Graphics processor monitoring method, system, device and electronic equipment

Abstract

Embodiments of the present application provide a graphics processor monitoring method, system, device, and electronic device, wherein the method includes: collecting first target parameters of a target graphics processor according to a preset period; determining whether the target graphics processor has an abnormality based on the first target parameters; if the target graphics processor has an abnormality, obtaining second target parameters of the target graphics processor through target instructions, wherein the second target parameters are all operating parameters of the target graphics processor detected during operation of the target graphics processor, and the second target parameters include the first target parameters; and determining the cause of the abnormality of the target graphics processor by analyzing the second target parameters. This application solves the problem of insufficient comprehensiveness and specificity of graphics processor monitoring in related technologies, thereby achieving the effect of ensuring the normal operation of the graphics processor.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of monitoring technology, and more specifically, to a graphics processor monitoring method, system, device, and electronic device. Background Art

[0002] In the era of big data and informatization, servers are required to perform a large amount of computing. To meet this demand, servers are often equipped with a graphics processing unit (GPU). GPUs primarily perform image processing data operations, freeing the server's CPU from graphics processing tasks, allowing it to perform more system tasks and improve overall server performance.

[0003] With the development of fields such as video encoding and decoding, scientific computing, and artificial intelligence, the demand for GPU servers is increasing. Therefore, monitoring and managing the core components of GPUs has become particularly important. In related technologies, maintenance personnel install GPU drivers and related monitoring software and then manually run the software to monitor the GPU. This results in a cumbersome and inefficient monitoring process, often resulting in incomplete and incomplete monitoring of the graphics processor.

[0004] To address the above-mentioned problems, no effective solutions have been proposed so far. Summary of the Invention

[0005] The embodiments of the present application provide a graphics processor monitoring method, system, device, and electronic device to at least solve the problem in the related art that graphics processor monitoring is not comprehensive and specific enough.

[0006] According to one embodiment of the present application, a graphics processor monitoring method is provided, comprising: collecting first target parameters of a target graphics processor according to a preset period, wherein the first target parameters include at least one of the following: a threshold parameter, a state parameter, and an asset information parameter; determining whether the target graphics processor has an abnormality based on the first target parameters; if the target graphics processor has an abnormality, obtaining second target parameters of the target graphics processor through a target instruction, wherein the second target parameters are all operating parameters of the target graphics processor detected during operation of the target graphics processor, and the second target parameters include the first target parameters; and determining the cause of the abnormality of the target graphics processor by analyzing the second target parameters.

[0007] In an exemplary embodiment, obtaining a second target parameter in a target graphics processor through a target instruction includes: determining a target interface corresponding to the target instruction, wherein the target instruction is an instruction in a preset format, and the preset format includes at least one of the following: a first instruction format and a second instruction format; and obtaining the second target parameter in the target graphics processor based on the target interface.

[0008] In an exemplary embodiment, obtaining a second target parameter in a target graphics processor according to a target interface includes: setting a parameter read status in the target graphics processor to allow reading; reading all operating parameters in the target graphics processor and encapsulating them through a target protocol to obtain the second target parameter, wherein the target protocol is a protocol supported by the target graphics processor.

[0009] In an exemplary embodiment, determining whether a target graphics processor has an abnormality based on a first target parameter includes: determining a preset threshold corresponding to a threshold-type parameter in the target graphics processor based on an asset information parameter, wherein the asset information parameter includes at least one of the following: a model of the target graphics processor, a manufacturer of the target graphics processor, and a serial number of the target graphics processor; comparing a parameter value of the threshold-type parameter with a preset threshold, wherein the threshold-type parameter includes at least one of the following: a temperature parameter in the target graphics processor, a voltage parameter in the target graphics processor, and a current parameter in the target graphics processor; and determining that the target graphics processor has an abnormality if the threshold-type parameter exceeds the preset threshold.

[0010] In an exemplary embodiment, judging whether the target graphics processor has an abnormality based on the first target parameter further includes: determining status information of a status parameter, wherein the status information includes normal and abnormal; and determining that the target graphics processor has an abnormality when the status information is abnormal.

[0011] In an exemplary embodiment, before obtaining the second target parameter in the target graphics processor through a target instruction in a preset format, the method further includes: sending an alarm message when an abnormality occurs in the target graphics processor, and sending the abnormal first target parameter to the front-end interface for display.

[0012] According to another embodiment of the present application, a graphics processor monitoring system is provided, comprising: a baseboard management controller and a target graphics processor, wherein the modules running in the baseboard management controller include: a functional application layer, a component management layer, and a protocol encapsulation layer, wherein the protocol encapsulation layer exchanges data with the target graphics processor to encapsulate parameter data in the target graphics processor and generate a target interface for invocation by the component management layer; the component management layer is configured to receive target instructions issued by the functional application layer and, based on the target instructions, call the target interface to collect parameter data in the target graphics processor; and the functional application layer collects parameter data for analysis by sending target instructions.

