System power management method and related device

Abstract

The application discloses a system power consumption management method and related equipment, relates to the edge computing technical field, and comprises the following steps: in response to detecting an access event of an expansion module, acquiring thermal characteristic identification information of the expansion module; based on preset thermal contribution level reference information and the thermal characteristic identification information, determining a thermal contribution level corresponding to the expansion module; based on the thermal contribution level and a preset system total power consumption constraint, determining an available power consumption budget of a processor; based on the available power consumption budget, adjusting a working state parameter of the processor, the working state parameter comprising at least one of a working frequency upper limit and a parallel processing task number upper limit; and based on the thermal contribution level, adjusting a control parameter of a heat dissipation device. The application can realize the collaborative adjustment among the thermal characteristics of the expansion module, the processor power consumption management and the heat dissipation control, improves the system operation reliability and the power consumption utilization efficiency under the condition of meeting the system total power consumption constraint.

Description

Technical Field

[0001] This application relates to the field of edge computing technology, and more specifically, to a system power management method and related equipment. Background Technology

[0002] With the rapid development of edge computing, artificial intelligence inference, and industrial IoT technologies, edge computing terminals, as crucial nodes connecting sensing devices and cloud computing platforms, are widely used in smart manufacturing, intelligent transportation, energy management, and intelligent monitoring. To meet the functional expansion needs of different application scenarios, edge computing terminals typically support various types of expansion modules, such as storage modules, communication modules, and AI acceleration modules. Since different expansion modules vary significantly in terms of power consumption, heat generation, and operating load, the integration of these modules alters the overall power and heat distribution of the edge computing terminal. Therefore, dynamic management of system power consumption and heat dissipation resources is of great importance.

[0003] In related technologies, power management for edge computing terminals typically employs preset power consumption strategies or fixed thermal management strategies. Examples include adjusting the operating frequency based on processor temperature thresholds, setting processor power consumption limits based on fixed load models, and using fixed thermal control curves. However, these methods often fail to adequately consider the impact of variations in expansion module types on system power consumption and thermal load. Using the same processor power consumption configuration and thermal management strategy when connecting expansion modules of different power levels can easily lead to unreasonable allocation of system power resources. Specifically, when connecting high-power expansion modules, simultaneous operation of the processor and expansion modules may increase overall system power consumption, triggering temperature protection mechanisms and causing frequent processor frequency reduction, thus affecting system performance. Conversely, when connecting low-power expansion modules, fixed power limits may result in underutilization of processor performance resources, leading to wasted system resources. In other words, related technologies suffer from the inability to dynamically adapt to the thermal characteristics of expansion modules when managing power consumption in edge computing terminals containing expansion modules, resulting in low system power utilization, insufficient thermal management flexibility, and poor operational performance stability. Summary of the Invention

[0004] In the summary section of this application, the relevant technical solutions are described in general terms, and a series of simplified concepts are introduced. These concepts will be further elaborated in the detailed embodiments section. This summary section should not be construed as limiting the key or essential technical features of the claimed solutions, nor is it intended to limit the scope of protection of the claimed solutions.

[0005] The system power consumption management method and related equipment provided in this application can achieve coordinated adjustment between the thermal characteristics of the expansion module, the power consumption management of the processor, and the heat dissipation control, thereby improving the system's operational reliability and power utilization efficiency while meeting the total system power consumption constraints.

[0006] In a first aspect, this application provides a system power consumption management method, comprising: in response to detecting an access event of an expansion module, acquiring thermal characteristic identification information of the expansion module; determining the thermal contribution level corresponding to the expansion module based on preset thermal contribution level comparison information and the thermal characteristic identification information; determining the available power consumption budget of the processor based on the thermal contribution level and a preset total system power consumption constraint; adjusting the operating state parameters of the processor based on the available power consumption budget, wherein the operating state parameters include at least one of an upper limit of operating frequency and an upper limit of the number of parallel processing tasks; and adjusting the control parameters of the heat dissipation device based on the thermal contribution level.

[0007] In some implementations, obtaining the thermal characteristic identification information of the expansion module includes: obtaining the manufacturer identifier, device identifier, and physical slot location identifier of the expansion module by reading the non-volatile storage area of ​​the expansion module.

[0008] In some implementations, determining the thermal contribution level of the expansion module based on preset thermal contribution level comparison information and thermal characteristic identification information includes: obtaining the initial rated power consumption parameter of the expansion module from a preset power consumption database according to the manufacturer identifier and the device identifier; correcting the initial rated power consumption parameter based on the physical slot location identifier and a preset location thermal influence coefficient to obtain a corrected rated power consumption parameter; and comparing the corrected rated power consumption parameter with at least one preset power consumption threshold to determine the thermal contribution level of the expansion module.

[0009] In some implementations, determining the processor's available power budget based on the thermal contribution level and a preset total system power consumption constraint includes: obtaining the corresponding basic safety margin value from a preset level margin mapping relationship based on the thermal contribution level; accumulating the corrected rated power consumption of all currently connected expansion modules to obtain a total expansion power consumption value; determining the interaction heat loss between modules based on the thermal coupling coefficient between each expansion module, so as to include the interaction heat loss in the total expansion power consumption value; and subtracting the total expansion power consumption value and the basic safety margin value from a preset total power consumption upper limit to obtain the available power budget.

[0010] In some implementations, adjusting the processor's operating state parameters based on the available power budget includes: obtaining the current ambient temperature; performing temperature compensation correction on the available power budget based on a preset ambient temperature-budget adjustment curve to obtain a target available power budget; and adjusting the upper limit of the processor's operating frequency and / or the upper limit of the number of parallel processing tasks based on the target available power budget.

[0011] In some implementations, adjusting the control parameters of the heat dissipation device based on the thermal contribution level includes: determining the target heat dissipation device corresponding to the expansion module from a preset thermal management area mapping relationship based on the physical slot location identifier; and adjusting the rotation speed control parameters of the target heat dissipation device based on the thermal contribution level so that the heat dissipation capacity of the target heat dissipation device matches the thermal contribution of the expansion module.

[0012] In some embodiments, the system power consumption management method further includes: during the operation of the expansion module, acquiring the real-time temperature value and temperature change rate of the expansion module; when the real-time temperature value exceeds a preset warning threshold and the temperature change rate exceeds a preset change rate threshold, applying a temporary power consumption reduction to the available power consumption budget, and increasing the target speed in the speed control parameters of the heat dissipation device until the temperature change rate falls back to the preset safe change rate range.

[0013] In some implementations, applying a temporary power reduction to the available power budget includes: obtaining the thermal time constant of the expansion module, wherein the thermal time constant characterizes the time required for the expansion module to reach a preset proportion of steady-state temperature rise under conditions without active cooling; predicting a temperature prediction value of the expansion module after a preset time step based on the real-time temperature value, the temperature change rate, and the thermal time constant; determining a base reduction amount based on the difference between the predicted temperature value and a preset danger temperature threshold; obtaining priority information of the parallel processing tasks currently being executed by the processor; determining a reduction exemption amount for the expansion module based on the priority information; superimposing the base reduction amount and the reduction exemption amount to obtain the temporary power reduction amount, and deducting the temporary power reduction amount from the available power budget.

[0014] Secondly, this application also provides a system power consumption management device, comprising: an information acquisition unit, configured to acquire thermal characteristic identification information of the expansion module in response to detecting an access event of the expansion module; a level determination unit, configured to determine the thermal contribution level corresponding to the expansion module based on preset thermal contribution level comparison information and the thermal characteristic identification information; a budget determination unit, configured to determine the available power consumption budget of the processor based on the thermal contribution level and a preset total system power consumption constraint; a first parameter adjustment unit, configured to adjust the working state parameters of the processor based on the available power consumption budget, the working state parameters including at least one of an upper limit of working frequency and an upper limit of the number of parallel processing tasks; and a second parameter adjustment unit, configured to adjust the control parameters of the heat dissipation device based on the thermal contribution level.

[0015] Thirdly, this application also provides an electronic device, including: a memory and a processor, wherein the processor is configured to implement the system power management method described in the first aspect when executing a computer program stored in the memory.

[0016] Fourthly, this application also provides a computer-readable storage medium storing computer-executable instructions or a computer program, wherein when the computer-executable instructions or the computer program are executed by a processor, the steps of the system power management method described in the first aspect are implemented.

[0017] Fifthly, this application also provides a computer program product, including a computer program or computer executable instructions, wherein when the computer program or computer executable instructions are executed by a processor, the steps of the system power management method provided in the embodiments of this application are implemented.

[0018] In summary, this application can obtain the corresponding thermal characteristic identification information after detecting an expansion module access event, and determine the thermal contribution level of the expansion module according to the preset thermal contribution level comparison information. Therefore, it can identify different expansion modules based on their own thermal characteristic differences, without adopting a fixed management strategy that is the same for all expansion modules. Thus, when different types of expansion modules are connected to the system, subsequent power consumption and heat dissipation control can be performed based on the actual thermal impact generated by the expansion modules, improving the matching between the system power consumption management strategy and the expansion module status. Since the system determines the processor's available power budget based on the determined thermal contribution level and the preset total system power consumption constraint, the processor's power allocation can be dynamically adjusted according to the power consumption brought by the current expansion module. When the expansion module has a high thermal contribution level, the system can correspondingly reduce the power budget available to the processor, avoiding the processor's operating power consumption and the expansion module's power consumption from exceeding the total system power consumption constraint. When the thermal contribution of the expansion module is low, the processor can obtain more available power resources, thereby improving the system resource utilization efficiency. Since the system adjusts the processor's operating state parameters based on the available power budget, including at least one of the upper limit of operating frequency and the upper limit of the number of parallel processing tasks, the processor's operating state can adapt to the current system thermal power conditions. By dynamically adjusting the processor performance parameters, the effective operating capability of the processor can be maintained while meeting system power constraints, avoiding problems such as underutilization of processor performance resources due to fixed processing strategies or system instability caused by excessive power consumption. Furthermore, since the system also adjusts the control parameters of the heat dissipation device based on the thermal contribution level of the expansion module, the heat dissipation strategy can be adapted to changes in the thermal load brought by the expansion module. When the thermal contribution level of the expansion module is high, control parameters with higher heat dissipation capacity can be used accordingly; when the thermal contribution level of the expansion module is low, the operating requirements of the heat dissipation device can be reduced. In summary, the system power management method provided in this application can achieve coordinated adjustment between the thermal characteristics of the expansion module, processor power management, and heat dissipation control, improving system reliability and power utilization efficiency while meeting the total system power constraints. Attached Figure Description

[0019] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit this specification. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0020] Figure 1 A flowchart illustrating a system power consumption management method provided in an embodiment of this application;

[0021] Figure 2A schematic diagram of the composition structure of a system power management device provided in this application embodiment;

[0022] Figure 3 This is a schematic diagram of the composition structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0023] The terms used in the specification, claims, and drawings of this application, such as "first," "second," "third," "fourth," etc. (if any), are used to distinguish similar objects and not to describe a specific order or sequence. Therefore, it is to be understood that these terms can be used interchangeably where appropriate, allowing the described embodiments to be used in different orders, unless specifically required by the illustrations or description. Furthermore, the terms "is" and "has," and any variations thereof, are intended to cover, non-exclusively, all possible constituent elements. For example, a process, method, system, product, or apparatus comprising several steps or units is not necessarily limited to the steps or units explicitly listed, but may also include other steps or units not explicitly listed, or steps or units inherent to the process, method, product, or apparatus.

