A server power consumption regulation method and electronic device

Abstract

The application discloses a kind of server power consumption regulation methods, electronic equipment, it is related to server technical field, including by substrate management controller out-of-band acquisition server power consumption data, first check data validity, invalid time is based on historical data prediction and with confidence degree determines whether to adopt historical peak value auxiliary, effective time generates first strategy including risk level and high-power component type position;When there is risk, respectively to high-speed peripheral component interconnect device according to regulation object Nonlinear power control is executed, to fan executes grouping gradient speed regulation, to hard disk according to state determines priority and power-off or power-on, while supporting operating system and substrate management controller double strategy;Power supply unit is detected twice after regulation, still alarm then directly to high-power component frequency reduction bottom. Thus, the technical problem that power consumption regulation is passive, extensive, dependent on host and prone to power-off is solved, and the technical effects of active prediction, layered accurate regulation and avoiding whole machine power-off are achieved.

Description

Technical Field

[0001] This application relates to the field of server technology, and in particular to a server power consumption control method and electronic device. Background Technology

[0002] With the popularization of artificial intelligence and big data technologies, high-computing power devices are increasing, and the power consumption of server machines is constantly rising. Server manufacturers usually provide power supply units according to the shipment configuration, but in scenarios such as user configuration changes, power supply unit failures, or component upgrades, it is still necessary to reasonably control the power consumption during operation. Currently, there are two main solutions: (1) Power control based on the server host: The operating system obtains the power of the central processing unit and directly controls it. Its shortcomings are: occupying host computing resources; relying on the operating system to run scripts, and being unable to respond when the operating system is fully loaded or not installed; and some components cannot be controlled by the operating system. (2) Dynamic power consumption capping based on out-of-band: dynamic control is achieved through out-of-band monitoring. Its shortcomings are: it only responds after the power supply unit triggers the throttling signal, tending to be passive; when the overcurrent is large, the power supply unit will shut down the output for a short time, causing the whole machine to lose power; and it only limits the power of the central processing unit and cannot control other components. In addition, there are problems with data silos and single data collection dimensions: energy-saving strategies are mostly fixed frequency adjustment of a single component, without combining the power consumption linkage relationship of each component for coordinated control, which can easily lead to an imbalance between energy saving and performance; and they mostly only collect the power consumption of core components such as central processing unit and memory, ignoring the power consumption contribution of auxiliary components such as high-speed peripheral component interconnect express (PCIe) devices, fans, and hard drives, which can easily cause deviations in peak power consumption prediction. Summary of the Invention

[0003] This application provides a server power consumption control method and electronic device to at least solve the problems of limited host power control resources, passive control, data silos and single acquisition dimensions, and inability to accurately control power consumption in related technologies.

[0004] This application provides a server power consumption control method, which includes: acquiring server power consumption data and determining the validity of the server power consumption data, wherein the server power consumption data includes power consumption data of at least one component; in response to the server power consumption data being valid, generating a first power consumption control strategy based on the server power consumption data, wherein the first power consumption control strategy includes a risk level of the server; in response to the risk level indicating no power consumption risk, re-executing the above steps; in response to the risk level indicating the existence of power consumption risk, determining a second power consumption control strategy based on the power consumption data of the component, and performing corresponding power consumption control operations on the component according to the second power consumption control strategy or a third power consumption control strategy issued by the operating system in the server; in response to the completion of all power consumption control operations, detecting whether the power supply unit in the server meets a preset alarm condition; in response to meeting the preset alarm condition, performing corresponding power consumption control operations on the components in a preset component list according to a preset fourth power consumption control strategy, and ending the current server power consumption control; in response to not meeting the preset alarm condition, re-executing the above steps.

[0005] This application also provides an electronic device, including: a memory for storing a computer program; and a processor for implementing the steps of any of the above-described server power consumption control methods when executing the computer program.

[0006] This application utilizes out-of-band acquisition of server power consumption data by the baseboard management controller. The data validity is first verified; if invalid, historical data is used for prediction, and the confidence level determines whether historical peak values ​​should be used for assistance. If valid, a first strategy is generated, including risk level and the location of high-power components. When risks exist, nonlinear power control is applied to high-speed peripheral interconnect devices, grouped gradient speed regulation is applied to fans, and hard drives are prioritized and powered off or on based on their status. Dual strategies from the operating system and baseboard management controller are supported. After regulation, a second detection of the power supply unit is performed; if alarms still occur, the frequency of high-power components is directly reduced as a fallback. This solves the technical problems of passive, coarse, host-dependent, and easily power-off power control, achieving proactive prediction, layered precise control, and prevention of system-wide power outages. Attached Figure Description

[0007] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 This application provides a structural block diagram of a server power consumption control system for implementing a server power consumption control method. Figure 2 This application provides a schematic diagram of power consumption control for a server power consumption regulation system. Figure 3 A schematic flowchart illustrating a server power consumption control method provided in an embodiment of this application; Figure 4 A detailed flowchart of a server power consumption control method provided in this application embodiment; Figure 5 An electronic device for implementing a server power consumption control method is provided in the embodiments of this application. Detailed Implementation

[0009] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this application.

[0010] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The terms "first," "second," etc., in this application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0011] To enable 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.

