CN113748398B

CN113748398B - Data processing and task scheduling method, device, system and storage medium

Info

Publication number: CN113748398B
Application number: CN201980095635.5A
Authority: CN
Inventors: 陶原; 卢毅军; 李栈; 宋军; 奉有泉; 赵旭; 陈钢
Original assignee: Alibaba Cloud Computing Ltd
Current assignee: Alibaba Cloud Computing Ltd
Priority date: 2019-09-06
Filing date: 2019-09-06
Publication date: 2023-01-17
Anticipated expiration: 2039-09-06
Also published as: CN113748398A; WO2021042373A1

Abstract

A data processing and task scheduling method, device, system and storage medium. The setting operation of kernel parameters related to a power consumption management mechanism is combined with artificial intelligence, performance data and power consumption data of the physical equipment under various value combinations of the kernel parameters are estimated based on a performance-power consumption estimation model trained by the artificial intelligence, furthermore, the obtained performance data and power consumption data are taken as the basis, appropriate values can be set for the kernel parameters, the power consumption and performance of the physical equipment are considered, and the parameter setting efficiency can be improved and the cost can be reduced by combining with the artificial intelligence.

Description

Data processing and task scheduling method, device, system and storage medium

Technical Field

The present application relates to the field of data processing technologies, and in particular, to a method, device, system, and storage medium for data processing and task scheduling.

Background

With the development of electronic technology, power consumption management of physical devices such as servers and computers is more and more important. In order to reduce power consumption on the basis of normal operation of equipment, a power consumption management mechanism is added in an operating system. Taking Linux operating system as an example, dynamic Voltage Dynamic Frequency Scaling (DVFS) and C-mode (C-state) are two main mechanisms used by a processor to reduce power consumption and ensure performance.

The power consumption management mechanisms are related to some equipment parameters, some equipment parameters may have various values, and the power consumption management mechanisms can generate different energy-saving effects under different parameter values. How to reasonably set the parameter value of the power consumption management mechanism in an efficient and low-cost manner so as to enable the power consumption management mechanism to generate a better or optimal energy-saving effect is a problem faced by the existing power consumption management mechanism.

Disclosure of Invention

Aspects of the present application provide a method, device, system, and storage medium for data processing and task scheduling, so as to efficiently and reasonably set a parameter value of a power consumption management mechanism and reduce cost.

An embodiment of the present application provides a data processing method, including: determining at least one core parameter related to a core-state power consumption management mechanism supported by a physical device from core parameters of the physical device; acquiring performance data and power consumption data of the physical equipment under various value combinations corresponding to the at least one kernel parameter, wherein the performance data and the power consumption data under at least part of the value combinations are estimated based on a performance-power consumption estimation model; determining a target value combination corresponding to the at least one kernel parameter according to the performance data and the power consumption data of the physical device under the various value combinations; and setting the at least one kernel parameter according to the target value combination so that the power consumption management mechanism operates according to the values in the target value combination.

An embodiment of the present application further provides a data processing method, including: determining at least one core parameter related to a power consumption management mechanism of a core state supported by a physical device from core parameters of the physical device; testing performance data and power consumption data of the physical equipment under partial value combination corresponding to the at least one kernel parameter by using a load testing tool; performing model training according to the performance data and the power consumption data of the physical equipment under the partial value combination to obtain a performance-power consumption estimation model; and estimating performance data and power consumption data of the physical equipment under other value combinations corresponding to the at least one kernel parameter by using the performance-power consumption estimation model.

An embodiment of the present application further provides a data processing method, including: determining at least one kernel parameter related to the device performance from kernel parameters of the physical device; acquiring performance data of the physical equipment under various value combinations corresponding to the at least one kernel parameter, wherein at least part of the performance data under the value combinations is estimated based on a performance estimation model; determining a target value combination corresponding to the at least one kernel parameter according to the performance data of the physical device under the plurality of value combinations; and setting the at least one kernel parameter according to the target value combination so that the physical equipment operates according to the values in the target value combination.

