Server quasi-system power consumption determination method and device and storage medium

Abstract

The invention relates to the technical field of servers, and discloses a power consumption determination method and device of a server quasi-system and a storage medium, a load of the server quasi-system under target configuration comprises a CPU and a memory bank, and the method comprises the steps that first power loss of a first power conversion module arranged on a mainboard is determined according to CPU power consumption and a first calculation model; determining first power consumption of the fan according to the rotating speed ratio of the fan and a second calculation model; determining whether the mainboard is provided with a second power conversion module or not, if not, determining second power consumption, and determining the sum value of the first power loss, the first power consumption and the second power consumption as the power consumption of the server quasi-system; and if yes, determining second power loss of the second power conversion module according to the memory bank power and a third calculation model, determining third power consumption, and determining a sum value of the first power loss, the second power loss, the first power consumption and the third power consumption as the power consumption of the server quasi-system. According to the method, the power consumption of the server quasi-system can be quickly determined.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the technical field of server, in particular to a method and device for determining power consumption of a server system and a storage medium. BACKGROUND

[0002] The power consumption of a server is an important technical parameter. Before configuring a power module for a server to provide power to the server, the power consumption of the server is usually evaluated to ensure that the output power of the configured power module can meet the power consumption requirement of the server and has a relatively high power utilization rate.

[0003] A server includes a server system and loads (i.e., accessories), wherein the server system usually only includes a server case, a mainboard, a backboard, a fan component, a power module, a central processing unit (CPU), a memory stick, a hard disk, and the like configured by a user, and the server system is used in combination with the loads. The total power consumption of the server system and all the loads is the overall power consumption of the server. Therefore, when evaluating the overall power consumption of the server, the power consumption of the server system and the power consumption of each load need to be evaluated. The power consumption of the load can usually be determined by the specification of the load, but in many scenarios, the power consumption data of the server system cannot be obtained from the specification of the server. How to quickly determine the power consumption of the server system is a problem to be solved. SUMMARY

[0004] In view of the above problems, embodiments of the present application provide a method and device for determining power consumption of a server system and a storage medium, to solve the problem that the power consumption of the server system cannot be quickly determined in the prior art.

[0005] According to one aspect of the embodiments of this application, a method for determining the power consumption of a server barebone system is provided. The load of the server barebone system under a target configuration includes a central processing unit (CPU) and memory modules. The method includes: determining a first power loss of a first power conversion module installed on the motherboard of the server barebone system based on the CPU power consumption and a preset first calculation model, wherein the first power conversion module is used to convert voltage and provide the converted voltage to the CPU; determining a first power consumption of a fan based on a fan speed ratio in the server barebone system and a preset second calculation model, wherein the speed ratio is the ratio of the fan's real-time speed to its maximum speed; determining whether a second power conversion module is installed on the motherboard, wherein the second power conversion module is used to convert voltage and provide the converted voltage to the memory modules; if the motherboard does not have a second power conversion module... The conversion module determines a second power consumption and uses the sum of the first power loss, the first power consumption, and the second power consumption as the power consumption of the server barebone system under the target configuration. The second power consumption is the total power consumption of the backplane and other devices on the motherboard in the server barebone system, excluding the first power conversion module. If the motherboard has a second power conversion module, the second power loss of the second power conversion module is determined based on the power consumption of the memory module and a preset third calculation model. A third power consumption is also determined, and the sum of the first power loss, the second power loss, the first power consumption, and the third power consumption is used as the power consumption of the server barebone system under the target configuration. The third power consumption is the total power consumption of the backplane and other devices on the motherboard, excluding the first and second power conversion modules.

[0006] In one alternative approach, determining the first power loss of the first power conversion module installed on the motherboard of the server barebone system based on the CPU's power consumption and a preset first calculation model includes: determining the first power loss of the first power conversion module installed on the motherboard of the server barebone system based on the CPU's power consumption P. cpu Based on the first calculation model, determine the first voltage conversion efficiency θ of the first power conversion module. cpu ; By formula PL1=P cpu / θ cpu -P cpu Determine the first power loss PL1.

[0007] In one alternative approach, the power consumption P of the CPU is... cpu Based on the first calculation model, determine the first voltage conversion efficiency θ of the first power conversion module. cpu This includes: through the first calculation model θ cpu =k0+k1×P cpu +k2×Pcpu 2 determining the first voltage conversion efficiency θ cpu wherein k0, k1 and k2 are constants.

[0008] In an alternative way, k0, k1 and k2 are determined by the following steps: controlling a server to work in a first working state and obtaining a first group of data, wherein the server comprises the server system, a power module, a test CPU, a test solid state disk and a test memory bar, in the first working state, the rotation speed of the fan is a first rotation speed and the utilization rate of the test CPU is a first utilization rate, the first group of data comprises a first output power consumption of the power module and a first CPU power consumption of the test CPU; controlling the server to work in a second working state and obtaining a second group of data, wherein in the second working state, the rotation speed of the fan is the first rotation speed and the utilization rate of the test CPU is a second utilization rate, the second group of data comprises a second output power consumption of the power module and a second CPU power consumption of the test CPU; controlling the server to work in a third working state and obtaining a third group of data, wherein in the third working state, the rotation speed of the fan is the first rotation speed and the utilization rate of the test CPU is a third utilization rate, the third group of data comprises a third output power consumption of the power module and a third CPU power consumption of the test CPU; controlling the server to work in a fourth working state and obtaining a fourth group of data, wherein in the fourth working state, the rotation speed of the fan is the first rotation speed and the utilization rate of the test CPU is a fourth utilization rate, the fourth group of data comprises a fourth output power consumption of the power module and a fourth CPU power consumption of the test CPU; controlling the server to work in a fifth working state and obtaining a fifth group of data, wherein in the fifth working state, the rotation speed of the fan is the first rotation speed and the utilization rate of the test CPU is a fifth utilization rate, the fifth group of data comprises a fifth output power consumption of the power module and a fifth CPU power consumption of the test CPU; substituting the first group of data, the second group of data, the third group of data, the fourth group of data and the fifth group of data into the formulas P out = P cpu ’ / θ cpu ’+ P 其他 and θ cpu ’= k0+k1×P cpu ’+ k2×P cpu ’ 2 respectively, to determine k0, k1 and k2, wherein P out is the output power consumption of the power module in each group of data, P cpu ’ is the power consumption of the test CPU in each group of data, θ cpu'P' represents the voltage conversion efficiency of the first power conversion module of the server in each operating state. 其他 This refers to the total power consumption of all devices in the server other than the test CPU.

[0009] In one alternative approach, the second computational model is P. 风扇 = a + b × η + c × η 2 +d×η 3 , where P 风扇 Let η be the first power consumption, η be the rotational speed ratio, and a, b, c, and d be constants.