[0012] The server power consumption control method described in this application is applied to, for example... Figure 1The diagram illustrates a server power consumption control system centered around a baseboard management controller (BMC). The BMC interacts with the various components shown in the diagram via an integrated circuit bus or a high-speed peripheral component interconnect bus. Specifically, the BMC interacts with the management engine or platform environment control interface via an intelligent platform management bus or an improved integrated circuit bus to obtain the CPU's power consumption; it also interacts with the hard drive via the integrated circuit bus and the complex programmable logic devices on the hard drive backplane, or directly with the non-volatile memory host controller interface on the hard drive backplane, to obtain the hard drive's power consumption; it interacts with the complex programmable logic devices on the fan board via the integrated circuit bus to obtain the fan power consumption; and it interacts with high-speed peripheral component interconnect devices via the integrated circuit bus or the high-speed peripheral component interconnect bus. Furthermore, the BMC can interact with the power supply unit via the power management bus to obtain the maximum power supported by the entire system and the current output power information of the power supply unit.

[0013] like Figure 2 As shown, the power supply unit is responsible for powering the entire server. The server motherboard has voltage regulation chips. The power supply unit's power is converted from AC to DC to power the high-speed peripheral component interconnect links. Voltage regulation chips are placed on each high-speed peripheral component interconnect link to monitor power. Each voltage regulation chip can be monitored and its data obtained by the baseboard management controller via the integrated circuit bus. All high-speed peripheral component interconnect device services are initiated by the host operating system. The host operating system can obtain the scheduling plan of each device and transmits the decision strategy to the baseboard management controller, which then makes the control decisions. The baseboard management controller obtains the maximum power and current total output power of the power supply unit through the power management bus. Then, based on the current power situation, its own strategy, and the strategy sent by the host operating system, it sends power control commands to the voltage regulation chips to achieve power control. The baseboard management controller monitors the power of the high-speed peripheral component interconnect devices in real time. For graphics processor devices, if a sudden surge in power causes excessive power consumption and power overload, the baseboard management controller can trigger a power braking signal, which is directly transmitted to the graphics processor device for power limiting.

[0014] Embodiments of this application provide a server power consumption control method, such as... Figure 3As shown, the server power consumption control method includes: acquiring server power consumption data and determining the validity of the server power consumption data, wherein the server power consumption data includes power consumption data of at least one component; in response to the server power consumption data being valid, generating a first power consumption control strategy based on the server power consumption data, wherein the first power consumption control strategy includes the server's risk level; in response to the risk level indicating no power consumption risk, re-executing the above steps; in response to the risk level indicating the presence of power consumption risk, determining a second power consumption control strategy based on the component's power consumption data, and performing corresponding power consumption control operations on the component according to the second power consumption control strategy or a third power consumption control strategy issued by the operating system in the server; in response to the completion of all power consumption control operations, detecting whether the power supply unit in the server meets a preset alarm condition; in response to meeting the preset alarm condition, performing corresponding power consumption control operations on the components in a preset component list according to a preset fourth power consumption control strategy, and ending the current server power consumption control; in response to not meeting the preset alarm condition, re-executing the above steps.

[0015] Specifically, the above method enables out-of-band active power consumption control without relying on the host operating system, and when there is still a risk after control, it directly caps the critical components to avoid server power outages.

[0016] Further, determining the validity of server power consumption data includes: obtaining validity determination information corresponding to the component, wherein the validity determination information includes at least one of the following: link status information of the data acquisition link corresponding to the component, preset power consumption range corresponding to the component, and acquisition timestamp corresponding to the power consumption data of the component; determining whether the validity determination information meets preset validity conditions; in response to the validity determination information meeting the preset validity conditions, determining that the power consumption data of the component is valid; in response to the validity determination information not meeting the preset validity conditions, determining that the power consumption data of the component is invalid.

[0017] Specifically, the above method verifies the validity of the power consumption data of each component by using at least one of the following validity judgment information: link status information, preset power consumption range, and collection timestamp, to prevent invalid data from entering the subsequent decision-making process and causing misjudgment.

[0018] Furthermore, determining whether the validity determination information meets the preset validity conditions includes: in response to the validity determination information being link status information, determining whether the link status information indicates that the data acquisition link is in a normal state; if yes, determining that the power consumption data of the component is valid; otherwise, determining that the power consumption data of the component is invalid; in response to the validity determination information being a preset power consumption range, determining whether the power consumption data of the component falls within the preset power consumption range; if yes, determining that the power consumption data of the component is valid; otherwise, determining that the power consumption data of the component is invalid; in response to the validity determination information being a collection timestamp, determining whether the difference between the collection timestamp and the current time is less than or equal to a preset time threshold; if yes, determining that the power consumption data of the component is valid; otherwise, determining that the power consumption data of the component is invalid.

[0019] Furthermore, determining whether the validity determination information meets the preset validity conditions also includes: in response to the validity determination information including link status information, preset power consumption range, and acquisition timestamp, judging the link status information, preset power consumption range, and acquisition timestamp respectively to obtain link judgment result, range judgment result, and time judgment result; in response to the link judgment result, range judgment result, and time judgment result all indicating that the preset validity conditions are met, the power consumption data of the component is determined to be valid; in response to any of the link judgment result, range judgment result, and time judgment result indicating that the preset validity conditions are not met, the power consumption data of the component is determined to be invalid.

