Data processing method, device and equipment, readable storage medium and program product

Abstract

The invention provides a data processing method and device, equipment, a readable storage medium and a program product, and the method comprises the steps: obtaining one or more pieces of operation index information of processing equipment comprising a storage unit and a calculation unit, the operation index information is any one of clock frequency information of the calculation unit, temperature information of the processing equipment, utilization rate change information of the processing equipment, utilization rate change information of the storage unit and operation proportion change information of the calculation unit; obtaining weight data corresponding to the operation index information meeting a weight statistics condition in the one or more pieces of operation index information, and determining a weight statistics result according to the obtained weight data; and determining computing power performance indication data of the processing equipment according to a weight statistics result. According to the method and the device, the computing power performance indication data of the processing equipment can be accurately determined according to the operation index information of one or more dimensions of the processing equipment, and accurate performance alarm can be performed on the processing equipment according to the computing power performance indication data.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a data processing method, a data processing apparatus, a computer device, a computer-readable storage medium, and a computer program product. Background Art

[0002] Artificial intelligence big models (also called AI big models) refer to large-scale, highly complex machine learning models built based on artificial intelligence technology using large-scale training data and computing resources. Since the training parameters of AI big models are large and the amount of data exchanged during training is high, the training process of AI big models is usually completed on large-scale computing clusters. A computing cluster can contain multiple computing nodes, each of which completes a part of the training task.

[0003] When a large-scale computing cluster is running, it is necessary to monitor the computing power performance of the processing devices in the computing nodes (e.g., graphics processing units (GPUs)) in order to timely discover and handle faults and ensure the operating efficiency of the processing devices. However, the existing computing power performance monitoring methods cannot accurately determine the performance indicator data of the processing devices. Therefore, how to accurately determine the computing power performance indicator data of the processing devices is a problem that needs to be solved urgently. Summary of the invention

[0004] The embodiments of the present application provide a data processing method, apparatus, device, readable storage medium and program product, which can accurately determine the computing power performance indication data of the processing device based on the operating indicator information of one or more dimensions of the processing device, so that accurate performance alarms can be issued to the processing device based on the accurate computing power performance indication data.

[0005] On the one hand, an embodiment of the present application provides a data processing method, the method comprising:

[0006] Acquire one or more operation indicator information of a processing device, the processing device including a storage unit and a computing unit, the operation indicator information being any one of clock frequency information of the computing unit, temperature information of the processing device, utilization change information of the processing device, utilization change information of the storage unit, and operation ratio change information of the computing unit, and any two of the operation indicator information have different information categories;

[0007] Determining, from the one or more pieces of operating indicator information, operating indicator information that meets a weighted statistical condition;

[0008] Obtaining weight data corresponding to each piece of operating indicator information satisfying the weight statistical condition, performing weight statistical processing on the weight data corresponding to each piece of operating indicator information satisfying the weight statistical condition, and obtaining a weight statistical result;

[0009] The computing power performance indicator data of the processing device is determined according to the weight statistical result.

[0010] On the one hand, an embodiment of the present application provides a data processing device, the device comprising:

[0011] an acquisition unit, configured to acquire one or more operation indicator information of a processing device, the processing device comprising a storage unit and a computing unit, the operation indicator information being any one of clock frequency information of the computing unit, temperature information of the processing device, utilization change information of the processing device, utilization change information of the storage unit, and operation ratio change information of the computing unit, and any two of the operation indicator information are of different information categories;

[0012] A processing unit, configured to determine, from the one or more pieces of operating indicator information, operating indicator information that meets a weighted statistical condition;

[0013] The acquisition unit is further used to acquire weight data corresponding to each piece of operating indicator information that meets the weight statistical condition;

[0014] The processing unit is further used to perform weight statistical processing on the weight data corresponding to each of the operation indicator information that meets the weight statistical condition to obtain a weight statistical result;

[0015] A determination unit is used to determine the computing power performance indication data of the processing device according to the weight statistical result.

[0016] On the one hand, an embodiment of the present application provides a computer device, comprising: a processor, a communication interface and a memory, wherein the processor, the communication interface and the memory are interconnected, wherein the memory stores executable program code, and the processor is used to call the executable program code to implement the data processing method provided in the embodiment of the present application.

[0017] Accordingly, an embodiment of the present application further provides a computer-readable storage medium, in which instructions are stored. When the computer-readable storage medium is executed on a computer, the computer implements the data processing method provided by the embodiment of the present application.

[0018] Accordingly, an embodiment of the present application further provides a computer program product, the computer program product comprising a computer program or computer instructions, the computer program or computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer program or computer instructions from the computer-readable storage medium, and the processor executes the computer program or computer instructions, so that the computer device implements the data processing method provided in the embodiment of the present application.

[0019] In the present application, one or more operating indicator information of a processing device including a storage unit and a computing unit is obtained, and the operating indicator information is any one of the clock frequency information of the computing unit, the temperature information of the processing device, the utilization change information of the processing device, the utilization change information of the storage unit and the operating proportion change information of the computing unit, and the information categories of any two operating indicator information are different; weight data corresponding to each operating indicator information that meets the weight statistical condition in the one or more operating indicator information is obtained, the weight statistical result is determined according to the obtained each weight data, and the computing power performance indication data of the processing device is determined according to the weight statistical result. Through the data processing method provided in the embodiment of the present application, the computing power performance indication data of the processing device can be determined according to one or more of the operating indicator information of the above five dimensions, so that the computing power performance indication data of the processing device can accurately reflect the current operating status of the processing device, and then the computing power performance of the processing device can be effectively monitored based on the accurate computing power performance indication data; the computing power performance indication data of the processing device can be determined according to the weight data corresponding to the operating indicator information that meets the weight statistical conditions, so that the computing power performance indication data can include operating indicator information of different dimensions, and can more accurately and realistically reflect the actual operating status of the processing device; in the model training process, the data processing method provided in the embodiment of the present application can be used to accurately determine the computing power performance indication data of the processing device, and further accurately perform performance alarms for the processing device based on the computing power performance indication data, which is conducive to operation and maintenance personnel to quickly determine the fault point of the processing equipment, reduce the time spent on resolving the fault, and thus improve the efficiency of model training. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 is a schematic diagram of the system architecture of a data processing system provided in an embodiment of the present application;

[0022] Figure 2 It is a flowchart of a data processing method provided in an embodiment of the present application;

[0023] Figure 3 It is a flowchart of another data processing method provided in an embodiment of the present application;

[0024] Figure 4 is a schematic diagram of a method for determining utilization rate change information provided by an embodiment of the present application;

[0025] Figure 5 A method for determining computing power performance indication data provided by an embodiment of the present application is shown;

[0026] Figure 6 It is a schematic diagram of a method for determining computing power performance indication data provided by an embodiment of the present application;

[0027] Figure 7 is a structural block diagram of a data processing device provided in an embodiment of the present application;

[0028] Figure 8 It is a structural block diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0030] It should be noted that the descriptions of "first", "second", etc. involved in the embodiments of the present application are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the technical features defined as "first" or "second" may explicitly or implicitly include at least one of the features.

[0031] Artificial intelligence big model (or AI big model) is a complex neural network built based on artificial intelligence technology and trained using a large number of parameters and computing resources. In order to improve the performance of AI big model, it is necessary to use a large amount of data to train it and produce high-quality prediction results. The training method of AI big model is usually distributed training, that is, the workload of the training model is split and shared to multiple processing devices. These processing devices can form a large-scale processing cluster. For example, the processing device can be a graphics processing unit (GPU). When training AI big model, it is necessary to accurately determine the computing power performance indicator data of the processing device in order to perform computing power performance alarm operation.

[0032] Based on this, a data processing method provided by an embodiment of the present application can obtain one or more operating indicator information of a processing device including a storage unit and a computing unit, the operating indicator information being any one of the clock frequency information of the computing unit, the temperature information of the processing device, the utilization change information of the processing device, the utilization change information of the storage unit, and the operating proportion change information of the computing unit, and the information categories of any two operating indicator information are different; obtain the weight data corresponding to each operating indicator information that meets the weight statistical condition in one or more operating indicator information, and determine the weight statistical result according to each weight data obtained; determine the computing power performance indication data of the processing device according to the weight statistical result. Through the method provided by the embodiment of the present application, the computing power performance indication data of the processing device can be accurately determined according to the operating indicator information of multiple dimensions, which is conducive to performing computing power alarm operations on the processing device according to the computing power performance indication data, thereby reducing the resources consumed by maintaining and monitoring the processing device.