[0010] In one alternative approach, a, b, c, and d are determined through the following steps: The control server operates in a sixth working state and acquires a sixth set of data, wherein the server includes the server barebone system, a power module, a test CPU, a test solid-state drive, and a test memory module. In the sixth working state, the fan speed is a second speed, and the test CPU and the test memory module are in a first state. The sixth set of data includes the sixth output power consumption of the power module. The control server operates in a seventh working state and acquires a seventh set of data, wherein in the seventh working state, the fan speed is a third speed, and the test CPU and the test memory module are in the first state. The seventh set of data includes the seventh output power consumption of the power module. The control server operates in an eighth working state and acquires an eighth set of data, wherein in the eighth working state, the... The fan speed is the fourth speed and the test CPU and the test memory module are in the first state. The eighth set of data includes the eighth output power consumption of the power module. The control server operates in the ninth working state and acquires the ninth set of data. In the ninth working state, the fan speed is the fifth speed and the test CPU and the test memory module are in the first state. The ninth set of data includes the ninth output power consumption of the power module. The control server operates in the tenth working state and acquires the tenth set of data. In the tenth working state, the fan speed is the sixth speed and the test CPU and the test memory module are in the first state. The tenth set of data includes the tenth output power consumption of the power module. Substitute the sixth set of data, the seventh set of data, the eighth set of data, the ninth set of data, and the tenth set of data into formula P. out '=P 风扇 '+P 其他 'and P 风扇 = a + b × η' + c × η' 2 +d×η' 3 To determine a, b, c, and d, where P out 'P' represents the output power consumption of the power module in each data set.风扇 ' is the power consumption of the fan, P 其他 ' represents the total power consumption of all components in the server other than the fan, and η' represents the ratio of the real-time speed of the fan to its maximum speed in each operating state.

[0011] In one alternative approach, determining the second power loss of the second power conversion module based on the power consumption P of the memory module includes: based on the power consumption P of the memory module... 内存条 and the third calculation model θ 内存条 =k3+k4×P 内存条 +k5×P 内存条 2 Determine the second voltage conversion efficiency θ of the second power conversion module. 内存条 Where k3, k4, and k5 are all constants; the formula PL2 = P 内存条 / θ 内存条 -P 内存条 Determine the second power loss PL2.

[0012] In an optional manner, k3, k4, and k5 are determined through the following steps: The control server operates in an eleventh working state and acquires an eleventh set of data, wherein the server includes the server barebone system, a power module, a test CPU, a test solid-state drive, and a test memory module. In the eleventh working state, the utilization rate of the test CPU is the sixth utilization rate, the fan speed is the first speed, and the utilization rate of the test memory module is the seventh utilization rate. The eleventh set of data includes the eleventh output power consumption of the power module and the first memory power consumption of the test memory module. The control server operates in a twelfth working state and acquires a twelfth set of data, wherein in the twelfth working state, the utilization rate of the test CPU is the sixth utilization rate, the fan speed is the first speed, and the utilization rate of the test memory module is the eighth utilization rate. The twelfth set of data includes the twelfth output power consumption of the power module and the second memory power consumption of the test memory module. The control server operates in a thirteenth working state and acquires a thirteenth set of data, wherein in the thirteenth working state, the utilization rate of the test CPU is the sixth utilization rate, the fan speed is the first speed, and the utilization rate of the test memory module is the eighth utilization rate. The twelfth set of data includes the twelfth output power consumption of the power module and the second memory power consumption of the test memory module. The fan speed is the first speed and the utilization rate of the test memory module is the ninth utilization rate. The thirteenth set of data includes the thirteenth output power consumption of the power module and the third memory power consumption of the test memory module. The control server operates in the fourteenth working state and acquires the fourteenth set of data. In the fourteenth working state, the utilization rate of the test CPU is the sixth utilization rate, the fan speed is the first speed and the utilization rate of the test memory module is the tenth utilization rate. The fourteenth set of data includes the fourteenth output power consumption of the power module and the fourth memory power consumption of the test memory module. The control server operates in the fifteenth working state and acquires the fifteenth set of data. In the fifteenth working state, the utilization rate of the test CPU is the sixth utilization rate, the fan speed is the first speed and the utilization rate of the test memory module is the eleventh utilization rate. The fifteenth set of data includes the fifteenth output power consumption of the power module and the fifth memory power consumption of the test memory module. Substitute the eleventh set of data, the twelfth set of data, the thirteenth set of data, the fourteenth set of data and the fifteenth set of data into formula P. out =P 内存条 ' / θ 内存条 '+P 其他 "and θ 内存条 =k3+k4×P 内存条 '+k5×P 内存条 ' 2 To determine k3, k4, and k5, where P out "P represents the output power consumption of the power module in each data set." 内存条P is the power consumption of the test memory stick in each group of data, θ 内存条 P is the voltage conversion efficiency of the second power conversion module of the server in each working state, P 其他 P is the total power consumption of the devices in the server other than the test memory stick.

[0013] According to another aspect of the embodiments of the present application, there is provided a power consumption determination device of a server system, comprising a memory, a processor and a computer program stored in the memory, the processor executes the computer program to implement the power consumption determination method of the server system as described above.

[0014] According to still another aspect of the embodiments of the present application, there is provided a computer readable storage medium having a computer program stored thereon, the computer program is executed by a processor to implement the power consumption determination method of the server system as described above.

[0015] In the embodiments of the present application, when it is needed to determine the maximum power consumption of the server system under the target configuration, so as to determine the maximum power consumption of the server based on the maximum power consumption of the server system, and then design the power supply and heat dissipation system for the server based on the maximum power consumption of the server, for the server system whose mainboard is not provided with the second power conversion module, the maximum power consumption of the CPU is determined through the specification book of the CPU, and the first power loss of the first power conversion module can be quickly determined based on the maximum power consumption of the CPU and the first calculation model, and the first power consumption of the fan can be quickly determined based on the second calculation model, and after the total power consumption (second power consumption) of the backboard and the devices other than the first power conversion module provided on the mainboard is determined, the power consumption of the server system under the target configuration can be determined based on the first power loss, the first power consumption and the second power consumption. For the server system whose mainboard is provided with the second power conversion module, the maximum power consumption of the memory stick is determined through the specification book of the memory stick, and the second power loss of the second power conversion module can be quickly determined based on the maximum power consumption of the memory stick and the third calculation model, and after the total power consumption (third power consumption) of the backboard and the devices other than the first power conversion module and the second power conversion module provided on the mainboard is determined, the power consumption of the server system under the target configuration can be determined based on the first power loss, the second power loss, the first power consumption and the third power consumption. Through the above-mentioned manner, the efficiency of determining the power consumption of the server system is improved.

[0016] Similarly, for the power consumption of the server system when the power consumption of the CPU is less than the maximum power consumption, the power consumption of the memory stick is less than the maximum power consumption, and the rotating speed of the fan is less than the maximum rotating speed, the power consumption of the server system under the target configuration can also be quickly determined through the above-mentioned manner.

[0017] The above description is only a summary of the technical solutions of the embodiments of the present application. In order to make the technical means of the embodiments of the present application more clearly understood, and to be implemented according to the content of the description, and in order to make the above and other purposes, characteristics and advantages of the embodiments of the present application more obvious and easy to understand, the following specific embodiments of the present application are described. BRIEF DESCRIPTION OF DRAWINGS

[0018] The accompanying drawings are included to provide a further understanding of the present application, and are incorporated herein and constitute a part of the detailed description. It should be noted that in the accompanying drawings, the same or similar elements are denoted by the same reference numerals. In the drawings:

[0019] Figure 1 A flowchart of a method for determining power consumption of a server system is shown;

[0020] Figure 2 A structural diagram of a device for determining power consumption of a server system is shown. DETAILED DESCRIPTION

[0021] The exemplary embodiments of the present application will be described more fully hereinafter with reference to the accompanying drawings. While the exemplary embodiments of the present application are shown in the drawings, it is understood that the present application can be embodied in many forms and should not be construed as limited to the embodiments set forth herein.