[0024] In this application, a "module" or "unit" refers to a computer program or part of a computer program that has a specific function and works in conjunction with other related parts to achieve a predetermined goal. These modules or units can be implemented by software, hardware (e.g., processing circuitry or memory), or a combination of both. One or more processors or memories can implement one or more modules or units. Furthermore, each module or unit can also be part of a larger module or unit.

[0025] The technical solutions of this application will be described in detail below with reference to the accompanying drawings of the embodiments. It should be noted that the described embodiments are only a part of this application, and not all embodiments. In the following description, the "some embodiments" mentioned are only a subset of all possible embodiments, which may be the same or different subsets, and different embodiments can be combined with each other without conflict.

[0026] Figure 1 This is a schematic flowchart illustrating a system power management method provided in an embodiment of this application. For example, see [link to example]. Figure 1 The system power consumption management method provided in this application embodiment may include the following steps 101 to 105:

[0027] Step 101: In response to the detection of the access event of the expansion module, obtain the thermal characteristic identification information of the expansion module.

[0028] In some examples, the expansion module is an M.2 specification expansion function module adapted to the hardware interface of the edge computing terminal. It is the core external hardware for expanding the computing power, communication and storage capabilities of the edge terminal, and it is also the main heat and power load component of the terminal. In the conventional deployment scenario of the edge computing terminal, the expansion module mainly includes three types of mainstream hardware: Non-Volatile Memory Host Controller Interface Specification (NVMe) solid-state drives, 5th Generation Mobile Communication Technology (5G) modules, and Artificial Intelligence (AI) accelerator cards. The thermal characteristics such as rated power consumption, heat generation rate and steady-state temperature rise of different types of expansion modules are significantly different, which are the core factors causing changes in the overall heat distribution and power load of the terminal.

[0029] Access events are triggering scenarios when the hardware topology of an edge computing terminal changes. They include two core operating conditions: the first is the terminal power-on startup phase, during which the system completes hardware initialization scanning and identifies all currently physically installed expansion modules; the second is the hot-swapping operation of modules during normal terminal operation, i.e., the installation of expansion modules while the device is powered on and not restarted. Both types of access events trigger the system's hardware detection mechanism to determine that the terminal's expansion hardware load structure has changed, automatically initiating the thermal characteristic identification information collection process to ensure that the system can adapt to hardware change conditions in real time.

[0030] Thermal characteristic identification information is a set of fundamental parameters characterizing the hardware identity, power consumption level, and heat generation characteristics of an expansion module. It is the core basis for distinguishing the thermal contribution capabilities of different modules and the core data foundation for subsequent thermal management strategy adaptation. This information may include a Vendor Identification (VID), a Device Identification (PID), module rated power consumption parameters, and module thermal resistance parameters. These parameters accurately correspond to the inherent thermal and power consumption attributes of expansion modules from different manufacturers and models, enabling a direct differentiation of heat generation differences between low, medium, and high power consumption modules and avoiding management deviations caused by homogenized hardware identification of different modules. It can be accessed via a high-speed serial computer expansion bus (Peripheral Component Interconnect Express, PCIe) configuration space read channel or an electrically erasable programmable read-only memory (EPROM) dedicated to the expansion module. In the read-only memory (EEPROM) storage area, the system batch collects the hardware and thermal parameters of the expansion modules. For example, when the edge computing terminal powers on and performs a self-test, the system traverses all M.2 hardware slots and initiates parameter reading commands for each of the connected NVMe solid-state drives, 5G communication modules, and AI accelerator cards. If a hot-plugging of an AI accelerator card is detected during terminal operation, the system immediately activates the hardware identification program, reads the unique manufacturer identification code, unique device identification code, rated power consumption, thermal resistance, and other core thermal characteristic identification information corresponding to the AI ​​accelerator card, and completely retains the basic parameter data of the newly added expansion module, thus completing the information collection work for this hardware access event.

[0031] By implementing step 101, the system actively acquires the corresponding thermal characteristic identification information when an expansion module is detected to be connected, enabling the system to identify the thermal-related attributes of the currently connected expansion module and provide a data basis for subsequent thermal contribution assessment and power consumption adjustment. Compared with a fixed thermal management strategy, the system can perform targeted management based on the actual status of the connected expansion module, thereby improving the matching degree between the thermal management strategy and the actual operating status of the system.

[0032] Step 102: Determine the thermal contribution level corresponding to the extended module based on the preset thermal contribution level comparison information and thermal characteristic identification information.

[0033] In some examples, the preset thermal contribution level comparison information is a standardized classification reference data table that is pre-fixed and stored in the local firmware within the edge computing terminal system. It serves as the benchmark for determining the thermal contribution level of the expansion module. The preset thermal contribution level comparison information uses the rated power consumption value of the expansion module as the judgment dimension and divides multiple standardized power consumption threshold ranges. Each threshold range corresponds to a fixed thermal contribution level, which can adapt to the power consumption and heat generation characteristics of various M.2 expansion hardware commonly used in edge terminals.

[0034] Thermal contribution level is a grading standard used to quantify the heat contribution and power consumption level of expansion modules in the overall system. It is used to intuitively distinguish the impact of different expansion modules on the overall thermal environment and power load of the edge terminal. The thermal contribution level can be divided into three standard levels based on preset power consumption thresholds: low power consumption level, medium power consumption level, and high power consumption level. Different levels correspond to different power consumption weights and heat generation impact weights. The higher the level, the more heat the corresponding expansion module generates during operation and the greater its impact on the overall power consumption and heat distribution. The collected rated power consumption parameters of the expansion module can be compared one by one with the power consumption thresholds in the preset thermal contribution level comparison information to match the corresponding threshold range and finally determine the thermal contribution level of the expansion module. For example, the system presets the power consumption threshold for low power consumption level as rated power consumption less than or equal to 5 watts, the threshold for medium power consumption level as rated power consumption greater than 5 watts and less than or equal to 10 watts, and the threshold for high power consumption level as rated power consumption greater than 10 watts.

[0035] For example, after obtaining the complete thermal characteristic identification information of the newly added access expansion module, the edge computing terminal automatically retrieves the preset thermal contribution level comparison information stored in the local firmware and extracts the standard power consumption threshold range corresponding to each level in the comparison table. The system uses the rated power consumption parameter of the expansion module parsed from the thermal characteristic identification information as the core comparison parameter and matches it with the power consumption threshold of each level in sequence. When a non-volatile memory host controller interface specification solid-state drive with a rated power of 6 watts is accessed, the system matches it to the threshold range corresponding to medium power consumption and determines that the module is at the medium power consumption thermal contribution level. When a fifth-generation mobile communication module with a rated power of 8 watts is accessed, the system also determines that it is at the medium power consumption thermal contribution level. When an artificial intelligence accelerator card with a rated power of 12 watts is accessed, the system matches it to the threshold range corresponding to high power consumption and determines that the module is at the high power consumption thermal contribution level.

[0036] By implementing step 102, the acquired thermal characteristic identification information is matched with the preset thermal contribution level comparison information, which can classify the degree of thermal impact of different expansion modules, thereby quantifying the contribution of expansion modules to the overall thermal load of the system. This allows the system to adopt different management strategies based on the differences in the thermal contribution levels of expansion modules, avoiding management mismatch problems caused by using the same control method for all expansion modules.

[0037] Step 103: Determine the processor's available power budget based on the thermal contribution level and the preset total system power consumption constraint.

[0038] In some examples, the preset total system power consumption constraint is the maximum power consumption limit of the entire device that is fixed in the firmware during the edge computing terminal hardware design stage. It is determined by the terminal power supply circuit, heat dissipation structure, and heat resistance specifications of the entire device casing. It represents the maximum power consumption allowed when all hardware components of the entire device are running synchronously and is the top-level constraint standard for allocating various hardware power consumption resources. For example, in the embodiments of this application, the preset total system power consumption constraint of the edge computing terminal is 40 watts.

[0039] The processor is the core computing chip of the edge terminal, namely the System on Chip (SoC). It integrates three types of computing cores: the Central Processing Unit (CPU), the Graphics Processing Unit (GPU), and the Neural Processing Unit (NPU). It undertakes all computing tasks such as data inference, service scheduling, and image processing at the edge. It is also another major source of power consumption and heat generation for the whole machine, apart from the expansion modules. The processor power consumption reading and calling are all completed through the underlying hardware registers. The system can read and write the processor power consumption and frequency operation related registers in real time for power budget allocation and computing power management.

[0040] The available power budget is the maximum allowable power consumption allocated to the on-chip system after deducting the power consumption required by all expansion modules and reserving a safety margin from the total power consumption limit of the entire system. It represents the power consumption range within which the processor can operate stably without overheating or exceeding power limits under the current expansion hardware configuration. The value directly determines the upper limit of the processor's available computing performance. The available power budget can be calculated by the system in real time. The calculation process is completed based on a fixed mathematical formula and does not require manual intervention. The value changes dynamically with the thermal contribution level of the connected expansion modules. For example, the higher the thermal contribution level of the expansion module, the greater the power consumption of the module itself, and the lower the final calculated available power budget value of the processor. The lower the thermal contribution level of the expansion module, the less power consumption of the module, and the higher the available power budget value that the processor can be allocated.

