A method and device for testing the performance stability of a smart phone

Abstract

The present invention discloses a performance stability test method and device for a smart phone, and relates to the technical field of mobile phone testing; the mobile phone is initially tested in multiple test categories to obtain an initial test score set, and the initial test total score is obtained according to the initial test score set; if there is a category of initial test scores in the initial test score set that exceeds a preset threshold, and the initial test total score is less than the preset total score, the weight corresponding to each category of scores is determined according to the initial test score set; the total relevant score is obtained according to the weight corresponding to each category of scores and the initial test score set, and if the total relevant score is less than the preset relevant total score, the mobile phone is subjected to in-depth testing. The intelligent screening mechanism of the preliminary test results ensures that only devices with unqualified performance require further in-depth testing, which reduces resource waste, and makes the evaluation process more accurate through adaptive weight adjustment, so that the system can ensure the accuracy of mobile phone quality testing without wasting resources.

Description

Technical Field

[0001] The present invention belongs to the technical field of mobile phone testing, and in particular relates to a method and device for testing the performance stability of a smart phone. Background Art

[0002] In the field of smartphone performance testing and quality control, existing technologies usually rely on detection methods based on a single hardware or a single indicator. These methods usually focus on the performance evaluation of a certain hardware. For example, common testing methods include load testing, frame rate testing, memory stress testing, and storage read and write latency testing. These detection methods evaluate the performance of the device under specific conditions by independently testing each hardware component.

[0003] However, the evaluation of a single indicator often cannot fully reflect the comprehensive performance of smartphones in actual use, and the hardware configurations and performance characteristics of different mobile phones vary greatly. This leads to misjudgments or missed judgments in detection methods based on simple thresholds. For example, a certain hardware such as the GPU may perform poorly, but because other hardware such as the CPU performs well, the overall performance is mistakenly judged as qualified.

[0004] In the existing technology, deep detection usually involves comprehensive performance testing of all devices. However, the performance of many mobile phones has been well evaluated through preliminary testing, so there is no need for deep testing. However, the existing technology still conducts testing during large-scale production and quality testing, resulting in a waste of testing time and resources. Summary of the invention

[0005] The purpose of the present invention is to solve the above-mentioned problem and to provide a method and device for testing the performance stability of a smart phone.

[0006] In a first aspect of the present invention, a method for testing the performance stability of a smart phone is first proposed, the method comprising:

[0007] For each mobile phone, perform initial detection of multiple detection categories on the mobile phone to obtain an initial detection score set, and calculate the sum of all scores in the initial detection score set to obtain an initial detection total score;

[0008] If there is a category of initial detection scores in the initial detection score set that exceeds a preset threshold, and the total score of the initial detection is less than the preset total score, then determining the weight corresponding to each category of scores according to the initial detection score set;

[0009] A total relevant score is obtained according to the weight corresponding to each type of score and the initial detection score set. If the total relevant score is less than the preset total relevant score, a deep detection is performed on the mobile phone.

[0010] Optionally, performing multiple detection category initial detection on the mobile phone to obtain an initial detection score set includes:

[0011] By running the preset task, recording the load change curve obtained by the CPU load changing over time, and calculating the proportion of the load change curve that does not exceed the preset load change curve to obtain the CPU score;

[0012] By running graphics-intensive tasks, recording the frame rate change curve obtained by the GPU rendering frame rate time change, and calculating the proportion of the frame rate change curve that does not exceed the preset frame rate change curve to obtain a GPU score;

[0013] By starting a multi-task application, recording the release time of each task when the task is closed, calculating the sum of all release times to obtain a total closing time, and calculating the ratio of the preset closing time to the total closing time to obtain a memory score;

[0014] Through file reading and writing operations, a delay variation curve obtained by recording the stored delay changing over time is obtained, and a storage score is obtained by calculating the proportion of the delay variation curve that does not exceed the preset delay variation curve;

[0015] An initial detection score set is obtained according to the CPU score, the GPU score, the memory score, and the storage score.

[0016] Optionally, determining the weight corresponding to each type of score according to the initial detection score set includes:

[0017] For each detection category in the initial detection score set, the difference between the score corresponding to the category and the preset score corresponding to the category is calculated to obtain a residual corresponding to each detection category;

[0018] Normalize the residual corresponding to each detection category to determine the initial weight corresponding to each detection category;

[0019] The initial weights corresponding to all detection categories are brought into the preset model to obtain the weight corresponding to each detection category.