[0022] The power consumption of a server system is a key to the evaluation of the overall power consumption. How to quickly and conveniently determine the power consumption of a server system under different load powers is one of the important problems faced by server users. For a manufacturer of a production server, it is usually required to provide the maximum power consumption of the server to the customer, so that the customer can design the power supply and heat dissipation system for the server based on the maximum power consumption of the server. However, different customers may have different load requirements for the server system. When different loads are sold to different customers using the same server system, if the power consumption of the server is determined by respectively controlling the server to work and measuring the power consumption of the server in the working state, the efficiency of this method is low. Alternatively, for a user, after replacing or adding or reducing the load of the server system, if the power consumption of the server in the working state is measured by the above method every time, the efficiency is also low.

[0023] Since the server system includes the server chassis, the mainboard, the backboard and the fan, the power consumption of the mainboard, the backboard and the fan is determined respectively, and the power consumption of the server system is determined. Generally, the first power conversion module for providing working voltage for the CPU is arranged on the mainboard, and the first power conversion module provides the converted voltage to the CPU after converting the voltage, and the converted voltage is the working voltage of the CPU. When the first power conversion module converts the voltage, power loss is usually generated, and the power loss is part of the mainboard power consumption, and the power loss is related to the CPU power consumption, that is, if the CPU power consumption is different, the power loss is also different.

[0024] Currently, two types of mainboards are usually included in the server, one is the mainboard provided with the second power conversion module for providing working voltage for the memory stick (for example, the mainboard not supporting DDR5 type memory), and the other is the mainboard not provided with the power conversion module (for example, the mainboard supporting DDR5 type memory). The second power conversion module is similar to the first power conversion module, and for the mainboard provided with the second power conversion module, the power loss generated by the second power conversion module when converting the voltage is also part of the mainboard power consumption.

[0025] In addition to the power conversion module, other devices such as the chip of the baseboard management controller (BMC) are arranged on the mainboard, and the power consumption of these devices does not change with the power consumption of the CPU and the memory stick. The backboard power consumption includes the hard disk power consumption and the backboard itself power consumption, and the backboard power consumption basically does not change with the hard disk power consumption, so the power consumption of the backboard is basically constant. The fan power consumption is usually only related to the fan speed.

[0026] In summary, the power consumption of the server system is not fixed, and when the same server system is configured with different loads, the power consumption of the server system will also be different.

[0027] For the first power conversion module mentioned above, the corresponding first power consumption loss is the difference between the input power consumption and the output power consumption of the first power conversion module. Since the first power conversion module is used to provide power supply for the CPU, the output power consumption of the first power conversion module is the power consumption of the CPU.

[0028] The inventor of the present application knows that the voltage conversion efficiency of the first power conversion module has a certain relationship with the output power, and the ratio of the output power to the voltage conversion efficiency is the input power of the first power conversion module. Therefore, after the relationship between the voltage conversion efficiency and the output power of the first power conversion module is determined in advance, if the server system needs to configure a CPU, the maximum input power of the first power conversion module can be determined according to the CPU maximum power consumption in the CPU specification, and then the first power consumption loss is determined.

[0029] The second power conversion module is similar to the first power conversion module, so the way to determine the maximum power consumption loss of the second power conversion module can refer to the way to determine the maximum power consumption loss of the first power conversion module, which will not be described here.

[0030] The fan power consumption is usually only related to the fan speed, and the fan power consumption usually has a certain relationship with the fan speed.

[0031] Based on this, the present application proposes a power consumption determination method of a server system. By pre-determining the relationship between the power output efficiency and the output power of the first power conversion module corresponding to the first power conversion module, and the maximum power consumption of the CPU configured by the server system, the maximum input power of the first power conversion efficiency can be quickly determined, and then the maximum power loss of the first power conversion module is determined. Similarly, for the mainboard provided with the second power conversion module, by pre-determining the relationship between the voltage conversion efficiency and the output power of the second power conversion module corresponding to the second power conversion module, and the maximum power consumption of the memory bank configured by the server system, the maximum input power of the second power conversion efficiency can be quickly determined, and then the maximum power loss of the second power conversion module is determined. Then the total power consumption of other devices on the mainboard and the backboard is determined, and by pre-determining the relationship between the fan power consumption and the fan speed, the maximum power consumption of the fan can be quickly determined by the relationship. Finally, the sum of the power loss of the power conversion module, the total power consumption of other devices on the mainboard and the backboard, and the maximum power consumption of the fan is determined as the maximum power consumption of the server system, thereby improving the efficiency of determining the power consumption of the server system.

[0032] Figure 1A flowchart of a method for determining power consumption of a server system is shown. The method is performed by a terminal device, which can be a terminal device including one or more processors, such as a touch phone, a smart phone, a tablet computer, a portable electronic device, or other electronic devices. The processor can be a CPU, or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present application, without limitation. The one or more processors included in the terminal device can be the same type of processor, such as one or more CPUs; or different types of processors, such as one or more CPUs and one or more ASICs, without limitation. As shown in FIG. 10, the method includes the following steps. Figure 1

[0033] Step 110: determining a first power loss of a first power conversion module disposed on a motherboard in the server system according to the power consumption of the CPU and a preset first calculation model.

[0034] As described above, for the same server system, if different loads are configured, the power consumption of the server system will also be different. For ease of introduction, the present application takes determining the power consumption of the server system under a target configuration as an example for introduction, wherein the load of the server system under the target configuration includes a CPU and a memory stick.

[0035] As described above, the first power conversion module is configured to convert a voltage and provide the converted voltage to the CPU, and the converted voltage provided by the first power conversion module to the CPU is the operating voltage of the CPU.

[0036] The first calculation model is a relationship between the voltage conversion efficiency of the first power conversion module and the output power thereof. According to the CPU chip data and test results, it is found that the output efficiency of the first power conversion module increases in a parabolic relationship with the output power, and therefore, the first calculation model is a quadratic polynomial between the first voltage conversion efficiency of the first power conversion module and the output power. Specifically, in embodiments of the present application, the first calculation model is the following formula (1).

[0037] θ cpu = k0+ k1x P cpu + k2x P cpu 2 (1)

[0038] wherein θ cpu is the first voltage conversion efficiency, and P cpu ​The k0, k1 and k2 are constants for the power consumption of the CPU. The specific value determination method of the k0, k1 and k2 will be described below.

[0039] In the embodiment of the present application, if the maximum power consumption of the server system under the target configuration needs to be determined, the maximum power consumption of the CPU in the load included in the server system under the target configuration is taken as P cpu Substituting the above formula (1), the corresponding first voltage conversion efficiency θ cpu .

[0040] Then, the first power loss PL1 is determined by the following formula (2).

[0041] PL1 = P cpu / θ cpu - P cpu (2)

[0042] Wherein, P cpu is the maximum power consumption of the CPU, and P cpu / θ cpu is the maximum power output of the first power conversion module.

[0043] Since the maximum power consumption P cpu of the CPU in the load included in the server system under the target configuration can be directly obtained from the CPU specification, by substituting the maximum power consumption of the CPU as P cpu into the above formula (1) and formula (2), the maximum power loss (i.e. the first power loss PL1) of the first power conversion module can be quickly and accurately determined.

[0044] The power consumption of the CPU will change with the change of its computing resource utilization. Therefore, in order to ensure the stable operation of the server, it is usually necessary to determine the maximum power consumption it can reach, and then design the power supply and heat dissipation system based on the maximum power consumption, so as to ensure that the power supply can still supply power stably and the heat dissipation system can effectively dissipate heat in time even in the case of the highest load and the maximum power consumption of the server, so as to maintain the normal work of the server. Therefore, the above method determines the first power loss PL1 by using the maximum power consumption of the CPU, so as to determine the maximum power loss of the first power conversion module, so that the maximum power consumption of the server system under the target configuration can be determined based on the maximum power loss subsequently.

