Load-aware server dynamic power adjustment method and electronic device

Abstract

The application discloses a load-aware server dynamic power consumption adjustment method and electronic equipment, and relates to the technical field of computer application, which comprises the following steps: obtaining the running state data of a server, determining the first power consumption adjustment mode of the server according to the running state data, further obtaining user behavior data, and generating the second power consumption adjustment mode adapting to the actual demand of the user in combination with the user behavior data. Finally, the hardware parameters in the server are dynamically adjusted according to the second power consumption adjustment mode, so that automatic, intelligent and fine adjustment of power consumption can be realized. Thus, the problem that the power consumption adjustment method of the server in the prior art cannot dynamically adjust the power consumption distribution according to the real-time load state and user behavior, and the power consumption strategy is generally inconsistent with the actual demand of the user, is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer application, and in particular, to a server dynamic power consumption adjustment method based on load sensing and an electronic device. BACKGROUND

[0002] In scenarios such as data centers, cloud computing platforms, enterprise servers, and high-performance computing (HPC), the running load of a server has significant dynamic fluctuations.

[0003] Traditional server power consumption adjustment mechanisms rely on static strategies (such as fixed performance modes or energy-saving modes) and cannot dynamically adjust power consumption allocation according to real-time load states. Moreover, existing server power consumption adjustment methods cannot adjust power consumption allocation according to user behavior, resulting in power consumption strategies that are often inconsistent with actual user needs. SUMMARY

[0004] The present application provides a server dynamic power consumption adjustment method based on load sensing and an electronic device to at least solve the problem that the power consumption adjustment method of the server in the related art cannot dynamically adjust power consumption allocation according to real-time load states and user behavior, resulting in power consumption strategies that are often inconsistent with actual user needs.

[0005] The present application provides a server dynamic power consumption adjustment method based on load sensing, which comprises:

[0006] Obtaining running state data of the server, the running state data comprising processor utilization, temperature sensor data, server wake-up trigger sources, and power supply states;

[0007] Determining a first power consumption adjustment mode of the server according to the processor utilization, the temperature sensor data, the server wake-up trigger sources, and the power supply states;

[0008] Obtaining user behavior data of the server, the user behavior data comprising user-pre-set power supply modes and pre-stored user-preferred usage modes;

[0009] Obtaining a second power consumption adjustment mode of the server according to the first power consumption adjustment mode, the user-pre-set power supply modes, and the pre-stored user-preferred usage modes;

[0010] Adjusting hardware parameters in the server according to the second power consumption adjustment mode to achieve automatic and intelligent adjustment of server power consumption.

[0011] The present application also provides an electronic device comprising a memory for storing a computer program and a processor for executing the computer program to implement the steps of any of the above server dynamic power consumption adjustment methods based on load sensing.

[0012] By the load-aware server dynamic power consumption adjustment method and the electronic device provided in the application, the running state data of the server is acquired, and a first power consumption adjustment mode of the server is determined accordingly. Then, the user behavior data is further acquired, and a second power consumption adjustment mode adapting to the actual demand of the user is generated in combination with the user behavior data. Finally, the hardware parameters in the server are dynamically adjusted according to the second power consumption adjustment mode, so that automatic, intelligent and fine adjustment of power consumption can be realized. Thus, the problem that the power consumption adjustment method of the server in the related art cannot dynamically adjust the power consumption distribution according to the real-time load state and user behavior, resulting in that the power consumption strategy is generally inconsistent with the actual demand of the user, is solved. BRIEF DESCRIPTION OF DRAWINGS

[0013] In order to more clearly illustrate the embodiments of the present application, the drawings needed in the embodiments will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0014] Figure 1 The flowchart of the load-aware server dynamic power consumption adjustment method provided in the embodiments of the present application is shown.

[0015] Figure 2 The structural diagram of the load-aware server dynamic power consumption adjustment device provided in the embodiments of the present application is shown.

[0016] Figure 3 The structural diagram of the electronic device provided in the present application is shown.

[0017] Among the above drawings, the following reference signs are included:

[0018] 210-acquisition module; 220-data processing module; 230-fusion processing module; 240-adjustment module; 310-processor; 320-memory; 330-communication component. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, each other embodiment obtained by those skilled in the art without creative labor belongs to the protection scope of the present application.

[0020] It should be noted that in the description of the present application, the term "comprising" or "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that processes, methods, articles or devices including a series of elements not only include those elements, but also include other elements not explicitly listed, or further include elements inherent in such processes, methods, articles or devices. The terms "first", "second" and the like in the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. Those skilled in the art can understand that "first", "second" and the like do not limit the number and execution order, and "first", "second" and the like do not limit the difference. It should be noted that in the embodiments of the present application, the words "exemplary" or "for example" are used to indicate an example, illustration or description. Any embodiment or design scheme described as "exemplary" or "for example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the use of "exemplary" or "for example" is intended to present the relevant concept in a specific manner. In the embodiments of the present application, "at least one" means one or more, and "multiple" means two or more.

[0021] It should be noted that "at the time" in the embodiments of the present application can be at the moment when a certain condition occurs, or within a period of time after a certain condition occurs, which is not limited in the embodiments of the present application. In addition, the load-aware server dynamic power consumption adjustment method provided in the embodiments of the present application is only as an example, and the load-aware server dynamic power consumption adjustment method can also include more or less content.

[0022] In typical application scenarios such as data centers, cloud computing platforms, enterprise servers and HPCs, the running load of servers usually presents significant dynamic fluctuation characteristics. For example, the virtual machine resource allocation demand in a cloud computing platform changes in real time; the load difference of an enterprise server between peak and trough periods can be several times, and the server system needs to handle a large number of concurrent requests during the peak period, while it may be in light load or idle state for a long time during the trough period; the HPC cluster needs to switch between computing-intensive and low-power standby states frequently when performing large-scale scientific computing tasks.

[0023] However, the power consumption adjustment mechanism commonly used by existing servers is mostly based on static configuration strategies, such as pre-set high-performance mode or energy-saving mode, whose power consumption and performance parameters are fixed after the server system is started and cannot be adaptively adjusted according to real-time CPU utilization, server system temperature, wake-up source, power state and other operating environment factors. Such static strategies often lead to low energy efficiency when the load changes dramatically: the server still maintains high power consumption when running light load, causing energy waste; and the performance is limited when there is a sudden high load, affecting the quality of service. For example, when a user executes a low-load task at night, the server may still maintain a high power consumption state, causing energy waste; and in a sudden high-load scenario, the power consumption strategy fails to respond in time, which may cause performance bottlenecks or instability of the server system.