[0020] Specifically, when the above method uses three criteria—link status, preset power consumption range, and collection timestamp—to determine data validity, all three criteria must be met. If any criterion is not met, the data is deemed invalid, thereby improving data reliability and reducing the risk of false and missed warnings.

[0021] Furthermore, such as Figure 4 As shown, the server power consumption control method further includes: in response to the presence of invalid data in the server power consumption data, identifying the invalid component corresponding to the invalid data and obtaining the historical power consumption data of the invalid component; predicting the invalid component based on the historical power consumption data to obtain the predicted power consumption data of the invalid component and the prediction confidence level corresponding to the predicted power consumption data; in response to the prediction confidence level being less than a preset confidence threshold, obtaining the historical peak statistics data of the invalid component, and generating a first power consumption control strategy based on the historical peak statistics data and the valid data in the server power consumption data; and in response to the prediction confidence level being greater than or equal to the preset confidence threshold, generating the first power consumption control strategy based on the predicted power consumption data and the valid data.

[0022] Specifically, when invalid data exists, the above method first performs historical data prediction on invalid components. If the prediction confidence is lower than the threshold, the first power consumption control strategy is generated by combining historical peak statistics with valid data. If the confidence meets the standard, the first power consumption control strategy is generated by combining predicted power consumption with valid data. This ensures that decision-making is not interrupted or out of control even when data is abnormal.

[0023] Furthermore, a first power consumption control strategy is generated based on server power consumption data, including: determining the overall power consumption of the server based on the server power consumption data; obtaining the maximum supported power of the power supply unit, and determining the risk level of the server based on the overall power consumption and the maximum supported power; identifying components in the server whose power consumption ratio is greater than or equal to a preset ratio threshold as high-power components, and identifying components in the server whose power consumption ratio is less than the preset ratio threshold and which are currently active or predicted to be active in the future as optimizable components; obtaining the data acquisition link corresponding to the high-power components, and determining the location of the high-power components based on the data acquisition link corresponding to the high-power components; and generating the first power consumption control strategy based on the risk level, the type and location of the high-power components, and the type of optimizable components.

[0024] Specifically, the above method determines the risk level based on the overall power consumption and the maximum supported power of the power supply unit, while identifying high-power components and optimizable components, providing clear targets for subsequent precise control.

[0025] Furthermore, in response to a prediction confidence level less than a preset confidence threshold, the total power consumption of the server is determined based on server power consumption data, including: determining the total power consumption based on historical peak statistics and valid data; in response to a prediction confidence level greater than or equal to a preset confidence threshold, the total power consumption of the server is determined based on server power consumption data, including: determining the total power consumption based on predicted power consumption data and valid data.

[0026] Furthermore, based on the total power consumption and maximum supported power, the risk level of the server is determined, including: calculating the power ratio of the total power consumption to the maximum supported power; in response to the power ratio being less than a preset ratio threshold, the risk level of the server is determined to be Level 1, wherein Level 1 is used to characterize that the server has no power consumption risk; in response to the power ratio being greater than or equal to the preset ratio threshold, the risk level of the server is determined to be Level 2, wherein Level 2 is used to characterize that the server has power consumption risk.

[0027] Further, a second power consumption control strategy is determined based on the power consumption data of the components, including: determining the control objects in the server based on the power consumption data of the components, the type and location of high-power components, and the type of optimizable components; in response to the control objects including high-speed peripheral component interconnect devices in the server, generating a second power consumption control strategy corresponding to the high-speed peripheral component interconnect devices based on the power demand corresponding to the high-speed peripheral component interconnect devices and the current output power of the voltage regulation chip connected to the high-speed peripheral component interconnect devices; in response to the control objects including fans in the server, generating a second power consumption control strategy corresponding to the fans based on the correspondence between the silkscreen number of the high-speed peripheral component interconnect devices and the fan group, and the correspondence between the temperature range and speed gradient of the fans in the fan group; in response to the control objects including hard drives in the server, generating a second power consumption control strategy corresponding to the hard drives based on the indicator light status information, health status information, and remaining lifespan information of the hard drives.

[0028] Specifically, the above method determines the control target based on component power consumption data, the type and location of high-power components, and the type of components that can be optimized, and generates corresponding second power consumption control strategies for PCIe devices, fans, and hard drives respectively, so as to achieve precise control at the component level.

[0029] Furthermore, according to the second power consumption control strategy or the third power consumption control strategy issued by the operating system in the server, the corresponding power consumption control operation is performed on the component, including: determining whether the third power consumption control strategy issued by the operating system has been received; in response to receiving the third power consumption control strategy, performing the corresponding power consumption control operation on the component according to the third power consumption control strategy; in response to not receiving the third power consumption control strategy, performing the corresponding power consumption control operation on the component according to the second power consumption control strategy.

[0030] Specifically, the above method prioritizes the execution of the third power control strategy when it receives the third power control strategy issued by the operating system, and only executes the second power control strategy of the baseboard management controller when it does not receive the third power control strategy. This forms a dual decision-making mechanism between the baseboard management controller and the operating system, which ensures both the real-time performance when the operating system is online and the independent control of the baseboard management controller when the operating system is offline or has not issued a strategy.