[0041] For example, after the edge computing terminal completes the thermal contribution level determination of the expansion modules, the system automatically retrieves the preset total system power consumption constraint value stored in the firmware. Simultaneously, based on the determined thermal contribution level, it matches the corresponding rated power consumption value of the expansion modules, sums the rated power consumption of all currently connected expansion modules to obtain the total power consumption of the expansion modules, and then reserves a fixed power consumption safety margin in conjunction with the overall system's operational safety requirements. The system substitutes the preset calculation relationship into the formula: the maximum power consumption of the on-chip system equals the upper limit of the overall system power consumption minus the sum of the power consumption of all expansion modules minus the safety power consumption margin, thus completing the processor's available power consumption budget calculation. When the terminal is only equipped with a non-volatile memory host controller interface standard solid-state drive with a rated power consumption of 6 watts, this... The module corresponds to a medium power consumption thermal contribution level. The total power consumption of the expansion module is 6 watts, and the whole machine reserves a 7-watt safety power consumption margin. The preset total power consumption constraint of the whole machine is 40 watts. After calculation, the processor's available power consumption budget is 27 watts. When the terminal connects to an AI accelerator card with a rated power consumption of 12 watts, the module corresponds to a high power consumption thermal contribution level. The total power consumption of the expansion module is 12 watts. Under the same conditions of a 7-watt safety margin and a 40-watt upper limit for the whole machine power consumption, the processor's available power consumption budget is reduced to 21 watts. This calculation logic can dynamically allocate the whole machine power consumption resources according to the heat dissipation level of different expansion modules. Without exceeding the whole machine power supply constraint, it can accurately define the upper limit of the processor power consumption, taking into account both the whole machine power supply safety and the processor's computing power release space.

[0042] By implementing step 103, the processor power allocation can take into account the additional power consumption and thermal load brought by the expansion module; when an expansion module with a high thermal contribution level is connected, the system can reasonably limit the power consumption that the processor can use, and avoid exceeding the system power consumption constraint when the processor and the expansion module are running together; when the thermal contribution of the expansion module is low, more power consumption space can be reserved for the processor, and the system power consumption resource utilization efficiency can be improved.

[0043] Step 104: Based on the available power consumption budget, adjust the processor's operating state parameters. The operating state parameters may include at least one of the upper limit of operating frequency and the upper limit of the number of parallel processing tasks.

[0044] In some examples, operating status parameters are a set of configuration parameters used to manage processor power consumption and computing power. They are direct control variables for the system to regulate processor power consumption and performance. Combined with the computing architecture characteristics of edge computing terminals, operating status parameters can include two core parameters: the upper limit of operating frequency and the upper limit of the number of parallel processing tasks. The system can adjust either parameter individually according to the actual power consumption scenario, or it can adjust both types of parameters simultaneously to adapt to different levels of available power budget and achieve a dynamic balance between power consumption and performance. The upper limit of the operating frequency is the maximum operating frequency allowed for the processor core, which determines the processor's single-cycle processing capability and instantaneous power consumption output level. In this embodiment, the processor is a system-on-a-chip (SoC), which contains a central processing unit (CPU), a neural network processing unit (NNF), and other processing cores. Each processing core is configured with an independent upper limit of its operating frequency. This parameter is configured and modified by the system through a dynamic voltage and frequency scaling (DVFS) mechanism, and the parameter baseline threshold is pre-stored in the frequency configuration table of the system firmware. For example, in the default state of the SoC, the upper limit of the CPU operating frequency can be set to 1.8 GHz, and the upper limit of the NNF operating frequency can be set to 1.2 GHz. When the available power budget decreases, the system can lower the upper limit of the frequency of the corresponding core to reduce power consumption. The upper limit of the number of parallel processing tasks is the maximum number of tasks that the parallel computing unit inside the processor can handle in a single run. It is mainly for AI inference and data parallel computing scenarios of neural network processing units, and is used to manage the processor's multi-task concurrent computing power and overall power consumption load. The baseline configuration range of this parameter is preset in the system task scheduling register, and the system can dynamically limit the maximum capacity of the concurrent task queue according to the power budget. For example, under standard power consumption conditions, the neural network processing unit supports a maximum of four parallel inference tasks running at the same time. When the power resources of the whole machine are occupied by a large number of high-power expansion modules, the system can reduce the number of parallel tasks and reduce the overall computing power consumption of the processor.

[0045] For example, after calculating the available power consumption budget for the processor, the edge computing terminal automatically retrieves a preset power parameter matching table from the firmware to establish a correspondence between the available power consumption budget and the processor's operating state parameters, thus achieving adaptive parameter adjustment. The system prioritizes using the currently calculated available power consumption budget as a benchmark, matching the corresponding dynamic voltage and frequency adjustment level to adaptively adjust the upper limit of the operating frequency of the on-chip system's central processing unit and neural network processing unit. Simultaneously, the system synchronously matches the maximum number of concurrent tasks for the neural network processing unit according to the power consumption budget level, limiting the upper limit of the number of parallel processing tasks. When the terminal connects to a non-volatile memory host controller interface specification solid-state drive with medium power consumption... When the hard drive is connected, the overall power consumption resources are sufficient, and the system maintains the upper limit of the default operating frequency of the on-chip system and the upper limit of the four-way parallel task processing of the neural network processing unit, thus fully releasing the terminal's computing performance. When the terminal is connected to a high-power AI accelerator card, the expansion module occupies a large amount of overall power consumption resources, and the available power budget of the processor decreases significantly. The system automatically lowers the upper limit of the operating frequency of each computing core of the on-chip system, and at the same time lowers the upper limit of the number of parallel processing tasks of the neural network processing unit from four to two, accurately reducing the overall power output of the processor. This ensures that the actual power consumption of the processor strictly matches the available power budget, avoiding the problems of overheating and frequency reduction due to excessive power consumption of the whole machine, and ensuring the continuous and stable operation of edge computing services.

[0046] By implementing step 104, processor computing resources can be configured reasonably while meeting power consumption constraints. By adjusting the upper limit of the operating frequency or the number of parallel processing tasks, the processor can avoid instability caused by power consumption exceeding the limit, while reducing resource waste caused by fixed performance configuration, thereby improving the overall operating efficiency of the system.

[0047] Step 105: Adjust the control parameters of the heat dissipation device based on the thermal contribution level.

[0048] In some examples, the heat dissipation device is a hardware component deployed inside the edge computing terminal to actively dissipate heat from the M.2 expansion module and the on-chip system. It is mainly a DC speed-controlled cooling fan built into the terminal, which is responsible for removing the heat generated by the expansion module and the core computing chip during operation through active air cooling. It is the hardware that executes the thermal management of the entire terminal. The heat dissipation device supports stepless speed adjustment and can achieve fine control of the operating level and speed through the system-level driver. It can adapt to the heat dissipation requirements of expansion modules with different heat levels and provide stable heat dissipation guarantee for the closed and compact hardware operating environment of the edge terminal.

[0049] Control parameters are configuration parameters used to define and constrain the operating status of the heat dissipation device and determine the overall heat dissipation capacity. They can include two key parameters: the initial operating temperature of the fan and the maximum operating speed of the fan. These two parameters together constitute the standardized fan speed control curve of the terminal. The initial operating temperature of the fan is used to define the temperature threshold for the heat dissipation device to start active heat dissipation intervention, and the maximum operating speed of the fan is used to limit the upper limit of the speed of the heat dissipation device when it is operating at full load. The baseline configuration set of control parameters is pre-stored in the heat dissipation strategy data table of the terminal system firmware, and the system can dynamically retrieve and adaptively modify it according to the heat contribution level.

[0050] For example, after determining the thermal contribution level of the access expansion module, the edge computing terminal immediately retrieves the pre-stored thermal level and heat dissipation parameter mapping table in the local firmware, matches the dedicated heat dissipation control parameters corresponding to the current module level, and completes the dynamic update and loading of the overall heat dissipation strategy. The system uses a continuous weight function to smoothly adjust the parameters, avoiding performance fluctuations and noise spikes caused by sudden changes in the speed of the heat dissipation device, and ensuring the stability of heat dissipation strategy switching. When a low-power expansion module is accessed, the system maintains a relatively high initial fan operating temperature, reduces the maximum fan operating speed, and minimizes ineffective heat dissipation power consumption under low load conditions. When a medium-power non-volatile memory host controller interface specification solid-state drive or a fifth-generation mobile communication module is connected, the system uses the default standard thermal curve to balance heat dissipation effect and heat dissipation power consumption. When a high-power artificial intelligence accelerator card is connected, the system actively lowers the fan's initial operating temperature, enabling the heat dissipation device to intervene in heat dissipation in advance. At the same time, the maximum operating speed of the fan is increased to enhance the overall heat dissipation limit, timely dissipate the large amount of heat generated by the high-power module, effectively suppress the generation of local hot spots on the terminal, avoid high temperature triggering the system to drastically reduce frequency, and ensure that the edge computing terminal has stable heat dissipation and operating performance under different extended hardware configurations.

[0051] By implementing step 105, the heat dissipation capacity can be matched with the current system thermal load. When the thermal contribution level of the expansion module is high, the heat dissipation control intensity can be increased to meet the higher heat release requirements. When the thermal contribution level of the expansion module is low, the heat dissipation control requirements can be reduced, enabling dynamic allocation of heat dissipation resources and improving the adaptability of the heat dissipation device and the efficiency of system thermal management.

[0052] In summary, the embodiments of this application can obtain the corresponding thermal characteristic identification information after detecting the access event of the expansion module, and determine the thermal contribution level of the expansion module according to the preset thermal contribution level comparison information. Therefore, it can identify the expansion modules based on their own thermal characteristic differences, without having to adopt a fixed management strategy that is the same for all expansion modules. Thus, when different types of expansion modules are accessed by the system, subsequent power consumption and heat dissipation control can be performed according to the actual thermal impact generated by the expansion modules, improving the matching between the system power consumption management strategy and the expansion module status. Since the system determines the available power budget of the processor based on the determined thermal contribution level and the preset total system power consumption constraint, the power allocation of the processor can be dynamically adjusted according to the power consumption of the current expansion module. When the expansion module has a high thermal contribution level, the system can correspondingly reduce the power budget available to the processor, avoiding the processor's operating power consumption and the expansion module's power consumption from exceeding the total system power consumption constraint. When the thermal contribution of the expansion module is low, the processor can obtain more available power resources, thereby improving the system resource utilization efficiency. Since the system adjusts the processor's operating state parameters based on the available power budget, including at least one of the upper limit of operating frequency and the upper limit of the number of parallel processing tasks, the processor's operating state can adapt to the current system thermal power conditions. By dynamically adjusting the processor performance parameters, the effective operating capability of the processor can be maintained while meeting system power constraints, avoiding problems such as underutilization of processor performance resources due to fixed processing strategies or system instability caused by excessive power consumption. Furthermore, since the system also adjusts the control parameters of the heat dissipation device based on the thermal contribution level of the expansion module, the heat dissipation strategy can be adapted to changes in the thermal load brought by the expansion module. When the thermal contribution level of the expansion module is high, control parameters with higher heat dissipation capacity can be used accordingly; when the thermal contribution level of the expansion module is low, the operating requirements of the heat dissipation device can be reduced. In summary, the system power management method provided in this application embodiment can achieve coordinated adjustment between the thermal characteristics of the expansion module, processor power management, and heat dissipation control, improving system reliability and power utilization efficiency while meeting the total system power constraints.