[0033] The data processing method provided in the embodiment of the present application can be applied to machine learning technology in the field of artificial intelligence. Artificial Intelligence (AI) is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. Machine Learning (ML) is the core of artificial intelligence and the fundamental way to make computers intelligent. Machine learning and deep learning usually include artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and teaching learning. AI big model is an application of machine learning technology. In the distributed training process of AI big model, it is necessary to monitor the computing power of the processing equipment that performs the model training operation to timely discover and handle abnormal situations, and the data processing method provided in the embodiment of the present application can be combined with the operating indicator information of multiple dimensions of the processing equipment to accurately determine the computing power performance indication data of the processing equipment, which is conducive to accurately performing the computing power performance alarm operation for the processing equipment according to the computing power indication data.

[0034] The data processing method provided in the embodiment of the present application can also be applied to the field of cloud computing. Cloud computing is a computing model that distributes computing tasks on a resource pool composed of a large number of computers, so that various application systems can obtain computing power, storage space and information services as needed. The cloud computing resource pool mainly includes: computing devices (virtualized machines, including operating systems), storage devices, and network devices. According to the data processing method provided in the embodiment of the present application, the computing power monitoring operation of the computing device in the cloud computing scenario can be implemented. Through the data processing method provided in the embodiment of the present application, the computing power performance indication data of the computing device can be accurately determined based on the operating indicator information of the computing device in multiple dimensions, so that the computing power performance alarm operation can be performed based on the obtained computing power performance indication data, which effectively ensures the operating efficiency of the computing device and reduces the resources consumed in maintaining the computing device.

[0035] The architecture of the data processing system provided in the embodiments of the present application will be introduced below in conjunction with the accompanying drawings.

[0036] See also Figure 1 , which is a schematic diagram of the system architecture of a data processing system provided by an embodiment of the present application, the data processing system includes a terminal device 101 and multiple computing nodes, any of which may include: a monitoring device 102 and a processing device 103, and the monitoring device 102 may interact with the terminal device 101 and the processing device 103 for data. Among them:

[0037] The terminal device 101 can interact with the object, and can also interact with the monitoring device 102. The terminal device 101 can be a handheld device with communication function (such as a smart phone, a tablet computer), a computing device (such as a personal computer (PC), a vehicle terminal, an intelligent voice interaction device, a wearable device or other intelligent device, but is not limited thereto.

[0038] The monitoring device 102 may be deployed in the same computing node as the processing device 103. The monitoring device 102 may be a device equipped with a central processing unit (CPU), or may be other types of devices with computing capabilities.

[0039] The processing device 103 may include a computing unit and a storage unit. The processing device 103 may be a graphics processing unit (GPU) capable of performing graphics-related computing tasks. For example, the processing device 103 may be a GPU that performs model training operations during the distributed training of a large AI model. It may also be other types of processing devices, for example, the processing device 103 may also be a central processing unit. The processing device 103 may interact with the monitoring device 102 for data.

[0040] The following will explain in detail Figure 1 The working principle of the data processing system shown:

[0041] The monitoring device 102 obtains one or more operation index information of the processing device 103, the processing device 103 includes a storage unit and a computing unit, the operation index information is any one of the clock frequency information of the computing unit, the temperature information of the processing device, the utilization change information of the processing device, the utilization change information of the storage unit, and the operation proportion change information of the computing unit, and the information categories of any two operation index information are different; determine the operation index information that meets the weight statistical condition from the one or more operation index information; obtain the weight data corresponding to each operation index information that meets the weight statistical condition, perform weight statistical processing on the weight data corresponding to each operation index information that meets the weight statistical condition, and obtain the weight statistical result; determine the computing power performance indication data of the processing device 103 according to the weight statistical result. After detecting that the computing power performance indication data of the processing device 103 meets the computing power performance alarm condition, the monitoring device 102 can generate computing power performance alarm prompt information, and send the computing power performance alarm prompt information to the terminal device 101. Through the data processing method provided in the embodiment of the present application, the computing power performance indication data of the processing equipment can be accurately determined based on the operating indicator information of multiple dimensions, which is conducive to accurate computing power performance alarm operations based on the computing power performance indication data, thereby helping to reduce the resources consumed in maintaining and monitoring the processing equipment.

[0042] It is to be understood that the schematic diagram of the data processing system described in the embodiment of the present application is to more clearly illustrate the data processing method of the embodiment of the present application, and does not constitute a limitation on the data processing method provided in the embodiment of the present application. For example, the data processing method provided in the embodiment of the present application can be executed by the monitoring device 102, and can also be executed by other devices that are different from the monitoring device 102 and can communicate with the terminal device 101 and the processing device 103. It is known to those skilled in the art that Figure 1 The number of terminal devices 101, computing nodes, monitoring devices 102, and processing devices 103 is merely illustrative. Any number of devices and nodes may be configured as required for business implementation. Moreover, with the evolution of system architecture and the emergence of new business scenarios, the data processing method provided in the embodiment of the present application is also applicable to similar technical problems.

[0043] See also Figure 2 , Figure 2 The following is a flow chart of a data processing method provided as an example in an embodiment of the present application. The data processing method can be implemented by the above-mentioned monitoring device 102, or by other devices. The flow chart of the data processing method provided in the embodiment of the present application includes but is not limited to:

[0044] S201. Obtain one or more operating indicator information of a processing device, wherein the processing device includes a storage unit and a computing unit, and the operating indicator information is any one of clock frequency information of the computing unit, temperature information of the processing device, utilization change information of the processing device, utilization change information of the storage unit, and operating ratio change information of the computing unit, and information categories of any two of the operating indicator information are different.

[0045] In the embodiment of the present application, the processing device may be a GPU that completes image and graphics related computing work, or may be other devices with computing functions, or may be a heterogeneous processing device different from a central processing unit. The processing device may include a storage unit and a computing unit, the storage unit is used to store computing instructions and data; the computing unit may read instructions and data from the storage unit and execute them. The clock frequency information of the computing unit can be used to measure the speed at which the computing unit executes instructions. The larger the clock frequency, the faster the computing unit executes instructions, that is, the faster the processing speed of the computing unit. The temperature information of the processing device can be used to reflect the current operating state of the processing device. If the temperature of the processing device is too high, it may cause damage to the processing device. The utilization change information of the processing device can indicate the change in the utilization of the overall resources of the processing device (including computing resources, storage resources, etc.) over a period of time. If the utilization change information of the processing device indicates that the degree of change (or fluctuation range) of the utilization of the processing device is too large, it can be considered that an abnormal situation has occurred. The utilization change information of the storage unit can indicate the change in the utilization of the storage resources of the processing device over a period of time. Since the computing unit needs to read instructions and data from the storage unit, if the storage unit fails, it will also affect the overall processing performance of the processing device. If the utilization change information of the storage unit indicates that the utilization change of the storage unit is too large, it can be considered that an abnormality has occurred in the storage unit. The operation ratio change information of the computing unit can be used to reflect the change of the operation ratio data of the computing unit over a period of time; the operation ratio data of the computing unit can be the ratio of the computing resources in operation to the total computing resources. If the operation ratio change information of the computing unit indicates that the operation ratio of the computing unit is too large, it can be considered that an abnormality has occurred in the computing unit. One or more operation indicator information of the processing device is obtained, and these operation indicator information can be any one of the five types of information mentioned above, and the information categories of any two operation indicator information are different. For example: if two operating indicator information of the processing device are obtained, the two operating indicator information may be the clock frequency information of the computing unit and the temperature information of the processing device, or the temperature information of the processing device and the utilization change information of the processing device, etc.; if three operating indicator information of the processing device are obtained, the three operating indicator information may be the utilization change information of the processing device, the utilization change information of the storage unit and the operating proportion change information of the computing unit, or the clock frequency information of the computing unit, the utilization change information of the storage unit and the operating proportion change information of the computing unit; if five operating indicator information of the processing device are obtained, the five operating indicator information may be the clock frequency information of the computing unit, the temperature information of the processing device, the utilization change information of the processing device, the utilization change information of the storage unit and the operating proportion change information of the computing unit.The method provided in the embodiment of the present application can obtain operating indicator information of multiple dimensions of the processing device, which is conducive to the subsequent comprehensive and accurate judgment and processing of the computing power performance of the processing device based on this operating indicator information.