[0031] Furthermore, according to the second power consumption control strategy, corresponding power consumption control operations are performed on the components, including: in response to the control object including high-speed peripheral component interconnect devices, determining the maximum output power percentage of the voltage regulation chip corresponding to the high-speed peripheral component interconnect device based on the required power of the high-speed peripheral component interconnect device and a preset nonlinear control relationship, and performing power control operations on the high-speed peripheral component interconnect device based on the maximum output power percentage, wherein the nonlinear control relationship is used to characterize that the maximum output power percentage first rises rapidly as the required power increases, then the growth rate slows down and approaches the upper limit value; in response to the control object including fans, performing speed control operations on the fans in the fan group based on the correspondence between the silkscreen number of the high-speed peripheral component interconnect device and the fan group, and the correspondence between the temperature range and speed gradient of the fans in the fan group; in response to the control object including hard drives, determining the control priority of the hard drive based on the indicator light status information, health status information and remaining lifespan information of the hard drive, and performing a power-off operation on the hard drive according to the control priority; and in response to the disappearance of the power consumption risk of the server, performing a power-on operation on the power-off hard drive.

[0032] Specifically, the maximum output power percentage of the voltage regulation chip is determined according to the nonlinear control relationship for PCIe devices and power control is performed. For fans, speed control is performed according to the correspondence between the silkscreen number and the fan group and the correspondence between the temperature range and the speed gradient. For hard drives, priority is determined according to the indicator light status, health status and remaining lifespan and power-off or power-on is performed. In this way, the control granularity is refined to each component, including PCIe devices, fans and hard drives.

[0033] In a specific implementation, as shown in Table 1 (Fan Power Consumption Control Strategy Table): Table 1

[0034] When the overall power consumption of the system is too high, the fan speed on the side corresponding to the PCIe device can be increased, rather than increasing the speed of all fans, to achieve precise heat dissipation during special periods. For example, a server has 6 fans, divided into two airflow channels, one for the PCIe devices under CPU0 and the other for the PCIe devices under CPU1. Assuming that the SLOT0 device on CPU0 is hot, the fan groups FAN-0, FAN-1, and FAN-2 corresponding to CPU0 will increase their speeds, with FAN-0 at 80% duty and FAN1 and FAN2 at 70% duty. If the SLOT1 device on CPU0 is hot, the fan groups corresponding to CPU0 will have FAN-0 and FAN2 at 70% duty, while FAN1 will be at 80% duty. To prevent airflow chaos caused by zoned speed adjustment, a shroud can be added to reduce turbulence and backflow.

[0035] In a specific implementation, when data is written to a hard drive on the backplane, the RAID card or operating system illuminates the activity indicator light for that hard drive location via the backplane. The baseboard management controller can access the registers of the complex programmable logic device (CPL) on the backplane to determine which hard drive's activity indicator light is illuminated, thus identifying which hard drive is currently being written to and determining its current activity status. Furthermore, for non-volatile memory host controller interface (NTCMI) hard drives and Serial AMI or Serial Advanced Technology (SATA) hard drives connected to the RAID card, the baseboard management controller can obtain the remaining lifespan or self-monitoring analysis and reporting technology (SAIT) information of the hard drives through out-of-band access. This information provides the historical access frequency of the disk, as well as its current temperature, health status, and read / write volume. In this solution, the baseboard management controller checks whether the disk's activity indicator light is illuminated via the CPL on the backplane. If it is, it indicates that the hard drive is in access mode, and to prevent data write failures, these hard drives will not be powered down by the management decision. Then, the health status of the hard drive is checked, including whether the hard drive is in a faulty or degraded state; therefore, the management decision will prioritize reducing the power consumption of these drives. Finally, the remaining lifespan of the hard drives is checked. A longer remaining lifespan indicates that the disks were previously accessed less frequently, thus giving them higher priority during power consumption limiting. For example, suppose a server has three backplanes, each with 20 hard drives, totaling 60 hard drives. If a power consumption risk exists, the baseboard management controller obtains the hard drive health status through the RAID card to determine which drives have high priority. Then, the baseboard management controller sends a command to the relevant backplane to disconnect the hard drives via electronic fuses. Once the power consumption risk disappears, the baseboard management controller can re-power the hard drives via the electronic fuses, achieving dynamic control.

[0036] In a specific implementation, the lower the hard drive priority value, the earlier the hard drive will be processed. Specifically, the priority is as follows: first priority is for hard drives in a faulty or degraded state, with the activity indicator light off and a high remaining lifespan; second priority is for hard drives in a faulty or degraded state, with the activity indicator light off and a low remaining lifespan; third priority is for hard drives in normal health, with the activity indicator light off and a high remaining lifespan; fourth priority is for hard drives in normal health, with the activity indicator light off and a low remaining lifespan; hard drives with the activity indicator light on will not be powered off.

[0037] Furthermore, the power control operation on the high-speed peripheral component interconnect device based on the maximum output power percentage includes: obtaining the power consumption growth rate of the high-speed peripheral component interconnect device between two adjacent polls; in response to the power consumption growth rate being less than or equal to a preset growth rate threshold, generating a power control command based on the maximum output power percentage and the current output power of the voltage regulation chip corresponding to the high-speed peripheral component interconnect device, and controlling the voltage regulation chip corresponding to the high-speed peripheral component interconnect device to perform the power control operation based on the power control command; and in response to the power consumption growth rate being greater than the preset growth rate threshold, sending a power braking signal to the high-speed peripheral component interconnect device.