[0024] In addition, current mainstream power consumption management schemes generally lack the ability to perceive user behavior characteristics and cannot combine user pre-set power modes and pre-stored user preference usage patterns for personalized power consumption scheduling. This makes the power consumption strategy executed by the server system often inconsistent with the actual use intention or business demand of the user, making it difficult to strike a reasonable balance between performance guarantee and energy efficiency optimization.

[0025] Therefore, there is an urgent need for a dynamic power consumption adjustment mechanism that can integrate real-time server system load state and user behavior characteristics to achieve more intelligent and accurate server energy efficiency management and achieve a balance between energy efficiency optimization and user experience improvement.

[0026] Based on this, the embodiments of the present application propose a load-aware server dynamic power consumption adjustment method and electronic device, which can be used in the field of computer application, and aims to solve the above technical problems. By obtaining the running state data of the server, the first power consumption adjustment mode of the server is determined, and then the user behavior data is combined to generate a second power consumption adjustment mode that adapts to the actual demand of the user. Finally, the hardware parameters in the server are dynamically adjusted according to the second power consumption adjustment mode, so as to realize automatic, intelligent and fine adjustment of power consumption.

[0027] The technical solutions of the present application and how the technical solutions of the present application solve the above technical problems will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of the present application will be described below with reference to the accompanying drawings.

[0028] In order for those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0029] Figure 1A flowchart of a load-aware server dynamic power consumption adjustment method provided by an embodiment of the present application is shown in Figure 1 As shown in the flowchart, an embodiment of the present application provides a load-aware server dynamic power consumption adjustment method, which is described in detail as follows.

[0030] S101, obtaining running state data of the server.

[0031] In an embodiment of the present application, a load-aware server dynamic power consumption adjustment method can be implemented by a Basic Input / Output System (BIOS) or a Unified Extensible Firmware Interface (UEFI). For the convenience of description, the execution subject of a load-aware server dynamic power consumption adjustment method is described as firmware below.

[0032] Specifically, the running state data includes processor utilization, temperature sensor data, server wake-up trigger source and power state. A load-aware server dynamic power consumption adjustment method can be executed by the firmware of the server in the server system startup phase or the Runtime Services phase.

[0033] Specifically, the running state data of the server can be collected in real time by accessing the underlying hardware interface. For example, the current processor utilization can be obtained by reading the processor performance monitoring register. The temperature sensor data of each key area (such as the processor, power management circuit, and memory slot) on the motherboard can be read through a Super Input / Output (Super I / O) or a Platform Environment Control Interface (PECI).

[0034] The wake-up trigger source of the server can be identified through a Real-Time Clock (RTC) wake-up state register, including an RTC Alarm, a Wake-on-LAN (WoL) or a PowerButton Event.

[0035] The current power state can be determined as Alternating Current (AC) power supply or Battery power supply and the battery remaining power information can be obtained through an Advanced Configuration and Power Interface (ACPI) power state object or a Power Management Bus (PMBus) interface.

[0036] S102, determining a first power consumption adjustment mode of the server according to the processor utilization rate, the temperature sensor data, the wake-up trigger source of the server and the power state.

[0037] Specifically, the running state data can be taken as input to query a decision rule table pre-stored in a non-volatile random-access memory (NVRAM) in a firmware non-volatile storage area, such as a serial peripheral interface flash (SPI Flash). The decision rule table can be indexed by a multi-dimensional condition combination (such as a temperature interval, a power supply type, a wake-up source type and a processor utilization rate threshold) and mapped to a predefined power consumption adjustment mode, including an overheat protection mode, an emergency energy saving mode, a normal energy saving mode, a balance mode or a rapid mode. The corresponding first power consumption adjustment mode can be output according to a matching result.

[0038] S103, obtaining user behavior data of the server.

[0039] In the embodiment of the application, the user behavior data includes a user preset power supply mode and a pre-stored user preference use mode.

[0040] Specifically, the user pre-configured behavior data can be read from a UEFI variable storage space (UEFI Variable Services), a complementary metal-oxide-semiconductor (CMOS) or an NVRAM.

[0041] Specifically, the user pre-configured behavior data includes a user preset power supply mode, such as a "extreme performance", a "default balance" or a "maximum energy saving" global strategy selected through a basic input output system setup interface (BIOS Setup).

[0042] and pre-stored user preference usage pattern. The user preference usage pattern can be automatically learned by firmware background service based on historical running logs (such as high load period distribution in the past 7 days, commonly used wake-up mode, etc.) and persistently stored, used to represent the user's tendency to performance response speed and energy efficiency.

[0043] S104, obtaining a second power consumption adjustment mode of the server according to the first power consumption adjustment mode, the user preset power mode, and the pre-stored user preference usage pattern.

[0044] Specifically, after obtaining the user behavior data, the fusion decision logic can be further executed. Specifically, if the first power consumption adjustment mode is the overheat protection mode or the emergency energy saving mode (i.e. involving server system safety or survivability constraints), it is directly taken as the final second power consumption adjustment mode, ignoring the user settings.

[0045] Otherwise, the first power consumption adjustment mode, the user preset power mode, and the user preference usage pattern are input to the built-in fusion decision engine (which can be implemented based on priority override, weighted scoring, or finite state machine) of the firmware, the server system state and user intention are comprehensively considered, and the coordinated second power consumption adjustment mode is output.

[0046] S105, adjusting the hardware parameters in the server according to the second power consumption adjustment mode, to realize automatic intelligent adjustment of the server power consumption.

[0047] Specifically, according to the second power consumption adjustment mode, the underlying hardware abstraction layer (Hardware Abstraction Layer, HAL) or ACPI control method can be called to dynamically adjust the following hardware parameters:

[0048] Set the performance state (P-state) and idle state (C-state) of the processor; control the start and stop of unused processor cores (Core Parking); adjust the memory working frequency (DRAM Frequency Scaling) or enter the self-refresh energy saving mode; configure the link power management state of the Peripheral Component Interconnect Express (PCIe) device; adjust the fan speed curve or enable active cooling strategy; on platforms supporting Dynamic Voltage and Frequency Scaling (DVFS), adjust the core voltage and frequency.

[0049] The embodiment of the application provides a load-aware server dynamic power consumption adjustment method, which has the following advantages: the dynamic power consumption adjustment logic is sunk to the BIOS or UEFI firmware layer, and thus can take effect earlier than the operating system (OS) starts, avoids OS layer power management delay or strategy conflict, can obtain more accurate and low-delay state data by directly accessing hardware registers, is compatible with ACPI and other standards, is convenient for cooperation with upper-layer management systems (such as a baseboard management controller and a cloud platform scheduler), and supports an operating system-free scene (such as firmware updating and remote diagnosis) and can improve management stability.