An embodiment of the present application further provides a data processing method, including: determining at least one kernel parameter related to the device performance from the kernel parameters of the physical device; testing the performance data of the physical equipment under the partial value combination corresponding to the at least one kernel parameter by using a load testing tool; performing model training according to the performance data of the physical equipment under the partial value combination to obtain a performance estimation model; and estimating the performance data of the physical equipment under other value combinations corresponding to the at least one kernel parameter by using the performance estimation model.

An embodiment of the present application further provides a data processing method, including: determining at least one core parameter related to the power consumption of the device from the core parameters of the physical device; acquiring power data of the physical equipment under various value combinations corresponding to the at least one kernel parameter, wherein at least part of the power data under the value combinations is estimated based on a power estimation model; determining a target value combination corresponding to the at least one kernel parameter according to the power data of the physical device under the plurality of value combinations; and setting the at least one kernel parameter according to the target value combination so that the physical equipment operates according to the values in the target value combination.

An embodiment of the present application further provides a data processing method, including: determining at least one core parameter related to the power consumption of the device from the core parameters of the physical device; testing power data of the physical equipment under the partial value combination corresponding to the at least one kernel parameter by using a load testing tool; performing model training according to the power data of the physical equipment under the partial value combination to obtain a power estimation model; and estimating power data of the physical equipment under other value combinations corresponding to the at least one kernel parameter by using the power estimation model.

An embodiment of the present application further provides a data processing method, including: determining at least one core parameter related to a power consumption management mechanism of a core state supported by a physical device from core parameters of the physical device; testing performance data and power consumption data of the physical equipment under the partial value combination corresponding to the at least one kernel parameter by using a load testing tool; performing model training according to the performance data and the power consumption data of the physical equipment under the partial value combination to obtain a performance-power consumption estimation model; estimating performance data and power consumption data of the physical equipment under various value combinations corresponding to the at least one kernel parameter by using the performance-power consumption estimation model; wherein the plurality of value combinations comprises the partial value combinations.

An embodiment of the present application further provides an apparatus management system, including: at least one physical device and at least one model computing device; wherein the at least one physical device supports a power consumption management mechanism of a kernel mode respectively; the at least one model computing device is configured to determine, from kernel parameters of a target device, at least one kernel parameter related to a kernel-state power consumption management mechanism supported by the target device, and obtain performance data and power consumption data of the target device under multiple value combinations corresponding to the at least one kernel parameter based on an artificial intelligence model; the target device is any one of the at least one physical device; the target device is configured to determine a target value combination corresponding to the at least one kernel parameter according to the performance data and the power consumption data of the target device under the multiple value combinations corresponding to the at least one kernel parameter, which are obtained by the model computing device, and set the at least one kernel parameter according to the target value combination.

An embodiment of the present application further provides a data center system, including: the system comprises model computing equipment and at least one machine room, wherein the at least one machine room comprises at least one piece of physical equipment, and the at least one piece of physical equipment respectively supports a power consumption management mechanism of a kernel mode; the model computing device is used for determining at least one kernel parameter related to a kernel-mode power consumption management mechanism supported by a target device from kernel parameters of the target device, and obtaining performance data and power consumption data of the target device under various value combinations corresponding to the at least one kernel parameter based on an artificial intelligence model; the target device is any one of the at least one physical device; the target device is configured to determine a target value combination corresponding to the at least one kernel parameter according to the performance data and the power consumption data of the target device under the multiple value combinations corresponding to the at least one kernel parameter, which are obtained by the model computing device, and set the at least one kernel parameter according to the target value combination.

An embodiment of the present application further provides a physical device, including: a memory and a processor; the memory for storing a computer program; the processor, coupled with the memory, to execute the computer program to: determining at least one core parameter related to a core-state power consumption management mechanism supported by the physical device from the core parameters of the physical device; acquiring performance data and power consumption data of the physical equipment under various value combinations corresponding to the at least one kernel parameter, wherein the performance data and the power consumption data under at least part of the value combinations are estimated based on a performance-power consumption estimation model; determining a target value combination corresponding to the at least one kernel parameter according to the performance data and the power consumption data of the physical device under the plurality of value combinations; and setting the at least one kernel parameter according to the target value combination so that the power consumption management mechanism operates according to the values in the target value combination.