[0013] In an exemplary embodiment, the functional application layer includes: a real-time monitoring module, a first diagnostic module, and a second diagnostic module, wherein the real-time monitoring module is used to collect first target parameters of a target graphics processor according to a preset period, and determine whether the target graphics processor has an abnormality based on the first target parameters, wherein the first target parameters include: threshold parameters, status parameters, and asset information parameters; the first diagnostic module is used to obtain second target parameters in the target graphics processor by sending a target instruction in a first instruction format, and analyze the second target parameters to obtain a specific cause of the abnormality in the target graphics processor; the second diagnostic module is used to obtain the second target parameters in the target graphics processor by sending a target instruction in a second instruction format, and analyze the second target parameters to obtain a specific cause of the abnormality in the target graphics processor.

[0014] According to another embodiment of the present application, a graphics processor monitoring device is provided, comprising: a periodic monitoring module for collecting first target parameters of a target graphics processor according to a preset period, wherein the first target parameters include at least one of the following: a threshold parameter, a state parameter, and an asset information parameter; an abnormality determination module for determining whether the target graphics processor has an abnormality based on the first target parameters; a parameter acquisition module for acquiring, when the target graphics processor has an abnormality, second target parameters of the target graphics processor through a target instruction, wherein the second target parameters are all operating parameters of the target graphics processor detected during operation of the target graphics processor, and the second target parameters include the first target parameters; and an abnormality analysis module for determining the cause of the abnormality of the target graphics processor by analyzing the second target parameters.

[0015] According to another embodiment of the present application, a computer-readable storage medium is provided, in which a computer program is stored. The computer program is configured to execute the steps of any of the above method embodiments when running.

[0016] According to another embodiment of the present application, an electronic device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0017] This application utilizes a GPU server monitoring model based on a BMC (Baseboard Management Controller) to monitor some key parameters of the GPU in real time. When a key parameter is abnormal, all operating parameters of the GPU are obtained for diagnosis. This solves the problem of insufficiently comprehensive and specific monitoring of the graphics processor in related technologies, thereby ensuring the normal operation of the graphics processor. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a flowchart of a graphics processor monitoring method provided according to an embodiment of the present application;

[0019] Figure 2 This is a hardware structure block diagram of a mobile terminal according to a graphics processor monitoring method provided in an embodiment of the present application;

[0020] Figure 3 is a structural diagram of a graphics processor monitoring system provided according to an embodiment of the present application;

[0021] Figure 4 1 is a schematic diagram of the structure of a graphics processor monitoring device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] The embodiments of the present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0023] To facilitate those skilled in the art to better understand the embodiments of the present application, some technical terms or nouns involved in the embodiments of the present application are explained as follows:

[0024] A graphics processing unit (GPU), also known as a display core, visual processor, or display chip, is a microprocessor specifically designed to perform image and graphics-related calculations on personal computers, workstations, game consoles, and some mobile devices (such as tablets and smartphones).

[0025] Baseboard Management Controller (BMC): A server remote management controller that can perform operations such as firmware upgrades and device monitoring even when the machine is not powered on.

[0026] IPMI: (Intelligent Platform Management Interface): The core of IPMI is a dedicated chip / controller (BMC), which is independent of the operating system, BIOS and processor, and is therefore an out-of-band management device. At the same time, IPMI is a set of computer interface specifications defined for autonomous computer subsystems. It is used to provide management and monitoring functions independent of the host system's CPU, firmware (BIOS or UEFI) and operating system, and other hardware and software. Developers can interact with BMC through IPMI OEM CMD.

[0027] PLDM: Platform Level Data Model. An internal-facing, low-level data model designed to be a valid data / control source for mappings under the Common Information Model (CIM).

[0028] RESTful: It is a set of architectural specifications, constraints, and principles. An architecture that complies with this specification is a RESTful architecture.

[0029] Uniform Resource Identifier (URI): A string of characters used to identify the name of an Internet resource.

[0030] Management Component Transport Protocol (MCTP): is a protocol for bidirectional communication between intelligent devices within the platform management subsystem of a computer system using one or more buses.

[0031] SMBUS (System Management Bus): A low-speed communication bus used in mobile and desktop PC systems. It uses a cheap and powerful two-wire bus to control devices on the motherboard and collect information.

[0032] It should be noted that the terms "first", "second", etc. in the description and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0033] The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, Figure 2 This is a hardware structure block diagram of a mobile terminal of a graphics processor monitoring method according to an embodiment of the present application. Figure 2 As shown, the mobile terminal may include one or more ( Figure 2Only one is shown) a processor 202 (the processor 202 may include but is not limited to a microprocessor MCU or a programmable logic device FPGA and other processing devices) and a memory 204 for storing data, wherein the mobile terminal may also include a transmission device 206 and an input and output device 208 for communication functions. It will be understood by those skilled in the art that Figure 2 The structure shown is only for illustration and does not limit the structure of the mobile terminal. Figure 2 More or fewer components than shown, or with Figure 2 Different configurations shown.