[0050] In a possible implementation, the running state data of the server is acquired, including: determining the processor utilization by reading a first register; acquiring temperature sensor data and power state through a super input-output integrated circuit; and acquiring the wakeup trigger source of the server through a real-time clock interface.

[0051] Specifically, the instruction execution period and the reference clock period of the processor in a preset time window (for example, 10 seconds) can be acquired by reading a first register (for example, a performance monitoring register) in the processor, and then the current processor utilization is obtained.

[0052] The temperature sensor data (for example, the temperature near the processor, the memory or the power supply module) of the key area of the server system and the power state of the whole machine, including whether the whole machine is powered by alternating current, whether the whole machine is in a battery power supply mode and the battery remaining capacity information, can be acquired by accessing the temperature sensor and the power state detection circuit integrated on the motherboard through the super input-output integrated circuit.

[0053] The wakeup trigger source of the whole machine that triggers the server to recover from the low-power-consumption state to the running state can be identified by reading the RTC wakeup event register or the server system wakeup source identifier defined by ACPI through the real-time clock interface, and the wakeup trigger source includes but is not limited to: a timing wakeup, a network wakeup or a power key wakeup.

[0054] Through the above method, the embodiment of the application can acquire multi-dimensional running state data in real time through the firmware layer, improve the perception accuracy of the server system to the load, the temperature, the power supply and the wakeup scene, support dynamic, accurate and low-delay power consumption adjustment, and thus the intelligence and the scene adaptability of the energy efficiency management can be significantly enhanced.

[0055] In a possible implementation, the user behavior data of the server is acquired, including: acquiring the user preset power supply mode and the pre-stored user preference use mode through a non-volatile random access memory.

[0056] Specifically, by accessing the non-volatile random access memory (NVRAM) in the firmware environment, the power mode (such as "high performance mode", "balanced mode" or "energy saving mode") pre-configured by the user in the basic input output system (BIOS) or unified extensible firmware interface (UEFI) setting interface is read; at the same time, the pre-stored user preference usage mode is loaded from the NVRAM, which can be generated by the firmware background learning module based on historical running data (such as commonly used wake-up mode, typical load period, performance sensitivity, etc.) and persistently stored, and is used to represent the long-term use tendency of the user to the performance and energy efficiency of the server system.

[0057] The above-mentioned user behavior data can provide key inputs for subsequent fusion of server system state and user intent, and generation of personalized power consumption strategy.

[0058] Through the above method, the power consumption adjustment strategy can combine the actual use habit and configuration intention of the user, improve the energy efficiency adaptability while ensuring the performance of the server system, so that the problem that the traditional static strategy is out of touch with the user's demand can be avoided.

[0059] In a possible implementation, the first power consumption adjustment mode of the server is determined according to the processor utilization rate, the temperature sensor data, the wake-up trigger source of the server and the power state, including: according to the processor utilization rate, the temperature sensor data, the wake-up trigger source of the server and the power state, matching the corresponding third power consumption adjustment mode from the pre-stored decision rule table; the third power consumption adjustment mode is taken as the first power consumption adjustment mode of the server.

[0060] Specifically, first, the real-time collected running state data can be normalized, including: dividing the processor utilization rate into multiple preset intervals (such as low, medium and high load), comparing the temperature sensor data with the pre-set first temperature threshold and second temperature threshold to determine the thermal state, identifying the specific type of the wake-up trigger source (such as timed wake-up, network wake-up or power key wake-up), and confirming the current power state as AC power supply or battery power supply.

[0061] Subsequently, based on the above-mentioned multi-dimensional state combination, the pre-stored decision rule table in the firmware is queried. The decision rule table can store the mapping relationship between different running scenarios and power consumption adjustment modes in the form of condition and action, for example, "high temperature and any wake-up source" corresponds to the overheat protection mode, "battery power supply, low power and any load" corresponds to the emergency energy saving mode, etc.

[0062] The corresponding third power consumption adjustment mode (such as an overheat protection mode, an emergency energy saving mode, a normal energy saving mode, a balance mode, or a rapid mode) is retrieved from the decision rule table by accurately matching the current state combination.

[0063] Finally, the third power consumption adjustment mode is directly used as the first power consumption adjustment mode of the server to generate a final regulation strategy by subsequent fusion with user behavior data. This process realizes automatic and rule-driven preliminary power consumption decision based on the real-time state of the server system, laying a foundation for intelligent energy efficiency management.

[0064] In a possible implementation, the wake-up trigger sources include a timing wake-up, a network wake-up, and a power key wake-up; and the corresponding third power consumption adjustment mode is matched from the pre-stored decision rule table according to the processor utilization, the temperature sensor data, the wake-up trigger source, and the power state of the server, including: when the temperature sensor data indicates that the temperature is greater than or equal to a first temperature threshold, the third power consumption adjustment mode of the server is determined as the overheat protection mode according to the pre-stored decision rule table; when the power state is battery power supply and the battery power is lower than a preset power threshold, the third power consumption adjustment mode of the server is determined as the emergency energy saving mode according to the pre-stored decision rule table; when the wake-up trigger source is the power key wake-up, the temperature sensor data indicates that the temperature is less than the first temperature threshold and greater than a second temperature threshold, the power state is alternating current power supply, and the processor utilization is less than a first utilization threshold and greater than a second utilization threshold, the third power consumption adjustment mode of the server is determined as the balance mode according to the pre-stored decision rule table; when the wake-up trigger source is the network wake-up, the temperature sensor data indicates that the temperature is less than or equal to the second temperature threshold, the power state is alternating current power supply, and the processor utilization is greater than or equal to the first utilization threshold, the third power consumption adjustment mode of the server is determined as the rapid mode according to the pre-stored decision rule table; and when the wake-up trigger source is the timing wake-up, the temperature sensor data indicates that the temperature is less than or equal to the second temperature threshold, the power state is alternating current power supply, and the processor utilization is less than or equal to the second utilization threshold, the third power consumption adjustment mode of the server is determined as the normal energy saving mode according to the pre-stored decision rule table.

[0065] Specifically, the wake-up trigger sources of the server can include three typical scenarios of a timing wake-up, a network wake-up, and a power key wake-up. Based on these wake-up sources, and in combination with the processor utilization, the temperature sensor data, and the power state, the corresponding third power consumption adjustment mode can be matched from the multi-dimensional decision rule table pre-stored in the firmware, and the specific rules are as follows:

[0066] When the temperature sensor data indicates that the current server system temperature is greater than or equal to the first temperature threshold (e.g. 85°C), regardless of other conditions, it is determined that there is an overheat risk, and the third power consumption adjustment mode can be set to the overheat protection mode according to the decision rule table to forcibly limit performance and enhance heat dissipation, ensuring hardware safety.