[0053] In some embodiments, obtaining the thermal characteristic identification information of the expansion module may include: obtaining the manufacturer identifier, device identifier, and physical slot location identifier of the expansion module by reading the non-volatile storage area of ​​the expansion module.

[0054] In some examples, the non-volatile storage area is a permanent storage hardware area configured locally on various M.2 specification expansion modules. The data stored inside will not be lost after power failure. It is specifically used to solidify the factory hardware parameters, identity information and basic characteristic parameters of the expansion module. The edge computing terminal can read the data in this area through the hardware underlying read and write protocol. It is the carrier for obtaining the native hardware information of the expansion module in the embodiments of this application.

[0055] The manufacturer identifier is a standardized identification code used to distinguish the manufacturers of expansion modules. It is one of the most basic identity parameters of the expansion module. The manufacturer identifier can be read by the terminal's underlying driver by initiating a read-only read command to traverse and read the manufacturer-specific code field in the non-volatile storage area of ​​the expansion module. For example, different types of edge computing expansion modules each have their own unique manufacturer identifier, which can be used to distinguish the same type of storage modules, communication modules and intelligent computing modules produced by different manufacturers.

[0056] Device identifiers are standardized codes pre-defined by manufacturers to distinguish different hardware models within the same manufacturer. They can accurately locate the specific hardware specifications, product series, and basic hardware attributes of expansion modules. Device identifiers work in conjunction with manufacturer identifiers to achieve precise matching of expansion module hardware models, preventing confusion between different models of modules from the same manufacturer and modules of the same specifications from different manufacturers. Device identifiers are obtained in the same way as manufacturer identifiers, directly by the terminal by reading the device-specific code field in the non-volatile storage area of ​​the expansion module. For example, non-volatile memory host controller interface specification solid-state drives and artificial intelligence accelerator cards from the same manufacturer have completely different device identifiers, and the specific type of module can be quickly distinguished through coding.

[0057] The physical slot location identifier of the expansion module is a positioning code used to characterize the specific installation position of the expansion module within the edge computing terminal chassis. The terminal pre-sets a corresponding location code for each M.2 hardware slot, which is the physical slot location identifier. This identifier can distinguish expansion modules in different positions under the terminal's multi-slot hardware architecture, making up for the deficiency that relying solely on module identity parameters cannot distinguish the installation position. The physical slot location identifier can be generated by the terminal system scanning its own hardware slot register, matching the slot number to which the current expansion module is physically connected, and generating the corresponding location identifier information. For example, if the edge computing terminal has M.2 slot No. 1 and M.2 slot No. 2, and the two slots correspond to independent physical slot location identifiers, even if identical expansion modules are connected, the system can still distinguish the module installation position through the location identifier.

[0058] For example, when the edge computing terminal detects a power-on or hot-plug event of an expansion module, the system immediately initiates a hardware information acquisition process. It establishes a communication link with the expansion module's non-volatile storage area through the underlying hardware read / write protocol. The system uses a read-only mode to sequentially read the manufacturer identification data and device identification data embedded in the storage area, while simultaneously scanning the terminal's local slot hardware registers to obtain the physical slot location identifier corresponding to the current installation point of the expansion module. After completing all data reading, the system binds and stores the three types of identification information, forming a complete set of thermal characteristic identification information for the accessed expansion module. Through this acquisition method, the system can not only accurately identify the manufacturer and specific model of the expansion module, but also accurately locate the module's physical installation position. This provides comprehensive and accurate data support for subsequent thermal characteristic analysis, power consumption classification, and heat dissipation control based on the module model and installation position, adapting to the complex operating scenarios of multiple modules and slots in edge computing terminals.

[0059] Through the implementation of the above embodiments, after detecting the access event of the expansion module, the thermal characteristic identification information of the expansion module is obtained, and the thermal impact degree corresponding to the expansion module is determined according to the thermal contribution level. Furthermore, the available power consumption budget of the processor is determined in combination with the total power consumption constraint of the system, so that the edge computing terminal can dynamically adjust the processor operating status and heat dissipation control parameters according to the power consumption and thermal characteristics of different expansion modules. Compared with fixed power consumption configuration and fixed heat dissipation strategy, it can match the system power consumption allocation with the actual thermal load of the expansion module, reduce the problem of unreasonable power consumption allocation caused by changes in expansion modules, improve the system power consumption utilization rate, and improve the operating stability of the edge computing terminal under different expansion configurations.

[0060] In some embodiments, step 102 may include: obtaining the initial rated power consumption parameters of the expansion module from a preset power consumption database based on the manufacturer identifier and the device identifier; correcting the initial rated power consumption parameters based on the physical slot location identifier and a preset location thermal influence coefficient to obtain the corrected rated power consumption parameters; and comparing the corrected rated power consumption parameters with at least one preset power consumption threshold to determine the thermal contribution level corresponding to the expansion module.

[0061] In some examples, the preset power consumption database is a structured parameter data table pre-embedded in the local firmware of the edge computing terminal. It is specifically used to store the standardized power consumption and thermal parameters of various mainstream M.2 expansion modules. The preset power consumption database can use manufacturer identification and device identification as index keywords to complete data classification storage and fast retrieval. It includes the factory rated power consumption data of various expansion modules commonly used in edge scenarios, such as non-volatile memory host controller interface standard solid-state drives, fifth-generation mobile communication modules, and artificial intelligence accelerator cards.

[0062] The initial rated power consumption parameter is the standard rated operating power consumption specified in the expansion module's manufacturer's specifications. It represents the theoretical power consumption value of the module under standard room temperature, standard airflow environment, and rated load conditions, and can characterize the basic power consumption and heat generation level of the expansion module. The manufacturer's identifier and device identifier can be used as search keywords to traverse the preset power consumption database and match the standard rated power consumption parameter of the corresponding expansion module model. For example, the database can match the initial rated power consumption parameter of a conventional non-volatile memory host controller interface specification solid-state drive as 6 watts, the initial rated power consumption parameter of a conventional fifth-generation mobile communication module as 8 watts, and the initial rated power consumption parameter of a conventional artificial intelligence accelerator card as 12 watts.

[0063] The preset location thermal impact coefficient is an environmental correction parameter calibrated during the terminal's overall structural design phase and pre-stored in the system firmware. It is used to quantify the impact of different physical slot locations on the heat dissipation and actual heat generation performance of the expansion module. Due to the compact internal space of the edge computing terminal chassis, the airflow, heat interference from surrounding devices, and heat dissipation clearance vary between different slots. The actual temperature rise and power consumption of the same module installed in different slots are not consistent. The preset location thermal impact coefficient is configured independently for each physical slot. The value is greater than zero. Slots with poor airflow and dense heat sources have a higher coefficient, while slots with good airflow and excellent heat dissipation conditions have a lower coefficient.

[0064] The corrected rated power consumption parameter is a true equivalent power consumption parameter obtained through location environment compensation calculation, closely matching the actual installation conditions of the terminal. It can be processed by combining the initial rated power consumption parameter with the preset position thermal influence coefficient of the corresponding slot to offset the environmental differences between standard operating conditions and actual equipment operating conditions, resulting in a power consumption value that truly reflects the module's current heat contribution capability. Compared to the initial rated power consumption parameter, the corrected rated power consumption parameter can accurately reflect the actual heat output level of the module at a specific chassis location, providing a more realistic calculation basis for subsequent heat level determination. For example, when a standard 6-watt storage module is installed in a slot with poor heat dissipation, after coefficient correction, an equivalent power consumption parameter slightly higher than 6 watts can be obtained, truly restoring the actual heat influence of the module at that location.

[0065] At least one preset power consumption threshold is a power consumption critical value pre-set by the system to divide different heat contribution levels, and it serves as the judgment benchmark for quantitative classification of heat generation levels. In this application embodiment, multiple sets of stepped preset power consumption thresholds can be set to form interval judgment standards to distinguish between three heat contribution levels: low power consumption, medium power consumption, and high power consumption. The preset power consumption thresholds are fixed in the system's thermal management configuration file and can be called and compared by the system in real time. The threshold values ​​are calibrated according to the power consumption distribution range of conventional expansion modules in edge terminals, which can cover the power consumption classification requirements of most commercial M.2 expansion modules. For example, the system sets two core preset power consumption thresholds of 5 watts and 10 watts to divide three complete power consumption ranges as boundary standards for level judgment.

[0066] The process of comparing the corrected rated power consumption parameters with at least one preset power consumption threshold to determine the thermal contribution level of the expansion module involves using the environmentally corrected equivalent power consumption value as the basis for judgment, comparing it sequentially with the preset power consumption thresholds of each level, accurately matching the corresponding power consumption range, and thus locking in the thermal contribution level of the expansion module.

[0067] For example, after obtaining the manufacturer's identifier, device identifier, and physical slot location identifier of the expansion module, the edge computing terminal first uses the manufacturer's identifier and device identifier as search indexes to quickly traverse the local preset power consumption database and accurately retrieve the initial rated power consumption parameters corresponding to that model of expansion module; then, based on the physical slot location identifier corresponding to the current module, the system retrieves the preset location thermal impact coefficient specific to that slot, and performs environmental compensation calculations on the initial rated power consumption parameters through a preset correction algorithm to eliminate operating condition deviations caused by chassis airflow and surrounding heat source interference, and calculates the corrected rated power consumption parameters; after completing the power consumption correction, the system retrieves the locally fixed multi-level preset power consumption thresholds and applies the corrected parameters... Rated power consumption parameters are sequentially compared and matched with thresholds to automatically determine the thermal contribution level based on the power consumption range of the parameters. In actual operation, when a non-volatile memory host controller interface standard solid-state drive with a standard initial power consumption of 6 watts is connected, and the module is installed in a conventional heat sink slot, the positional thermal impact coefficient approaches 1. The corrected rated power consumption parameter is basically maintained at around 6 watts, matching a power consumption range greater than 5 watts and less than or equal to 10 watts, and the system determines that the module is at the medium power consumption thermal contribution level. When an AI accelerator card with a standard initial power consumption of 12 watts is connected, the power consumption value is still greater than 10 watts after positional coefficient correction, and the system directly determines that the module is at the high power consumption thermal contribution level. Through this correction and comparison classification logic, the deviation caused by fixed parameter classification can be effectively avoided, allowing thermal management and power consumption control strategies to be fully adapted to the actual hardware installation environment of the edge terminal, improving the overall power utilization efficiency and thermal management flexibility.