[0034] The memory 204 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the graphics processor monitoring method in the embodiment of the present application. The processor 202 executes various functional applications and data processing by running the computer program stored in the memory 204, that is, implementing the above-mentioned method. The memory 204 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 204 may further include a memory remotely located relative to the processor 202, and these remote memories may be connected to the mobile terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0035] Transmission device 206 is used to receive or transmit data via a network. A specific example of the aforementioned network may include a wireless network provided by the mobile terminal's communications provider. In one embodiment, transmission device 206 includes a network interface controller (NIC), which can be connected to other network devices via a base station to enable communication with the Internet. In another embodiment, transmission device 206 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0036] In this embodiment, a method running on the above mobile terminal is provided. Figure 1 is a flowchart of a graphics processor monitoring method provided according to an embodiment of the present application, such as Figure 1 As shown, the process includes the following steps:

[0037] Step S102: collecting first target parameters of a target graphics processor according to a preset period, wherein the first target parameters include at least one of the following: a threshold parameter, a state parameter, and an asset information parameter;

[0038] In this embodiment, the target graphics processor is a GPU, the first target parameter is a key parameter of the GPU, the threshold parameter is mainly a numerical parameter of the GPU, including but not limited to: a temperature parameter of the target graphics processor, a voltage parameter of the target graphics processor, and a current parameter of the target graphics processor; the status parameter is an identification parameter used to indicate whether the overall operating status of the GPU is normal; and the asset information parameter is a parameter used to indicate the production information of the GPU, including but not limited to: the model of the target graphics processor, the manufacturer of the target graphics processor, the serial number of the target graphics processor, etc.

[0039] In this embodiment, the above-mentioned preset period can be adjusted according to actual needs.

[0040] Step S104, determining whether the target graphics processor has an abnormality based on the first target parameter;

[0041] In this embodiment, whether a target graphics processor has an abnormality can be determined based on threshold parameters. Specifically, determining whether a target graphics processor has an abnormality based on a first target parameter includes the following steps: determining a preset threshold corresponding to the threshold parameter in the target graphics processor based on an asset information parameter, wherein the asset information parameter includes at least one of the following: the model of the target graphics processor, the manufacturer of the target graphics processor, and the serial number of the target graphics processor; comparing a parameter value of the threshold parameter with a preset threshold, wherein the threshold parameter includes at least one of the following: a temperature parameter in the target graphics processor, a voltage parameter in the target graphics processor, and a current parameter in the target graphics processor; and determining that the target graphics processor has an abnormality if the threshold parameter exceeds the preset threshold.

[0042] As an optional implementation, the BMC sends read commands to the GPU at pre-set intervals. Upon receiving these commands, the GPU begins collecting its current GPU information. This information not only reflects the GPU's current performance but also other non-performance information. Once the GPU has collected its current GPU information, it sends it to the BMC. After receiving this information, the BMC monitors the GPU based on the received information. Because the BMC periodically sends read commands to the GPU, it can periodically monitor the GPU, enabling timely detection of GPU issues and timely maintenance of GPU anomalies.

[0043] For example, the temperature preset threshold of GPU model A is 100° C., and the temperature parameter detected by the temperature sensor inside the GPU is 120° C., which exceeds the preset temperature preset threshold. At this time, it is determined that the GPU is abnormal.

[0044] For example, when the power parameter detected by the internal power sensor of GPU model A exceeds the preset power threshold corresponding to the GPU model A, it is determined that the GPU is abnormal.

[0045] In this embodiment, whether the target GPU has an abnormality can also be determined based on the state parameter. Specifically, based on the first target parameter, determining whether the target GPU has an abnormality further includes the following steps: determining status information of the state parameter, where the status information includes normal and abnormal; and determining that the target GPU has an abnormality if the status information is abnormal.

[0046] Step S106, when the target graphics processor is abnormal, obtaining second target parameters of the target graphics processor through the target instruction, wherein the second target parameters are all operating parameters of the target graphics processor detected during the operation of the target graphics processor, and the second target parameters include the first target parameters;

[0047] In this embodiment, the second target parameters are all parameters obtained after detecting the GPU operation process, and the target protocol is a protocol type that can be supported by the target graphics processor.

[0048] In some embodiments of the present application, obtaining a second target parameter in a target graphics processor through a target instruction includes the following steps: determining a target interface corresponding to the target instruction, wherein the target instruction is an instruction in a preset format, and the preset format includes at least one of the following: a first instruction format and a second instruction format; and obtaining the second target parameter in the target graphics processor based on the target interface.

[0049] In this embodiment, the first instruction format is an IPMI OEM command format, and the second instruction format is a restful URL command format. Both can be used to obtain all GPU parameters. The appropriate command format can be selected based on actual needs. This application solution enhances the flexibility of GPU detection by providing an interface function that supports two different command formats. It should be noted that the above two command formats are only examples, and this solution is not limited to the above two command formats for obtaining all GPU parameters.

[0050] In some embodiments of the present application, obtaining a second target parameter in a target graphics processor according to a target interface includes the following steps: setting a parameter read status in the target graphics processor to allow reading; reading all operating parameters in the target graphics processor and encapsulating them using a target protocol to obtain the second target parameter, wherein the target protocol is a protocol supported by the target graphics processor.