[0020] Optionally, the total relevance score is obtained according to the weight corresponding to each category score and the initial detection score set, including:

[0021] By formula Get the total relevance score;

[0022] in, is the total correlation score, is the CPU score weight, is the CPU score weight, is the CPU memory weight, is the CPU score weight, A is the CPU score, B is the GPU score, C is the memory score, and D is the storage score.

[0023] Optionally, after obtaining the total relevance score according to the weight corresponding to each category score and the initial detection score set, the following is further included:

[0024] If the total correlation score is greater than or equal to the preset total correlation score, a corresponding depth detection is performed for the detection category exceeding the preset threshold.

[0025] In a second aspect of the present invention, a performance stability testing device for a smart phone is provided, comprising:

[0026] An initial detection module is used to perform multiple detection category initial detection on each mobile phone to obtain an initial detection score set, and calculate the sum of all scores in the initial detection score set to obtain an initial detection total score;

[0027] A weight determination module, configured to determine the weight corresponding to each category of scores according to the initial detection score set if there is a category of initial detection scores in the initial detection score set that exceeds a preset threshold and the total score of the initial detection is less than a preset total score;

[0028] The first depth detection module is used to obtain a total correlation score according to the weight corresponding to each type of score and the initial detection score set, and if the total correlation score is less than a preset total correlation score, perform a depth detection on the mobile phone.

[0029] Optionally, the initial detection module includes:

[0030] A CPU detection module is used to record a load change curve obtained by running a preset task and recording the CPU load over time, and calculate the proportion of the load change curve that does not exceed the preset load change curve to obtain a CPU score;

[0031] A GPU detection module is used to record a frame rate change curve obtained by running a graphics-intensive task, and calculate the proportion of the frame rate change curve that does not exceed a preset frame rate change curve to obtain a GPU score;

[0032] A memory detection module is used to start a multi-task application, record the release time of each task when the task is closed, calculate the sum of all release times to obtain a total closing time, and calculate the ratio of a preset closing time to the total closing time to obtain a memory score;

[0033] A storage detection module, used to record the delay change curve obtained by the storage delay changing over time through file read and write operations, and calculate the proportion of the delay change curve that does not exceed the preset delay change curve to obtain a storage score;

[0034] An initial detection score set determination module is used to obtain an initial detection score set according to the CPU score, the GPU score, the memory score and the storage score.

[0035] Optionally, the weight determination module includes:

[0036] A residual determination module, configured to calculate, for each detection category in the initial detection score set, a difference between the score corresponding to the category and a preset score corresponding to the category to obtain a residual corresponding to each detection category;

[0037] An initial weight determination module is used to normalize the residual corresponding to each detection category to determine the initial weight corresponding to each detection category;

[0038] The weight adjustment module is used to bring the initial weights corresponding to all detection categories into the preset model to obtain the weight corresponding to each detection category.

[0039] Optionally, the first depth detection module includes:

[0040] By formula Get the total relevance score;

[0041] in, is the total correlation score, is the CPU score weight, is the CPU score weight, is the CPU memory weight, is the CPU score weight, A is the CPU score, B is the GPU score, C is the memory score, and D is the storage score.

[0042] Optionally, the first depth detection module further includes:

[0043] The second depth detection module is configured to perform a corresponding depth detection on a detection category exceeding a preset threshold if the total correlation score is greater than or equal to a preset total correlation score.

[0044] Beneficial effects of the present invention:

[0045] The present invention proposes a performance stability test method for a smart phone. For each mobile phone, multiple detection categories are initially detected to obtain an initial detection score set, and the sum of all scores in the initial detection score set is calculated to obtain the initial detection total score; if there is a category of initial detection scores in the initial detection score set that exceeds a preset threshold, and the initial detection total score is less than the preset total score, the weight corresponding to each category of scores is determined according to the initial detection score set; the total relevant score is obtained according to the weight corresponding to each category of scores and the initial detection score set, and if the total relevant score is less than the preset relevant total score, the mobile phone is subjected to a deep detection. Through the intelligent screening mechanism of the preliminary detection results, it is ensured that only those devices with unqualified performance need further deep detection, thereby significantly improving the test efficiency, reducing resource waste, and reducing costs. Through adaptive weight adjustment, the evaluation process is made more accurate and reasonable, and the overall process of mobile phone quality detection is further optimized, so that the system can ensure the accuracy of mobile phone quality detection without wasting resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The present invention will be further described below in conjunction with the accompanying drawings.