[0038] In a specific implementation, this paper adopts The function represents the control mode of the voltage regulator chip's output power. Among them, This is the final approximation value of the curve, which is 1 here; For the required power, and This represents a percentage of the maximum output power. Specifically, when the PCIe device demand is zero, the voltage regulator chip provides basic power output, which in this case is half of the maximum power. At this point, inserting any PCIe device will provide a certain amount of power. When PCIe device demand increases due to business needs, the voltage regulator chip rapidly increases the power supply to quickly respond to the demand. Conversely, when the PCIe device demand approaches a certain threshold, the voltage regulator chip slows down the power output to prevent power overruns.

[0039] Furthermore, the detection of whether the power supply unit in the server meets the preset alarm conditions includes: acquiring the output current value and output power value of the power supply unit; determining whether the output current value is greater than a preset current threshold and whether the output power value is greater than a preset power threshold; determining that the power supply unit meets the preset alarm conditions in response to the output current value being greater than the preset current threshold or the output power value being greater than the preset power threshold; and determining that the power supply unit does not meet the preset alarm conditions in response to the output current value being less than or equal to the preset current threshold and the output power value being less than or equal to the preset power threshold.

[0040] Specifically, when the power supply unit meets the alarm conditions, the above method sends a frequency reduction command to the graphics processor and controls it to actively reduce its frequency according to the preset component list including the graphics processor and the fourth power consumption control strategy, until the power consumption risk of the power supply unit is eliminated, as the last line of defense to prevent the server from losing power due to overcurrent.

[0041] Furthermore, according to the preset fourth power consumption control strategy, the corresponding power consumption control operation is performed on the components in the preset component list, including: obtaining the preset component list, wherein the preset component list includes a graphics processor; sending a frequency reduction instruction to the graphics processor according to the fourth power consumption control strategy; and in response to the graphics processor receiving the frequency reduction instruction, controlling the graphics processor to perform an active frequency reduction operation until the power consumption risk of the power supply unit is eliminated.

[0042] Furthermore, determining the second power consumption control strategy based on the power consumption data of the components also includes: obtaining the percentage of hard drive power consumption in the total system power consumption, and the power consumption growth rate of the graphics processor; responding to the hard drive power consumption percentage being less than a first percentage threshold, returning to the step of obtaining server power consumption data; responding to the hard drive power consumption percentage being greater than or equal to the first percentage threshold, and the power consumption growth rate of the graphics processor being greater than a preset growth rate threshold, generating a second power consumption control strategy corresponding to the high-speed peripheral component interconnect device based on the required power of the high-speed peripheral component interconnect device and the current output power of the voltage regulation chip corresponding to the high-speed peripheral component interconnect device; responding to the hard drive power consumption percentage being greater than or equal to the first percentage threshold, and the power consumption growth rate of the graphics processor being less than or equal to the preset growth rate threshold, generating a second power consumption control strategy corresponding to the hard drive based on the hard drive's indicator light status information, health status information, and remaining lifespan information.

[0043] Specifically, the above method obtains the proportion of hard disk power consumption to total system power consumption and the power consumption growth rate of the graphics processor. When the proportion of hard disk power consumption is less than a first proportion threshold, it returns to re-collect data without calling the hard disk power consumption control scheme. When the proportion of hard disk power consumption meets the standard but the power consumption growth rate of the graphics processing unit exceeds a preset growth rate threshold, it switches to the PCIe device power consumption control scheme. The hard disk power consumption control scheme is only called when the proportion of hard disk power consumption meets the standard and the power consumption growth rate of the graphics processing unit is slow. This avoids the situation where adjusting the hard disk on the server has little effect on data writing, and achieves the selection of the correct control object according to the scenario.

[0044] In specific implementations, since the central processing unit (CPU), graphics processing unit (GPU), and fans account for the vast majority of power consumption on servers, while disk power consumption is negligible, invoking a disk power management decision scheme would not only affect data writing but also have little effect. Therefore, the decision scheme should be invoked based on different scenarios. When the overall hard drive power consumption is less than 5%, the impact of hard drive power consumption need not be considered. When the overall hard drive power consumption is greater than 5%, but the GPU power consumption increase is greater than 3% per polling data cycle, then the problem cannot be solved by invoking the disk power management decision scheme; instead, the power control scheme for high-speed peripheral component interconnect devices should be used directly. If the overall power consumption on the storage server is greater than 10% or even 20%, then a disk power management scheme should be considered.

[0045] Furthermore, based on the correspondence between the silkscreen serial numbers of the high-speed peripheral component interconnect devices and the fan groups, as well as the correspondence between the temperature ranges and speed gradients of the fans within the fan groups, a second power consumption control strategy is generated for each fan. This strategy includes: predicting the first high-speed peripheral component interconnect device and the second high-speed peripheral component interconnect device to be released and activated in the future based on historical power consumption data from the server; determining the target fan group corresponding to the first high-speed peripheral component interconnect device, and determining the theoretical speed of the target fan group based on the correspondence between the temperature ranges and speed gradients of the fans within the target fan group; and responding to the prediction that the heat dissipation demand of the first high-speed peripheral component interconnect device and the power consumption increase demand of the second high-speed peripheral component interconnect device will meet at a target time point, increasing the fan speed of the first high-speed peripheral component interconnect device to a target speed higher than the theoretical speed based on the correspondence between the silkscreen serial numbers of the first high-speed peripheral component interconnect device and the fan groups, so that the first high-speed peripheral component interconnect device will be reduced to the target temperature value before the target time point, thereby releasing the power demand in advance.