[0046] S202: Determine, from the one or more pieces of operating indicator information, operating indicator information that meets a weighted statistical condition.

[0047] In an embodiment of the present application, after obtaining one or more operating index information of the processing device, the operating index information that meets the weight statistical condition can be determined from the one or more operating index information. The weight statistical conditions corresponding to the operating index information of different information categories may be different. For example: the weight statistical condition corresponding to the temperature information of the processing device may be different from the weight statistical condition corresponding to the utilization change information of the processing device. The method provided in the embodiment of the present application can use the weight statistical condition to perform preliminary screening on the acquired operating index information, so as to facilitate subsequent processing using the operating index information that meets the conditions.

[0048] It should be noted that the weight statistical conditions corresponding to the clock frequency information of the computing unit, the temperature information of the processing device, the utilization change information of the processing device, the utilization change information of the storage unit and the operation ratio change information of the computing unit can be modified according to different actual application requirements.

[0049] S203, obtaining weight data corresponding to each piece of operating indicator information satisfying the weight statistical condition, performing weight statistical processing on the weight data corresponding to each piece of operating indicator information satisfying the weight statistical condition, and obtaining a weight statistical result.

[0050] In the embodiment of the present application, the weight data corresponding to each operation index information that meets the weight statistical condition can be obtained. The weight data corresponding to the operation index information of different information categories may be different. For example: the operation index information that meets the weight statistical condition is the temperature information of the processing device and the utilization change information of the processing device. Assuming that the temperature information of the processing device is more important than the utilization change information of the processing device in reflecting the computing power performance of the processing device (because the temperature of the processing device is too high and will directly cause damage to the processing device), the weight data corresponding to the temperature information of the processing device can be greater than the weight data corresponding to the utilization change information of the processing device. The weight data corresponding to each operation index information that meets the weight statistical condition is subjected to weight statistical processing, and a weight statistical result can be obtained. The data processing method provided in the embodiment of the present application can use weight data to quantify each operation index information that meets the weight statistical condition, thereby unifying the operation index information of different dimensions into one dimension, and obtaining a weight statistical result, which is convenient for determining the computing power performance indicator data of the processing device according to the weight statistical result later.

[0051] It should be noted that the weight data corresponding to the clock frequency information of the computing unit, the temperature information of the processing device, the utilization change information of the processing device, the utilization change information of the storage unit and the operation ratio change information of the computing unit can all change according to different actual application requirements.

[0052] S204: Determine computing power performance indicator data of the processing device according to the weight statistical result.

[0053] In an embodiment of the present application, the computing power performance indicator data of the processing device can be determined based on the weight statistical results. The computing power performance of the processing device refers to the ability of the device to process information, and also refers to the ability of computer hardware and software to work together to perform certain computing requirements. The computing power performance indicator data can accurately reflect the current ability of the processing device to process information. Based on the computing power performance indicator data, computing power monitoring related operations for the processing device can be adaptively performed. Through the method provided in the embodiment of the present application, accurate computing power performance indicator data of the processing device can be obtained, which is conducive to accurate performance alarms for the processing device based on the computing power performance indicator data.

[0054] In one embodiment, the data processing method provided in the embodiment of the present application can be applied to a heterogeneous computing cluster. A heterogeneous computing cluster usually includes multiple heterogeneous computing nodes, which can be configured with different systems and different hardware devices. Heterogeneous computing clusters can be used to process tasks that require large-scale data processing, such as AI large model training. When a heterogeneous computing cluster performs model training, it is necessary to monitor the performance of each processing device in the heterogeneous computing cluster. The data processing method provided in the embodiment of the present application can be used to accurately determine the computing power performance indication data of the processing device, and determine whether the computing power performance of the processing device has decreased according to the computing power performance indication data; if the computing power performance decreases, a computing power performance alarm message can be sent, so that the operation and maintenance personnel can quickly discover and solve the problem of computing power performance decrease of the processing device, thereby improving the efficiency of model training in the heterogeneous computing cluster; at the same time, since the computing power performance indication data can accurately reflect the processing performance of the processing device, the problem of erroneously sending computing power performance alarm information is also reduced, saving the resources consumed by operation and maintenance, and improving the efficiency of discovering and handling computing power performance decrease problems.

[0055] Based on the above embodiments, the beneficial effects of the present application are: the data processing method provided in the embodiments of the present application can obtain operating indicator information of multiple dimensions of the processing equipment, and the operating indicator information of each dimension can reflect the operating status of the processing equipment from different angles to a certain extent; the operating indicator information can be screened according to weight statistical conditions, and the weight statistical results can be determined according to the operating indicator information that meets the conditions, so that the operating indicator information of different dimensions can be uniformly mapped to the same dimension, and the obtained weight statistical results can comprehensively express the content reflected by each operating indicator information; the computing power performance indication data of the processing equipment is determined according to the weight statistical results, so that the computing power performance indication data can accurately reflect the current operating status of the processing equipment, which is conducive to realizing computing power performance monitoring of the processing equipment, reducing the resources consumed by maintaining and monitoring the processing equipment, and improving the efficiency of computing power performance monitoring.

[0056] See also Figure 3 , Figure 3 A flow chart of another data processing method provided as an example in an embodiment of the present application. The data processing method can be implemented by the above-mentioned monitoring device 102, or by other devices. The flow of the data processing method provided in the embodiment of the present application includes but is not limited to:

[0057] S301. Obtain one or more operating indicator information of a processing device, wherein the processing device includes a storage unit and a computing unit, and the operating indicator information is any one of clock frequency information of the computing unit, temperature information of the processing device, utilization change information of the processing device, utilization change information of the storage unit, and operating ratio change information of the computing unit, and information categories of any two of the operating indicator information are different.

[0058] In the embodiment of the present application, the processing device may be a GPU device that completes image and graphics related computing work, or may be other devices with computing functions. The processing device may include a storage unit and a computing unit, the storage unit is used to store computing instructions and data; the computing unit can read instructions and data from the storage unit and execute them. The clock frequency information of the computing unit can be used to measure the speed at which the computing unit executes instructions. The larger the clock frequency, the faster the computing unit executes instructions, that is, the faster the processing speed of the computing unit. The temperature information of the processing device can be used to reflect the current operating state of the processing device. If the temperature of the processing device is too high, it may cause damage to the processing device. The utilization change information of the processing device can indicate the changes in the utilization of the overall resources of the processing device (including computing resources, storage resources, etc.) over a period of time. If the utilization change information of the processing device indicates that the degree of change (or fluctuation range) of the utilization of the processing device is too large, it can be considered that an abnormal situation has occurred. The utilization change information of the storage unit can indicate the changes in the utilization of the storage resources of the processing device over a period of time. Since the computing unit needs to read instructions and data from the storage unit, if the storage unit fails, it will also affect the overall processing performance of the processing device. If the utilization change information of the storage unit indicates that the utilization change of the storage unit is too large, it can be considered that an abnormality has occurred in the storage unit. The operation ratio change information of the computing unit can be used to reflect the change of the operation ratio data of the computing unit over a period of time; the operation ratio data of the computing unit can be the ratio of the computing resources in operation to the total computing resources. If the operation ratio change information of the computing unit indicates that the operation ratio of the computing unit is too large, it can be considered that an abnormality has occurred in the computing unit. One or more operation indicator information of the processing device is obtained, and these operation indicator information can be any one of the five types of information mentioned above, and the information categories of any two operation indicator information are different.