[0051] As an optional implementation, parameter acquisition is accomplished by calling the following interfaces. Specifically, the SetNumericSensorEnable interface is called to enable GPU internal threshold parameters, i.e., to set the read status of the threshold parameters in the target GPU to allow reading; the SetStateSensorEnables interface is called to enable GPU internal state parameters, i.e., to set the read status of the state parameters in the target GPU to allow reading; the GetStateSensorReadings interface and the GetSensorReading interface are called to read the raw data of the state parameters and threshold parameters, respectively; finally, the GetPDR interface is called to obtain the specific semantics of the threshold parameters and state parameters, including the units, calculation method of the true value, and coefficients.

[0052] As an optional implementation, the asset information parameters are obtained by calling the GetFRURecordTable interface.

[0053] In order to improve the real-time and comprehensiveness of GPU monitoring, the following steps are also included before obtaining the second target parameter in the target graphics processor through a target instruction in a preset format: when there is an abnormality in the target graphics processor, an alarm message is sent, and the first target parameter with the abnormality is sent to the front-end interface for display.

[0054] Specifically, some key parameters of the GPU are monitored in real time and displayed externally through sensors (Sensor) or IPMI OEM commands. When an abnormality is detected, an alarm message is sent to avoid data processing errors caused by GPU abnormalities.

[0055] Step S108: Determine the cause of the abnormality of the target graphics processor by analyzing the second target parameter.

[0056] Specifically, the second target parameters, ie, all parameters of the GPU, include component status parameters in addition to the key parameters, wherein the component status parameters are used to characterize the current status of each component in the GPU.

[0057] As an optional embodiment, abnormal components in the GPU may be determined by detecting component status parameters, and the component information may be fed back to the front-end interface.

[0058] This application solution does not require maintenance personnel to install GPU drivers and related monitoring software, which simplifies the monitoring process and improves monitoring efficiency. Moreover, the BMC can automatically obtain GPU key parameter information in real time to automatically monitor the status of the GPU. In addition, when there is an abnormality in the key parameters, all parameters of the GPU are obtained to further analyze the cause of the abnormality.

[0059] Through the above steps, the problem of insufficiently comprehensive and specific monitoring of the graphics processor in related technologies is solved, thereby achieving the effect of ensuring the normal operation of the graphics processor. It can be used to assist in locating problems and reduce operation and maintenance costs.

[0060] The execution subject of the above steps may be a controller, a terminal, etc., but is not limited thereto.

[0061] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus the necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0062] An embodiment of the present application further provides a graphics processor monitoring system, which can be used to execute the steps in any of the above method embodiments.

[0063] Figure 3 FIG. 1 is a structural diagram of a graphics processor monitoring system provided according to an embodiment of the present application. Figure 3 As shown, the graphics processor monitoring system includes: a baseboard management controller 30 and a target graphics processor 32, wherein the modules running in the baseboard management controller 30 include: a functional application layer 302, a component management layer 304 and a protocol encapsulation layer 306, and,

[0064] The protocol encapsulation layer 306 exchanges data with the target graphics processor 32 to encapsulate parameter data in the target graphics processor 32 and generate a target interface, wherein the target interface is used for calling by the component management layer 304;

[0065] In this embodiment, the above-mentioned protocol encapsulation layer includes: PLDM protocol layer and Mctp protocol layer

[0066] Specifically, the PLDM protocol layer implements the PLDM Base, PLDM Control and Monitoring, and PLDM FRU protocols; the MCTP protocol layer implements the MCTP Base / Binding protocols. This application adopts a layered design concept, with each layer providing an API for the layer above and invoking the API of the layer below.

[0067] The component management layer 304 is used to receive the target instruction issued by the functional application layer 302, and call the target interface to collect parameter data in the target graphics processor 32 according to the target instruction;

[0068] Specifically, the component management layer provides a unified interface for the functional application layer, internally encapsulates an API interface, and interacts with the PLDM protocol layer through a dual FIFO (first-in-first-out) method to ensure stable communication and decoupling from the PLDM and MCTP protocol layers.

[0069] The functional application layer 302 collects parameter data for analysis by sending target instructions.

[0070] In some embodiments of the present application, the functional application layer 302 includes: a real-time monitoring module 3022, a first diagnosis module 3024, and a second diagnosis module 3026, wherein:

[0071] The real-time monitoring module 3022 is configured to collect first target parameters of the target GPU 32 according to a preset period, and determine whether the target GPU 32 has an abnormality based on the first target parameters, wherein the first target parameters include threshold parameters, status parameters, and asset information parameters;

[0072] Specifically, the real-time monitoring module 3022 is a module for monitoring the key parameters of the GPU in real time. Parameter acquisition can be completed by calling the following interfaces: specifically, calling the SetNumericSensorEnable interface to enable the internal threshold parameters of the GPU, that is, setting the reading status of the threshold parameters in the target graphics processor to allow reading; calling the SetStateSensorEnables interface to enable the internal state parameters of the GPU, that is, setting the reading status of the state parameters in the target graphics processor to allow reading; calling the GetStateSensorReadings interface and the GetSensorReading interface to read the raw data of the state parameters and threshold parameters respectively; finally, calling the GetPDR interface to obtain the specific semantics of the threshold parameters and state parameters, including units, calculation methods of real values, coefficients, etc.