[0046] Specifically, the above method predicts the first PCIe device with resources to be released and the second PCIe device with resources to be activated in the future period based on historical power consumption data, determines the target fan group and its theoretical speed corresponding to the first PCIe device, and when it is predicted that the heat dissipation demand of the first PCIe device and the power consumption increase demand of the second PCIe device will meet at the target time point, the fan speed corresponding to the first PCIe device is increased to the target speed higher than the theoretical speed in advance, so that the first PCIe device drops to the target temperature value before the target time point to release the power demand in advance, thereby avoiding the power limitation of the second PCIe device due to the remaining heat dissipation demand of the first PCIe device when the fan speed is increased later.

[0047] In a specific implementation, taking a graphics processing unit (GPU) as an example, a method is used to avoid power peaks by increasing fan speed. Based on historical data from current operations, it's possible to predict which GPUs will be used in the future and which GPU resources will be released. To prevent a situation where older GPUs are still running at high speeds while future GPUs will increase their speeds and power consumption, a power peak will occur at a certain point. To avoid this, the fan speed of the GPUs whose resources have been released can be increased, exceeding the calculated theoretical speed. This allows for rapid cooling, reaching a lower temperature value earlier, and thus releasing power demands sooner. Therefore, when subsequent GPUs increase their fan speeds, they will not experience power limitations due to the cooling needs of the GPUs whose resources have been released.

[0048] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method.

[0049] Embodiments of this application also provide an electronic device, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above-described server power consumption control method embodiments.

[0050] When the program instructions are read and executed by one or more processors, they can also perform operations corresponding to the steps in the above method embodiments, as described above, and will not be repeated here. Reference Figure 5 This exemplifies the architecture of an electronic device, which may include a processor 510, a video display adapter 511, a disk drive 512, an input / output interface 513, a network interface 514, and a memory 520. The processor 510, video display adapter 511, disk drive 512, input / output interface 513, network interface 514, and memory 520 can communicate with each other via a communication bus 530.

[0051] The processor 510 can be implemented using a general-purpose central processing unit (CPU), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solution provided in this application.

[0052] The memory 520 can be implemented as a read-only memory (ROM), random access memory (RAM), static storage device, dynamic storage device, etc. The memory 520 can store an operating system 521 for controlling the operation of the electronic device 500, and a basic input / output system (BIOS) 522 for controlling the low-level operations of the electronic device 500. Additionally, it can store a web browser 523, data storage management 524, and an icon / font processing system 525. The icon / font processing system 525 can be the application program that specifically implements the aforementioned steps in this embodiment. In summary, when implementing the technical solution provided in this application through software or firmware, the relevant program code is stored in the memory 520 and executed by the processor 510.

[0053] Input / output interface 513 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touch screens, microphones, various sensors, etc., and output devices may include displays, speakers, vibrators, indicator lights, etc.

[0054] Network interface 514 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0055] Bus 530 includes a pathway for transmitting information between various components of the device, such as processor 510, video display adapter 511, disk drive 512, input / output interface 513, network interface 514, and memory 520.

[0056] In addition, the electronic device 500 can also obtain information on specific acquisition conditions from the virtual resource object acquisition condition information database (not shown in the figure) for condition judgment.

[0057] It should be noted that although the above-described electronic device 500 only shows a processor 510, a video display adapter 511, a disk drive 512, an input / output interface 513, a network interface 514, a memory 520, and a bus 530, in specific implementations, the electronic device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the solution of this application, and does not necessarily include all the components shown in the figures.

[0058] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause an electronic device (which may be a personal computer, cloud server, or network device, etc.) to execute the methods of various embodiments or some parts of the embodiments of this application.

[0059] Embodiments of this application also provide a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above-described server power consumption control method embodiments when running.

[0060] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0061] The embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above-described server power consumption control method embodiments.

[0062] Embodiments of this application also provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps in any of the above-described server power consumption control method embodiments.

[0063] Any of the components, modules, units, parts, methods, and operations described herein can be implemented using software, firmware, hardware (e.g., fixed logic circuitry), manual processing, or any combination thereof. Alternatively or additionally, any functionality described herein can be performed at least in part by one or more hardware logic components, such as, but not limited to, a central processing unit (CPU), a field-programmable gate array (FPGA), an application-specific integrated circuit (ASIC), an application-specific standard product (ASSP), a system-on-a-chip (SoC), a complex programmable logic device (CPLD), a microprocessor (MCU), etc. The terms "system," "computing device," or "apparatus" as used herein encompass various means, devices, and machines for processing data, including, for example, one or more programmable processors, computers, SoCs, or combinations thereof. The apparatus may also include code that creates an execution environment for the computer program in question, such as code constituting processor firmware, a protocol stack, a database management system, an operating system, a cross-platform runtime environment, a virtual machine, or one or more combinations thereof. The aforementioned computer program (also known as a program, software, software application, app, script, or code) can be written in any form of programming language, including compiled or interpreted languages, declarative or procedural languages, and can be deployed in any form, including as a standalone program or as a module, component, subroutine, object, or other unit suitable for a computing environment.

