors; and The sequence of interactively fused reading and writing speed local timing feature vectors and the sequence of interactively fused working sound signal segment waveform feature vectors are concatenated to obtain a sequence of working sound segment-reading and writing speed local timing interactive fusion feature vectors.

5. The hard disk testing device according to claim 4, characterized in that: The global semantic interaction association encoding module is used to: pass the sequence of the working sound segment-reading and writing speed local temporal interaction fusion feature vectors through a converter-based context encoder to obtain a global interaction semantic feature vector as the global interaction semantic feature.

6. The hard disk testing device according to claim 5, characterized in that: The hard disk performance detection module is used to: pass the global interactive semantic feature vector through a classifier to obtain a classification result, and the classification result is used to indicate whether the performance of the tested hard disk meets a predetermined standard.

7. A hard disk testing method, characterized in that: include: Obtaining the read and write speed values ​​of the tested hard disk at multiple predetermined time points within a predetermined time period; Acquiring a working sound signal of the hard disk under test collected by a sound sensor during the predetermined time period; Arranging the read / write speed values ​​at the plurality of predetermined time points into a read / write speed timing input vector according to the time dimension; Performing local timing feature analysis on the read / write speed timing input vector to obtain a sequence of read / write speed local timing feature vectors; Performing local signal analysis on the working sound signal to obtain a sequence of waveform feature vectors of working sound signal segments; Using a granularity-by-granularity interactive matching module to process the sequence of the read / write speed local time series feature vectors and the sequence of the working sound signal segment waveform feature vectors to obtain a sequence of working sound segment-read / write speed local time series interactive fusion feature vectors; Performing context-associative coding on the sequence of the working sound segment-reading and writing speed local temporal interaction fusion feature vectors to obtain a global interaction semantic feature; Based on the global interactive semantic features, it is determined whether the performance of the hard disk under test meets a predetermined standard.

Images