[0045] However, for the case that the application scenario of the server is single (e.g. special computing task) and the CPU power consumption is stable, if the CPU power consumption is less than the maximum power consumption, the actual power consumption of the CPU can be substituted into the above formula (1) and formula (2) to determine the first power loss of the first power conversion module in actual application, and then the actual power consumption of the server in the working state is determined according to the first power loss, so as to design the power supply and the heat dissipation system based on the actual power consumption of the server, avoid the resource waste caused by the design based on the maximum power consumption of the server (e.g. configure a power module with too large power supply capacity or design a heat dissipation system with too strong heat dissipation capacity), and thus optimize the cost and energy efficiency under the premise of meeting the performance requirement.

[0046] It is worth noting that if the load of the server system under the target configuration includes multiple CPUs, the power loss of the power conversion module corresponding to each CPU can be determined by the above method, and finally the sum of the power loss of the power conversion module of all the CPUs is used to determine the power consumption of the server system.

[0047] As introduced above, the power consumption of the fan is related to the rotation speed thereof. For the convenience of introduction, the ratio of the real-time rotation speed of the fan to the maximum rotation speed is defined as the rotation speed ratio in the following.

[0048] Step 120: determining the first power consumption of the fan according to the rotation speed ratio of the fan in the server system and the preset second calculation model.

[0049] The second calculation model is the relationship between the power consumption of the fan and the rotation speed ratio. According to the relevant data and test results of the fan, it is found that the power consumption of the fan and the rotation speed ratio present a cubic polynomial relationship, and therefore the second calculation model in the present application is a cubic polynomial between the power consumption of the fan and the rotation speed ratio of the fan. Specifically, in the embodiments of the present application, the second calculation model is the following formula (3).

[0050] P 风扇 = a + b x η + c x η 2 + d x η 3 (3)

[0051] Wherein, P 风扇 is the first power of the fan, η is the rotation speed ratio of the fan, and a, b, c and d are all constants. The specific value determination method of a, b, c and d will be introduced in detail below.

[0052] In the embodiments of the present application, if it is necessary to determine the maximum power consumption of the server system under the target configuration, the maximum power consumption of the fan should be determined, and then η is directly substituted into the above formula (3).

[0053] As the first power conversion module, for the application scenario of the server is relatively single (such as dedicated computing task) and the rotating speed of the fan is relatively fixed, if the rotating speed of the fan is less than the maximum rotating speed, the ratio of the actual rotating speed of the fan to the maximum rotating speed can be substituted into the above formula (3) to determine the first power consumption of the fan in actual application, and then the actual power consumption of the server in the working state is determined according to the first power consumption, so as to design the power supply based on the actual power consumption of the server, and avoid resource waste caused by designing according to the maximum power consumption of the server. For example, if the rotating speed of the fan is fixedly set to 60% of the maximum rotating speed, then 60% of η is substituted into the above formula (3).

[0054] Step 130: Determine whether the second power conversion module is arranged on the mainboard. If not, go to step 140; if yes, go to step 160.

[0055] As introduced before, the second power conversion module is used for converting voltage and providing the converted voltage to the memory stick, and the converted voltage provided by the second power conversion module to the memory stick is the working voltage of the memory stick.

[0056] Step 140: Determine the second power consumption.

[0057] Among them, since the second power conversion module is not arranged on the mainboard, there is no power loss of the second power conversion module, so the second power consumption is the total power consumption of the backboard and other devices arranged on the mainboard except the first power conversion module in the server system.

[0058] Specifically, the mainboard is arranged with the first power conversion module and the chip such as BMC, the power consumption of such chip does not change with the change of the load of the server system, and the total power consumption of such chip is P 其他器件 . And since the backboard power consumption P 背板 does not change with the change of the load power consumption, it is basically constant. Therefore, in this step, the second power consumption P2 can be determined by the following formula (4).

[0059] P2=P 其他器件 +P 背板 (4)

[0060] Step 150: Determine the sum of the first power loss, the first power consumption and the second power consumption as the power consumption of the server system under the target configuration.

[0061] Among them, for the mainboard of the server system under the target configuration, if the second power conversion module is not arranged, the power consumption of the server system under the target configuration is the sum of the first power loss, the first power consumption and the second power consumption.

[0062] Step 160: Determine the second power loss of the second power conversion module based on the power consumption of the memory module and the preset third calculation model.

[0063] This step is similar to step 110.

[0064] In this step, the third calculation model is the relationship between the voltage conversion efficiency of the second power conversion module and its output power. Based on memory module data and test results, it was found that the output efficiency of the second power conversion module exhibits a parabolic relationship with increasing output power. Therefore, the third calculation model is a quadratic polynomial between the second voltage conversion efficiency and the output power of the second power conversion module. Specifically, in this embodiment, the first calculation model is the following formula (5).

[0065] θ 内存条 =k3+k4×P 内存条 +k5×P 内存条 2 (5)

[0066] Where, θ 内存条 For the second voltage conversion efficiency, P 内存条 For the power consumption of the memory module, k3, k4, and k5 are all constants. The specific methods for determining the values ​​of k3, k4, and k5 will be explained in detail below.

[0067] In this embodiment of the application, if it is necessary to determine the maximum power consumption of the server barebone system under the target configuration, the maximum power consumption of the memory modules in the load included in the server barebone system under the target configuration is taken as P. 内存条 Substituting into formula (5) above, the corresponding second voltage conversion efficiency θ can be determined. 内存条 .

[0068] The second power loss PL2 can then be determined using the following formula (6).

[0069] PL2 = P 内存条 / θ 内存条 -P 内存条 (6)

[0070] Among them, P 内存条 When P is the maximum power consumption of the memory module, 内存条 / θ 内存条 This is the maximum power output of the second power conversion module.

[0071] Due to the maximum power consumption P of the memory modules in the load included in the server barebone system under the target configuration. 内存条 This information can be obtained directly from the memory module's specifications; therefore, the maximum power consumption of this memory module can be used as P. 内存条Substitute the above formula (5) and formula (6), that is, the maximum power loss of the second power conversion module (i.e. the second power loss PL2) can be quickly and accurately determined.

[0072] Similar to the CPU, the power consumption of the memory bank changes with the change of its resource utilization. Therefore, in order to ensure the stable operation of the server, it is usually necessary to determine the maximum power consumption that it can reach, and then design the power supply and heat dissipation system based on the maximum power consumption, so as to ensure that the power supply can still be stably powered even in the case of the highest server load and the maximum power consumption, and the heat dissipation system can also effectively dissipate heat in time to maintain the normal work of the server. Therefore, the above-mentioned method uses the maximum power consumption of the memory bank to determine the second power loss PL2 to determine the maximum power loss of the second power conversion module, so that the maximum power consumption of the server system under the target configuration can be determined based on the maximum power loss subsequently.

[0073] However, for the case where the application scenario of the server is relatively single (such as a dedicated computing task) and the power consumption of the memory bank is relatively stable, if the power consumption of the memory bank is less than its maximum power consumption, the actual power consumption of the memory bank can be substituted into the above-mentioned formula (5) and formula (6) to determine the second power loss of the second power conversion module in actual application, and then the actual power consumption of the server in the working state is determined according to the second power loss, so as to design the power supply and heat dissipation system based on the actual power consumption of the server, avoiding the waste of resources caused by designing according to the maximum power consumption of the server.