[0047] Figure 1 A flowchart of a method for testing the performance stability of a smart phone is provided for an embodiment of the present invention;

[0048] Figure 2 A schematic structural diagram of a device for testing the performance stability of a smart phone is provided in accordance with an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. The term "and / or" herein is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the present invention, the description of "first", "second", etc. is only used for descriptive purposes, and cannot be understood as indicating or implying its relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" can explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0050] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of the present invention.

[0051] The embodiment of the present invention provides a method for testing the performance stability of a smart phone. Figure 1 , Figure 1 A flowchart of a method for testing the performance stability of a smart phone provided by an embodiment of the present invention. The method comprises the following steps:

[0052] S101, for each mobile phone, performing initial detection of multiple detection categories on the mobile phone to obtain an initial detection score set, and calculating the sum of all scores in the initial detection score set to obtain an initial detection total score;

[0053] S102, if there is a category of initial detection scores in the initial detection score set that exceeds a preset threshold, and the total score of the initial detection is less than the preset total score, then determine the weight corresponding to each category of scores according to the initial detection score set;

[0054] S103, obtaining a total relevant score according to the weight corresponding to each type of score and the initial detection score set, and if the total relevant score is less than a preset total relevant score, performing a deep detection on the mobile phone.

[0055] A performance stability testing method for a smart phone provided in an embodiment of the present invention ensures that only those devices with unqualified performance require further in-depth testing through an intelligent screening mechanism of preliminary test results, thereby significantly improving test efficiency, reducing resource waste, and lowering costs. The evaluation process is made more accurate and reasonable through adaptive weight adjustment, further optimizing the overall process of mobile phone quality testing, so that the system can ensure the accuracy of mobile phone quality testing without wasting resources.

[0056] In one implementation, through the total score evaluation after preliminary testing, only those mobile phones that fail the preliminary test results will be further deeply tested, effectively avoiding repeated and time-consuming deep testing of mobile phones whose performance has already passed, thereby greatly improving the overall test efficiency.

[0057] In one implementation, the weight of each type of score is dynamically adjusted by calculating the difference between the initial detection score and the preset score and normalizing the residual. Differences in hardware configurations of different mobile phones are automatically taken into account, making the evaluation results more reasonable and accurate, and avoiding the problem of a single hardware failure affecting the overall evaluation results.

[0058] In one implementation, if none of the initial detection scores exceeds the corresponding preset threshold, it means that the mobile phone detection is qualified; if there is a category of initial detection scores that exceeds the preset threshold, but the total initial detection score is greater than or equal to the preset total score, then the category that exceeds the threshold is re-tested separately. If the re-test is still not qualified, it means that there is a problem with this part alone.

[0059] In one implementation, multiple detection categories include CPU load detection, GPU rendering frame rate detection, release time detection when multi-tasking applications are closed, and storage delay time detection; each detection category corresponds to a score threshold, and the preset threshold, preset total score and preset related total score are determined by technical personnel. In-depth detection is a comprehensive performance test of all devices.

[0060] In one embodiment, performing multiple detection category initial detection on the mobile phone to obtain an initial detection score set includes:

[0061] By running the preset tasks, recording the load change curve obtained by CPU load changing over time, and calculating the proportion of the load change curve that does not exceed the preset load change curve to obtain the CPU score;

[0062] By running graphics-intensive tasks, recording the frame rate change curve obtained by the GPU rendering frame rate time change, and calculating the proportion of the frame rate change curve that does not exceed the preset frame rate change curve to obtain the GPU score;

[0063] By starting a multi-tasking application, recording the release time of each task when the task is closed, calculating the sum of all release times to get the total closing time, and calculating the ratio of the preset closing time to the total closing time to get the memory score;

[0064] Through file read and write operations, a delay change curve obtained by recording the storage delay changing over time is obtained, and the proportion of the delay change curve that does not exceed the preset delay change curve is calculated to obtain a storage score;

[0065] An initial detection score set is obtained based on the CPU score, GPU score, memory score, and storage score.