[0068] By implementing the above embodiments, the non-volatile storage area of ​​the expansion module is read to obtain the manufacturer's identifier, device identifier, and physical slot location identifier of the expansion module. This enables the system to determine the corresponding basic thermal characteristic data based on the identification information carried by the expansion module itself and the installation location information. Since different expansion modules and different installation locations may correspond to different power consumption and heat dissipation effects, the embodiments of this application can improve the accuracy of expansion module identification and provide more reliable data for subsequent thermal contribution level assessment and power consumption management. This improves the problem of insufficient thermal management flexibility caused by the inability of traditional systems to perceive differences in expansion modules.

[0069] In some embodiments, step 103 may include: obtaining the corresponding basic safety margin value from a preset level margin mapping relationship based on the thermal contribution level; accumulating the corrected rated power consumption of all currently connected expansion modules to obtain the total expansion power consumption value; determining the interaction heat loss between modules based on the thermal coupling coefficient between each expansion module, so as to include the interaction heat loss in the total expansion power consumption value; and subtracting the total expansion power consumption value and the basic safety margin value from the preset total power consumption upper limit value to obtain the available power consumption budget.

[0070] In some examples, the preset level margin mapping relationship is a corresponding matching data table pre-embedded in the terminal system firmware, used to establish exclusive association rules between the thermal contribution level of the extension module and the safe power consumption margin. This mapping relationship configures independent margin standards for the three thermal contribution levels of low power, medium power, and high power, fully adapting to the temperature rise characteristics, load fluctuation amplitude, and thermal stability differences of modules of different levels, realizing the hierarchical adaptation configuration of the safe margin, rather than using a uniform fixed margin value. For example, low power level modules correspond to a smaller basic safe margin, and high power level modules correspond to a larger basic safe margin, thereby adapting to the temperature rise fluctuation risk of high heat generation modules.

[0071] The basic safety margin is a dynamic power consumption buffer value reserved to avoid the risk of power consumption exceeding limits caused by instantaneous load fluctuations, voltage fluctuations, and slight temperature differences in the environment of edge computing terminals. It is used to ensure that the power consumption and thermal operation of the whole machine are always within a safe range. After determining the thermal contribution level of the expansion module, the preset level margin mapping relationship can be directly queried to obtain the exclusive margin value corresponding to the current operating condition. For example, the system can preset the basic safety margin value to be 2 watts for low power consumption level, 4 watts for medium power consumption level, and 7 watts for high power consumption level, adapting to the safety protection requirements of different heat dissipation scenarios through graded margins.

[0072] The total extended power consumption value is the sum of the rated power consumption of all currently connected and running extended modules of the edge computing terminal after environmental location correction. It is a parameter that characterizes the basic power consumption overhead of the entire extended hardware. It can traverse all currently online extended modules, retrieve the corrected rated power consumption parameters of each module, and perform arithmetic summation on all parameters to obtain the basic extended total power consumption without taking into account the thermal coupling effect, providing basic calculation data for subsequent power consumption correction and budget calculation.

[0073] The thermal coupling coefficient between each expansion module is a calibration parameter used to quantify the degree of mutual thermal interference when multiple expansion modules are running on the same machine. The internal space of the edge computing terminal is compact, and multiple M.2 expansion modules are arranged close together. The heat generated by a single module will have a cumulative effect on the temperature rise of other surrounding modules through air convection and heat conduction on the circuit board. The thermal coupling coefficient is pre-calibrated and solidified in the system firmware according to the module arrangement spacing, the normal heat generation range, and the chassis air duct structure. Different module combinations correspond to different thermal coupling coefficients. The higher the power consumption of the modules and the denser the arrangement, the larger the corresponding thermal coupling coefficient value and the more significant the mutual thermal interference effect.

[0074] Interactive heat loss is the additional equivalent power consumption increase caused by the thermal coupling effect between multiple expansion modules when they work simultaneously. The thermal interference between multiple modules will lead to the deterioration of the overall heat dissipation conditions and the actual temperature rise of the modules compared to independent operating conditions, which is equivalent to the additional hidden power consumption of the whole machine. Interactive heat loss cannot be directly reflected by the rated power consumption of a single module. It must be obtained by joint calculation based on the thermal coupling coefficient and the corrected power consumption of each module, which can accurately restore the real heat consumption level of the whole machine when multiple modules are running concurrently.

[0075] The process of including interactive heat loss in the total extended power consumption value involves the system adding the interactive heat loss generated between multiple modules to the power consumption accumulation result of each independent extended module, and then performing a second correction on the basic total extended power consumption value to obtain the final total extended power consumption value that closely reflects the actual operating conditions of multi-module coupled heat generation. This operation can compensate for the deficiency of ignoring thermal interference loss in the single power consumption accumulation calculation, solve the problems of underestimation of power consumption and unreasonable budget allocation in multi-module mixed insertion scenarios, and make the overall power consumption statistics more consistent with the actual operating state of the equipment.

[0076] The preset total power consumption limit is the maximum safe power consumption threshold of the entire machine as specified in the terminal hardware design. It is the top-level constraint on the power consumption of all hardware in the entire machine. By deducting the actual equivalent power consumption of all expansion modules, including thermal coupling losses, and the basic safety margin adapted to the current module level from the maximum allowable power consumption of the entire machine, the remaining power consumption resources are the maximum power consumption space that the on-chip system can freely schedule and operate safely, which is the available power consumption budget of the processor.

[0077] For example, after the edge computing terminal completes the thermal contribution level determination and rated power consumption correction of all connected expansion modules, it first queries the preset level margin mapping relationship in the system firmware according to the thermal contribution level of each module, and matches and obtains the basic safety margin value corresponding to the current operating condition; then the system traverses all connected valid expansion hardware, summarizes and accumulates the corrected rated power consumption parameters of all modules, and obtains the initial total expansion power consumption value; the system further retrieves the preset module thermal coupling coefficient, combines the combination type and arrangement position of the currently online modules, calculates the interactive heat loss generated by the collaborative work of multiple modules, and adds this part of the loss to the initial total expansion power consumption value to complete the final correction of the total expansion power consumption value; finally, the system reads the preset total power consumption upper limit value of the whole machine. By subtracting the corrected total extended power consumption value and the basic safety margin value from the total power consumption limit, the system accurately calculates the processor's available power consumption budget under the current operating conditions. For example, when a terminal is equipped with a medium-power non-volatile memory host controller interface standard solid-state drive and a high-power artificial intelligence accelerator card, the two types of modules working in close proximity generate a positive thermal coupling effect. The system calculates the corresponding interactive heat loss and includes it in the total extended power consumption. At the same time, it matches the high-level corresponding large value basic safety margin, and finally calculates the processor's available power consumption budget that fits the actual heat and power consumption conditions. This avoids the risk of overheating and frequency reduction caused by insufficient power reserve, and also prevents excessive margin from wasting computing resources, effectively improving the accuracy and stability of the edge terminal's overall power consumption management.

[0078] Through the implementation of the above embodiments, the initial rated power consumption parameters of the expansion module are obtained from the preset power consumption database based on the manufacturer's identifier and the device identifier. The rated power consumption is then corrected by combining the positional thermal influence coefficient corresponding to the physical slot location. This allows the system to comprehensively consider the power consumption characteristics of the expansion module itself and the impact of the installation location on heat distribution. On this basis, the thermal contribution level is determined by comparing the corrected rated power consumption parameters with the preset power consumption threshold. This makes the thermal contribution assessment of the expansion module more consistent with the actual operating conditions, avoids thermal load estimation deviations caused by judging solely based on fixed power consumption parameters, and improves the accuracy of the power management strategy.

[0079] In some embodiments, the aforementioned adjustment of the processor's operating state parameters based on the available power budget may include: obtaining the current ambient temperature; performing temperature compensation correction on the available power budget based on a preset ambient temperature-budget adjustment curve to obtain a target available power budget; and adjusting the upper limit of the processor's operating frequency and / or the upper limit of the number of parallel processing tasks based on the target available power budget.

[0080] In some examples, the current ambient temperature is the air temperature of the operating environment of the edge computing terminal in real time. It is a physical parameter characterizing the external heat dissipation conditions of the terminal and directly determines the heat dissipation efficiency of the terminal's heat dissipation device and the upper limit of the steady-state temperature rise of the hardware. The terminal can rely on the high-precision onboard temperature sensor chip built into it to collect the ambient temperature data inside the chassis in real time and upload it to the system's underlying driver. The system periodically reads the sensor data to complete the accurate sampling of the current ambient temperature. For example, the ambient temperature of the terminal in a normal indoor deployment is 25 degrees Celsius. Under high load conditions in a closed cabinet, the ambient temperature can rise to more than 40 degrees Celsius. In outdoor low-temperature scenarios, the ambient temperature can drop to below zero degrees Celsius. Different temperature environments correspond to completely different overall heat dissipation conditions.

[0081] The preset ambient temperature-budget adjustment curve is a standardized mapping curve calibrated through whole-machine thermal testing and embedded in the system firmware before the device leaves the factory. It is used to characterize the correspondence between changes in ambient temperature and the power consumption budget correction coefficient. This curve covers the entire operating temperature range of the terminal in the form of a continuous function, abandoning the segmented abrupt adjustment method, and can achieve smooth fine-tuning of the power consumption budget, avoiding performance jitter caused by parameter jumps. For example, when the ambient temperature is in the normal temperature range, the correction coefficient of the curve approaches 1, basically not changing the basic available power consumption budget; when the ambient temperature continues to rise and exceeds the normal temperature threshold, the correction coefficient of the curve gradually decreases, and the power consumption budget is moderately contracted to compensate.

[0082] The target available power budget is the final effective power consumption constraint value of the processor obtained after environmental temperature compensation correction. It serves as the benchmark used by the system to ultimately regulate the processor's operating parameters. The target available power budget uses the basic available power budget as the calculation basis and combines it with the curve correction coefficient corresponding to the real-time environmental temperature to complete dynamic calibration, accurately adapting to the maximum power consumption limit that the processor can safely operate under the current environmental heat dissipation capacity. Compared with the uncorrected basic available power budget, the target available power budget takes into account both the inherent power consumption constraints of the hardware and the external environmental heat dissipation constraints, and the power allocation result is more in line with the real-time operating conditions of the terminal. For example, in high-temperature environments, the target available power budget will be slightly lower than the basic available power budget to avoid overheating and frequency throttling caused by insufficient heat dissipation; in low-temperature environments, the target available power budget can be moderately higher than the basic available power budget to fully release the processor's remaining computing power performance.