An embodiment of the present application further provides a model computing device, including: a memory and a processor; the memory for storing a computer program; the processor, coupled with the memory, to execute the computer program to: determining at least one core parameter related to a core-state power consumption management mechanism supported by a physical device from core parameters of the physical device; testing performance data and power consumption data of the physical equipment under partial value combination corresponding to the at least one kernel parameter by using a load testing tool; performing model training according to the performance data and the power consumption data of the physical equipment under the partial value combination to obtain a performance-power consumption estimation model; and estimating performance data and power consumption data of the physical equipment under other value combinations corresponding to the at least one kernel parameter by using the performance-power consumption estimation model.

Embodiments of the present application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to implement the steps in the various data processing methods provided by the embodiments of the present application.

An embodiment of the present application further provides a task scheduling method, including: acquiring a task to be scheduled and a performance requirement of the task to be scheduled; selecting resource equipment which meets the performance requirement and the kernel-mode power consumption parameter value meets the set power consumption requirement from at least one resource equipment; scheduling the task to be scheduled to the resource equipment which meets the performance requirement and the kernel-mode power consumption parameter value meets the set power consumption requirement; the kernel-mode power consumption parameter value refers to a value combination of at least one kernel parameter related to a kernel-mode power consumption management mechanism supported by the resource device.

An embodiment of the present application further provides a task scheduling device, including: a memory and a processor; the memory for storing a computer program; the processor, coupled with the memory, to execute the computer program to: acquiring a task to be scheduled and a performance requirement of the task to be scheduled; selecting resource equipment which meets the performance requirement and a kernel-mode power consumption parameter value which meets a set power consumption requirement from at least one resource equipment; scheduling the task to be scheduled to the resource equipment which meets the performance requirement and the kernel-mode power consumption parameter value meets the set power consumption requirement; the kernel-mode power consumption parameter value refers to a value combination of at least one kernel parameter related to a kernel-mode power consumption management mechanism supported by the resource device.

Embodiments of the present application further provide a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to implement the steps in the task scheduling method provided in the embodiments of the present application.

In the embodiment of the application, the setting operation of the kernel parameters related to the power consumption management mechanism is combined with artificial intelligence, the performance data and the power consumption data of the physical equipment under various value combinations corresponding to the kernel parameters are obtained based on the performance-power consumption estimation model trained by the artificial intelligence, and then the obtained performance data and the power consumption data are taken as the basis, the appropriate values can be set for the kernel parameters, the power consumption and the performance of the physical equipment are considered, and the combination with the artificial intelligence can improve the parameter setting efficiency and reduce the cost.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:

fig. 1 is a schematic structural diagram of a data center system according to an exemplary embodiment of the present application;

fig. 2 is a schematic diagram illustrating a principle of kernel parameter setting performed by a data center system according to an exemplary embodiment of the present application;

FIG. 3 is a schematic diagram illustrating an operating principle of a performance-power consumption estimation model according to an exemplary embodiment of the present application;

fig. 4a is a schematic structural diagram of a device management system according to an exemplary embodiment of the present application;

fig. 4b is a schematic structural diagram of an edge cloud network system according to an exemplary embodiment of the present application;

fig. 5a is a schematic flowchart of a data processing method according to an exemplary embodiment of the present application;

FIG. 5b is a schematic flow chart diagram illustrating another data processing method according to an exemplary embodiment of the present application;

FIG. 6a is a schematic flow chart diagram illustrating yet another data processing method according to an exemplary embodiment of the present application;

FIG. 6b is a schematic flow chart diagram illustrating yet another data processing method according to an exemplary embodiment of the present application;

FIG. 7a is a schematic flowchart of another data processing method provided in an exemplary embodiment of the present application;

FIG. 7b is a schematic flow chart diagram illustrating yet another data processing method according to an exemplary embodiment of the present application;

FIG. 7c is a schematic flow chart diagram illustrating yet another data processing method according to an exemplary embodiment of the present application;

FIG. 7d is a flowchart illustrating a task scheduling method according to an exemplary embodiment of the present disclosure;

FIG. 8a is a schematic structural diagram of a physical device according to an exemplary embodiment of the present application;

FIG. 8b is a schematic diagram of a model computing device according to an exemplary embodiment of the present application;

fig. 8c is a schematic structural diagram of a task scheduling apparatus according to an exemplary embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be described in detail and completely with reference to the following specific embodiments of the present application and the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application, and not all of the embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application without making any creative effort belong to the protection scope of the present application.