[0066] In one implementation, by independently and carefully evaluating the performance of multiple hardware modules such as CPU, GPU, memory and storage, we can fully understand the performance of the mobile phone in different usage scenarios. The multi-dimensional testing method can avoid the performance of a single indicator or hardware affecting the overall evaluation, thereby more accurately reflecting the actual performance of the mobile phone.

[0067] In one implementation, the preset tasks may be decompression, video decoding, complex numerical calculations, etc.; the preset load change curve is a qualified curve recorded under the same preset task under the same mobile phone parameters; the preset frame rate change curve is a qualified curve recorded under the same mobile phone parameters by running the same graphics-intensive task; the preset shutdown time is a qualified time corresponding to the release time when the same multi-tasking application is closed under the same mobile phone parameters; the preset delay change curve is a qualified curve recorded under the same mobile phone parameters by the same read and write operations; in order to ensure that this data can be measured at the same time, the mobile phone must implement all the above functions for measurement. For example, when executing a task, it is necessary to record the CPU load of the task, the GPU rendering frame rate time change, the delay time change during reading and writing, and the release time when closing.

[0068] In one implementation, since the scores of each type of hardware are calculated independently, if a problem occurs in a hardware module (such as the CPU or GPU), the specific hardware component can be quickly located without the need for tedious troubleshooting in the global test. For mobile phones with performance bottlenecks, the problem can be identified more quickly, helping production and R&D teams to quickly resolve the problem.

[0069] In one implementation, by evaluating the performance of each hardware module separately and determining whether to conduct a deeper test based on the actual score, redundant comprehensive testing of devices with good hardware performance can be avoided. This not only improves the testing efficiency, but also effectively saves testing resources and time, and is particularly suitable for scenarios of large-scale production or batch testing.

[0070] In one embodiment, determining the weight corresponding to each type of score according to the initial detection score set includes:

[0071] For each detection category in the initial detection score set, the difference between the score corresponding to the category and the preset score corresponding to the category is calculated to obtain the residual corresponding to each detection category;

[0072] Normalize the residual corresponding to each detection category to determine the initial weight corresponding to each detection category;

[0073] The initial weights corresponding to all detection categories are brought into the preset model to obtain the weight corresponding to each detection category.

[0074] In one implementation, by calculating the difference (residual) between the detection category score and the preset score, the performance deviation of the device in each detection category can be accurately reflected. If the gap between the score of a category and the preset score is large, it means that there are performance problems or abnormalities in the category during the test. The weight will be adjusted accordingly to ensure that the weight of each detection category can better reflect the deviation between actual performance and expected performance, thereby having a more accurate impact on the overall evaluation. The preset model is the residual-based adaptive inertia weight model commonly found on the market.

[0075] In one implementation, the residuals of each detection category are normalized so that the influence range of each detection category is consistent, which not only ensures fairness between different categories, but also makes weight adjustment more flexible and adaptive. The weight can be automatically adjusted according to the actual performance of the device without relying on fixed standards, thereby improving the adaptability of the system; the normalization process can be minimum-maximum normalization, Z-Score standardization, etc.

[0076] In one embodiment, obtaining a total relevance score according to the weight corresponding to each category score and the initial detection score set includes:

[0077] By formula Get the total relevance score;

[0078] in, is the total correlation score, is the CPU score weight, is the CPU score weight, is the CPU memory weight, is the CPU score weight, A is the CPU score, B is the GPU score, C is the memory score, and D is the storage score.

[0079] In one implementation, the total relevance score is calculated by combining the scores of each category and their corresponding weights. The comprehensive calculation method can comprehensively reflect the performance of the mobile phone in all aspects. The impact of different hardware (such as CPU, GPU, memory, storage) is considered uniformly, thus avoiding one-sided evaluation under a single dimension and ensuring that the final score is not just a reflection of a single aspect, but a comprehensive reflection of the performance of the entire machine.

[0080] In one embodiment, after obtaining the total relevance score according to the weight corresponding to each category score and the initial detection score set, the method further includes:

[0081] If the total correlation score is greater than or equal to the preset total correlation score, a corresponding depth detection is performed for the detection category exceeding the preset threshold.

[0082] In one implementation, during large-scale production, if the total correlation score is high enough, there is no need to perform in-depth testing on all detection categories. Only those categories that exceed the preset threshold are subjected to corresponding in-depth testing, thereby avoiding redundant testing, effectively saving testing time and resources, reducing unnecessary repeated testing, and improving production efficiency and resource utilization.