[0059] In one embodiment, the one or more operation index information of the processing device obtained may include first operation index information, and the first operation index information is utilization change information of the processing device. The specific implementation method of obtaining the first operation index information may be: obtaining the current utilization information of the processing device, and obtaining the historical utilization information of the processing device, the information collection time of the historical utilization information is earlier than the information collection time of the current utilization information; according to the information collection time of the current utilization information and the information collection time of the historical utilization information, the current utilization information and the historical utilization information are subjected to curve fitting processing to obtain the utilization change curve of the processing device; according to the utilization change curve of the processing device, the utilization change information of the processing device is determined, and the utilization change information of the processing device is determined as the first operation index information. The current utilization information of the processing device can be obtained by performing collection processing on the processing device; however, the utilization change range of the processing device is large, and the operation status of the processing device cannot be accurately judged only by a single current utilization information. Therefore, the current utilization information of the processing device is combined with the historical utilization information of the processing device, so as to obtain the utilization change information of the processing device that can accurately reflect the utilization fluctuation (change) of the processing device. The historical utilization information of the processing equipment refers to the utilization information of the processing equipment collected within a period of time before the information collection time of the current utilization information. Since the information collection time of the historical utilization information is earlier than the information collection time of the current utilization information, the historical utilization information of the processing equipment can be determined based on the information collection time of the current utilization information of the processing equipment. For example, the utilization information with an information collection time within half an hour before the information collection time of the current utilization information can be determined as the historical utilization information. Assuming that the information collection time of the current utilization information is 12:00 on March 1, 2023, then the utilization information with an information collection time within the time period from 12:00 on March 1, 2023 to 11:30 on March 1, 2023 is the historical utilization information of the processing equipment. After determining the historical utilization information, the current utilization information and the historical utilization information can be subjected to curve fitting processing in the order of the information collection time of the current utilization information and the information collection time of the historical utilization information to obtain the utilization change curve of the processing equipment; and the utilization change information of the processing equipment can be determined according to the utilization change curve of the processing equipment, and the utilization change information of the processing equipment can be determined as the first operation indicator information. The utilization change curve of the processing equipment obtained by performing curve fitting processing on the current utilization information and the historical utilization information can be as follows Figure 4 shown. Figure 4 In the figure, the ordinate represents the utilization information of the processing equipment, and the abscissa represents the information collection time of the utilization information. As can be seen from the figure, the current utilization information of the processing equipment is 80% (i.e. Figure 4The vertical coordinate value of point A in the figure), the current utilization information collection time is 12:00, and the utilization change information of the processing equipment determined according to the change rate curve of the processing equipment is 0.5 (i.e. Figure 4 The absolute value of the slope of the tangent line at point A in the graph is the absolute value of the slope of the tangent line at point A in the graph. It can be seen that the utilization change information of the processing device can better reflect the actual operation status of the processing device than the current utilization information of the processing device. According to the method provided in the embodiment of the present application, more comprehensive and accurate operation index information can be obtained by combining historical data and current data, which is conducive to accurately determining the computing power performance indicator data of the processing device based on these operation index information.

[0060] It should be noted that when the first operation indicator information is the utilization change information of the storage unit or the operation ratio change information of the computing unit, the information determination method provided in the above embodiment is also applicable. For example: the current utilization information of the storage unit can be obtained, and the historical utilization information of the storage unit can be obtained, and the information collection time of the historical utilization information is earlier than the information collection time of the current utilization information; according to the information collection time of the current utilization information of the storage unit and the information collection time of the historical utilization information of the storage unit, the current utilization information of the storage unit and the historical utilization information of the storage unit are subjected to curve fitting processing to obtain the utilization change curve of the storage unit; the utilization change information of the storage unit is determined according to the utilization change curve of the storage unit. For another example: the current operation ratio information of the computing unit and the historical operation ratio information of the computing unit can be obtained, and the information collection time of the historical operation ratio information is earlier than the information collection time of the current operation ratio information; according to the information collection time of the current operation ratio information of the computing unit and the information collection time of the historical operation ratio information of the computing unit, the current operation ratio information of the computing unit and the historical operation ratio information of the computing unit are subjected to curve fitting processing to obtain the operation ratio change curve of the computing unit; the operation ratio change information of the computing unit is determined according to the operation ratio change curve of the computing unit.

[0061] In one embodiment, the frequency of obtaining the clock frequency information of the computing unit, the temperature information of the processing device, the current utilization information of the processing device, the current utilization information of the storage unit and the current operation ratio change information of the computing unit may be once per minute.

[0062] In one embodiment, the method for the monitoring device to obtain one or more operating indicator information of the processing device can be that the client program deployed on the processing device collects information at each collection time point and sends the collected information to the monitoring device. The client program in the processing device can use a script to call an existing application programming interface (Application Programming Interface, API). For example: Taking the processing device as a GPU as an example, the current utilization information of the GPU can be obtained through the API interface of the NVIDIA Collective Communications Library (NCCL).

[0063] S302: Determine, from the one or more pieces of operating indicator information, operating indicator information that meets a weighted statistical condition.

[0064] In an embodiment of the present application, operating indicator information that satisfies the weighted statistical conditions can be determined from one or more pieces of operating indicator information of the acquired processing device, and the weighted statistical conditions corresponding to operating indicator information of different information categories may be different.

[0065] In one embodiment, one or more operation index information may include second operation index information, and the second operation index information is the temperature information of the processing device. Then, the implementation method of determining the operation index information that meets the weighted statistical condition from the one or more operation index information may be: comparing the temperature information of the processing device with the temperature threshold; if the temperature value indicated by the temperature information is greater than or equal to the temperature threshold, then determining the second operation index information as the operation index information that meets the weighted statistical condition. The temperature threshold may be determined according to the weighted statistical condition corresponding to the temperature information of the processing device, for example: the weighted statistical condition corresponding to the temperature information of the processing device may be "the temperature of the processing device cannot exceed 85 degrees Celsius", and the temperature threshold may be determined to be 85. If the temperature value indicated by the temperature information (assuming 85.5 degrees Celsius) is greater than or equal to the temperature threshold, then the second operation index information may be determined as the operation index information that meets the weighted statistical condition.

[0066] In one embodiment, the weight statistical conditions corresponding to the information of the five dimensions of the clock frequency information of the computing unit, the temperature information of the processing device, the utilization change information of the processing device, the utilization change information of the storage unit, and the operation ratio change information of the computing unit can be shown in the following Table 1:

[0067] Table 1

[0068]

[0069] In Table 1, the description of "the clock frequency information of the computing unit" is the "current clock frequency information" of the computing unit, and its corresponding weight statistical condition is "the clock frequency is reduced"; the description of "the temperature information of the processing device" is the "current temperature information" of the processing device, and its corresponding weight statistical condition is "the temperature exceeds 85 degrees Celsius"; the description of "the utilization change information of the processing device" is "the utilization fluctuation of the processing device", and its corresponding weight statistical condition is "the utilization of the processing device fluctuates"; the description of "the utilization change information of the storage unit" is "the utilization fluctuation of the storage unit", and its corresponding weight statistical condition is "the utilization of the storage unit fluctuates"; the description of "the operation proportion change information of the computing unit" is "the operation proportion fluctuation of the computing unit", and its corresponding weight statistical condition is "the operation proportion of the computing unit fluctuates". According to the weight statistical conditions shown in Table 1, more detailed and specific weight statistical conditions can be further determined, and the thresholds corresponding to the information of the above five dimensions can be determined according to these weight statistical conditions. The determination result can be shown in Table 2 below:

[0070] Table 2

[0071]

[0072] In Table 2, the weighted statistical condition corresponding to "clock frequency information of computing unit" is "clock frequency value drops below 1K", and its corresponding threshold is "1K". The reason for setting the threshold is that "clock frequency below 1K affects performance"; the weighted statistical condition corresponding to "temperature information of processing equipment" is "exceeding 85 degrees Celsius", and its corresponding threshold is "85". The reason for setting the threshold is that "temperature above 85 degrees Celsius affects performance"; the weighted statistical condition corresponding to "utilization change information of processing equipment" is "utilization fluctuation exceeds 20%", and its corresponding threshold is "20%". The reason for setting the threshold is that "there may be data waiting situation" during the operation of the processing equipment. , so the utilization fluctuation of the processing equipment is usually large; the weight statistical condition corresponding to the "storage unit utilization change information" is "the utilization fluctuation exceeds 5%", and its corresponding threshold is "5%". The reason for setting the threshold in this way is that during the use of the storage unit, the storage unit "has a low probability of actively releasing after loading data", so the utilization fluctuation of the storage unit is generally small; the weight statistical condition corresponding to the "computing unit operation ratio change information" is "the operation ratio fluctuation exceeds 30%", and its corresponding threshold is "30%". The reason for setting the threshold in this way is that during the operation of the computing unit, the computing unit "may have data waiting conditions", so the operation ratio of the computing unit is usually large. The data processing method provided in the embodiment of the present application can use the corresponding weight statistical conditions to discriminate and process the operation index information according to the information category of the operation index information, which is conducive to the subsequent further processing of the operation index information that meets the weight statistical conditions, reducing the amount of data calculations and improving processing efficiency.