In order to solve the problems in the prior art, in the embodiment of the application, setting operation of kernel parameters related to a power consumption management mechanism is combined with artificial intelligence, performance data and power consumption data of a physical device under various value combinations corresponding to the kernel parameters can be obtained based on a performance-power consumption estimation model trained by the artificial intelligence, and further, appropriate values can be set for the kernel parameters according to the obtained performance data and power consumption data, so that power consumption and performance of the physical device are considered, and the combination with the artificial intelligence can improve parameter setting efficiency and reduce cost.

The technical solutions provided by the embodiments of the present application are described in detail below with reference to the accompanying drawings.

Fig. 1 is a schematic structural diagram of a data center system according to an exemplary embodiment of the present application. As shown in fig. 1, the data center system 100 includes: a model computing device 101 and at least one machine room 102; at least one machine room 102 comprises at least one physical device 103.

The machine room 102 is a physical place where the machine equipment is stored, and may be, for example, a room or a factory building. In this embodiment, the number of the physical devices 103 in each machine room 102 is not limited, and each machine room 102 may include one physical device 103 or may include a plurality of physical devices 103. Generally, a computer room 102 will include a plurality of physical devices 103.

In this embodiment, the physical device 103 is a physical device that is installed with an operating system and supports a kernel-mode power consumption management mechanism. Of course, the computer room 102 may include, in addition to the physical device 103 which is installed with an operating system and supports the kernel-mode power consumption management mechanism, some devices which do not need the operating system and some devices which are installed with the operating system but do not support the kernel-mode power consumption management mechanism, which is not limited to this. In this embodiment as well as in other embodiments, the focus is on the physical device 103 that is installed with an operating system and supports kernel-mode power consumption management mechanisms.

In this embodiment, the device form of the physical device 103 is not limited, and any device form that is installed with an operating system and supports a kernel-state power consumption management mechanism is applicable to the embodiment of the present application. Alternatively, the physical device 103 may be an IT device that is installed with an operating system and supports a kernel-mode power consumption management mechanism, but is not limited thereto. For example, the physical device 103 may include, but is not limited to, at least one of: various server devices, computer devices, and various network switching devices, etc. The server device may be, including but not limited to: a regular server, an array of servers or a cloud server, etc. Optionally, various applications or services may run on these physical devices 103, such as a cloud computing service, a game service, an instant messaging service, a mail service, or an online transaction service, etc. Of course, no applications or services may be run on these physical devices 103. Depending on whether the physical device 103 runs an application or a service, and the type and the number of the running applications or services, a power consumption management mechanism on the physical device 103 may be triggered and may perform a corresponding function.

Essentially, an operating system is software that is responsible for controlling the hardware resources of a physical device and providing an environment for upper layer applications to run. The operating system provides two CPU running states, a kernel state and a user state. The user mode is an active space of an upper application program, and the execution of the application program must depend on resources provided by an operating system, including CPU resources, storage resources, I/O resources and the like. The power consumption management mechanism of this embodiment is a mechanism provided by an operating system to manage power consumption of a physical device, is a power consumption management mechanism of an operating system level, and needs to operate in a kernel mode, which is simply referred to as a kernel-mode power consumption management mechanism. It should be noted that the power consumption management mechanisms of the kernel modes supported by different physical devices 103 may be the same or different. The power management mechanism of the kernel mode is related to the os, and if the os of the physical device 103 is different, the power management mechanism of the kernel mode supported by the physical device 103 is also different. For example, taking the physical device 103 as an example of a Linux operating system, the supported kernel-mode power consumption management mechanisms include, but are not limited to: DVFS and C-state.