[0083] Based on the same inventive concept, the embodiment of the present invention also provides a performance stability testing device for a smart phone. Figure 2 , Figure 2 A schematic diagram of the structure of a performance stability testing device for a smart phone provided by an embodiment of the present invention includes:

[0084] An initial detection module is used to perform multiple detection category initial detection on each mobile phone to obtain an initial detection score set, and calculate the sum of all scores in the initial detection score set to obtain an initial detection total score;

[0085] A weight determination module, configured to determine the weight corresponding to each category of scores according to the initial detection score set if there is a category of initial detection scores in the initial detection score set that exceeds a preset threshold and the total score of the initial detection is less than a preset total score;

[0086] The first depth detection module is used to obtain a total correlation score according to the weight corresponding to each type of score and the initial detection score set, and if the total correlation score is less than a preset total correlation score, perform a depth detection on the mobile phone.

[0087] A performance stability testing device for a smart phone provided in an embodiment of the present invention ensures that only those devices with unqualified performance require further in-depth testing through an intelligent screening mechanism of preliminary test results, thereby significantly improving test efficiency, reducing resource waste, and lowering costs. The evaluation process is made more accurate and reasonable through adaptive weight adjustment, further optimizing the overall process of mobile phone quality testing, so that the system can ensure the accuracy of mobile phone quality testing without wasting resources.

[0088] In one embodiment, the initial detection module includes:

[0089] The CPU detection module is used to run preset tasks, record the load change curve obtained by CPU load changing over time, and calculate the proportion of the load change curve that does not exceed the preset load change curve to obtain the CPU score;

[0090] The GPU detection module is used to record the frame rate change curve obtained by the GPU rendering frame rate time change by running graphics-intensive tasks, and calculate the proportion of the frame rate change curve that does not exceed the preset frame rate change curve to obtain the GPU score;

[0091] The memory detection module is used to start a multi-task application, record the release time of each task when the task is closed, calculate the sum of all release times to obtain the total closing time, and calculate the ratio of the preset closing time to the total closing time to obtain the memory score;

[0092] A storage detection module is used to record the delay change curve obtained by the storage delay changing over time through file read and write operations, and calculate the proportion of the delay change curve that does not exceed the preset delay change curve to obtain a storage score;

[0093] The initial detection score set determination module is used to obtain an initial detection score set according to the CPU score, GPU score, memory score and storage score.

[0094] In one embodiment, the weight determination module includes:

[0095] A residual determination module, for calculating the difference between the score corresponding to each detection category in the initial detection score set and the preset score corresponding to the category to obtain the residual corresponding to each detection category;

[0096] An initial weight determination module is used to normalize the residual corresponding to each detection category to determine the initial weight corresponding to each detection category;

[0097] The weight adjustment module is used to bring the initial weights corresponding to all detection categories into the preset model to obtain the weight corresponding to each detection category.

[0098] In one embodiment, the first depth detection module includes:

[0099] By formula Get the total relevance score;

[0100] in, is the total correlation score, is the CPU score weight, is the CPU score weight, is the CPU memory weight, is the CPU score weight, A is the CPU score, B is the GPU score, C is the memory score, and D is the storage score.

[0101] In one embodiment, the first depth detection module further includes:

[0102] The second depth detection module is configured to perform a corresponding depth detection on a detection category exceeding a preset threshold if the total correlation score is greater than or equal to a preset total correlation score.