[0073] It should be noted that the weight statistical conditions shown in Table 1 and Table 2 above are merely exemplary. In actual application, the weight statistical conditions corresponding to the operating indicator information of each dimension can be modified according to actual application requirements.

[0074] S303: Obtain weight data corresponding to each piece of operating indicator information satisfying the weight statistical condition, perform weight statistical processing on the weight data corresponding to each piece of operating indicator information satisfying the weight statistical condition, and obtain a weight statistical result.

[0075] In an embodiment of the present application, for one or more operating indicator information that meet the weight statistical conditions, the weight data corresponding to each operating indicator information that meets the weight statistical conditions can be obtained, and the weight data corresponding to each operating indicator information that meets the weight statistical conditions can be subjected to weight statistical processing to obtain a weight statistical result.

[0076] In one embodiment, the method for obtaining the weight data corresponding to each of the operating indicator information that meets the weight statistical condition can be: determining the weight data corresponding to each of the operating indicator information that meets the weight statistical condition according to the weight data form. The weight data form can be shown in the following Table 3:

[0077] Table 3

[0078] Dimensions Weight data Clock frequency information of the computing unit 50％ Processing equipment temperature information 30％ Processing equipment utilization change information 5％ Storage unit utilization change information 5％ Compute unit operation ratio change information 10％

[0079] In Table 3, the weight data corresponding to "clock frequency information of computing unit" is "50%"; the weight data corresponding to "temperature information of processing device" is "30%"; the weight data corresponding to "utilization change information of processing device" is "5%"; the weight data corresponding to "utilization change information of processing device" is "5%"; the weight data corresponding to "operation proportion change information of computing unit" is "10%". The method provided in the embodiment of the present application can evaluate and judge the computing power performance of the processing device by using temperature information, utilization change information of processing device and operation proportion change information of computing unit in combination with the clock frequency information of computing unit, effectively improve the accuracy of evaluation and judgment results, and reduce false alarms or missed alarms for computing power performance of processing device.

[0080] It should be noted that the weight data shown in Table 3 above are only exemplary. In actual application scenarios, the weight data corresponding to each operation index information can be modified according to actual needs. The method provided in the embodiment of the present application can determine different weight data for operation index information of different dimensions, thereby obtaining reasonable and accurate computing power performance indication data of the processing device.

[0081] In one embodiment, the second operating index information is the temperature information of the processing device, and the second operating index information is the operating index information that satisfies the weight statistical condition. Then, the implementation method of obtaining the weight data corresponding to each of the operating index information that satisfies the weight statistical condition can be: obtaining the reference weight data corresponding to the temperature information of the processing device; obtaining the temperature difference between the temperature value indicated by the temperature information and the temperature threshold, and determining the target weight data according to the reference weight data and the temperature difference; determining the target weight data as the weight data corresponding to the second operating index information that satisfies the weight statistical condition. The temperature value corresponding to the temperature information of the processing device is linearly related to the computing power performance of the processing device, that is, the larger the temperature value corresponding to the temperature information of the processing device, the lower the computing power performance of the processing device. Therefore, the reference weight data corresponding to the temperature information of the processing device can be obtained, and the temperature difference between the temperature value indicated by the temperature information and the temperature threshold is obtained, and the target weight data is determined according to the reference weight data and the temperature difference. For example: the reference weight data corresponding to the temperature information of the processing device is 0.25, the temperature value indicated by the temperature information is 86 degrees Celsius, and the temperature threshold is 85 degrees Celsius, then the temperature difference is 1, then the target weight data is determined according to the reference weight data and the temperature difference as 1*0.25*K (at this time, K is 2) = 0.5. K can be adaptively modified according to actual needs. The target weight data can be determined as the weight data corresponding to the second operating index information that meets the weight condition. The method provided in the embodiment of the present application can associate the weight data with the numerical value corresponding to the operating index information, improve the correlation between the weight data and the operating index information, and is more conducive to improving the accuracy of the weight data.

[0082] In one embodiment, the weight data corresponding to each of the operating indicator information that meets the weight statistical condition is subjected to weight statistical processing to obtain the weight statistical result. The implementation method can be: the weight data corresponding to each of the operating indicator information that meets the weight statistical condition is subjected to weight accumulation processing to obtain the weight accumulation result; and the weight accumulation result is determined as the weight statistical result. For example: five operating indicator information are obtained, and the five operating indicator information all meet the weight statistical condition, then the weight data corresponding to each of the operating indicator information that meets the weight statistical condition can be subjected to weight accumulation processing to obtain the weight accumulation result, as shown in the following formula (1):

[0083] D=∑(SM_clock,temp,rate,mem,SM_active) (1)

[0084] In formula (1), SM_clock represents the weight data corresponding to the clock frequency information of the computing unit, temp represents the weight data corresponding to the temperature information of the processing device, rate represents the weight data corresponding to the utilization change information of the processing device, mem represents the weight data corresponding to the utilization change information of the storage unit, SM_active represents the weight data corresponding to the operation ratio change information of the computing unit, and D represents the weight accumulation result. The weight accumulation result can be determined as a weight statistical result. The data processing method provided in the embodiment of the present application can accumulate the weight data corresponding to the operation index information of different information types, so that the weight statistical result can be determined by comprehensively analyzing the operation index information of each information type, so that the weight statistical result can more fully express the current operation status of the processing device.

[0085] S304: Determine computing power performance indicator data of the processing device according to the weight statistical result.

[0086] In an embodiment of the present application, the computing power performance indication data of the processing device can be determined based on the weight statistical results. For example: the weight statistical results can be directly determined as the computing power performance indication data of the processing device, or the weight statistical results can be multiplied with the relevant data to obtain the computing power performance indication data of the processing device, or the weight statistical results can be summed with the relevant data to obtain the computing power performance indication data of the processing device. The data processing method provided in the embodiment of the present application can accurately determine the computing power performance indication data of the processing device in combination with the operation indicator information of multiple dimensions, so as to facilitate the subsequent performance alarm processing according to the computing power performance indication data.

[0087] S305. Compare the computing power performance indication data of the processing device with the set threshold data to obtain a comparison result. If the comparison result indicates that the computing power performance indication data is greater than or equal to the set threshold data, it is determined that the computing power performance alarm condition is met.

[0088] In the embodiment of the present application, the computing power performance indication data of the processing device can be compared with the set threshold data to obtain a comparison result. The set threshold data can be modified according to actual needs. For example, at the beginning of the operation of the processing device, it is necessary to closely monitor the computing power performance of the processing device, so the set threshold data can be set to a smaller value (such as 0.3), so that when only a small amount of operating index information meets the weight statistical condition, the computing power performance indication data determined according to the operating index information is also greater than or equal to the set threshold data; after the processing device has been running stably for a period of time, the computing power performance monitoring of the processing device can be appropriately relaxed, so the set threshold data can be set to a larger value (such as 0.5), so that only when a large number of operating index information meets the weight statistical condition, the computing power performance indication data determined according to the operating index information can be greater than or equal to the set threshold data. If the comparison result indicates that the computing power performance indication data is greater than or equal to the set threshold data, it is determined that the computing power performance alarm condition is met, and subsequent operations can be performed. The method provided in the embodiment of the present application can flexibly judge the size of the computing power performance indication data of the processing device, thereby determining whether the computing power performance indication data meets the computing power performance alarm condition.

[0089] S306: Store and process the computing power performance indication data, the equipment information of the processing equipment, and the information collection time corresponding to the operation index information.