[0073] The GPU in a server is a microprocessor specifically designed for image processing. To enable BMC monitoring of the GPU, the BMC sends read commands to the GPU at pre-set intervals. Upon receiving the read commands, the GPU begins collecting its current GPU information. This information not only reflects the GPU's current performance, but also other non-performance information. Once the GPU has collected its current GPU information, it sends it to the BMC. After receiving the information, the BMC monitors the GPU based on the received information. Because the BMC periodically sends read commands to the GPU, it can periodically monitor the GPU, promptly identifying any GPU issues and enabling timely maintenance of any GPU anomalies.

[0074] For example, the temperature preset threshold of model A GPU is 100° C., while the temperature parameter detected by the GPU internal temperature sensor is 120° C., which exceeds the preset temperature preset threshold. At this time, it is determined that the GPU is abnormal.

[0075] As an optional implementation method, the current temperature of the GPU, that is, the temperature comparison result, can be sent to the front-end interactive interface for display.

[0076] The first diagnostic module 3024 is configured to obtain a second target parameter in the target graphics processor 32 by sending a target instruction in the first instruction format, and analyze the second target parameter to obtain a specific cause of an abnormality in the target graphics processor 32;

[0077] Specifically, the first diagnosis module 3024 is configured to diagnose the GPU by sending instructions in the IPMI OEM command format.

[0078] The second diagnostic module 3026 is configured to obtain a second target parameter in the target graphics processor 32 by sending a target instruction in a second instruction format, and analyze the second target parameter to obtain a specific cause of the abnormality in the target graphics processor 32 .

[0079] Specifically, the first diagnosis module 3024 is used to diagnose the GPU by sending instructions in the restful URL command format.

[0080] In this embodiment, both the first diagnostic module and the second diagnostic module can obtain all parameters of the GPU. The appropriate command format can be selected based on actual needs. This solution enhances the flexibility of GPU detection by providing an interface function that supports two different command formats. It should be noted that the above two command formats are merely examples, and this solution is not limited to obtaining all GPU parameters using these two command formats.

[0081] In this embodiment, the baseboard management controller is a BMC, and the target graphics processor 32 is a PVC OAM GPU, which is a GPU from Intel. Its out-of-band management physical channel is SMBUS, and the application layer implements a partial subset of the PLDM over Mctp protocol stack. The baseboard management controller (BMC) can manage the PVC OAM GPU through PLDM over Mctp over SMBUS.

[0082] Specifically, the BMC is a management controller unique to the server, which can automatically monitor the current operating status of the GPU in the server, such as monitoring the status of sensors in the GPU server, accessing the BIOS configuration or system operation console, etc., and can timely regulate the GPU server based on the current operating status.

[0083] In addition to real-time monitoring of key GPU parameters and displaying them via sensors or IPMI OEM commands, the solution in this application also provides an interface for diagnosing the GPU, such as an IPMI OEM command or a restful URL. When a GPU anomaly occurs, this interface can be used to obtain all GPU parameters, helping to locate the problem and reduce maintenance costs. This solves the problem of insufficiently comprehensive and detailed GPU monitoring in related technologies, thereby ensuring the normal operation of the GPU and helping to locate problems and reduce maintenance costs.

[0084] This embodiment also provides a graphics processor monitoring device for implementing the aforementioned embodiments and preferred implementations. Details already described will not be repeated. As used below, the term "module" may refer to a combination of software and / or hardware that implements a predetermined function. While the devices described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and contemplated.

[0085] Figure 4 FIG. 1 is a structural diagram of a graphics processor monitoring device provided according to an embodiment of the present application. Figure 4 As shown, the graphics processor monitoring device includes:

[0086] The periodic monitoring module 40 is configured to collect a first target parameter of the target graphics processor according to a preset period, wherein the first target parameter includes at least one of the following: a threshold parameter, a state parameter, and an asset information parameter;

[0087] An abnormality determination module 42 is used to determine whether a target graphics processor has an abnormality based on the first target parameter;

[0088] a parameter acquisition module 44 configured to acquire, when an abnormality occurs in the target graphics processor, second target parameters of the target graphics processor through a target instruction, wherein the second target parameters are all operating parameters of the target graphics processor detected during the operation of the target graphics processor, and the second target parameters include the first target parameters;

[0089] The abnormality analysis module 46 is configured to determine the cause of the abnormality of the target graphics processor by analyzing the second target parameter.

[0090] In this embodiment, whether a target graphics processor has an abnormality can be determined based on threshold parameters. Specifically, determining whether a target graphics processor has an abnormality based on a first target parameter includes the following steps: determining a preset threshold corresponding to the threshold parameter in the target graphics processor based on an asset information parameter, wherein the asset information parameter includes at least one of the following: the model of the target graphics processor, the manufacturer of the target graphics processor, and the serial number of the target graphics processor; comparing a parameter value of the threshold parameter with a preset threshold, wherein the threshold parameter includes at least one of the following: a temperature parameter in the target graphics processor, a voltage parameter in the target graphics processor, and a current parameter in the target graphics processor; and determining that the target graphics processor has an abnormality if the threshold parameter exceeds the preset threshold.