[0067] When the power supply state is battery power supply and the remaining battery power is lower than the preset power threshold (e.g. 10%), the energy emergency state is entered, and the third power consumption adjustment mode can be set to the emergency energy saving mode according to the decision rule table to minimize power consumption and prolong the available running time.

[0068] When the wake-up trigger source is the power key wake-up, the temperature is between the first temperature threshold and the second temperature threshold (e.g. 60°C-85°C), the power supply state is AC power supply, and the processor utilization rate is at a medium level (i.e. less than the first utilization threshold such as 70% but greater than the second utilization threshold such as 30%), it is determined that it is a regular interactive use scenario, and the third power consumption adjustment mode can be set to the balanced mode according to the decision rule table to achieve a compromise between performance response and energy efficiency.

[0069] When the wake-up trigger source is network wake-up (indicating that there may be a remote management or service request), the temperature is not higher than the second temperature threshold (e.g. ≤ 60°C), the power supply is AC power supply, and the processor utilization rate is high (≥ the first utilization threshold, such as ≥ 70%), it is determined that high-load tasks will be processed, and the third power consumption adjustment mode can be set to the urgent mode according to the decision rule table to release maximum computing performance and ensure service response quality.

[0070] When the wake-up trigger source is a timed wake-up (commonly used for scheduled tasks or periodic maintenance), the temperature is low (≤ the second temperature threshold, such as ≤ 60°C), the power supply is AC power supply, and the processor utilization rate is low (≤ the second utilization threshold, such as ≤ 30%), it is identified as a light-load background task scenario, and the third power consumption adjustment mode can be set to the normal energy saving mode according to the decision rule table to actively reduce energy consumption while meeting task execution.

[0071] The above rules map the multi-dimensional running state combination to a semantically clear power consumption mode, allowing the server to automatically select the most suitable initial power consumption strategy under different wake-up contexts and environmental conditions, providing a precise basis for generating final control instructions based on subsequent user behavior fusion.

[0072] In a possible implementation, the second power consumption adjustment mode of the server is obtained according to the first power consumption adjustment mode, the user preset power supply mode, and the pre-stored user preference use mode, and includes: when the first power consumption adjustment mode is the overheat protection mode or the emergency energy saving mode, the first power consumption adjustment mode is taken as the second power consumption adjustment mode of the server; when the first power consumption adjustment mode is any one of the balance mode, the rapid mode, or the normal energy saving mode, the first power consumption adjustment mode, the user preset power supply mode, and the user preference use mode are all input to a pre-stored fusion decision engine; and the second power consumption adjustment mode is obtained according to the fusion decision engine.

[0073] Specifically, first, it is judged whether the first power consumption adjustment mode belongs to a server system safety priority mode. If the first power consumption adjustment mode is the overheat protection mode (triggered because the temperature exceeds a safety threshold) or the emergency energy saving mode (triggered because the battery power is seriously insufficient), it indicates that the current server system is in a thermal runaway danger or an energy critical state, and at this time, no matter how the user configures, the hardware safety and basic running ability need to be prioritized. Therefore, the first power consumption adjustment mode can be directly taken as the final second power consumption adjustment mode, and the user intention fusion link is skipped to ensure that the mandatory protection policy takes effect immediately.

[0074] Secondly, if the first power consumption adjustment mode is any one of the balance mode, the rapid mode, or the normal energy saving mode, it indicates that the server system is in a normal running interval and has a regulation space for considering performance, energy efficiency, and user experience. In this case, the first power consumption adjustment mode, the user preset power supply mode, and the pre-stored user preference use mode are input to the fusion decision engine pre-stored in the firmware.

[0075] The first power consumption adjustment mode can reflect the objective demand of the current server system based on the load and the environment state. The user preset power supply mode can be a global strategy (such as extreme performance, default balance, or maximum energy saving) explicitly selected by the user in the BIOS / UEFI setting interface. The pre-stored user preference use mode can be an implicit preference model generated based on the user historical behavior (such as a high-frequency use period, a commonly used wake-up mode, a tolerance to response delay, etc.) through a lightweight learning algorithm, and is usually stored in the NVRAM.

[0076] The fusion decision engine can adopt a multi-level priority coverage mechanism or a weighted scoring strategy for comprehensive determination. For example, when the user preset is extreme performance and the current first power consumption adjustment mode is the normal energy saving mode, the fusion decision engine can promote the second power consumption adjustment mode to the balance mode or the rapid mode.

[0077] When the user preference shows a long-term low-load usage habit, even if the first power consumption adjustment mode is the rapid mode, the fusion decision engine can downgrade the second power consumption adjustment mode to the balance mode to avoid unnecessary energy consumption.

[0078] In addition, if the user setting conflicts less with the first power consumption adjustment mode (server system state), such as the first power consumption adjustment mode is the balance mode and the user preset is also the balance, the balance mode can be directly adopted as the second power consumption adjustment mode.

[0079] Finally, the fusion decision engine can output a second power consumption adjustment mode that coordinates the server system state and the user intention as the only basis for subsequent hardware parameter adjustment. This method can significantly improve the personalization and intelligence level of the power consumption strategy under the premise of ensuring the stability of the server system, and effectively solves the problem of disconnection between the traditional static strategy and the actual use demand of the user.

[0080] In one possible implementation, according to the second power consumption adjustment mode, the hardware parameters in the server are adjusted to realize automatic intelligent adjustment of the server power consumption, including: according to the second power consumption adjustment mode, querying a pre-stored hardware control strategy mapping table to obtain a corresponding hardware adjustment parameter set; and according to the hardware adjustment parameter set, adjusting the hardware parameters in the server to realize automatic intelligent adjustment of the server power consumption.

[0081] Specifically, first, based on the determined second power consumption adjustment mode (such as the overheat protection mode, the emergency energy saving mode, the normal energy saving mode, the balance mode or the rapid mode), a pre-stored hardware control strategy mapping table can be queried in a non-volatile storage area (such as a strategy configuration area in SPI Flash) of firmware (such as BIOS or UEFI). The hardware control strategy mapping table can be indexed by the second power consumption adjustment mode and associated with a set of refined hardware adjustment parameters, which can specifically include control instructions of multiple dimensions such as processor, memory, input / output (Input / Output, I / O) subsystem and heat dissipation unit.