[0074] It is worth noting that if the load of the server system under the target configuration includes multiple memory banks, the power loss of the power conversion module corresponding to each memory bank can be determined by the above-mentioned method, and finally the sum of the power losses of the power conversion modules of all the memory banks is used to determine the power consumption of the server system.

[0075] Step 170: Determine the third power consumption.

[0076] The third power consumption is the total power consumption of the backboard and other devices disposed on the mainboard except the first power conversion module and the second power conversion module.

[0077] Step 180: Determine the sum of the first power loss, the second power loss, the first power consumption and the third power consumption as the power consumption of the server system under the target configuration.

[0078] The steps 170-180 are similar to the steps 140-150, and therefore the principles and implementation manners of the steps 170-180 can be referred to the steps 140-150, which will not be described herein again.

[0079] In the embodiments of the present application, when it is needed to determine the maximum power consumption of the server system under the target configuration, so as to determine the maximum power consumption of the server based on the maximum power consumption of the server system, and then design the power supply and heat dissipation system for the server based on the maximum power consumption of the server, for the server system without the second power conversion module on the mainboard, the maximum power consumption of the CPU is determined through the specification book of the CPU, and the first power loss of the first power conversion module can be quickly determined based on the maximum power consumption of the CPU and the first calculation model, and the first power consumption of the fan can be quickly determined based on the second calculation model, and after the second power consumption is determined, the power consumption of the server system under the target position can be determined based on the first power loss, the first power consumption and the second power consumption. For the server system with the second power conversion module on the mainboard, the maximum power consumption of the memory bank is determined through the specification book of the memory bank, and the second power loss of the second power conversion module can be quickly determined based on the maximum power consumption of the memory bank and the third calculation model, and after the third power consumption is determined, the power consumption of the server system under the target position can be determined based on the first power loss, the second power loss, the first power consumption and the third power consumption. Through the above-mentioned manner, the efficiency of determining the power consumption of the server system is improved.

[0080] Similarly, for the power consumption of the server system when it is needed to determine the power consumption of the CPU to be a certain power consumption less than the maximum power consumption, the power consumption of the memory bank to be a certain power consumption less than the maximum power consumption, and the rotating speed of the fan to be a certain rotating speed less than the maximum rotating speed, the power consumption of the server system under the target configuration can also be quickly determined through the above-mentioned manner.

[0081] In order to quickly and accurately determine k0, k1 and k2 in the first calculation model, in the embodiments of the present application, k0, k1 and k2 are determined through steps a1-a6 as follows.

[0082] Step a1: control the server to work in the first working state and obtain the first group of data.

[0083] The server includes a server system, a power module, a test CPU, a test solid state disk and a test memory bar. The server system included in the server is a server system whose power consumption under a target configuration needs to be determined. The power module is used to provide working voltage for the server. In the embodiment of the application, the power module is a power supply unit (PC Powersupply unit, PSU). The test CPU and the test memory bar can be the same as or different from the CPU and the memory bar included in the server system under the target configuration. In the embodiment of the application, the test CPU can be determined based on thermal design power (Thermal Design Power, TDP). Preferably, the test CPU is a CPU with a maximum power of TDP power, so as to ensure that the test data can cover all CPU power loads. For example, if the Whitley platform is used, the test CPU can be a CPU with a maximum power of 270 W; if the EGS platform is used, the test CPU can be a CPU with a maximum power of 350 W.

[0084] In the first working state, the rotation speed of the fan is the first rotation speed and the utilization rate of the test CPU is the first utilization rate. The first rotation speed can be set as needed, for example, the first rotation speed can be set as the maximum rotation speed. The first utilization rate can be zero, that is, the test CPU is in an IDLE state. The first group of data includes the first output power consumption of the power module and the first CPU power consumption of the test CPU. Specifically, when the server is in the IDLE state, the fan is set to 100% rotation speed by the BMC, and the output power consumption P out0 and the power consumption P cpu0 of the test CPU are obtained by using an intelligent platform management interface (Intelligent Platform Management Interface, IPMI) tool.

[0085] Step a2: controlling the server to work in a second working state and obtaining a second group of data.

[0086] In the second working state, the rotation speed of the fan is the first rotation speed and the utilization rate of the test CPU is the second utilization rate, and the second group of data includes the second output power consumption of the power module and the second CPU power consumption of the test CPU. The second utilization rate is greater than the first utilization rate, for example, the second utilization rate is 25%.

[0087] Specifically, the utilization rate of the test CPU is set to 25% by using a Stress tool, and the corresponding PSU output power consumption P out25％ and the test CPU power consumption P cpu25％Stress is a stress testing tool of Linux, which is mainly used to simulate the scene when the system load is high, and is used for stress testing of CPU, IO, memory, load, disk, etc. of the system. Stress tool can simulate different CPU utilization, memory utilization, etc.

[0088] Step a3: the control server works in the third working state and obtains a third set of data.

[0089] In the third working state, the fan speed is the first speed and the test CPU utilization is a third utilization, the third utilization is greater than the second utilization, for example, the third utilization is 50%.

[0090] This step is similar to step a2, so the specific implementation of this step can refer to step a2, which will not be repeated here. In this step, the obtained third set of data includes the PSU output power P out50％ and the test CPU power P cpu50％ .

[0091] Step a4: the control server works in the fourth working state and obtains a fourth set of data.

[0092] In the fourth working state, the fan speed is the first speed and the test CPU utilization is a fourth utilization, the fourth utilization is greater than the third utilization, for example, the fourth utilization is 75%.

[0093] This step is similar to step a2, so the specific implementation of this step can refer to step a2, which will not be repeated here. In this step, the obtained fourth set of data includes the PSU output power P out75％ and the test CPU power P cpu75％ .

[0094] Step a5: the control server works in the fifth working state and obtains a fifth set of data.

[0095] In the fifth working state, the fan speed is the first speed and the test CPU utilization is a fifth utilization, the fifth utilization is greater than the fourth utilization, for example, the fifth utilization is 100%.

[0096] This step is similar to step a2, so the specific implementation of this step can refer to step a2, which will not be repeated here. In this step, the obtained fifth set of data includes the PSU output power P out100％ and the test CPU power P cpu100％ .

[0097] Step a6: substitute the first set of data, the second set of data, the third set of data, the fourth set of data and the fifth set of data into the following formula (7) and formula (8) respectively, and determine k0, k1 and k2.

[0098] P out = P cpu ' / θ cpu ' + P 其他 (7)

[0099] θ cpu ' = k0 + k1 x P cpu ' + k2 x P cpu ' 2 (8)

[0100] In the above formula (7) and formula (8), P 其他 , k0, k1 and k2 are unknown coefficients; P 其他 is the total power consumption of the other devices in the server except the test CPU; θ cpu ' is the voltage conversion efficiency of the first power conversion module of the server in each working state. For convenience, θ cpu0 ', θ cpu25％ ', θ cpu50％ ', θ cpu75％ ' and θ cpu100％ ' represent the voltage conversion efficiency of the first power conversion module of the server in the first working state to the fifth working state, respectively.

[0101] In this step, specifically, the first group of data P out0 and P cpu0 obtained in step a1 are substituted into P out and P cpu in formula (7) and formula (8) respectively to obtain a group of equations, in which only θ cpu0 ', P 其他 , k0, k1 and k2 are unknown numbers. Similarly, the second group of data to the fifth group of data obtained in steps a2 to a5 are substituted into P out and P cpu in formula (7) and formula (8) respectively to obtain four groups of equations. Finally, the five groups of equations obtained are combined to determine the specific values of k0, k1 and k2, so that formula (1) can be determined.