[0090] In an embodiment of the present application, after determining that the computing power performance alarm condition is met according to the computing power performance indication data of the processing device, the computing power performance indication data, the device information of the processing device, and the information collection time corresponding to the operation index information can be stored and processed. The device information of the processing device may include the physical identification information of the processing device (such as the device number), the virtual identification information of the processing device (such as the network address of the processing device), etc. When the operation index information is the clock frequency information of the computing unit or the temperature information of the processing device, the information collection time corresponding to the operation index information may be the time for collecting the clock frequency information of the computing unit or the temperature information of the processing device; when the operation index information is the utilization change information of the processing device, the information collection time corresponding to the operation index information may be the time for collecting the current utilization information of the processing device; when the operation index information is the utilization change information of the storage unit, the information collection time corresponding to the operation index information may be the time for collecting the current utilization information of the storage unit; when the operation index information is the operation ratio change information of the computing unit, the information collection time corresponding to the operation index information may be the time for collecting the current operation ratio information of the computing unit. Performing storage processing may be storing the data in a database. Storing the computing power performance indicator data and related information in a database is conducive to subsequently determining the resource profile curve of the processing equipment based on these data, and analyzing the resource profile curve to determine the equipment with frequent computing power performance failures, and perform special maintenance operations on the equipment. The method provided in the embodiment of the present application can store data such as the computing power performance indicator data of the processing equipment in a database, which is convenient for subsequent abnormal judgment of the computing power performance of the processing equipment.

[0091] See also Figure 5 , this figure shows a method for determining computing power performance indication data of a processing device provided in an embodiment of the present application. Figure 5In the process, the processing equipment information collection module can collect the operation-related information of the processing equipment, can collect the computing unit clock frequency information of the processing equipment, and the temperature information of the processing equipment, and send the operation index information of these two dimensions to the computing power performance judgment module; the processing equipment information collection module can also collect the current utilization information of the processing equipment, the current utilization information of the storage unit, and the current operation ratio information of the computing unit, and send the information of these three dimensions to the information database for data storage and processing; the curve fitting module can obtain the current utilization change information of the processing equipment and the historical utilization change information of the processing equipment from the information database, and perform curve fitting processing, and determine the utilization change information of the processing equipment according to the obtained curve; the curve fitting module can also use the same method to determine the utilization change information of the storage unit and the operation ratio change information of the computing unit; the curve fitting module sends the determined utilization change information of the processing equipment, the utilization change information of the storage unit, and the operation ratio change information of the computing unit to the computing power performance judgment module. The computing power performance judgment module can determine the computing power performance indication data of the processing equipment according to the received information of the five dimensions combined with the weight statistical conditions. The computing power performance judgment module can compare and process the computing power performance indication data. If the computing power performance indication data meets the computing power performance alarm condition, the computing power performance indication data and other related data can be stored in the information database. The method provided in the embodiment of the present application can periodically collect the operation-related data of the processing device and determine the computing power performance indication data of the processing device, thereby realizing the periodic monitoring of the computing power performance of the processing device, while effectively ensuring the operating performance of the processing device, reducing the resources consumed by computing power performance monitoring.

[0092] In one embodiment, the computing power performance judgment module determines the implementation method of the computing power performance indication data of the processing device according to the input five-dimensional information, which can be as follows: Figure 6 As shown ( Figure 6 The weight data corresponding to the operating indicator information of each dimension is shown in Table 3 above). Figure 6 In the calculation, the five-dimensional operating indicator information is input into the computing power performance judgment module, and the initial value of the computing power performance indicator data can be set to 0 (i.e. Figure 6), determine whether the clock frequency information of the computing unit meets the corresponding weight statistical conditions, if not, N=N, if yes, N=N+50%; determine whether the temperature information of the processing device meets the corresponding weight statistical conditions (the weight statistical conditions can be as shown in Table 2 above), if not, N=N, if yes, N=N+30%; determine whether the utilization change information of the processing device meets the corresponding weight statistical conditions, if not, N=N, if yes, N=N+5%; determine whether the utilization change information of the storage unit meets the corresponding weight statistical conditions, if not, N=N, if yes, N=N+5%; determine whether the operation proportion change information of the computing unit meets the corresponding weight statistical conditions, if not, N=N, if yes, N=N+10%; output the value of N, at this time, N is the computing power performance indication data of the processing device; determine whether N is greater than or equal to the set threshold data, if N is greater than or equal to the set threshold data, the computing power performance indication data and related data of the processing device can be stored in the database, and the relevant alarm operation is performed. The method provided in the embodiment of the present application can accurately determine the computing power performance indication data of the processing device, thereby facilitating accurate computing power performance alarm operations based on the computing power performance indication data of the processing device.

[0093] S307: Generate computing power performance alarm prompt information, and send the computing power performance alarm prompt information to the client corresponding to the processing device, so that the client outputs the computing power performance alarm prompt information.

[0094] In an embodiment of the present application, when the computing power performance indication data of the processing device meets the computing power performance alarm condition, a computing power performance alarm prompt information can be generated. For example, the computing power performance alarm prompt information can be "The computing power performance of the processing device (device number: 2001, network address: X) may be abnormal, please confirm in time." The device that performs the computing power performance monitoring operation of the processing device can send the computing power performance alarm prompt information to the client corresponding to the processing device, so that the client outputs the computing power performance alarm prompt information. The data processing method provided in the embodiment of the present application can help improve the accuracy of the computing power performance monitoring of the processing device, avoid the waste of operation and maintenance resources, and avoid the low efficiency of operation and maintenance; at the same time, the method provided in the embodiment of the present application can focus on the fault problem of the processing device, which is conducive to helping professional and technical personnel quickly locate the fault point of the processing device, alleviate the pressure of following up the problem when the processing device fails, and is conducive to quickly and effectively solving the problems in the processing device. During the model training process, the data processing method provided in the embodiment of the present application can accurately determine the computing power performance indication data of the processing equipment related to the model training, which is conducive to the reasonable allocation of model training tasks according to the computing power performance indication data. It can also determine accurate computing power performance alarms based on the computing power performance indication data, thereby reducing the idle waiting of processing equipment caused by false computing power performance alarms, thereby effectively improving the efficiency of model training.

[0095] Based on the above embodiments, the beneficial effects of the present application are as follows: the data processing method provided by the embodiment of the present application can periodically obtain operation-related information of multiple dimensions of the processing equipment, so as to accurately determine the operation status of the processing equipment based on this information; a reasonable and effective method is used to process the collected operation-related information (such as the current utilization information of the processing equipment, the current utilization information of the storage unit, or the current operation ratio of the computing unit) to obtain the operation index information, which effectively ensures the validity and accuracy of the operation index information; the operation index information can be distinguished and processed according to the weight statistical conditions corresponding to the operation index information of different dimensions, so as to realize the screening of the operation index information and improve the efficiency of data processing; the weight statistical results can be determined according to the weight data of each operation index information that meets the weight statistical conditions, and the operation index information of different dimensions is unified into one dimension, which is convenient for the subsequent overall judgment of the computing power performance of the processing equipment; the determined computing power performance indication data of the processing equipment can accurately reflect the current operation status of the processing equipment, and whether to perform the computing power performance alarm operation can be determined according to the computing power performance indication data of the processing equipment, which can effectively realize the computing power performance monitoring of the processing equipment, effectively reduce the resources consumed by computing power performance monitoring, and help the operation and maintenance personnel of the processing equipment to quickly locate the processing equipment with computing power performance failures, thereby improving the operation and maintenance efficiency.

[0096] See also Figure 7 , Figure 7 This is a structural block diagram of a data processing device provided in an embodiment of the present application. The device includes:

[0097] An acquiring unit 701 is configured to acquire one or more pieces of operating indicator information of a processing device, wherein the processing device includes a storage unit and a computing unit, and the operating indicator information is any one of clock frequency information of the computing unit, temperature information of the processing device, utilization change information of the processing device, utilization change information of the storage unit, and operation ratio change information of the computing unit, and any two pieces of the operating indicator information are of different information categories;

[0098] The processing unit 702 is used to determine the operation indicator information that meets the weight statistical condition from the one or more operation indicator information;

[0099] The acquisition unit 701 is further configured to acquire weight data corresponding to each piece of operating indicator information satisfying the weight statistical condition;

[0100] The processing unit 702 is further configured to perform weight statistical processing on the weight data corresponding to each of the operation indicator information satisfying the weight statistical condition to obtain a weight statistical result;

[0101] The determination unit 703 is used to determine the computing power performance indication data of the processing device according to the weight statistical result.