[0082] Subsequently, the firmware can dynamically adjust the key hardware parameters of the server according to the obtained hardware adjustment parameter set through a bottom-layer hardware abstraction interface, for example, through ACPI method, P-State / C-State control register, system management bus (System Management Bus, SMBus), power management bus (Power Management Bus, PMBus) or dedicated power management controller. Specifically, it includes but is not limited to:

[0083] Processor subsystem: e.g. set the processor's performance state (P-state), idle state (C-state), enable or disable Turbo Boost, on platforms supporting dynamic voltage and frequency scaling, coordinate core voltage and frequency, or perform Core Offlining for inactive cores.

[0084] Memory subsystem: adjust memory operating frequency, enable Self-Refresh, or reduce memory channel power.

[0085] I / O and peripherals: configure the power management state of PCIe links, e.g. enable Active State Power Management (ASPM), power off unused Universal Serial Bus (USB), Serial Advanced Technology Attachment (SATA) or network controllers.

[0086] Cooling system: adjust fan speed curve according to the requirements of the second power consumption adjustment mode, e.g. increase cooling intensity in turbo mode, or adopt silent temperature control strategy in energy saving mode.

[0087] Power management unit: on platforms supporting multi-phase power supply, dynamically start and stop Voltage Regulator Module (VRM) phases to match the current load requirements.

[0088] The above adjustment operations can be continuously performed by UEFI Runtime Services during the runtime of the server system (S0 state), or can be synchronized and take effect as part of the initialization process during the wake-up process from the low-power state (such as S3 / S4). At the same time, the server system can combine sensor feedback (such as the adjusted temperature and the measured power consumption value) to perform closed-loop verification, and if necessary, trigger strategy fine-tuning, thereby realizing safe, efficient and adaptive intelligent power consumption management.

[0089] Through the implementation, the server can automatically match the optimal energy efficiency strategy under different business scenarios, balance performance guarantee, thermal safety and energy saving, and significantly improve the overall Power Usage Effectiveness (PUE) of the data center and the consistency of user experience.

[0090] In a possible implementation, the hardware parameters in the server are adjusted according to the set of hardware adjustment parameters, to achieve automatic intelligent adjustment of the power consumption of the server, including: when the second power consumption adjustment mode is the overheating protection mode, at least one of the following operations is performed: reducing the processor core frequency to a preset safe running frequency, turning off part of the peripheral component interconnect express interface device, and increasing the fan speed to above a first preset speed; when the second power consumption adjustment mode is the emergency energy saving mode, at least one of the following operations is performed: reducing the performance state of the processor, turning off the unused processor core, disabling part of the peripheral power supply, and enabling dynamic voltage adjustment; when the second power consumption adjustment mode is the urgent mode, at least one of the following operations is performed: removing the upper limit of the processor frequency, enabling turbo boost, controlling each processor core to be in an active state, disabling the energy saving state, and increasing the memory frequency to a first preset frequency; when the second power consumption adjustment mode is the balance mode, at least one of the following operations is performed: dynamically adjusting the performance state of the processor according to the current load, enabling the first level of idle state for the processor, and maintaining the fan speed in the middle section of the temperature control curve; and when the second power consumption adjustment mode is the normal energy saving mode, at least one of the following operations is performed: reducing the performance state of the processor, enabling the second level of idle state for part of the processor, reducing the memory working frequency to a second preset frequency, enabling the peripheral component interconnect express interface link power management, and reducing the fan speed to below a second preset speed. The first level of idle state corresponds to a power consumption greater than that of the second level of idle state, the first preset frequency is greater than the second preset frequency, and the first preset speed is greater than the second preset speed.

[0091] Specifically, the hardware parameters in the server are adjusted according to the set of hardware adjustment parameters to achieve automatic intelligent adjustment of the power consumption of the server, which can specifically include the following fine control strategies for different second power consumption adjustment modes.

[0092] When the second power consumption adjustment mode is the overheating protection mode, it indicates that the temperature of the server system has reached or exceeded the safety threshold, and there is a risk of thermal runaway. At this time, the firmware can immediately perform forced cooling and power limiting measures, including at least one of the following operations:

[0093] The processor core frequency is dynamically reduced to a preset safe running frequency, for example, 50% of the base frequency or the lower limit of the thermal throttling defined by the manufacturer, to quickly reduce the heat source.

[0094] Shut down non-essential Peripheral Component Interconnect Express (PCIe) devices, such as redundant network cards, unused Non-Volatile Memory Express (NVMe) Solid State Drive (SSD) controllers, or expansion cards, to reduce local heat generation and bus power consumption.

[0095] Increase server system fan speed above a first preset speed, such as 90% duty cycle or maximum Revolutions Per Minute (RPM), to enhance active cooling capability and ensure core component temperature falls back to a safe range.

[0096] When the second power consumption adjustment mode is the emergency power saving mode, which is usually triggered when the battery power is below a critical value (e.g., <10%), the goal is to maximize the available time of the server system. At this time, deep power saving operations can be performed, including at least one of the following:

[0097] Switch the processor to the lowest performance state (i.e., the lowest P-state, such as Pn), significantly reducing the power consumption of the computing unit.

[0098] Turn off unused processor cores and put them into a deep sleep state such as C6 / C10 to avoid idling energy consumption.

[0099] Disable power supply for unnecessary peripherals (such as USB controllers, serial ports, auxiliary sensors, etc.), cutting off the power supply path for secondary loads.

[0100] Enable Dynamic Voltage Scaling (DVS) to cooperatively reduce core voltage and frequency, further compressing power consumption.

[0101] When the second power consumption adjustment mode is the emergency mode, the server system is expected to handle high-priority or high-throughput tasks (such as service response after network wake-up). At this time, the goal is to maximize performance, and at least one of the following operations can be performed:

[0102] Remove the upper limit of the processor frequency and enable turbo acceleration to allow the core to run at a higher frequency for a short time.

[0103] Force each processor core to remain in an active state (C0 state) and prohibit entry into any sleep state.

[0104] Disable various energy-saving features, including C-states, P-states automatic frequency reduction, and memory power saving mode.

[0105] The memory operating frequency is raised to a first preset frequency, such as 3200 Mega Transfers per second (MT / s) or the Joint Electron Device Engineering Council (JEDEC) nominal maximum frequency, to raise the data bandwidth to match the computing demand.

[0106] When the second power consumption adjustment mode is the balanced mode, the server system is usually in a regular interactive load scenario, and the response speed and energy efficiency need to be considered. At this time, a moderate adjustment strategy can be adopted, including at least one of the following:

[0107] The P-state is dynamically adjusted according to the real-time processor utilization, and the smooth transition between performance and power consumption is realized.

[0108] The first level of idle state (such as C1 or C3) is enabled for the processor, and moderate power saving is realized when the core is idle, but the faster wake-up delay is maintained.