[0102] In the embodiment of the application, the server is controlled to work in different working states, and in different working states, only the utilization rate of the test CPU is different, that is, only the power consumption of the test CPU in the server is a variable, and the power consumption of the rest of the devices is fixed. In this way, only one variable is controlled, and multiple groups of data are obtained, so that the specific values of k0, k1 and k2 can be quickly and accurately determined according to the multiple groups of data.

[0103] It is worth mentioning that in order to quickly determine k0, k1 and k2, the symbolic solution of the Solve function of the Matlab software can also be used to determine k0, k1 and k2.

[0104] In order to quickly and accurately determine a, b, c and d in the second calculation model, in the embodiment of the application, a, b, c and d are determined by the following steps b1 to step b6.

[0105] Step b1: the control server works in the sixth working state and acquires the sixth group of data.

[0106] In the sixth working state, the fan speed is the second speed and the test CPU and the test memory bar are in the first state. The second speed can be determined as needed, for example, the second speed is 10% (i.e. 10% of the maximum speed). The first state is the IDLE state, that is, the test CPU and the test memory bar are in the IDLE state. The sixth group of data includes the output power P out0 ’.

[0107] Step b2: the control server works in the seventh working state and acquires the seventh group of data.

[0108] In the seventh working state, the fan speed is the third speed and the test CPU and the test memory bar are in the first state. The third speed is greater than the second speed, for example, the third speed is 30% (i.e. 30% of the maximum speed). The seventh group of data includes the seventh output power P out30％ ’.

[0109] Step b3: the control server works in the eighth working state and acquires the eighth group of data.

[0110] In the eighth working state, the fan speed is the fourth speed and the test CPU and the test memory bar are in the first state. The fourth speed is greater than the third speed, for example, the fourth speed is 50% (i.e. 50% of the maximum speed). The eighth group of data includes the eighth output power P out50％ ’.

[0111] Step b4: the control server works in the ninth working state and acquires the ninth group of data.

[0112] In the ninth working state, the fan speed is the fifth speed and the test CPU and the test memory bar are in the first state. The fifth speed is greater than the fourth speed, for example, the fifth speed is 70% (i.e. 70% of the maximum speed). The ninth group of data includes the ninth output power P out70％ ’.

[0113] Step b5: the control server works in the tenth working state and acquires a tenth group of data.

[0114] In the tenth working state, the fan speed is the sixth speed and the test CPU and the test memory bar are in the first state, wherein the sixth speed is greater than the fifth speed, for example, the sixth speed is 100% (i.e. the maximum speed). The tenth group of data includes the tenth output power P out100％ ’ of the PSU read by the IPMI.

[0115] Step b6: the sixth group of data, the seventh group of data, the eighth group of data, the ninth group of data and the tenth group of data are substituted into the following formula (9) and formula (10) respectively, and a, b, c and d are determined.

[0116] P out ’ = P 风扇 ’ + P 其他 ’ (9)

[0117] P 风扇 ’ = a + b x η’ + c x η’ 2 + d x η’ 3 (10)

[0118] In the above formula (9) and formula (10), a, b, c and d are unknown coefficients; P 风扇 ’ is the power consumption of the fan, which is related to the speed of the fan; P 其他 ’ is the total power consumption of other devices in the server except the fan; η’ is the ratio of the real-time speed of the fan to the maximum speed in each working state. In the sixth working state, i.e. when the test CPU and the test memory bar are both in the IDLE state, η’ is 10%; in the seventh working state, η’ is 30%; in the eighth working state, η’ is 50%; in the ninth working state, η’ is 70%; and in the tenth working state, η’ is 100%.

[0119] In this step, specifically, the sixth group of data P out0 ’ acquired in step b1 and η’ corresponding to the sixth working state (η’ is 10% at this time) are substituted into P out ’ and η’ in formula (9) and formula (10) respectively, to obtain a group of equations, wherein P 风扇 ’, P 其他 ’, a, b, c and d are unknown numbers. Similarly, the seventh group of data to the tenth group of data acquired in steps b2 to b6 and the corresponding η’ are substituted into formula (9) and formula (10) respectively, to obtain four groups of equations. Finally, the five groups of equations obtained in combination can determine a, b, c and d, so that formula (3) can be determined.

[0120] In the embodiments of the present application, the server is controlled to work in different working states, and in different working states, only the rotating speed of the fan is different, that is, only the power consumption of the fan in the server is a variable, and the power consumption of the rest of the devices is fixed. In this way, only one variable is controlled, and multiple sets of data are obtained, so that the specific values of a, b, c and d can be quickly and accurately determined according to the multiple sets of data.

[0121] In order to quickly and accurately determine k3, k4 and k5 in the third calculation model, in the embodiments of the present application, k3, k4 and k5 are determined through steps c1-c6.

[0122] Step c1: control the server to work in the eleventh working state and obtain the eleventh set of data.

[0123] In the eleventh working state, the utilization rate of the test CPU is the sixth utilization rate, the rotating speed of the fan is the first rotating speed, and the utilization rate of the test memory bank is the seventh utilization rate. The sixth utilization rate and the seventh utilization rate can be set as needed, for example, the sixth utilization rate and the seventh utilization rate are both zero. The eleventh set of data includes the eleventh output power P out0 of the PSU read by the IPMI and the first memory power P 内存条0 of the test memory bank.

[0124] Step c2: control the server to work in the twelfth working state and obtain the twelfth set of data.

[0125] In the twelfth working state, the utilization rate of the test CPU is the sixth utilization rate, the rotating speed of the fan is the first rotating speed, and the utilization rate of the test memory bank is the eighth utilization rate. The eighth utilization rate is greater than the seventh utilization rate, for example, the eighth utilization rate is 25%. The twelfth set of data includes the twelfth output power P out25％ of the PSU read by the IPMI and the second memory power P 内存条25％ of the test memory bank.

[0126] Step c3: control the server to work in the thirteenth working state and obtain the thirteenth set of data.

[0127] In the thirteenth working state, the utilization rate of the test CPU is the sixth utilization rate, the rotating speed of the fan is the first rotating speed, and the utilization rate of the test memory bank is the ninth utilization rate. The ninth utilization rate is greater than the eighth utilization rate, for example, the ninth utilization rate is 50%. The thirteenth set of data includes the thirteenth output power P out50％ of the PSU read by the IPMI and the third memory power P 内存条50％ of the test memory bank.

[0128] Step c4: control the server to work in the fourteenth working state and obtain the fourteenth set of data.

[0129] In the fourteenth working state, the utilization rate of the test CPU is the sixth utilization rate, the rotating speed of the fan is the first rotating speed, and the utilization rate of the test memory bank is the tenth utilization rate. The tenth utilization rate is greater than the ninth utilization rate, for example, the tenth utilization rate is 75%. The fourteenth set of data includes the fourteenth output power P out75％ of the PSU read by using the IPMI and the fourth memory power P 内存条75％ of the test memory bank.

[0130] Step c5: controlling the server to work in the fifteenth working state and obtaining a fifteenth set of data.

[0131] In the fifteenth working state, the utilization rate of the test CPU is the sixth utilization rate, the rotating speed of the fan is the first rotating speed, and the utilization rate of the test memory bank is the eleventh utilization rate. The eleventh utilization rate is greater than the tenth utilization rate, for example, the eleventh utilization rate is 100%. The fifteenth set of data includes the fifteenth output power P out100％ of the PSU read by using the IPMI and the fifth memory power P 内存条100％ of the test memory bank.