[0091] In this embodiment, whether the target GPU has an abnormality can also be determined based on the state parameter. Specifically, based on the first target parameter, determining whether the target GPU has an abnormality further includes the following steps: determining status information of the state parameter, where the status information includes normal and abnormal; and determining that the target GPU has an abnormality if the status information is abnormal.

[0092] In some embodiments of the present application, obtaining a second target parameter in a target graphics processor through a target instruction includes the following steps: determining a target interface corresponding to the target instruction, wherein the target instruction is an instruction in a preset format, and the preset format includes at least one of the following: a first instruction format and a second instruction format; and obtaining the second target parameter in the target graphics processor based on the target interface.

[0093] In this embodiment, the first instruction format is an IPMI OEM command format, and the second instruction format is a restful URL command format. Both can be used to obtain all GPU parameters. The appropriate command format can be selected based on actual needs. This application solution enhances the flexibility of GPU detection by providing an interface function that supports two different command formats. It should be noted that the above two command formats are only examples, and this solution is not limited to the above two command formats for obtaining all GPU parameters.

[0094] In some embodiments of the present application, obtaining a second target parameter in a target graphics processor according to a target interface includes the following steps: setting a parameter read status in the target graphics processor to allow reading; reading all operating parameters in the target graphics processor and encapsulating them using a target protocol to obtain the second target parameter, wherein the target protocol is a protocol supported by the target graphics processor.

[0095] In order to improve the real-time and comprehensiveness of GPU monitoring, the following steps are also included before obtaining the second target parameter in the target graphics processor through a target instruction in a preset format: when there is an abnormality in the target graphics processor, an alarm message is sent, and the first target parameter with the abnormality is sent to the front-end interface for display.

[0096] This application solution does not require maintenance personnel to install GPU drivers and related monitoring software, which simplifies the monitoring process and improves monitoring efficiency. Moreover, the BMC can automatically obtain GPU key parameter information in real time to automatically monitor the status of the GPU. In addition, when there is an abnormality in the key parameters, all parameters of the GPU are obtained to further analyze the cause of the abnormality.

[0097] In addition to real-time monitoring of some key parameters of the GPU and displaying them to the outside world through Sensor or IPMIOEM commands, the solution in this application also provides an interface function for diagnosing the GPU: such as IPMI OEM commands or restful URLs. When the GPU has an abnormality, all the parameters of the GPU can be obtained through this interface function to assist in locating the problem and reduce operation and maintenance costs.

[0098] Through the above steps, the problem of insufficiently comprehensive and specific monitoring of the graphics processor in related technologies is solved, thereby achieving the effect of ensuring the normal operation of the graphics processor. It can be used to assist in locating problems and reduce operation and maintenance costs.

[0099] It should be noted that the above modules can be implemented through software or hardware. For the latter, it can be implemented in the following ways, but not limited to: the above modules are all located in the same processor; or the above modules are located in different processors in any combination.

[0100] An embodiment of the present application further provides a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the steps of any of the above-mentioned method embodiments when run: collecting first target parameters of a target graphics processor according to a preset period, wherein the first target parameters include at least one of the following: a threshold parameter, a state parameter, and an asset information parameter; judging whether the target graphics processor has an abnormality based on the first target parameters; if the target graphics processor has an abnormality, obtaining second target parameters of the target graphics processor through a target instruction, wherein the second target parameters are all operating parameters of the target graphics processor detected during the operation of the target graphics processor, and the second target parameters include the first target parameters; and determining the cause of the abnormality of the target graphics processor by analyzing the second target parameters.

[0101] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk, or an optical disk.

[0102] An embodiment of the present application further provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and a computer terminal comprises one or more processors (the processors may include, but are not limited to, a microprocessor (MCU) or a processing device such as a programmable logic device (FPGA). The processors are configured to execute the computer program to perform the steps of the graphics processor monitoring method in any of the above-mentioned method embodiments: collecting first target parameters of a target graphics processor according to a preset period, wherein the first target parameters include at least one of the following: a threshold parameter, a status parameter, and an asset information parameter; determining whether the target graphics processor has an abnormality based on the first target parameters; if the target graphics processor has an abnormality, obtaining second target parameters of the target graphics processor through a target instruction, wherein the second target parameters are all operating parameters of the target graphics processor detected during operation of the target graphics processor, and the second target parameters include the first target parameters; and determining the cause of the abnormality of the target graphics processor by analyzing the second target parameters. In this embodiment, the processor may be a baseboard management controller (BMC).

[0103] Specifically, the modules run by the baseboard management controller 3 include: a functional application layer, a component management layer, and a protocol encapsulation layer. In this embodiment, the protocol encapsulation layer includes: a PLDM protocol layer and a Mctp protocol layer.

[0104] Specifically, the PLDM protocol layer implements the PLDM Base, PLDM Control and Monitoring, and PLDM FRU protocols; the MCTP protocol layer implements the MCTP Base / Binding protocols. This application adopts a layered design concept, with each layer providing an API for the layer above and invoking the API of the layer below.