[0109] The fan speed is maintained in the middle section of the temperature control curve (such as 40%-60% RPM), and the noise and heat dissipation efficiency are balanced.

[0110] When the second power consumption adjustment mode is the ordinary energy saving mode, it is suitable for light load background tasks (such as timing maintenance tasks), and energy saving is the primary goal. At this time, further energy saving measures can be performed, including at least one of the following:

[0111] The processor performance state is reduced to a middle-low gear, and the peak performance is limited.

[0112] The second level of idle state (such as C6 or C8) is enabled for part of the processor cores, and the power consumption is significantly lower than that of the first level of idle state.

[0113] The memory operating frequency is reduced to a second preset frequency (such as 2133 MT / s), and the memory subsystem energy consumption is reduced;

[0114] The PCIe link power management (ASPM L1 / L1.2) is enabled, and the link automatically enters a low-power state when it is idle.

[0115] The fan speed is controlled below a second preset speed (such as 30% RPM or a silent threshold), and the noise and fan power consumption are reduced.

[0116] The power consumption corresponding to the first level of idle state is higher than that corresponding to the second level of idle state, the first preset frequency is higher than the second preset frequency, and the first preset speed is higher than the second preset speed. The above parameters are pre-configured in the hardware control strategy mapping table of the firmware and can be customized according to the platform model or user strategy.

[0117] Through the above multi-granularity, mode-driven hardware parameter adjustment mechanism, the server can automatically match the optimal energy efficiency strategy under different running scenarios, while ensuring business continuity and server system security, and realizing fine and intelligent management of power consumption.

[0118] In a possible implementation, after adjusting the hardware parameters in the server according to the second power consumption adjustment mode, the method further includes: writing the power consumption state of the server into an Advanced Configuration and Power Interface list; and passing the power consumption state of the server to an operating system through the Advanced Configuration and Power Interface list, so that the operating system takes over the power consumption adjustment function of the server.

[0119] Specifically, writing the power consumption state of the server into the Advanced Configuration and Power Interface list can be implemented by updating a specific field in an ACPI table.

[0120] After completing the adjustment of the server hardware parameters according to the second power consumption adjustment mode, the following cooperative operations are further included to realize seamless connection between firmware and the operating system in power consumption management.

[0121] First, the firmware can write the currently effective server power consumption state information (including the adopted power consumption adjustment mode, the corresponding P-state / C-state configuration, the memory frequency level, the PCIe power management state, etc.) into a predefined data structure in the Advanced Configuration and Power Interface (ACPI), for example, a Performance Supported States (PSS) table, a C-State States Table (CST), a Thermal State Dependency (TSD), or a Device-Specific Method (DSM) object, and update the related entries in the ACPI system description table, such as a Differentiated System Description Table (DSDT) or a Secondary System Description Table (SSDT).

[0122] Subsequently, the updated power consumption state and policy intent are delivered to the operating system (OS) through standard ACPI mechanisms, such as system control interrupt (SCI) or operating system-directed power management (OSPM) notification. The operating system can resolve the above information through ACPI driver during booting process or runtime, and learn the initial power consumption context set by the firmware, and take over the subsequent fine-grained power consumption adjustment function accordingly. For example, the Linux operating system kernel can enable the CPU frequency scaling subsystem (CPUFreq) scheduler based on the received P-state list, and the Windows operating system can dynamically adjust the ACPI thermal policy and power plan.

[0123] The mechanism ensures the consistency and continuity of power consumption management policies from the firmware layer to the operating system layer. The firmware is responsible for quickly establishing a safe and reasonable power consumption baseline in the early stage of the system (such as S0 initialization or S3 wake-up), and the operating system then combines application load, scheduling queue and user power policy for more fine-grained real-time control. Both work together to avoid policy conflicts and fully leverage the advantages of firmware bottom-layer control capabilities and operating system high-level scheduling intelligence, thereby achieving full-stack, high-efficiency server dynamic power consumption management.

[0124] The embodiments of the present application can divide the functional modules of the electronic device or the host device according to the above-mentioned method examples. For example, each functional module can be divided according to each function, or two or more functions can be integrated in one processing unit. The integrated unit can be realized in the form of hardware or in the form of a software functional module. It should be noted that the division of modules in the embodiments of the present application is illustrative, and is only a logical functional division. When actually implemented, another division method can be used.

[0125] Figure 2 The structure diagram of the server dynamic power consumption adjustment device based on load awareness provided by the embodiments of the present application is shown in FIG. 1. As shown in FIG. 1, the embodiments of the present application also provide a server dynamic power consumption adjustment device based on load awareness. The device includes an acquisition module 210, a data processing module 220, a fusion processing module 230, and an adjustment module 240. Figure 2

[0126] ​The obtaining module 210 is configured to obtain running state data of the server, the running state data comprising processor utilization, temperature sensor data, a wakeup trigger source of the server, and a power supply state.

[0127] The data processing module 220 is configured to determine a first power consumption adjustment mode of the server according to the processor utilization, the temperature sensor data, the wakeup trigger source of the server, and the power supply state.

[0128] The obtaining module 210 is further configured to obtain user behavior data of the server, the user behavior data comprising a user preset power supply mode and a pre-stored user preference use mode.

[0129] The fusion processing module 230 is configured to obtain a second power consumption adjustment mode of the server according to the first power consumption adjustment mode, the user preset power supply mode, and the pre-stored user preference use mode.

[0130] The adjustment module 240 is configured to adjust a hardware parameter in the server according to the second power consumption adjustment mode, so as to realize automatic intelligent adjustment of power consumption of the server.

[0131] In a possible implementation, the obtaining module 210 is further configured to determine the processor utilization by reading a first register.

[0132] The obtaining module 210 is further configured to obtain the temperature sensor data and the power supply state through a super input-output integrated circuit.

[0133] The obtaining module 210 is further configured to obtain the wakeup trigger source of the server through a real-time clock interface.

[0134] In a possible implementation, the obtaining module 210 is further configured to obtain the user preset power supply mode and the pre-stored user preference use mode through a non-volatile random access memory.

[0135] In a possible implementation, the data processing module 220 is further configured to match a corresponding third power consumption adjustment mode from a pre-stored decision rule table according to the processor utilization, the temperature sensor data, the wakeup trigger source of the server, and the power supply state.

[0136] The data processing module 220 is further configured to take the third power consumption adjustment mode as the first power consumption adjustment mode of the server.

[0137] In a possible implementation, the wakeup trigger source comprises a timing wakeup, a network wakeup, and a power key wakeup.