[0064] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0065] The above provides a detailed description of a server power consumption control method and electronic device provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A server power consumption control method, characterized in that, The method includes: Obtain server power consumption data and determine the validity of the server power consumption data, wherein the server power consumption data includes power consumption data of at least one component; In response to the fact that all the server power consumption data are valid, a first power consumption control strategy is generated based on the server power consumption data, wherein the first power consumption control strategy includes the risk level of the server; If the risk level indicates that there is no power consumption risk, repeat the above steps; In response to the risk level indicating the presence of power consumption risk, a second power consumption control strategy is determined based on the power consumption data of the component, and a corresponding power consumption control operation is performed on the component based on the second power consumption control strategy or a third power consumption control strategy issued by the operating system in the server. In response to the completion of all power consumption control operations, it is detected whether the power supply unit in the server meets the preset alarm conditions. In response to the fulfillment of the preset alarm conditions, according to the preset fourth power consumption control strategy, the corresponding power consumption control operation is performed on the components in the preset component list, and the current server power consumption control ends. If the preset alarm conditions are not met, the above steps are repeated.

2. The server power consumption control method according to claim 1, characterized in that, Determining the validity of the server power consumption data includes: Obtain validity determination information corresponding to the component, wherein the validity determination information includes at least one of the following: link status information of the data acquisition link corresponding to the component, preset power consumption range corresponding to the component, and acquisition timestamp corresponding to the power consumption data of the component; Determine whether the validity determination information meets the preset validity conditions; In response to the validity determination information satisfying the preset validity condition, the power consumption data of the component is determined to be valid; If the validity determination information does not meet the preset validity conditions, the power consumption data of the component is determined to be invalid.

3. The server power consumption control method according to claim 2, characterized in that, The determination of whether the validity information meets the preset validity conditions includes: In response to the validity determination information being the link status information, determine whether the link status information indicates that the data acquisition link is in a normal state. If yes, determine that the power consumption data of the component is valid; otherwise, determine that the power consumption data of the component is invalid. In response to the validity determination information being the preset power consumption range, it is determined whether the power consumption data of the component falls within the preset power consumption range. If yes, the power consumption data of the component is determined to be valid; otherwise, the power consumption data of the component is determined to be invalid. In response to the validity determination information being the collection timestamp, it is determined whether the difference between the collection timestamp and the current time is less than or equal to a preset time threshold. If yes, the power consumption data of the component is determined to be valid; otherwise, the power consumption data of the component is determined to be invalid.

4. The server power consumption control method according to claim 2, characterized in that, The step of determining whether the validity information meets the preset validity conditions further includes: In response to the validity determination information including the link status information, the preset power consumption range, and the collection timestamp, the link status information, the preset power consumption range, and the collection timestamp are judged respectively to obtain the link judgment result, the range judgment result, and the time judgment result; In response to the link determination result, the range determination result, and the time determination result all indicating that the preset valid condition is met, the power consumption data of the component is determined to be valid; If any of the link determination result, the range determination result, and the time determination result indicates that the preset valid condition is not met, the power consumption data of the component is determined to be invalid.

5. The server power consumption control method according to any one of claims 1-4, characterized in that, The method further includes: In response to the presence of invalid data in the server power consumption data, the invalid component corresponding to the invalid data is identified, and the historical power consumption data of the invalid component is obtained; Based on the historical power consumption data, the invalid component is predicted to obtain the predicted power consumption data of the invalid component and the prediction confidence level corresponding to the predicted power consumption data; In response to the prediction confidence being less than a preset confidence threshold, historical peak statistics of the invalid component are obtained, and the first power consumption control strategy is generated based on the historical peak statistics and valid data in the server power consumption data. In response to the prediction confidence being greater than or equal to the preset confidence threshold, the first power consumption control strategy is generated based on the predicted power consumption data and the valid data.

6. The server power consumption control method according to claim 5, characterized in that, The step of generating a first power consumption control strategy based on the server power consumption data includes: Based on the server power consumption data, determine the overall power consumption value of the server; Obtain the maximum supported power of the power supply unit, and determine the risk level of the server based on the total power consumption value and the maximum supported power; Components in the server whose power consumption ratio is greater than or equal to a preset threshold are identified as high-power components, and components in the server whose power consumption ratio is less than the preset threshold and which are currently active or predicted to be active in the future are identified as optimizable components. Obtain the data acquisition link corresponding to the high-power component, and determine the location of the high-power component based on the data acquisition link corresponding to the high-power component; The first power consumption control strategy is generated based on the risk level, the type and location of the high-power components, and the type of the optimizable components.

7. The server power consumption control method according to claim 6, characterized in that, In response to the prediction confidence level being less than the preset confidence threshold, determining the total power consumption of the server based on the server power consumption data includes: The power consumption of the entire machine is determined based on the historical peak statistics and the valid data. In response to the prediction confidence level being greater than or equal to the preset confidence threshold, determining the total power consumption of the server based on the server power consumption data includes: The total power consumption value is determined based on the predicted power consumption data and the valid data.