[0105] Specifically, the component management layer provides a unified interface for the functional application layer, internally encapsulates an API interface, and interacts with the PLDM protocol layer through a dual FIFO (first-in-first-out) method to ensure stable communication and decoupling from the PLDM and MCTP protocol layers.

[0106] In some embodiments of the present application, the functional application layer includes: a real-time monitoring module, a first diagnostic module, and a second diagnostic module.

[0107] Specifically, the above-mentioned real-time monitoring module is used to monitor the key parameter modules of the GPU in real time. The parameters can be acquired by calling the following interfaces. Specifically, the SetNumericSensorEnable interface is called to enable the internal threshold parameters of the GPU, that is, to set the reading status of the threshold parameters in the target graphics processor to allow reading; the SetStateSensorEnables interface is called to enable the internal state parameters of the GPU, that is, to set the reading status of the state parameters in the target graphics processor to allow reading; the raw data of the state parameters and threshold parameters are read by calling the GetStateSensorReadings interface and the GetSensorReading interface respectively; finally, the GetPDR interface is called to obtain the specific semantics of the threshold parameters and state parameters, including units, calculation methods of the true values ​​and coefficients, etc. The first diagnostic module is used to diagnose the GPU by sending instructions in the IPMI OEM command format. The first diagnostic module is used to diagnose the GPU by sending instructions in the restful URL command format.

[0108] In this embodiment, both the IPMI OEM command format and the restful URL command format can be used to obtain all GPU parameters. The appropriate command format can be selected based on actual needs. This solution enhances the flexibility of GPU detection by providing an interface function that supports two different command formats. It should be noted that the above two command formats are merely examples, and this solution is not limited to using these two command formats to obtain all GPU parameters.

[0109] In an exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0110] In this embodiment, the input and output devices may be front-end interactive pages;

[0111] Specifically, the target object can view and modify the above target instructions according to actual needs in the front-end interactive page; in addition, the front-end interactive page can also display information such as key parameters and all parameters of the obtained target graphics processor.

[0112] As an optional implementation, before obtaining the second target parameter in the target graphics processor through a target instruction in a preset format, the following steps are also included: when an abnormality occurs in the target graphics processor, an alarm message is sent, and the first target parameter with the abnormality is sent to the front-end interface for display.

[0113] This application solution does not require maintenance personnel to install GPU drivers and related monitoring software, which simplifies the monitoring process and improves monitoring efficiency. Moreover, the BMC can automatically obtain GPU key parameter information in real time to automatically monitor the status of the GPU. In addition, when there is an abnormality in the key parameters, all parameters of the GPU are obtained to further analyze the cause of the abnormality.

[0114] In addition to real-time monitoring of key GPU parameters and displaying them via sensors or IPMI OEM commands, the solution in this application also provides an interface for diagnosing the GPU, such as an IPMI OEM command or a restful URL. When a GPU anomaly occurs, this interface can be used to obtain all GPU parameters, helping to locate the problem and reduce maintenance costs. This solves the problem of insufficiently comprehensive and detailed GPU monitoring in related technologies, thereby ensuring the normal operation of the GPU and helping to locate problems and reduce maintenance costs.

[0115] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail here.

[0116] Obviously, those skilled in the art should understand that the modules or steps of the present application described above can be implemented using a general-purpose computing device, they can be concentrated on a single computing device, or distributed across a network composed of multiple computing devices, they can be implemented using program code executable by the computing device, and thus, they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be performed in a different order than herein, or they can be fabricated into separate integrated circuit modules, or multiple modules or steps can be fabricated into a single integrated circuit module for implementation. Thus, the present application is not limited to any specific combination of hardware and software.

[0117] The above description is merely a preferred embodiment of the present application and is not intended to limit the present application. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present application shall be included within the scope of protection of the present application.

Claims

1. A graphics processor monitoring method, characterized in that: include: Collecting first target parameters of a target graphics processor according to a preset period, wherein the first target parameters include at least one of the following: a threshold parameter, a state parameter, and an asset information parameter; determining, based on the first target parameter, whether the target graphics processor has an abnormality; When an abnormality occurs in the target graphics processor, obtaining second target parameters of the target graphics processor through a target instruction, wherein the second target parameters are all operating parameters of the target graphics processor detected during the operation of the target graphics processor, and the second target parameters include the first target parameters; determining a cause of the abnormality of the target graphics processor by analyzing the second target parameter; Wherein, judging whether the target graphics processor has an abnormality based on the first target parameter includes: determining a preset threshold value corresponding to the threshold-type parameter in the target graphics processor based on the asset information parameter, wherein the asset information parameter includes at least one of the following: the model of the target graphics processor, the manufacturer of the target graphics processor, and the serial number of the target graphics processor; comparing a parameter value of the threshold-type parameter with the preset threshold value, wherein the threshold-type parameter includes at least one of the following: a temperature parameter in the target graphics processor, a voltage parameter in the target graphics processor, and a current parameter in the target graphics processor; and determining that the target graphics processor has an abnormality when the threshold-type parameter exceeds the preset threshold value; The second target parameter further includes a component state parameter, and the component state parameter is used to indicate the current state of each component in the target graphics processor, and the target graphics processor includes multiple components.