[0083] For example, after completing the basic available power consumption budget calculation, the edge computing terminal immediately initiates the ambient temperature compensation and correction process. The system collects the current ambient temperature of the device in real time through the onboard temperature sensor chip, accurately obtains the real-time temperature value, and performs data filtering to eliminate instantaneous temperature interference errors. Subsequently, the system retrieves the preset ambient temperature-budget adjustment curve in the firmware, matches the standard correction coefficient corresponding to the current ambient temperature, and uses a preset compensation algorithm to scale and correct the basic available power consumption budget, calculating the target available power consumption budget adapted to the current heat dissipation environment. After completing the power consumption budget calibration, the system adaptively adjusts the on-chip system's operating status based on the magnitude of change in the target available power consumption budget. The system adjusts the upper limit of the operating frequency of the central processing unit and the neural network processing unit through a dynamic voltage and frequency adjustment mechanism, while simultaneously adjusting the upper limit of the number of parallel processing tasks of the neural network processing unit. When the terminal is deployed in a high-temperature, enclosed environment, the target available power consumption budget is slightly reduced, and the system appropriately lowers the upper limit of the processor frequency and the number of parallel tasks to avoid local hot spots and overheating of the whole machine caused by high temperature and high load. When the terminal is deployed in a low-temperature, ventilated environment, the target available power consumption budget is appropriately relaxed, and the system maintains a higher processor frequency and upper limit of parallel tasks to fully tap the potential of processor computing power and maximize the processing efficiency of edge computing services while ensuring the safe and stable operation of the equipment.

[0084] Through the implementation of the above embodiments, a basic safety margin value is obtained based on the thermal contribution level, and the corrected rated power consumption of each expansion module is accumulated. At the same time, the interactive heat loss is determined by combining the thermal coupling coefficient between expansion modules. This allows the system to comprehensively consider the impact of multiple expansion modules operating together when calculating the available power consumption budget of the processor. Compared with the method of allocating power consumption based only on the power consumption of a single module, this method can more accurately reflect the overall thermal power consumption status of the edge computing terminal. While meeting the total power consumption constraints of the system, it can reasonably allocate the available power consumption resources of the processor and improve the reliability of system power consumption management.

[0085] In some embodiments, step 105 may include: determining the target heat dissipation device corresponding to the expansion module from a preset thermal management area mapping relationship based on the physical slot location identifier; and adjusting the rotation speed control parameters of the target heat dissipation device based on the thermal contribution level so that the heat dissipation capacity of the target heat dissipation device matches the thermal contribution of the expansion module.

[0086] In some examples, the preset thermal management area mapping relationship is a partition correspondence data table that is pre-calibrated and solidified in the system firmware after the overall device structure design is completed. It is used to establish the association matching rules between each physical slot position of the terminal and the dedicated thermal management area and corresponding heat dissipation device. The mapping relationship is divided according to the chassis airflow layout, the coverage of heat dissipation devices, and the hardware stacking structure, dividing the inside of the whole machine into multiple independent thermal management areas. Each physical slot position uniquely corresponds to a set of target heat dissipation devices. For example, the front slot of the terminal corresponds to the front airflow heat dissipation device, and the rear slot corresponds to the rear airflow heat dissipation device. Different slot positions lock independent heat dissipation execution hardware through the preset thermal management area mapping relationship.

[0087] The target cooling device is an active cooling hardware specifically responsible for the heat dissipation management of an expansion module, matched according to the physical slot location of the expansion module through a preset thermal management area mapping relationship. The target cooling device is an adjustable-speed DC cooling fan built into the edge computing terminal. A single terminal is equipped with multiple cooling fans to cover different chassis areas. Each fan can independently receive system speed adjustment commands and complete adaptive speed adjustment. For example, when the expansion module is connected to the high slot of the terminal, the system matches the cooling fan in the corresponding high area as the target cooling device, and only adjusts the parameters of the fan in that area without changing the operating status of the fans in the low area.

[0088] Speed ​​control parameters are a set of core configuration parameters used by the system to manage the operating conditions of the target heat dissipation device and determine its heat dissipation output capacity. They mainly include key control parameters such as fan initial temperature threshold, fan stepless speed range, and maximum operating speed, which together form a dynamic heat dissipation curve adapted to different heat levels. For example, low-power modules are matched with higher fan initial temperature and lower maximum speed, medium-power modules are matched with conventional speed control parameters, and high-power modules are matched with lower fan initial temperature and higher maximum speed, thus adapting to different levels of heat demand.

[0089] For example, after the edge computing terminal completes the identification of the physical slot location of the expansion module and the determination of its thermal contribution level, it automatically retrieves the preset thermal management area mapping relationship stored in the system firmware. Using the physical slot location of the current expansion module as the retrieval basis, it accurately matches the dedicated target heat dissipation device corresponding to the slot and determines the hardware object to be executed for this heat dissipation control. Based on the determined thermal contribution level of the expansion module, the system retrieves the standard heat dissipation configuration strategy corresponding to the level and adaptively adjusts the speed control parameters of the target heat dissipation device. The system uses a continuous weight adjustment method to complete the smooth update of parameters, avoiding fluctuations in the overall operating noise caused by sudden changes in fan speed. To mitigate system performance fluctuations; when connecting low-power expansion modules, the system raises the startup temperature of the target heat dissipation device and lowers its maximum speed to reduce ineffective heat dissipation power consumption under low load conditions; when connecting medium-power non-volatile memory host controller interface standard solid-state drives or 5G mobile communication modules, the system enables conventional speed control parameters to balance heat dissipation effect and heat dissipation power consumption; when connecting high-power artificial intelligence accelerator cards, the system proactively lowers the fan's initial operating temperature and increases the fan's maximum operating speed, allowing the target heat dissipation device to intervene in heat dissipation in advance and improve its extreme heat dissipation capacity, fully dissipating the dense heat generated by the operation of high-power modules.

[0090] Through the implementation of the above embodiments, the current ambient temperature is obtained, and the available power consumption budget is corrected for temperature compensation according to the preset ambient temperature and budget adjustment curve, so that the processor power consumption budget can adapt to the thermal management requirements under different ambient temperature conditions; further, the upper limit of the processor's operating frequency and the upper limit of the number of parallel processing tasks are adjusted based on the target available power consumption budget, so that the processor's computing resources can be dynamically adjusted according to the actual thermal environment, avoiding instability caused by excessive power consumption in high temperature environments, and fully releasing processing power when temperature conditions permit, thereby improving the stability of the edge computing terminal's operating performance.

[0091] In some embodiments, the aforementioned system power consumption management method may further include: during the operation of the expansion module, acquiring the real-time temperature value and temperature change rate of the expansion module; when the real-time temperature value exceeds a preset warning threshold and the temperature change rate exceeds a preset change rate threshold, applying a temporary power consumption reduction amount to the available power consumption budget, and increasing the target speed in the speed control parameters of the heat dissipation device until the temperature change rate falls back to the preset safe change rate range.

[0092] In some examples, the real-time temperature value is the surface operating temperature of the expansion module at the current moment during operation. It is an absolute temperature parameter that directly characterizes the instantaneous thermal state of the module and can intuitively reflect the current heat accumulation level of the expansion module. The real-time temperature value can be obtained by the terminal system through the onboard temperature acquisition circuit and the temperature measurement unit built into the expansion module through periodic sampling. The system continuously reads the raw temperature data at a millisecond sampling frequency and removes sampling noise through data smoothing processing to ensure that the temperature data is true and valid. For example, under normal load conditions, the real-time temperature value of various expansion modules is maintained below 60 degrees Celsius, while under high load instantaneous impact conditions, the real-time temperature value can rise rapidly to above 70 degrees Celsius.

[0093] The temperature change rate is a parameter characterizing the dynamic trend of temperature change in an expansion module. It is used to quantify the temperature rise of the expansion module in real time per unit time, and can predict the overheating trend of the module temperature rising continuously in advance, thus overcoming the deficiency of a single static temperature sample that can only reflect the instantaneous temperature and cannot predict the temperature rise trend. The temperature change rate can be obtained by the system through differential calculation of real-time temperature values ​​over multiple consecutive frames. The system records the temperature data at consecutive sampling times and calculates the temperature rise per unit time by combining it with a fixed sampling time interval to obtain the real-time temperature change rate. For example, the temperature change rate of the expansion module approaches zero under low and stable load, but will show a significant positive increase when the instantaneous load suddenly increases.

[0094] The preset warning threshold is a critical safety temperature value for extended modules that is pre-defined in the thermal management strategy file. It serves as the criterion for distinguishing between normal and warning temperature conditions of the modules. Based on the hardware specifications and temperature tolerance of various extended modules, the preset warning threshold is lower than the module's hardware heat resistance limit. It is used to identify the risk of excessive temperature rise in advance, allowing sufficient time for system intervention and adjustment. The preset rate of change threshold is a critical standard for the rate of temperature rise preset by the system. It is used to determine whether the extended module exhibits a rapid temperature rise or a trend of continuous deterioration in thermal condition. This threshold is used to screen for sudden abnormal temperature rise conditions, distinguishing between steady-state small temperature fluctuations and rapid temperature rises caused by abnormal loads, preventing the system from erroneously intervening in normal temperature fluctuations. When the measured rate of change is lower than the preset rate of change threshold, it indicates that the module temperature is in a stable fluctuation state and no intervention is required. When the measured rate of change exceeds the preset rate of change threshold, it indicates that the module is accumulating heat rapidly, posing a short-term overheating risk, and requiring the activation of dynamic protection strategies.

[0095] The core execution logic of dynamic thermal intervention in this application embodiment is to apply a temporary power reduction to the available power budget and increase the target speed in the speed control parameters of the heat dissipation device until the temperature change rate falls back to the preset safe change rate range. This strategy is a short-term, reversible dynamic adjustment mechanism, which is different from the conventional steady-state power budget and heat dissipation parameter adjustment. It only takes effect temporarily when the expansion module experiences an abnormally rapid temperature rise. The temporary power reduction is used to moderately constrain the instantaneous computing power output of the processor, reduce the overall power consumption and heat generation pressure of the whole machine, and increase the target speed of the heat dissipation device to enhance the instantaneous heat dissipation capability. The two-way cooperation suppresses the continuous rise of the module temperature. The system continuously monitors the temperature change rate index. When the temperature rise rate falls back to the preset safe change rate range, the temporary power reduction and heat dissipation enhancement strategy is automatically released, and the conventional power budget and heat dissipation speed parameters are restored to avoid performance loss caused by long-term power reduction.