[0103] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A method for testing the performance stability of a smart phone, characterized in that: The method comprises: For each mobile phone, perform initial detection of multiple detection categories on the mobile phone to obtain an initial detection score set, and calculate the sum of all scores in the initial detection score set to obtain an initial detection total score; If there is a category of initial detection scores in the initial detection score set that exceeds a preset threshold, and the total score of the initial detection is less than the preset total score, then determining the weight corresponding to each category of scores according to the initial detection score set; A total relevant score is obtained according to the weight corresponding to each type of score and the initial detection score set. If the total relevant score is less than the preset total relevant score, a deep detection is performed on the mobile phone; The initial detection score set obtained by performing multiple detection category initial detection on the mobile phone includes: By running the preset task, recording the load change curve obtained by the CPU load changing over time, and calculating the proportion of the load change curve that does not exceed the preset load change curve to obtain the CPU score; By running graphics-intensive tasks, recording the frame rate change curve obtained by the GPU rendering frame rate time change, and calculating the proportion of the frame rate change curve that does not exceed the preset frame rate change curve to obtain a GPU score; By starting a multi-task application, recording the release time of each task when the task is closed, calculating the sum of all release times to obtain a total closing time, and calculating the ratio of the preset closing time to the total closing time to obtain a memory score; Through file reading and writing operations, a delay variation curve obtained by recording the stored delay changing over time is obtained, and a storage score is obtained by calculating the proportion of the delay variation curve that does not exceed the preset delay variation curve; Obtaining an initial detection score set according to the CPU score, the GPU score, the memory score, and the storage score; Determining the weight corresponding to each type of score according to the initial detection score set includes: For each detection category in the initial detection score set, the difference between the score corresponding to the category and the preset score corresponding to the category is calculated to obtain a residual corresponding to each detection category; Normalize the residual corresponding to each detection category to determine the initial weight corresponding to each detection category; Bring the initial weights corresponding to all detection categories into the preset model to obtain the weight corresponding to each detection category; The total relevant scores obtained based on the weights corresponding to each category of scores and the initial detection score set include: By formula Get the total relevance score; in, is the total correlation score, is the CPU score weight, is the CPU score weight, is the CPU memory weight, is the CPU score weight, A is the CPU score, B is the GPU score, C is the memory score, and D is the storage score.

2. A method for testing the performance stability of a smart phone according to claim 1, characterized in that: After obtaining the total relevance score based on the weights corresponding to each category of scores and the initial detection score set, it also includes: If the total correlation score is greater than or equal to the preset total correlation score, a corresponding depth detection is performed for the detection category exceeding the preset threshold.

3. A performance stability testing device for a smart phone, characterized in that: The device comprises: An initial detection module is used to perform multiple detection category initial detection on each mobile phone to obtain an initial detection score set, and calculate the sum of all scores in the initial detection score set to obtain an initial detection total score; A weight determination module, configured to determine the weight corresponding to each category of scores according to the initial detection score set if there is a category of initial detection scores in the initial detection score set that exceeds a preset threshold and the total score of the initial detection is less than a preset total score; A first depth detection module, configured to obtain a total correlation score according to the weight corresponding to each type of score and the initial detection score set, and perform a depth detection on the mobile phone if the total correlation score is less than a preset total correlation score; The initial detection module comprises: A CPU detection module is used to record a load change curve obtained by running a preset task and recording the CPU load over time, and calculate the proportion of the load change curve that does not exceed the preset load change curve to obtain a CPU score; A GPU detection module is used to record a frame rate change curve obtained by running a graphics-intensive task, and calculate the proportion of the frame rate change curve that does not exceed a preset frame rate change curve to obtain a GPU score; A memory detection module is used to start a multi-task application, record the release time of each task when the task is closed, calculate the sum of all release times to obtain a total closing time, and calculate the ratio of a preset closing time to the total closing time to obtain a memory score; A storage detection module, used to record the delay change curve obtained by the storage delay changing over time through file read and write operations, and calculate the proportion of the delay change curve that does not exceed the preset delay change curve to obtain a storage score; an initial detection score set determination module, configured to obtain an initial detection score set according to the CPU score, the GPU score, the memory score, and the storage score; The weight determination module comprises: A residual determination module, configured to calculate, for each detection category in the initial detection score set, a difference between the score corresponding to the category and a preset score corresponding to the category to obtain a residual corresponding to each detection category; An initial weight determination module is used to normalize the residual corresponding to each detection category to determine the initial weight corresponding to each detection category; A weight adjustment module is used to bring the initial weights corresponding to all detection categories into the preset model to obtain the weight corresponding to each detection category; The first depth detection module includes: By formula Get the total relevance score; in, is the total correlation score, is the CPU score weight, is the CPU score weight, is the CPU memory weight, is the CPU score weight, A is the CPU score, B is the GPU score, C is the memory score, and D is the storage score.

4. A performance stability testing device for a smart phone according to claim 3, characterized in that: The first depth detection module also includes: The second depth detection module is configured to perform a corresponding depth detection on a detection category exceeding a preset threshold if the total correlation score is greater than or equal to a preset total correlation score.

Images