DVFS is a dynamic technique that dynamically adjusts the operating frequency and voltage of a chip according to different requirements of an application program run by the chip (e.g., CPU) on computing power, thereby achieving the purpose of saving energy. For the same chip, the higher the operating frequency, the higher the required voltage and the greater the power consumption. C-state is a low power consumption mechanism which can enable the CPU to enter a low power consumption state when in an idle state, C modes contained in the C-state are started from C0 to Cn, the C0 is a normal working mode of the CPU, and the CPU is in a 100% running state; the higher the value of n behind C is, the deeper the CPU sleeps, the smaller the power consumption of the CPU is, and certainly, more time is required for returning to the C0 mode; wherein n is a positive integer.

In this embodiment, the core-state power consumption management mechanism is related to some core parameters, which are parameters of the core-state power consumption management mechanism, and values of the core parameters may affect an energy saving effect of the core-state power consumption management mechanism. In this embodiment, the kernel parameter generally refers to parameters in source codes of various operating systems, such as a kernel parameter of a Linux operating system, a kernel parameter of a Windows operating system, a kernel parameter of a UNIX operating system, or a kernel parameter of a MAC operating system.

Taking Linux operating system as an example, the kernel parameters related to DVFS include but are not limited to: the method comprises the following steps of (1) recording the lowest working frequency (namely scaling _ min _ freq) capable of being operated by a CPU (central processing unit), the highest working frequency (namely scaling _ max _ freq) capable of being operated by the CPU and an adjusting mode (namely scaling _ governor) of the working frequency of the CPU; for the physical device 103, the energy saving effect of the DVFS can be changed by adjusting the value of at least one of the three parameters. Taking the Linux operating system as an example, the kernel parameters related to C-states include but are not limited to: an entry time threshold (marked as target _ latency) corresponding to each level of C mode; wherein, the entry time threshold represents the time that the physical device 103 needs to be kept in the corresponding C mode at least after entering the mode, and is a time condition that the physical device 103 needs to meet when entering the corresponding C mode; for the physical device 103, the difficulty level of the CPU entering the corresponding C mode may be changed by adjusting the entry time threshold corresponding to the corresponding C mode, and the energy saving effect of the C-states mechanism is changed.

Regardless of the core-state power management mechanism, the core parameters associated therewith may have a variety of values. For example, the highest operating frequency scaling _ max _ freq at which the processor can operate may be set to 2.4GHZ, 3.6GHZ, etc. The values of kernel parameters related to the power consumption management mechanism are different, and the energy saving effect which can be generated by the power consumption management mechanism is different. In practical applications, the power consumption of the physical device 103 has a certain relationship with the performance, and generally, the lower the power consumption, the worse the performance, so that the low power consumption cannot be pursued at once, and ideally, the balance between the power consumption and the performance should be pursued according to the application requirements. In different application scenarios, the requirements of the physical device 103 on power consumption and performance may be different, and how to set values of kernel parameters related to a kernel-mode power consumption management mechanism supported by the physical device 103, so that the power consumption management mechanism generates a better or optimal energy saving effect, and the requirements of the physical device 103 on performance are met at the same time, which is a problem to be solved.

The kernel parameters related to the power consumption management mechanism belong to system-level parameters, and need to be set in the source code of the operating system, and after the values of the kernel parameters are reset each time, the operating system needs to be reinstalled, and only after the operating system runs for a certain time, whether the current values can meet the requirements of the equipment on power consumption and performance or not can be judged, and whether readjustment is needed or not can be judged. If there are more kernel parameters associated with the power management mechanism, the number of value combinations of these kernel parameters will be larger, and if an attempt is made to select the best value combination, it will take a long time.

Based on the above consideration, in this embodiment, the setting operation of the kernel parameter is combined with the artificial intelligence, and based on the artificial intelligence model, the performance data and the power consumption data of the physical device 103 under the multiple value combinations corresponding to the relevant kernel parameters are obtained, and further, based on the performance data and the power consumption data of the physical device 103 under the multiple value combinations corresponding to the relevant kernel parameters, the setting of the kernel parameter is performed, so that not only can a proper value be set for the kernel parameter, but also the requirements of the physical device 103 on power consumption and performance can be considered at the same time, and the combination with the artificial intelligence can improve the parameter setting efficiency and reduce the cost.