8. The server power consumption control method according to claim 6, characterized in that, The step of determining the risk level of the server based on the overall power consumption value and the maximum supported power includes: Calculate the power ratio between the total power consumption and the maximum supported power. In response to the power ratio being less than a preset ratio threshold, the risk level of the server is determined to be Level 1, wherein Level 1 is used to characterize that the server does not have power consumption risk; In response to the power ratio being greater than or equal to the preset ratio threshold, the risk level of the server is determined to be Level 2, wherein Level 2 is used to characterize the server as having a power consumption risk.

9. The server power consumption control method according to claim 8, characterized in that, The step of determining the second power consumption control strategy based on the power consumption data of the component includes: Based on the power consumption data of the components, the type and location of the high-power components, and the type of the optimizable components, the control objects in the server are determined; In response to the fact that the control object includes the high-speed peripheral component interconnect device in the server, a second power consumption control strategy corresponding to the high-speed peripheral component interconnect device is generated based on the power demand corresponding to the high-speed peripheral component interconnect device and the current output power of the voltage regulation chip connected to the high-speed peripheral component interconnect device. In response to the fact that the controlled object includes the fan in the server, a second power consumption control strategy corresponding to the fan is generated based on the correspondence between the silkscreen number of the high-speed peripheral component interconnect device and the fan group, and the correspondence between the temperature range and speed gradient of the fan in the fan group. In response to the fact that the controlled object includes the hard drive in the server, a second power consumption control strategy corresponding to the hard drive is generated based on the indicator light status information, health status information and remaining lifespan information of the hard drive.

10. The server power consumption control method according to claim 9, characterized in that, The step of performing corresponding power control operations on the component according to the second power control strategy or the third power control strategy issued by the operating system in the server includes: Determine whether the third power consumption control strategy issued by the operating system has been received; In response to receiving the third power consumption control strategy, the corresponding power consumption control operation is performed on the component according to the third power consumption control strategy; In response to the failure to receive the third power consumption control strategy, the corresponding power consumption control operation is performed on the component according to the second power consumption control strategy.

11. The server power consumption control method according to claim 10, characterized in that, The step of performing a corresponding power control operation on the component according to the second power control strategy includes: In response to the fact that the controlled object includes the high-speed peripheral component interconnect device, the maximum output power percentage of the voltage regulation chip corresponding to the high-speed peripheral component interconnect device is determined according to the power demand corresponding to the high-speed peripheral component interconnect device and the preset nonlinear control relationship, and power control operation is performed on the high-speed peripheral component interconnect device according to the maximum output power percentage. The nonlinear control relationship is used to characterize that the maximum output power percentage first rises rapidly as the power demand increases, then the growth rate slows down and approaches the upper limit value. In response to the fact that the controlled object includes the fan, the fan speed control operation is performed on the fan in the fan group according to the correspondence between the silkscreen number of the high-speed peripheral component interconnect device and the fan group, and the correspondence between the temperature range and speed gradient of the fan in the fan group; In response to the fact that the controlled object includes the hard drive, the control priority of the hard drive is determined according to the indicator light status information, health status information and remaining lifespan information of the hard drive, and the hard drive is powered off according to the control priority. In response to the disappearance of the power consumption risk of the server, the power-on operation is performed on the powered-off hard drive.

12. The server power consumption control method according to claim 11, characterized in that, The step of performing power control operations on the high-speed peripheral component interconnect device based on the maximum output power percentage includes: Obtain the power consumption growth rate of the high-speed peripheral component interconnect device between two consecutive polling sessions; In response to the power consumption growth rate being less than or equal to a preset growth rate threshold, a power control command is generated based on the maximum output power percentage and the current output power of the voltage regulation chip corresponding to the high-speed peripheral component interconnect device, and the voltage regulation chip corresponding to the high-speed peripheral component interconnect device is controlled to perform power control operations based on the power control command. In response to the power consumption growth rate exceeding the preset growth rate threshold, a power braking signal is sent to the high-speed peripheral component interconnect device.

13. The server power consumption control method according to claim 12, characterized in that, The step of detecting whether the power supply unit in the server meets the preset alarm conditions includes: Obtain the output current and output power values ​​of the power supply unit; Determine whether the output current value is greater than a preset current threshold, and determine whether the output power value is greater than a preset power threshold; In response to the output current value being greater than the preset current threshold, or the output power value being greater than the preset power threshold, it is determined that the power supply unit meets the preset alarm condition; In response to the output current value being less than or equal to the preset current threshold and the output power value being less than or equal to the preset power threshold, it is determined that the power supply unit does not meet the preset alarm condition.

14. The server power consumption control method according to claim 13, characterized in that, The step of performing corresponding power control operations on components in a preset component list according to a preset fourth power control strategy includes: Obtain the preset component list, wherein the preset component list includes a graphics processor; According to the fourth power consumption control strategy, a frequency reduction command is sent to the graphics processor; In response to the graphics processor receiving the frequency reduction command, the graphics processor is controlled to perform an active frequency reduction operation until the power consumption risk of the power supply unit is eliminated.

15. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the server power consumption control method as described in any one of claims 1-14 when executing the computer program.

Images