[0138] The data processing module 220 is further configured to determine, according to the pre-stored decision rule table, that the third power consumption adjustment mode of the server is an overheating protection mode when the temperature sensor data indicates that the temperature is greater than or equal to a first temperature threshold.

[0139] The data processing module 220 is further configured to determine, according to the pre-stored decision rule table, the third power consumption adjustment mode of the server as an emergency power saving mode when the power supply state is battery power supply and the battery power is lower than the pre-set power threshold.

[0140] The data processing module 220 is further configured to determine, according to the pre-stored decision rule table, the third power consumption adjustment mode of the server as a balance mode when the wake-up trigger source is the power key wake-up, the temperature sensor data indicates that the temperature is less than the first temperature threshold and greater than the second temperature threshold, the power supply state is AC power supply, and the processor utilization rate is less than the first utilization threshold and greater than the second utilization threshold.

[0141] The data processing module 220 is further configured to determine, according to the pre-stored decision rule table, the third power consumption adjustment mode of the server as a rapid mode when the wake-up trigger source is network wake-up, the temperature sensor data indicates that the temperature is less than or equal to the second temperature threshold, the power supply state is AC power supply, and the processor utilization rate is greater than or equal to the first utilization threshold.

[0142] The data processing module 220 is further configured to determine, according to the pre-stored decision rule table, the third power consumption adjustment mode of the server as a normal power saving mode when the wake-up trigger source is a timing wake-up, the temperature sensor data indicates that the temperature is less than or equal to the second temperature threshold, the power supply state is AC power supply, and the processor utilization rate is less than or equal to the second utilization threshold.

[0143] In a possible implementation, the fusion processing module 230 is further configured to, when the first power consumption adjustment mode is the overheat protection mode or the emergency power saving mode, take the first power consumption adjustment mode as the second power consumption adjustment mode of the server.

[0144] The fusion processing module 230 is further configured to, when the first power consumption adjustment mode is any one of the balance mode, the rapid mode or the normal power saving mode, input the first power consumption adjustment mode, the user pre-set power mode and the user preference use mode into the pre-set fusion decision engine.

[0145] The fusion processing module 230 is further configured to obtain the second power consumption adjustment mode according to the fusion decision engine.

[0146] In a possible implementation, the adjustment module 240 is further configured to query the pre-stored hardware regulation strategy mapping table according to the second power consumption adjustment mode, and obtain a corresponding hardware adjustment parameter set.

[0147] The adjustment module 240 is further configured to adjust the hardware parameters in the server according to the hardware adjustment parameter set, so as to realize automatic intelligent adjustment of the power consumption of the server.

[0148] In a possible implementation, the adjusting module 240 is further configured to, when the second power consumption adjustment mode is the overheating protection mode, perform at least one of the following operations: reducing a processor core frequency to a preset safe operation frequency, shutting down part of peripheral component interconnect express (PCIe) devices, and increasing a fan rotating speed to be higher than a first preset rotating speed.

[0149] The adjusting module 240 is further configured to, when the second power consumption adjustment mode is the emergency energy saving mode, perform at least one of the following operations: reducing a performance state of the processor, shutting down unused processor cores, disabling power supply for part of peripherals, and enabling dynamic voltage adjustment.

[0150] The adjusting module 240 is further configured to, when the second power consumption adjustment mode is the urgent mode, perform at least one of the following operations: removing an upper limit of a processor frequency, enabling turbo boost, controlling each processor core to be in an active state, disabling an energy saving state, and increasing a memory frequency to be a first preset frequency.

[0151] The adjusting module 240 is further configured to, when the second power consumption adjustment mode is the balance mode, perform at least one of the following operations: dynamically adjusting a performance state of the processor according to a current load, enabling a first level of idle state for the processor, and maintaining the fan rotating speed to be in a middle section of a temperature control curve.

[0152] The adjusting module 240 is further configured to, when the second power consumption adjustment mode is the normal energy saving mode, perform at least one of the following operations: reducing a performance state of the processor, enabling a second level of idle state for part of the processor, reducing a memory working frequency to be a second preset frequency, enabling power management of a peripheral component interconnect express (PCIe) link, and reducing the fan rotating speed to be lower than a second preset rotating speed. The first level of idle state corresponds to a power consumption that is greater than a power consumption corresponding to the second level of idle state, the first preset frequency is greater than the second preset frequency, and the first preset rotating speed is greater than the second preset rotating speed.

[0153] In a possible implementation, the apparatus further includes a delivery module.

[0154] The delivery module is configured to write, to an advanced configuration and power interface (ACPI) list, a power consumption state of the server.

[0155] The delivery module is further configured to deliver, to an operating system, the power consumption state of the server through the ACPI list, so that the operating system takes over a power consumption adjustment function of the server.

[0156] The features of the embodiments of the server dynamic power consumption adjustment apparatus based on load sensing can be understood in the light of the related descriptions of the embodiments of the server dynamic power consumption adjustment method based on load sensing, which will not be repeated here.

[0157] Figure 3A structural schematic diagram of an electronic device is provided in the present application. As shown in Figure 3 The electronic device provided in the present embodiment comprises at least one processor 310 and a memory 320. Optionally, the electronic device further comprises a communication component 330. The processor 310, the memory 320 and the communication component 330 are connected through a bus.

[0158] In the implementation process, the at least one processor 310 executes the computer execution instructions stored in the memory 320, so that the at least one processor 310 executes the above-mentioned load-aware server dynamic power consumption adjustment method embodiment.

[0159] The specific implementation process of the processor 310 can refer to the above-mentioned method embodiments, which have similar implementation principles and technical effects, and will not be described here in detail.

[0160] In the above-mentioned embodiments, it should be understood that the processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC) and the like. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the application can be directly embodied as the execution of the hardware processor, or executed by the combination of hardware and software modules in the processor.

[0161] The memory can contain a random access memory (RAM), and can also include a non-volatile memory (NVM), for example, at least one disk memory.

[0162] The bus can be an industry standard architecture (ISA) bus, a peripheral component (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, the bus in the drawings of the present application does not limit to only one bus or one type of bus.

[0163] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is configured to execute the steps in any of the above load-aware server dynamic power consumption adjustment method embodiments when running.

[0164] In an example embodiment, the above computer readable storage medium can include, but is not limited to, a U disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store computer programs.

[0165] The embodiment of the present application further provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the steps in any of the above load-aware server dynamic power consumption adjustment method embodiments.

[0166] The embodiment of the present application further provides another computer program product, which comprises a non-volatile computer readable storage medium, and the non-volatile computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps in any of the above load-aware server dynamic power consumption adjustment method embodiments.