[0096] For example, during the entire normal operation of various expansion modules, the edge computing terminal continuously maintains a temperature monitoring mechanism, constantly collecting real-time temperature values ​​from each connected expansion module and calculating the real-time temperature change rate of the module through continuous temperature data. The system compares the collected and calculated temperature parameters with local preset warning thresholds and preset change rate thresholds in real time. When the real-time temperature value of an expansion module has not yet triggered hardware high-temperature protection but has exceeded the preset warning threshold, and the temperature change rate is simultaneously detected to continuously exceed the preset change rate threshold, the system determines that the current module has a risk of rapid temperature rise and overheating, and immediately initiates a dynamic thermal protection process. The system adds a corresponding temporary power reduction amount to the currently available processor power budget, appropriately tightening the processor's instantaneous power consumption. The system sets an upper limit for power consumption output to suppress the accumulation of heat sources caused by overload of the entire machine's computing power. At the same time, the system actively increases the target speed value inside the speed control parameters of the heat dissipation device to improve the instantaneous heat exchange efficiency of air cooling and accelerate the dissipation of heat on the module surface. The system continuously monitors the temperature change rate index of the expansion module and verifies the temperature rise and fall status in real time until the temperature change rate steadily falls and stabilizes within the preset safe change rate range. After confirming that the module temperature rise trend is stable and the thermal state has returned to the safe range, the temporary power consumption reduction constraint is removed, the standard available power consumption budget of the processor is restored, and the target speed of the heat dissipation device is reduced to the normal matching level. While effectively avoiding the risk of instantaneous thermal runaway, the system maximizes the continuity of edge computing terminal business operation and the integrity of computing power performance.

[0097] Through the implementation of the above embodiments, the corresponding target heat dissipation device is determined according to the physical slot location of the expansion module, and the rotation speed control parameters of the target heat dissipation device are adjusted based on the thermal contribution level of the expansion module, so that the heat dissipation resources can be matched and adjusted for specific heat sources. Compared with the method of controlling the unified operation of the heat dissipation device of the whole machine, it can realize precise heat dissipation control for the area where the expansion module is located, so that the heat dissipation capacity is matched with the actual thermal contribution of the expansion module, reduce the consumption of ineffective heat dissipation resources, and improve the flexibility and efficiency of heat dissipation management of edge computing terminals.

[0098] In some embodiments, the aforementioned application of a temporary power reduction to the available power budget may include: obtaining the thermal time constant of the expansion module, wherein the thermal time constant characterizes the time required for the expansion module to reach a preset proportion of steady-state temperature rise under conditions without active cooling; predicting the temperature prediction value of the expansion module after a preset time step based on the real-time temperature value, the temperature change rate, and the thermal time constant; determining the basic reduction amount based on the difference between the temperature prediction value and a preset dangerous temperature threshold; obtaining the priority information of the parallel processing tasks currently being executed by the processor; determining the reduction exemption amount for the expansion module based on the priority information; superimposing the basic reduction amount and the reduction exemption amount to obtain the temporary power reduction amount, and deducting the temporary power reduction amount from the available power budget.

[0099] In some examples, the thermal time constant is an inherent hardware parameter characterizing the thermal response of an expansion module. Specifically, it refers to the time required for the expansion module to reach a steady-state temperature rise of a preset percentage under conditions without active cooling. It accurately reflects the speed of temperature rise and heat accumulation capacity of the expansion module. Expansion modules of different materials and power consumption specifications exhibit significant differences in thermal time constants. The smaller the thermal time constant value, the faster the module's temperature rises, the more concentrated the heat accumulation, and the higher the risk of thermal runaway. For example, AI accelerator cards have high power density, resulting in smaller thermal time constant values ​​and faster temperature rise response. Non-volatile memory host controller interface standard solid-state drives have relatively larger thermal time constant values, indicating a smoother temperature rise process.

[0100] The preset time step is a fixed calculation time interval used by the system to predict temperature trends. It is a benchmark time dimension for quantifying short-term temperature changes in the future, and is used to ensure the regularity and consistency of temperature prediction. For example, the system can preset the time step to 100 milliseconds to predict the temperature change trend of the module in a short time interval, ensuring the real-time performance and accuracy of temperature prediction, and capturing potential overheating risks in advance.

[0101] The predicted temperature value is the estimated temperature of the module after a preset time step, calculated by the system based on the current thermal state and inherent thermal characteristics of the extended module. This parameter can predict the trend of module temperature rise in advance, compensate for the lag in temperature monitoring, and enable early intervention for overheating risks. It can integrate three types of parameters: real-time temperature value, real-time temperature change rate, and module thermal time constant, and calculate the predicted module temperature value at future moments through a preset heat conduction extrapolation algorithm. For example, when the module's real-time temperature is high, the temperature change rate is rising rapidly in a positive direction, and the thermal time constant is small, the calculated predicted temperature value will be significantly higher than the current real-time temperature value, indicating that the module has a risk of rapid overheating.

[0102] The preset dangerous temperature threshold is a critical temperature value that is pre-fixed in the system to determine whether the expansion module has entered a high-risk overheating state. It is the criterion for distinguishing between normal temperature rise, warning temperature rise and dangerous temperature rise of the module. The preset dangerous temperature threshold is based on the hardware tolerance limit temperature calibration of the expansion module. If the value is higher than the aforementioned preset warning threshold, it means that the module is about to reach the upper limit of the hardware safe operation temperature. If not intervened in time, it will trigger hardware protection, forced frequency reduction or even equipment shutdown failure.

[0103] The base reduction is the initial temporary power consumption reduction determined by the system based on the module's future overheating risk level. It forms the basis for the temporary power consumption reduction. The system can calculate the difference between the predicted temperature value and the preset dangerous temperature threshold. The smaller the difference, the closer the module is to the dangerous temperature, and the larger the corresponding base reduction value. Conversely, the base reduction value is appropriately reduced. For example, if the predicted temperature value is close to the preset dangerous temperature threshold, the system outputs a large base reduction to quickly suppress the combined power consumption and heat source. If the predicted temperature value only slightly exceeds the warning threshold and is far from the dangerous temperature threshold, the system outputs a small base reduction to reduce performance loss while controlling the temperature.

[0104] Priority information is the level identifier information of various parallel processing tasks currently being executed by the processor. It is used to distinguish the importance of different edge services and provide a basis for determining differentiated exemptions for power consumption reduction. The business tasks run by the edge computing terminal include different types such as core inference tasks, data acquisition tasks, and background operation and maintenance tasks. Each type of task corresponds to a preset fixed priority level. The system can traverse the current processor concurrent task queue in real time, read the preset priority identifier of each task, and complete the summary statistics of task priority information. For example, real-time artificial intelligence inference and core data processing tasks are high-priority tasks, while background log organization and idle inspection tasks are low-priority tasks.

[0105] The reduction exemption amount is a computing power limit determined by the system based on task priority information that does not require power reduction. It is used to protect the stable computing power output of high-priority core businesses and avoid damage to core business performance caused by a one-size-fits-all power reduction. The system can quantify the exemptable power consumption limit by calling a preset priority-exempt computing power mapping table based on the number of current high-priority tasks, computing load and computing power occupancy ratio. For example, when the processor is executing multiple high-priority artificial intelligence inference tasks, the system can match a higher reduction exemption amount to reduce the overall power reduction and ensure the stable operation of core businesses.

[0106] The calculation and execution logic of the refined dynamic power consumption constraint in this application embodiment is to superimpose the basic reduction amount and the reduction exemption amount to obtain the temporary power consumption reduction amount, and then deduct the temporary power consumption reduction amount from the available power consumption budget. The superposition operation is a positive difference adaptation operation, which uses the basic reduction amount as the benchmark reduction range required for cooling, deducts the reduction exemption amount corresponding to high priority tasks, and finally obtains the optimal temporary power consumption reduction amount that takes into account both cooling requirements and service protection. The system deducts the calculated temporary power consumption reduction amount from the current available power consumption budget of the processor in real time to form a short-term effective temporary power consumption constraint upper limit, thereby achieving a two-way balance between dynamic temperature control and service protection.

[0107] For example, after the edge computing terminal detects an abnormal temperature rise rate and triggers a warning from an expansion module, it immediately initiates a refined temporary power reduction calculation process. First, based on the hardware identifier of the currently connected abnormally overheating expansion module, the system retrieves the corresponding module's thermal time constant pre-stored in the firmware to accurately obtain the module's inherent thermal response characteristic parameters. Simultaneously, the system collects the module's real-time temperature value and real-time temperature change rate, and, combined with a preset time step, calculates the predicted temperature value of the module at the next moment using a thermal trend extrapolation algorithm. The system compares the predicted temperature value with a preset dangerous temperature threshold, and, based on the degree to which the temperature approaches the dangerous threshold, quantifies and calculates the basic reduction amount adapted to the current overheating risk. On this basis, the system reads the priority information of the processor's current parallel task queue in real time to identify the currently running tasks. For high-priority core services and low-priority backend services, the corresponding reduction exemption amount is matched according to the computing power usage of high-priority tasks. The system adds the basic reduction amount and the reduction exemption amount, removes the exemption computing power quota corresponding to the core service, and obtains the final accurate temporary power reduction amount. Finally, the system deducts the temporary power reduction amount from the current available power budget of the processor to form a short-term processor power constraint, which, together with the operation of increasing the speed of the heat dissipation device, suppresses the rapid temperature rise of the module. When the temperature change rate of the module falls back to the safe range, the system automatically cancels this temporary power reduction and restores the normal available power budget of the processor. This effectively avoids the risk of overheating damage to the expansion module and preserves the computing performance of the edge core services to the greatest extent, realizing the coordinated optimization of terminal thermal management and service stability.

[0108] Through the implementation of the above embodiments, real-time temperature values ​​and temperature change rates are acquired during the operation of the expansion module. When the temperature exceeds the warning threshold and the temperature change rate exceeds the set conditions, the available power consumption budget of the processor is temporarily reduced, and the rotation speed of the target heat dissipation device is increased. This enables the system to actively adjust according to the real-time thermal change trend of the expansion module. It can take power limiting and heat dissipation enhancement measures in advance during the rapid temperature rise of the expansion module, avoiding the system from triggering protection mechanisms or performance abnormalities due to continuous temperature rise, thereby improving the thermal safety and stability of the edge computing terminal during operation.