[0102] In one embodiment, one or more operation indicator information includes first operation indicator information, and the first operation indicator information is utilization change information of the processing device. The acquisition unit 701 is also used to: obtain current utilization information of the processing device, and obtain historical utilization information of the processing device, the information collection time of the historical utilization information is earlier than the information collection time of the current utilization information; perform curve fitting processing on the current utilization information and the historical utilization information according to the information collection time of the current utilization information and the information collection time of the historical utilization information to obtain the utilization change curve of the processing device; determine the utilization change information of the processing device according to the utilization change curve of the processing device, and determine the utilization change information of the processing device as the first operation indicator information.

[0103] In one embodiment, one or more operating indicator information includes second operating indicator information, and the second operating indicator information is temperature information of the processing device. When the processing unit 702 determines the operating indicator information that meets the weighted statistical condition from the one or more operating indicator information, it is specifically used to: compare the temperature information of the processing device with a temperature threshold; if the temperature value indicated by the temperature information is greater than or equal to the temperature threshold, determine that the second operating indicator information is the operating indicator information that meets the weighted statistical condition.

[0104] In one embodiment, the second operating indicator information is the operating indicator information that satisfies the weight statistical condition. When the acquisition unit 701 acquires the weight data corresponding to each of the operating indicator information that satisfies the weight statistical condition, it is specifically used to: acquire reference weight data corresponding to the temperature information of the processing device; acquire the temperature difference between the temperature value indicated by the temperature information and the temperature threshold, and determine the target weight data according to the reference weight data and the temperature difference; and determine the target weight data as the weight data corresponding to the second operating indicator information that satisfies the weight statistical condition.

[0105] In one embodiment, when the processing unit 702 performs weight statistical processing on the weight data corresponding to each piece of operating indicator information that meets the weight statistical conditions to obtain a weight statistical result, it is specifically used to: perform weight accumulation processing on the weight data corresponding to each piece of operating indicator information that meets the weight statistical conditions to obtain a weight accumulation result; and determine the weight accumulation result as the weight statistical result.

[0106] In one embodiment, the determination unit 703 is also used for: if it is determined that the computing power performance alarm condition is met based on the computing power performance indication data of the processing device, then the computing power performance indication data, the device information of the processing device and the information collection time corresponding to the operating index information are stored and processed; computing power performance alarm prompt information is generated, and the computing power performance alarm prompt information is sent to the client corresponding to the processing device, so that the client outputs the computing power performance alarm prompt information.

[0107] In one embodiment, the determination unit 703 is also used to: compare the computing power performance indication data of the processing device with the set threshold data to obtain a comparison result; if the comparison result indicates that the computing power performance indication data is greater than or equal to the set threshold data, it is determined that the computing power performance alarm condition is met.

[0108] It can be understood that the functions of each functional unit of the data processing device of the embodiment of the present application can be specifically implemented according to the data processing method in the above method embodiment, and its specific implementation process can refer to the relevant description in the above data processing method embodiment, which will not be repeated here.

[0109] Based on the above embodiments, the beneficial effects of the present application are as follows: the data processing method provided by the embodiment of the present application can periodically obtain operation-related information of multiple dimensions of the processing equipment, so as to accurately determine the operation status of the processing equipment based on this information; a reasonable and effective method is used to process the collected operation-related information (such as the current utilization information of the processing equipment, the current utilization information of the storage unit, or the current operation ratio of the computing unit) to obtain the operation index information, which effectively ensures the validity and accuracy of the operation index information; the operation index information can be distinguished and processed according to the weight statistical conditions corresponding to the operation index information of different dimensions, so as to realize the screening of the operation index information and improve the efficiency of data processing; the weight statistical results can be determined according to the weight data of each operation index information that meets the weight statistical conditions, and the operation index information of different dimensions is unified into one dimension, which is convenient for the subsequent overall judgment of the computing power performance of the processing equipment; the determined computing power performance indication data of the processing equipment can accurately reflect the current operation status of the processing equipment, and whether to perform the computing power performance alarm operation can be determined according to the computing power performance indication data of the processing equipment, which can effectively realize the computing power performance monitoring of the processing equipment, effectively reduce the resources consumed by computing power performance monitoring, and help the operation and maintenance personnel of the processing equipment to quickly locate the processing equipment with computing power performance failures, thereby improving the operation and maintenance efficiency.

[0110] See also Figure 8 , Figure 8 A block diagram of a computer device provided in an embodiment of the present application. The computer device described in the embodiment of the present application includes: a processor 801, a communication interface 802, and a memory 803. The processor 801, the communication interface 802, and the memory 803 may be connected via a bus or other means, and the embodiment of the present application takes the connection via a bus as an example.

[0111] Among them, the processor 801 (or CPU (Central Processing Unit)) is the computing core and control core of the computer device, which can parse various instructions in the computer device and process various data of the computer device. For example, the CPU can be used to parse the power on and off instructions sent by the user to the computer device, and control the computer device to perform power on and off operations; for another example, the CPU can transmit various interactive data between the internal structures of the computer device, and so on. The communication interface 802 can optionally include a standard wired interface, a wireless interface (such as Wi-Fi, a mobile communication interface, etc.), which is controlled by the processor 801 to send and receive data. The memory 803 (Memory) is a memory device in the computer device for storing programs and data. It can be understood that the memory 803 here can include both the built-in memory of the computer device and the extended memory supported by the computer device. The memory 803 provides a storage space, which stores the operating system of the computer device, which may include but is not limited to: Android system, iOS system, Windows Phone system, etc., and this application does not limit this.

[0112] In the embodiment of the present application, the processor 801 performs the following operations by running the executable program code in the memory 803:

[0113] Acquire one or more operation indicator information of a processing device, the processing device including a storage unit and a computing unit, the operation indicator information being any one of clock frequency information of the computing unit, temperature information of the processing device, utilization change information of the processing device, utilization change information of the storage unit, and operation ratio change information of the computing unit, and any two of the operation indicator information have different information categories;

[0114] Determining, from the one or more pieces of operating indicator information, operating indicator information that meets a weighted statistical condition;

[0115] Obtaining weight data corresponding to each piece of operating indicator information satisfying the weight statistical condition, performing weight statistical processing on the weight data corresponding to each piece of operating indicator information satisfying the weight statistical condition, and obtaining a weight statistical result;

[0116] The computing power performance indicator data of the processing device is determined according to the weight statistical result.

[0117] In one embodiment, one or more operation indicator information includes first operation indicator information, and the first operation indicator information is utilization change information of the processing device. The processor 801 is also used to: obtain current utilization information of the processing device, and obtain historical utilization information of the processing device, the information collection time of the historical utilization information is earlier than the information collection time of the current utilization information; perform curve fitting processing on the current utilization information and the historical utilization information according to the information collection time of the current utilization information and the information collection time of the historical utilization information to obtain the utilization change curve of the processing device; determine the utilization change information of the processing device according to the utilization change curve of the processing device, and determine the utilization change information of the processing device as the first operation indicator information.

[0118] In one embodiment, one or more operating indicator information includes second operating indicator information, and the second operating indicator information is temperature information of the processing device. When the processor 801 determines the operating indicator information that meets the weight statistical condition from the one or more operating indicator information, it is specifically used to: compare the temperature information of the processing device with a temperature threshold; if the temperature value indicated by the temperature information is greater than or equal to the temperature threshold, determine that the second operating indicator information is the operating indicator information that meets the weight statistical condition.

[0119] In one embodiment, the second operating indicator information is the operating indicator information that satisfies the weight statistical condition. When the processor 801 obtains the weight data corresponding to each of the operating indicator information that satisfies the weight statistical condition, it is specifically used to: obtain reference weight data corresponding to the temperature information of the processing device; obtain the temperature difference between the temperature value indicated by the temperature information and the temperature threshold, and determine the target weight data according to the reference weight data and the temperature difference; determine the target weight data as the weight data corresponding to the second operating indicator information that satisfies the weight statistical condition.