[0132] Step c6: substituting the eleventh set of data, the twelfth set of data, the thirteenth set of data, the fourteenth set of data and the fifteenth set of data into the following formula (11) and formula (12) respectively, and determining k3, k4 and k5.

[0133] P out ”=P 内存条 ’ / θ 内存条 ’+P 其他 ” (11)

[0134] θ 内存条 ’=k3+k4×P 内存条 ’+k5×P 内存条 ’ 2 (12)

[0135] In the above formula (11) and formula (12), k3, k4 and k5 are unknown coefficients; θ 内存条 ’ is the voltage conversion efficiency of the second power conversion module, and P 其他 ” is the total power consumption of other devices in the server except the test memory bank. For convenience of introduction, θ 内存条0 ’, θ 内存条25％ ’, θ 内存条50％ ’, θ 内存条75％ ’ and θ 内存条100％ ’ represent the voltage conversion efficiency of the second power conversion module of the server in the eleventh working state to the fifteenth working state respectively.

[0136] In this step, specifically, the eleventh set of data P obtained in step c1 is... out0 "and P 内存条0 Substitute P into formulas (11) and (12) respectively. out "and P 内存条 We obtain a set of equations, in which only θ is present. 内存条0 '、P 其他 k1, k2, k3, k4, and k5 are unknowns. Similar to the above method, substitute the data from the twelfth to the fifteenth groups obtained in steps c2 to c5 into formulas (11) and (12), respectively, for P. out "and P 内存条 ', thus obtaining four sets of equations. Finally, by combining the five sets of equations, the specific values ​​of k3, k4 and k5 can be determined, thereby determining formula (5).

[0137] In this embodiment, the server is controlled to operate under different working states. Under these different states, only the CPU utilization rate differs; that is, only the CPU's power consumption is variable, while the power consumption of other components remains constant. In this way, a single variable is controlled, and multiple sets of data are obtained. Therefore, the specific values ​​of k3, k4, and k5 can be quickly and accurately determined based on these multiple sets of data.

[0138] It should be noted that in the above formula (4), the formula P2 = P out -(P cpu / θ cpu )-(P 内存条 / θ 内存条 )-P 风扇 To determine P2. out This represents the output power consumption of the PSU as read via IPMI.

[0139] Figure 2 The diagram shows a schematic of the power consumption determination device for a server quasi-system provided in an embodiment of this application. The specific implementation of the power consumption determination device for the server quasi-system in this application is not limited.

[0140] like Figure 2 As shown, the power consumption determination device 200 of the server barebone system may include a processor 202 and a memory 204.

[0141] The memory 204 is used to store the computer program 206. The memory 204 may include high-speed RAM, and may also include non-volatile memory, such as at least one disk storage device. The computer program 206 may include computer-executable instructions.

[0142] The processor 202 is configured to execute the computer program 206 to implement the method embodiments of determining power consumption of a server system.

[0143] The processor 202 can be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application. The one or more processors included in the power consumption determination device 200 of the server system can be the same type of processors, such as one or more CPUs; or can be different types of processors, such as one or more CPUs and one or more ASICs.

[0144] The embodiments of the present application provide a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the method embodiments of determining power consumption of a server system.

[0145] The embodiments of the present application provide a computer program, which can be executed by a processor to implement the method embodiments of determining power consumption of a server system.

[0146] The embodiments of the present application provide a computer program product, which includes a computer program. The computer program is executed by a processor to implement the method embodiments of determining power consumption of a server system.

[0147] In several embodiments provided in the present application, any function realized in the form of a software function module / unit and sold or used as an independent product can be stored in a computer readable storage medium. Based on this understanding, all or part of the technical solutions of the present application can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions for causing a computer device (which can be a personal computer, a server, or an electronic device) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various other media capable of storing computer program codes.

[0148] The algorithms and displays presented herein are not inherently related to any particular computer, virtual system, or other apparatus. Various general purpose systems can be used with programs in accordance with the teachings herein, or it can prove convenient to construct more specialized apparatus to perform the required method steps. The required structure for a variety of these systems will be apparent from the description above. In addition, the present embodiments are not described with reference to any particular programming language. It will be appreciated that a variety of programming languages can be used to implement the teachings of the present application as described herein, and any references below to specific languages are provided for disclosure of enablement only.

[0149] It should be noted that the above-mentioned embodiments illustrate rather than limit the application, and that one skilled in the art will be able to design many alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word 'comprising' does not exclude the presence of elements or steps other than those listed in a claim. The word 'a' or 'an' preceding an element does not exclude the presence of a plurality of such elements. The application can be implemented by means of both hardware and software, and any combination thereof. In a claim enumerating several means, several of these means can be embodied by one and the same item of hardware. The use of the words 'first','second' and 'third', etc. do not imply any ordering, but such words are used to name the elements and do not constitute a limitation. The steps of the methods described herein do not have to be performed in the order described, unless otherwise specified.

[0150] The embodiments described above are merely meant to illustrate the present application and not to limit the scope of the patent. It should be noted that for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present application, and these should all fall within the scope of the present application. Therefore, the scope of the present application should be defined by the appended claims.

Claims

1. A method for determining power consumption of a server farm system, the method comprising: The load of the server system under the target configuration includes a central processing unit (CPU) and a memory bank, and the method comprises: determining a first power loss of a first power conversion module arranged on a motherboard in the server system according to a power consumption of the CPU and a preset first calculation model, wherein the first power conversion module is configured to convert voltage and provide the converted voltage to the CPU; determining a first power consumption of a fan in the server system according to a speed ratio of the fan and a preset second calculation model, wherein the speed ratio is a ratio of a real-time speed of the fan to a maximum speed; determining whether a second power conversion module is arranged on the motherboard, wherein the second power conversion module is configured to convert voltage and provide the converted voltage to the memory bank; if the second power conversion module is not arranged on the motherboard, determining a second power consumption and determining a sum of the first power loss, the first power consumption and the second power consumption as a power consumption of the server system under the target configuration, wherein the second power consumption is a total power consumption of a backboard in the server system and other devices arranged on the motherboard except the first power conversion module; if the second power conversion module is arranged on the motherboard, determining a second power loss of the second power conversion module according to a power of the memory bank and a preset third calculation model, determining a third power consumption, and determining a sum of the first power loss, the second power loss, the first power consumption and the third power consumption as the power consumption of the server system under the target configuration, wherein the third power consumption is a total power consumption of the backboard and other devices arranged on the motherboard except the first power conversion module and the second power conversion module.

2. The method of claim 1, wherein, The determining of the first power loss of the first power conversion module arranged on the motherboard in the server system according to the power consumption of the CPU and the preset first calculation model comprises: According to the power consumption P of the CPU cpu And the first calculation model, determine the first voltage conversion efficiency θ of the first power conversion module cpu ; By the formula PL1 = P cpu / θ cpu -P cpu determining the first power loss PL1.

3. The method of claim 2, wherein, The first voltage conversion efficiency θ of the first power conversion module is determined according to the power consumption P of the CPU cpu and the first calculation model cpu , comprising: by the first calculation model θ cpu = k0 + k1 x P cpu + k2 x P cpu 2 determining the first voltage conversion efficiency θ cpu where k0, k1 and k2 are constants.