[0109] Furthermore, as an implementation of the foregoing method embodiments, this application also provides a system power management device for implementing the foregoing method embodiments. This device embodiment corresponds to the foregoing method embodiments. For ease of reading, this system power management device embodiment will not repeat the details of the foregoing method embodiments one by one, but it should be understood that the device in this application embodiment can correspondingly implement all the contents of the foregoing method embodiments. For example... Figure 2 As shown, the power management device 20 of the system includes: an information acquisition unit 201, a level determination unit 202, a budget determination unit 203, a first parameter adjustment unit 204, and a second parameter adjustment unit 205. The information acquisition unit 201 is used to acquire the thermal characteristic identification information of the aforementioned expansion module in response to the detection of an access event of the expansion module. The level determination unit 202 is used to determine the thermal contribution level corresponding to the expansion module based on preset thermal contribution level comparison information and thermal characteristic identification information. The budget determination unit 203 is used to determine the available power consumption budget of the processor based on the thermal contribution level and a preset total system power consumption constraint. The first parameter adjustment unit 204 is used to adjust the processor's operating state parameters based on the available power consumption budget; the operating state parameters may include at least one of an upper limit for operating frequency and an upper limit for the number of parallel processing tasks. The second parameter adjustment unit 205 is used to adjust the control parameters of the heat dissipation device based on the thermal contribution level.

[0110] In some embodiments, the information acquisition unit 201 is further configured to obtain the manufacturer identifier, device identifier, and physical slot location identifier of the expansion module by reading the non-volatile storage area of ​​the expansion module.

[0111] In some embodiments, the rating determination unit 202 is further configured to obtain the initial rated power consumption parameters of the expansion module from a preset power consumption database based on the manufacturer identifier and the device identifier; correct the initial rated power consumption parameters based on the physical slot location identifier and the preset location thermal influence coefficient to obtain the corrected rated power consumption parameters; and compare the corrected rated power consumption parameters with at least one preset power consumption threshold to determine the thermal contribution level corresponding to the expansion module.

[0112] In some embodiments, the budget determination unit 203 is further configured to obtain the corresponding basic safety margin value from the preset level margin mapping relationship based on the thermal contribution level; accumulate the corrected rated power consumption of all currently connected expansion modules to obtain the total expansion power consumption value; determine the interaction heat loss between modules based on the thermal coupling coefficient between each expansion module, so as to include the interaction heat loss in the total expansion power consumption value; and subtract the total expansion power consumption value and the basic safety margin value from the preset total power consumption upper limit value to obtain the available power consumption budget.

[0113] In some embodiments, the first parameter adjustment unit 204 is further configured to obtain the current ambient temperature; perform temperature compensation correction on the available power consumption budget based on a preset ambient temperature-budget adjustment curve to obtain a target available power consumption budget; and adjust the upper limit of the processor's operating frequency and / or the upper limit of the number of parallel processing tasks based on the target available power consumption budget.

[0114] In some embodiments, the second parameter adjustment unit 205 is further configured to determine the target heat dissipation device corresponding to the expansion module from a preset thermal management area mapping relationship based on the physical slot location identifier; and adjust the rotation speed control parameter of the target heat dissipation device based on the thermal contribution level so that the heat dissipation capacity of the target heat dissipation device matches the thermal contribution of the expansion module.

[0115] In some embodiments, the system power management device 20 further includes a third parameter adjustment unit, which is used to obtain the real-time temperature value and temperature change rate of the expansion module during the operation of the expansion module; when the real-time temperature value exceeds a preset warning threshold and the temperature change rate exceeds a preset change rate threshold, a temporary power reduction amount is applied to the available power budget, and the target speed in the speed control parameter of the heat dissipation device is increased until the temperature change rate falls back to the preset safe change rate range.

[0116] In some embodiments, the third parameter adjustment unit is further configured to obtain the thermal time constant of the expansion module, wherein the thermal time constant characterizes the time required for the temperature change of the expansion module to reach a preset proportion of steady-state temperature rise under conditions without active heat dissipation; predict the temperature prediction value of the expansion module after a preset time step based on the real-time temperature value, the temperature change rate, and the thermal time constant; determine the basic reduction amount based on the difference between the temperature prediction value and the preset dangerous temperature threshold; obtain the priority information of the parallel processing task currently being executed by the processor; determine the reduction exemption amount of the expansion module based on the priority information; superimpose the basic reduction amount and the reduction exemption amount to obtain the temporary power consumption reduction amount, and deduct the temporary power consumption reduction amount from the available power consumption budget.

[0117] This application also provides a computer-readable storage medium storing computer-executable instructions or computer programs, which, when executed by a processor, will cause the processor to perform any step of the system power management method provided in this application.

[0118] In some embodiments, the computer-readable storage medium may be a random access memory (RAM), a read-only memory (ROM), flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM); or it may be a variety of devices that include one or any combination of the above-mentioned memories.

[0119] In some embodiments, computer-executable instructions may take the form of programs, software, software modules, scripts, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as stand-alone programs or as modules, components, subroutines, or other units suitable for use in a computing environment.

[0120] In some embodiments, computer-executable instructions may, but do not necessarily, correspond to files in a file system, and may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple co-located files (e.g., files that store one or more modules, subroutines, or code sections).

[0121] In some embodiments, computer-executable instructions may be deployed to execute on an electronic device, or on multiple electronic devices located at one location, or on multiple electronic devices distributed across multiple locations and interconnected via a communication network.

[0122] like Figure 3 As shown, this application also provides an electronic device 30, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor. When the processor 320 executes the computer program 311, it implements any step of the above-described system power management method.

[0123] This application also provides a computer program product comprising a computer program or computer-executable instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer program or computer-executable instructions from the computer-readable storage medium and executes the computer program or computer-executable instructions, causing the electronic device to perform any step of the system power management method described above.

[0124] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A system power consumption management method, characterized in that, include: In response to the detection of an access event of an expansion module, the thermal characteristic identification information of the expansion module is obtained; Based on the preset thermal contribution level comparison information and the thermal characteristic identification information, the thermal contribution level corresponding to the extended module is determined; Based on the thermal contribution level and the preset total system power consumption constraint, the available power consumption budget of the processor is determined; Based on the available power consumption budget, the processor's operating state parameters are adjusted, including at least one of an upper limit on the operating frequency and an upper limit on the number of parallel processing tasks; Based on the aforementioned thermal contribution level, adjust the control parameters of the heat dissipation device.

2. The system power consumption management method according to claim 1, characterized in that, The step of obtaining the thermal characteristic identification information of the extension module includes: By reading the non-volatile storage area of ​​the expansion module, the manufacturer identifier, device identifier, and physical slot location identifier of the expansion module can be obtained.

3. The system power consumption management method according to claim 2, characterized in that, The step of determining the thermal contribution level corresponding to the extended module based on the preset thermal contribution level comparison information and the thermal characteristic identification information includes: Based on the manufacturer identifier and the device identifier, the initial rated power consumption parameters of the expansion module are obtained from the preset power consumption database; Based on the physical slot location identifier and the preset location thermal influence coefficient, the initial rated power consumption parameter is corrected to obtain the corrected rated power consumption parameter. The corrected rated power consumption parameter is compared with at least one preset power consumption threshold to determine the thermal contribution level of the expansion module.

4. The system power consumption management method according to claim 3, characterized in that, The process of determining the processor's available power budget based on the thermal contribution level and the preset total system power consumption constraint includes: Based on the thermal contribution level, the corresponding basic safety margin value is obtained from the preset level margin mapping relationship; The total extended power consumption is obtained by summing the corrected rated power consumption of all currently connected expansion modules. The interaction heat loss between modules is determined based on the thermal coupling coefficient between each expansion module, so that the interaction heat loss is included in the total expansion power consumption value. The available power budget is obtained by subtracting the total extended power consumption value and the basic safety margin value from the preset total power consumption upper limit value.

5. The system power consumption management method according to claim 4, characterized in that, The step of adjusting the processor's operating state parameters based on the available power consumption budget includes: Get the current ambient temperature; Based on the preset ambient temperature-budget adjustment curve, the available power consumption budget is corrected by temperature compensation to obtain the target available power consumption budget; Based on the target available power consumption budget, adjust the upper limit of the processor's operating frequency and / or the upper limit of the number of parallel processing tasks.

6. The system power consumption management method according to claim 2, characterized in that, The adjustment of control parameters of the heat dissipation device based on the thermal contribution level includes: Based on the physical slot location identifier, the target heat dissipation device corresponding to the expansion module is determined from the preset thermal management area mapping relationship; Based on the thermal contribution level, the rotation speed control parameters of the target heat dissipation device are adjusted so that the heat dissipation capacity of the target heat dissipation device matches the thermal contribution of the expansion module.

7. The system power consumption management method according to claim 1, characterized in that, The system power consumption management method also includes: During the operation of the expansion module, the real-time temperature value and temperature change rate of the expansion module are acquired; When the real-time temperature value exceeds a preset warning threshold and the temperature change rate exceeds a preset change rate threshold, a temporary power reduction is applied to the available power consumption budget, and the target speed in the speed control parameters of the heat dissipation device is increased until the temperature change rate falls back to the preset safe change rate range.

8. The system power consumption management method according to claim 7, characterized in that, Applying a temporary power reduction to the available power budget includes: The thermal time constant of the expansion module is obtained, wherein the thermal time constant represents the time required for the temperature change of the expansion module to reach a preset proportion of steady-state temperature rise under conditions without active heat dissipation. Based on the real-time temperature value, the temperature change rate, and the thermal time constant, the temperature prediction value of the extended module is predicted after a preset time step. The basic reduction amount is determined based on the difference between the predicted temperature value and the preset dangerous temperature threshold. Obtain the priority information of the parallel processing task currently being executed by the processor; Based on the priority information, the reduction exemption amount of the extension module is determined; The basic reduction amount is added to the reduction exemption amount to obtain the temporary power consumption reduction amount, and the temporary power consumption reduction amount is deducted from the available power consumption budget.

9. A system power consumption management device, characterized in that, include: The information acquisition unit is used to acquire the thermal characteristic identification information of the expansion module in response to the detection of the access event of the expansion module; The grade determination unit is used to determine the thermal contribution level corresponding to the extended module based on the preset thermal contribution level comparison information and the thermal characteristic identification information; The budget determination unit is used to determine the available power consumption budget of the processor based on the thermal contribution level and the preset total system power consumption constraint. The first parameter adjustment unit is used to adjust the processor's operating state parameters based on the available power consumption budget. The operating state parameters include at least one of an upper limit for operating frequency and an upper limit for the number of parallel processing tasks. The second parameter adjustment unit is used to adjust the control parameters of the heat dissipation device based on the heat contribution level.

10. An electronic device, comprising: A memory and a processor, characterized in that the processor, when executing a computer program stored in the memory, implements the steps of the system power management method as described in any one of claims 1 to 8.

Images