2. The method according to claim 1, characterized in that Obtaining, by the target instruction, a second target parameter in the target graphics processor includes: Determining a target interface corresponding to the target instruction, wherein the target instruction is an instruction in a preset format, and the preset format includes at least one of the following: a first instruction format and a second instruction format; According to the target interface, a second target parameter in the target graphics processor is obtained.

3. The method according to claim 2, characterized in that Acquiring a second target parameter in the target graphics processor according to the target interface includes: Setting the read status of the parameters in the target graphics processor to allow reading; All the operating parameters in the target graphics processor are read and encapsulated through a target protocol to obtain the second target parameters, wherein the target protocol is a protocol supported by the target graphics processor.

4. The method according to claim 1, wherein Determining whether the target graphics processor has an abnormality according to the first target parameter further includes: Determine status information of the status parameter, wherein the status information includes: normal and abnormal; When the status information is abnormal, it is determined that an abnormality exists in the target graphics processor.

5. The method according to claim 1, wherein Before obtaining the second target parameter in the target graphics processor using a target instruction in a preset format, the method further includes: When an abnormality occurs in the target graphics processor, an alarm message is sent, and the abnormal first target parameter is sent to a front-end interface for display.

6. A graphics processor monitoring system, characterized in that: include: A baseboard management controller and a target graphics processor, wherein the modules running in the baseboard management controller include: a functional application layer, a component management layer, and a protocol encapsulation layer, and, The protocol encapsulation layer exchanges data with the target graphics processor, encapsulates parameter data in the target graphics processor, and generates a target interface, wherein the target interface is used for calling by the component management layer; The component management layer is used to receive target instructions issued by the functional application layer and, based on the target instructions, call the target interface to collect parameter data in the target graphics processor; The functional application layer collects parameter data for analysis by sending target instructions; The functional application layer includes: a real-time monitoring module, a first diagnostic module, and a second diagnostic module. The real-time monitoring module is used to collect first target parameters of a target graphics processor according to a preset period, and determine whether the target graphics processor has an abnormality based on the first target parameters, wherein the first target parameters include threshold parameters, status parameters, and asset information parameters; the first diagnostic module is used to obtain second target parameters in the target graphics processor by sending a target instruction in a first instruction format, and analyze the second target parameters to obtain a specific cause of the abnormality in the target graphics processor; the second diagnostic module is used to obtain second target parameters in the target graphics processor by sending a target instruction in a second instruction format, and analyze the second target parameters to obtain a specific cause of the abnormality in the target graphics processor; Determining whether the target graphics processor has an abnormality based on the first target parameter includes: determining a preset threshold corresponding to a threshold-type parameter in the target graphics processor based on the asset information parameter, wherein the asset information parameter includes at least one of the following: a model of the target graphics processor, a manufacturer of the target graphics processor, and a serial number of the target graphics processor; comparing a parameter value of the threshold-type parameter with the preset threshold, wherein the threshold-type parameter includes at least one of the following: a temperature parameter of the target graphics processor, a voltage parameter of the target graphics processor, and a current parameter of the target graphics processor; and determining that the target graphics processor has an abnormality when the threshold-type parameter exceeds the preset threshold. The second target parameter further includes a component state parameter, and the component state parameter is used to indicate the current state of each component in the target graphics processor, and the target graphics processor includes multiple components.

7. A graphics processor monitoring device, characterized in that: include: A periodic monitoring module, configured to collect a first target parameter of a target graphics processor according to a preset period, wherein the first target parameter includes at least one of the following: a threshold parameter, a state parameter, and an asset information parameter; an abnormality determination module, configured to determine whether a target graphics processor has an abnormality based on the first target parameter; a parameter acquisition module, configured to acquire, when an abnormality occurs in the target graphics processor, second target parameters of the target graphics processor through a target instruction, wherein the second target parameters are all operating parameters of the target graphics processor detected during the operation of the target graphics processor, and the second target parameters include the first target parameters; an abnormality analysis module, configured to determine a cause of an abnormality of the target graphics processor by analyzing the second target parameter; The abnormality determination module is further configured to: determine a preset threshold value corresponding to a threshold-type parameter in the target graphics processor based on an asset information parameter, wherein the asset information parameter includes at least one of the following: a model of the target graphics processor, a manufacturer of the target graphics processor, and a serial number of the target graphics processor; compare a parameter value of the threshold-type parameter with a preset threshold value, wherein the threshold-type parameter includes at least one of the following: a temperature parameter of the target graphics processor, a voltage parameter of the target graphics processor, and a current parameter of the target graphics processor; and determine that an abnormality exists in the target graphics processor when the threshold-type parameter exceeds the preset threshold value; The second target parameter further includes a component state parameter, and the component state parameter is used to indicate the current state of each component in the target graphics processor, and the target graphics processor includes multiple components.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the steps of the method for monitoring a graphics processor according to any one of claims 1 to 5 are implemented.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the graphics processor monitoring method described in any one of claims 1 to 5 are implemented.

Images