[0167] The skilled in the art can further realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized in electronic hardware, computer software or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been described in general terms in the above description. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. The skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0168] The above provides a detailed description of a load-aware server dynamic power consumption adjustment method and an electronic device. The principles and implementation modes of the present application are described by applying specific examples in this paper, and the above example description is only used to help understand the method and its core idea of the present application. It should be pointed out that for ordinary skilled in the art, some improvements and modifications can be made to the present application without departing from the principles of the present application, and these improvements and modifications also fall within the protection scope of the claims of the present application.

Claims

1. A load-aware dynamic power consumption adjustment method for servers, characterized in that, include: Obtain server operating status data, including processor utilization, temperature sensor data, server wake-up trigger source, and power status; Based on the processor utilization, temperature sensor data, server wake-up trigger source, and power status, a corresponding first power consumption adjustment mode is matched from a pre-stored decision rule table; wherein, the first power consumption adjustment mode is an overheat protection mode, an emergency energy-saving mode, a normal energy-saving mode, a balanced mode, or a rapid mode. Obtain user behavior data from the server, including user-preset power modes and pre-stored user preference usage modes, which are generated based on historical operation logs; When the first power consumption adjustment mode is the overheat protection mode or the emergency energy saving mode, the first power consumption adjustment mode is used as the second power consumption adjustment mode of the server. When the first power consumption adjustment mode is any one of the balanced mode, the rapid mode, or the normal energy-saving mode, the first power consumption adjustment mode, the user-preset power mode, and the user-preferred usage mode are all input to the preset fusion decision engine. The second power consumption adjustment mode of the server is obtained based on the output of the fusion decision engine; Based on the second power consumption adjustment mode and the set of hardware adjustment parameters corresponding to the second power consumption adjustment mode, the hardware parameters in the server are adjusted to achieve automatic and intelligent adjustment of server power consumption. The hardware parameters in the server include at least one of the processor core frequency, processor performance status, and memory operating frequency.

2. The method according to claim 1, characterized in that, The acquisition of server operating status data includes: The processor utilization rate is determined by reading the first register; The temperature sensor data and the power status are acquired through a super input / output integrated circuit; The wake-up trigger source of the server is obtained through the real-time clock interface.

3. The method according to claim 1, characterized in that, The acquisition of user behavior data from the server includes: The user-preset power mode and the pre-stored user preference usage mode are obtained through non-volatile random access memory.

4. The method according to claim 1, characterized in that, The wake-up trigger sources include timed wake-up, network wake-up, and power button wake-up; The step of matching the corresponding first power consumption adjustment mode from a pre-stored decision rule table based on the processor utilization, temperature sensor data, server wake-up trigger source, and power status includes: When the temperature sensor data indicates that the temperature is greater than or equal to the first temperature threshold, the server's first power consumption adjustment mode is determined to be the overheat protection mode according to the pre-stored decision rule table. When the power supply is battery powered and the battery level is lower than a preset power threshold, the first power consumption adjustment mode of the server is determined to be emergency energy-saving mode according to the pre-stored decision rule table. When the wake-up trigger source is the power button wake-up, the temperature sensor data indicates that the temperature is less than the first temperature threshold and greater than the second temperature threshold, the power state is AC power supply and the processor utilization is less than the first utilization threshold and greater than the second utilization threshold, the first power consumption adjustment mode of the server is determined to be the balanced mode according to the pre-stored decision rule table. When the wake-up trigger source is the network wake-up, the temperature sensor data indicates that the temperature is less than or equal to the second temperature threshold, the power supply status is AC power supply and the processor utilization is greater than or equal to the first utilization threshold, the first power consumption adjustment mode of the server is determined to be rapid mode according to the pre-stored decision rule table. When the wake-up trigger source is the timed wake-up, the temperature sensor data indicates that the temperature is less than or equal to the second temperature threshold, the power supply status is AC power supply and the processor utilization is less than or equal to the second utilization threshold, the server's first power consumption adjustment mode is determined to be the normal energy-saving mode according to the pre-stored decision rule table.

5. The method according to claim 1, characterized in that, The step of adjusting the hardware parameters in the server according to the second power consumption adjustment mode and the set of hardware adjustment parameters corresponding to the second power consumption adjustment mode to achieve automatic and intelligent adjustment of the server power consumption includes: According to the second power consumption adjustment mode, query the pre-stored hardware control strategy mapping table to obtain the corresponding set of hardware adjustment parameters; Based on the set of hardware adjustment parameters, the hardware parameters in the server are adjusted to achieve automatic and intelligent adjustment of the server's power consumption.

6. The method according to claim 5, characterized in that, The step of adjusting the hardware parameters in the server according to the set of hardware adjustment parameters and the set of hardware adjustment parameters corresponding to the second power consumption adjustment mode to achieve automatic and intelligent adjustment of the server power consumption includes: When the second power consumption adjustment mode is the overheat protection mode, at least one of the following operations is performed: reduce the processor core frequency to a preset safe operating frequency, shut down some peripheral component interconnect high-speed interface devices, and increase the fan speed to above the first preset speed; If the second power consumption adjustment mode is emergency power saving mode, then at least one of the following operations will be performed: reduce the performance state of the processor, shut down unused processor cores, disable power supply to some peripherals, and enable dynamic voltage regulation. When the second power consumption adjustment mode is rapid mode, at least one of the following operations is performed: remove the processor frequency limit and enable Turbo Boost, control each processor core to be in an active state, disable power saving state, and increase the memory operating frequency to the first preset frequency. When the second power consumption adjustment mode is the balanced mode, at least one of the following operations is performed: dynamically adjust the processor performance state according to the current load, enable the first level of idle state for the processor, and maintain the fan speed in the middle of the temperature control curve; When the second power consumption adjustment mode is the normal power saving mode, at least one of the following operations is performed: reducing the performance state of the processor, enabling the second level of idle state for some processors, reducing the memory operating frequency to the second preset frequency, enabling power management of the high-speed interface link for peripheral component interconnection, and reducing the fan speed to below the second preset speed. Wherein, the power consumption corresponding to the first level of idle state is greater than the power consumption corresponding to the second level of idle state, the first preset frequency is greater than the second preset frequency, and the first preset speed is greater than the second preset speed.

7. The method according to any one of claims 1 to 6, characterized in that, After adjusting the hardware parameters in the server according to the second power consumption adjustment mode, the method further includes: Write the power consumption status of the server into the advanced configuration and power interface list; The power consumption status of the server is transmitted to the operating system through the advanced configuration and power interface list, so that the operating system can take over the power consumption regulation function of the server.

8. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the load-aware server dynamic power consumption adjustment method as described in any one of claims 1 to 7 when executing the computer program.

Images