[0120] In one embodiment, when the processor 801 performs weight statistical processing on the weight data corresponding to each piece of operating indicator information that meets the weight statistical conditions to obtain a weight statistical result, it is specifically used to: perform weight accumulation processing on the weight data corresponding to each piece of operating indicator information that meets the weight statistical conditions to obtain a weight accumulation result; and determine the weight accumulation result as the weight statistical result.

[0121] In one embodiment, the processor 801 is also used for: if it is determined that the computing power performance alarm condition is met based on the computing power performance indication data of the processing device, then the computing power performance indication data, the device information of the processing device and the information collection time corresponding to the operating index information are stored and processed; computing power performance alarm prompt information is generated, and the computing power performance alarm prompt information is sent to the client corresponding to the processing device, so that the client outputs the computing power performance alarm prompt information.

[0122] In one embodiment, the processor 801 is also used to: compare the computing power performance indication data of the processing device with the set threshold data to obtain a comparison result; if the comparison result indicates that the computing power performance indication data is greater than or equal to the set threshold data, it is determined that the computing power performance alarm condition is met.

[0123] In a specific implementation, the processor 801, communication interface 802 and memory 803 described in the embodiment of the present application can execute the implementation method of the computer device described in a data processing method provided in an embodiment of the present application, and can also execute the implementation method described in a data processing device provided in an embodiment of the present application, which will not be repeated here.

[0124] Based on the above embodiments, the beneficial effects of the present application are as follows: the data processing method provided by the embodiment of the present application can periodically obtain operation-related information of multiple dimensions of the processing equipment, so as to accurately determine the operation status of the processing equipment based on this information; a reasonable and effective method is used to process the collected operation-related information (such as the current utilization information of the processing equipment, the current utilization information of the storage unit, or the current operation ratio of the computing unit) to obtain the operation index information, which effectively ensures the validity and accuracy of the operation index information; the operation index information can be distinguished and processed according to the weight statistical conditions corresponding to the operation index information of different dimensions, so as to realize the screening of the operation index information and improve the efficiency of data processing; the weight statistical results can be determined according to the weight data of each operation index information that meets the weight statistical conditions, and the operation index information of different dimensions is unified into one dimension, which is convenient for the subsequent overall judgment of the computing power performance of the processing equipment; the determined computing power performance indication data of the processing equipment can accurately reflect the current operation status of the processing equipment, and whether to perform the computing power performance alarm operation can be determined according to the computing power performance indication data of the processing equipment, which can effectively realize the computing power performance monitoring of the processing equipment, effectively reduce the resources consumed by computing power performance monitoring, and help the operation and maintenance personnel of the processing equipment to quickly locate the processing equipment with computing power performance failures, thereby improving the operation and maintenance efficiency.

[0125] The present application also provides a computer-readable storage medium, in which a computer program is stored, and when the computer program is run on a computer, the computer executes the data processing method described in the present application. The specific implementation method can be referred to the above description, and will not be repeated here.

[0126] The embodiment of the present application also provides a computer program product, which includes a computer program or a computer instruction, and the computer program or the computer instruction is stored in a computer-readable storage medium. The processor of the computer device reads the computer program or the computer instruction from the computer-readable storage medium, and the processor executes the computer program or the computer instruction, so that the computer device performs the data processing method as described in the embodiment of the present application. The specific implementation method can be referred to the above description, which will not be repeated here.

[0127] It should be noted that, for the above-mentioned various method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by this application.

[0128] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0129] The above disclosure is only part of the embodiments of the present application, which certainly cannot be used to limit the scope of rights of the present application. Therefore, equivalent changes made according to the claims of the present application are still within the scope covered by the present application.

Claims

1. A data processing method, characterized in that: The method comprises: Acquire one or more operation indicator information of a processing device, the processing device including a storage unit and a computing unit, the operation indicator information being any one of clock frequency information of the computing unit, temperature information of the processing device, utilization change information of the processing device, utilization change information of the storage unit, and operation ratio change information of the computing unit, and any two of the operation indicator information have different information categories; Determining, from the one or more pieces of operating indicator information, operating indicator information that meets a weighted statistical condition; Obtaining weight data corresponding to each piece of operating indicator information satisfying the weight statistical condition, performing weight statistical processing on the weight data corresponding to each piece of operating indicator information satisfying the weight statistical condition, and obtaining a weight statistical result; The computing power performance indicator data of the processing device is determined according to the weight statistical result.

2. The method according to claim 1, characterized in that The one or more operation index information includes first operation index information, wherein the first operation index information is utilization change information of the processing device, and the method further includes: Acquire current utilization information of the processing device, and acquire historical utilization information of the processing device, wherein information collection time of the historical utilization information is earlier than information collection time of the current utilization information; According to the information collection time of the current utilization information and the information collection time of the historical utilization information, curve fitting processing is performed on the current utilization information and the historical utilization information to obtain a utilization change curve of the processing device; The utilization change information of the processing equipment is determined according to the utilization change curve of the processing equipment, and the utilization change information of the processing equipment is determined as the first operation index information.

3. The method according to claim 1, characterized in that The one or more operation indicator information include second operation indicator information, where the second operation indicator information is temperature information of the processing device, and determining the operation indicator information satisfying the weighted statistical condition from the one or more operation indicator information includes: comparing the temperature information of the processing device with a temperature threshold; If the temperature value indicated by the temperature information is greater than or equal to the temperature threshold, it is determined that the second operating indicator information is operating indicator information that meets the weighted statistical condition.

4. The method according to claim 3, characterized in that The second operating indicator information is operating indicator information that satisfies the weighted statistical condition, and the obtaining of weight data corresponding to each piece of operating indicator information that satisfies the weighted statistical condition includes: Acquire reference weight data corresponding to the temperature information of the processing device; Acquire a temperature difference between a temperature value indicated by the temperature information and the temperature threshold, and determine target weight data according to the reference weight data and the temperature difference; The target weight data is determined as the weight data corresponding to the second operating indicator information that meets the weight statistical condition.

5. The method according to any one of claims 1 to 4, characterized in that: The weight statistical processing is performed on the weight data corresponding to each of the operation indicator information that meets the weight statistical condition to obtain the weight statistical result, including: Perform weight accumulation processing on the weight data corresponding to each of the operation indicator information that meets the weight statistical condition to obtain a weight accumulation result; The weight accumulation result is determined as the weight statistical result.

6. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: If it is determined that the computing power performance alarm condition is met according to the computing power performance indication data of the processing device, the computing power performance indication data, the device information of the processing device and the information collection time corresponding to the operation index information are stored and processed; Generate computing power performance alarm prompt information, and send the computing power performance alarm prompt information to the client corresponding to the processing device, so that the client outputs the computing power performance alarm prompt information.

7. The method according to claim 6, characterized in that The method further comprises: Comparing the computing power performance indication data of the processing device with the set threshold data to obtain a comparison result; If the comparison result indicates that the computing power performance indication data is greater than or equal to the set threshold data, it is determined that the computing power performance alarm condition is met.

8. A data processing device, characterized in that: The device comprises: an acquisition unit, configured to acquire one or more operation indicator information of a processing device, the processing device comprising a storage unit and a computing unit, the operation indicator information being any one of clock frequency information of the computing unit, temperature information of the processing device, utilization change information of the processing device, utilization change information of the storage unit, and operation ratio change information of the computing unit, and any two of the operation indicator information are of different information categories; A processing unit, configured to determine, from the one or more pieces of operating indicator information, operating indicator information that meets a weighted statistical condition; The acquisition unit is further used to acquire weight data corresponding to each piece of operating indicator information that meets the weight statistical condition; The processing unit is further used to perform weight statistical processing on the weight data corresponding to each of the operation indicator information that meets the weight statistical condition to obtain a weight statistical result; A determination unit is used to determine the computing power performance indication data of the processing device according to the weight statistical result.

9. A computer device, characterized in that: include: A processor, a communication interface and a memory, wherein the processor, the communication interface and the memory are interconnected, wherein the memory stores an executable program code, and the processor is used to call the executable program code to implement the data processing method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, which, when executed on a computer, enable the computer to implement the data processing method according to any one of claims 1 to 7.

11. A computer program product, characterized in that The computer program product comprises a computer program or a computer instruction, and when the computer program or the computer instruction is executed by a processor, the data processing method according to any one of claims 1 to 7 is implemented.

Images