4. The method of claim 3, wherein, k0, k1 and k2 are determined by the following steps: controlling a server to work in a first working state and obtaining a first group of data, wherein the server comprises the server system, a power module, a test CPU, a test solid state disk and a test memory bank, in the first working state, a speed of the fan is a first speed and a utilization rate of the test CPU is a first utilization rate, and the first group of data comprises a first output power consumption of the power module and a first CPU power consumption of the test CPU; controlling the server to work in a second working state and obtaining a second group of data, wherein in the second working state, the speed of the fan is the first speed and the utilization rate of the test CPU is a second utilization rate, and the second group of data comprises a second output power consumption of the power module and a second CPU power consumption of the test CPU; controlling the server to work in a third working state and obtaining a third set of data, wherein in the third working state, the rotation speed of the fan is the first rotation speed and the utilization rate of the test CPU is a third utilization rate, and the third set of data comprises a third output power consumption of the power module and a third CPU power consumption of the test CPU; controlling the server to work in a fourth working state and obtaining a fourth set of data, wherein in the fourth working state, the rotation speed of the fan is the first rotation speed and the utilization rate of the test CPU is a fourth utilization rate, and the fourth set of data comprises a fourth output power consumption of the power module and a fourth CPU power consumption of the test CPU; controlling the server to work in a fifth working state and obtaining a fifth set of data, wherein in the fifth working state, the rotation speed of the fan is the first rotation speed and the utilization rate of the test CPU is a fifth utilization rate, and the fifth set of data comprises a fifth output power consumption of the power module and a fifth CPU power consumption of the test CPU; The first group of data, the second group of data, the third group of data, the fourth group of data and the fifth group of data are substituted into the formula P out = P cpu ’ / θ cpu ’+ P 其他 and θ cpu ’= k0+k1×P cpu ’+k2×P cpu ’ 2 to determine k0, k1 and k2, wherein P out is the output power consumption of the power module in each group of data, P cpu ’ is the power consumption of the test CPU in each group of data, θ cpu ’ is the voltage conversion efficiency of the first power conversion module of the server under each working state, P 其他 is the total power consumption of other devices in the server except the test CPU.

5. The method of claim 1, wherein, The second calculation model is P 风扇 = a + b x η + c x η 2 + d x η 3 , wherein P 风扇 is the first power consumption, η is the speed ratio value, a, b, c, and d are all constants.

6. The method of claim 5, wherein, a, b, c and d are determined by the following steps: controlling the server to work in a sixth working state and obtaining a sixth set of data, wherein the server comprises the server system, the power module, the test CPU, the test solid state disk and the test memory bar, in the sixth working state, the rotation speed of the fan is the second rotation speed and the test CPU and the test memory bar are in a first state, and the sixth set of data comprises a sixth output power consumption of the power module; controlling the server to work in a seventh working state and obtaining a seventh set of data, wherein in the seventh working state, the rotation speed of the fan is the third rotation speed and the test CPU and the test memory bar are in the first state, and the seventh set of data comprises a seventh output power consumption of the power module; controlling the server to work in an eighth working state and obtaining an eighth set of data, wherein in the eighth working state, the rotation speed of the fan is the fourth rotation speed and the test CPU and the test memory bar are in the first state, and the eighth set of data comprises an eighth output power consumption of the power module; controlling the server to work in a ninth working state and obtaining a ninth set of data, wherein in the ninth working state, the rotation speed of the fan is the fifth rotation speed and the test CPU and the test memory bar are in the first state, and the ninth set of data comprises a ninth output power consumption of the power module; controlling the server to work in a tenth working state and obtaining a tenth set of data, wherein in the tenth working state, the rotation speed of the fan is the sixth rotation speed and the test CPU and the test memory bar are in the first state, and the tenth set of data comprises a tenth output power consumption of the power module; The sixth group of data, the seventh group of data, the eighth group of data, the ninth group of data and the tenth group of data are substituted into the formula P out ’ = P 风扇 ’ + P 其他 ’ and P 风扇 ’ = a + b x η’ + c x η’ 2 +d x η’ 3 , respectively, to determine a, b, c and d, wherein P out ’ is the output power consumption of the power module in each group of data, P 风扇 ’ is the power consumption of the fan, P 其他 ’ is the total power consumption of other devices in the server except the fan, and η’ is the ratio of the real-time rotating speed of the fan to the maximum rotating speed under each working state.

7. The method of claim 1, wherein, the second power loss of the second power conversion module is determined according to the power of the memory bar and a third calculation model; According to the power consumption P of the memory bank 内存条 and the third calculation model θ 内存条 = k3 + k4 × P 内存条 + k5 × P 内存条 2 , determine the second voltage conversion efficiency θ 内存条 of the second power conversion module, wherein k3, k4 and k5 are all constants; By the formula PL2 = P 内存条 / θ 内存条 -P 内存条 determining the second power loss PL2.

8. The method of claim 7, wherein, k3, k4 and k5 are determined by the following steps: The control server works in the eleventh working state and obtains an eleventh group of data, wherein the server includes the server benchmark system, a power module, a test CPU, a test solid state disk and a test memory bar, in the eleventh working state, the utilization rate of the test CPU is a sixth utilization rate, the rotating speed of the fan is a first rotating speed and the utilization rate of the test memory bar is a seventh utilization rate, the eleventh group of data includes an eleventh output power consumption of the power module and a first memory power consumption of the test memory bar; The control server works in the twelfth working state and obtains a twelfth group of data, wherein in the twelfth working state, the utilization rate of the test CPU is the sixth utilization rate, the rotating speed of the fan is the first rotating speed and the utilization rate of the test memory bar is an eighth utilization rate, the twelfth group of data includes a twelfth output power consumption of the power module and a second memory power consumption of the test memory bar; The control server works in the thirteenth working state and obtains a thirteenth group of data, wherein in the thirteenth working state, the utilization rate of the test CPU is the sixth utilization rate, the rotating speed of the fan is the first rotating speed and the utilization rate of the test memory bar is a ninth utilization rate, the thirteenth group of data includes a thirteenth output power consumption of the power module and a third memory power consumption of the test memory bar; The control server works in the fourteenth working state and obtains a fourteenth group of data, wherein in the fourteenth working state, the utilization rate of the test CPU is the sixth utilization rate, the rotating speed of the fan is the first rotating speed and the utilization rate of the test memory bar is a tenth utilization rate, the fourteenth group of data includes a fourteenth output power consumption of the power module and a fourth memory power consumption of the test memory bar; The control server works in the fifteenth working state and obtains a fifteenth group of data, wherein in the fifteenth working state, the utilization rate of the test CPU is the sixth utilization rate, the rotating speed of the fan is the first rotating speed and the utilization rate of the test memory bar is an eleventh utilization rate, the fifteenth group of data includes a fifteenth output power consumption of the power module and a fifth memory power consumption of the test memory bar; The eleventh group of data, the twelfth group of data, the thirteenth group of data, the fourteenth group of data and the fifteenth group of data are substituted into the formula P out ”=P 内存条 ’ / θ 内存条 ’+P 其他 ”and θ 内存条 ’=k3+k4×P 内存条 ’+k5×P 内存条 ’ 2 , to determine k3, k4 and k5, wherein P out ” is the output power consumption of the power module in each group of data, P 内存条 ’ is the power consumption of the test memory bank in each group of data, θ 内存条 ’ is the voltage conversion efficiency of the second power conversion module of the server in each working state, P 其他 ” is the total power consumption of other devices in the server except the test memory bank.

9. A power consumption determination device for a server farm system, comprising a memory, a processor, and a computer program stored on the memory, wherein, The processor executes the computer program to implement the power consumption determination method of the server benchmark system of any one of claims 1-8.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the power consumption determination method of the server benchmark system of any one of claims 1-8.

Images