In order to achieve the above purpose, in the data center system 100 of this embodiment, a model computing device 101 is additionally provided, and is mainly responsible for obtaining performance data and power consumption data of the physical device 103 under various value combinations corresponding to relevant kernel parameters. For the model computing device 101, for each physical device 103, at least one core parameter related to a core-state power consumption management mechanism supported by the physical device 103 may be determined from the core parameters of the physical device 103, and performance data and power consumption data of the physical device 103 under multiple value combinations corresponding to the at least one core parameter are provided based on an artificial intelligence model, so that the physical device 103 completes setting of the related core parameters accordingly.

The model computing device 101 may be disposed in a certain computer room 102, or may be disposed separately from each computer room 102, for example, may be disposed in the cloud. In addition, the present embodiment does not limit the device form of the model computing device 101, and may be any computing device having a certain computing capability and communication capability. As shown in fig. 1, model computing device 101 may be a conventional server, a cloud server, a server with a GPU, a server or array of servers with a specific Ai chip, or the like.

Alternatively, the model computing device 101 may be communicatively connected to each physical device 103, and after obtaining the performance data and the power consumption data of each physical device 103 under the multiple value combinations corresponding to the relevant core parameters, the model computing device may provide the performance data and the power consumption data of each physical device 103 under the multiple value combinations corresponding to the relevant core parameters to each physical device 103 based on the communicative connection between the model computing device and each physical device 103. Wherein, the model computing device 101 and each physical device 103 can be connected in a wireless or wired mode. Alternatively, the physical device 103 may be communicatively connected to the model computing device 101 through a mobile network. The network standard of the mobile network may be any one of 2G (GSM), 2.5G (GPRS), 3G (WCDMA, TD-SCDMA, CDMA2000, UTMS), 4G (LTE), 4G + (LTE +), 5G, wiMax, or a new network standard that will appear in the future. Optionally, the physical device 103 may also be communicatively connected to the model computing device 101 via bluetooth, wiFi, infrared, zigbee, NFC, or the like.

It should be noted that the model computing device 101 does not necessarily have to establish a communication connection with the physical device 103. In a scenario where a communication connection is not established between the model computing device 101 and the physical device 103, performance data and power consumption data of the physical device 103 under various value combinations corresponding to relevant kernel parameters may also be copied from the model computing device 101 to the physical device 103 through some mobile storage devices such as a mobile hard disk and a usb disk.

For convenience of description, the present embodiment takes any one physical device 103 as an example, and the operation principle of the present system is explained with reference to the flowchart shown in fig. 2. For convenience of description and distinction, the following description will be made by taking the target device as an example, and the target device represents any one of the physical devices 103.

For example, in a case where a new physical device is added in a certain computer room 102, the added physical device may be used as a target device, before an operating system is installed for the target device, kernel parameters related to a power consumption management mechanism in an operating system source code may be set according to application requirements of the target device, and the operating system may be installed on the target device according to the operating system source code after parameter setting. For another example, when load information (e.g., an application or a service) on the target device changes, in order to enable the target device to meet requirements of new load information on power consumption and performance, values of kernel parameters related to a power consumption management mechanism in the source code of the operating system may be modified again, and the operating system may be reinstalled for the target device according to the modified source code of the operating system. For another example, when a certain device is rented and sold to a certain client, the device rented and sold to the client may be used as a target device, and in order to ensure that the target device can successfully provide corresponding services for the client, an operating system may be installed for the target device according to the requirements of the client, and kernel parameters related to the power consumption management mechanism in the source code of the operating system may be set before the operating system is installed.

Regardless of the above requirements, the model computing device 101 may determine, from the core parameters of the target device, at least one core parameter associated with a core-state power management mechanism supported by the target device. The number of the kernel parameters may be one or more, depending on the power management mechanism.

Alternatively, the model computing device 101 may have human-computer interaction capability, and may then perform human-computer interaction with an administrator of the data center system 100 to determine at least one kernel parameter associated with the kernel-mode power consumption management mechanism supported by the target device. For example, the model computing device 101 may provide a human-machine interface or a command window, and an administrator may input, to the model computing device 101, an identifier of the target device and information of the power consumption management mechanism of the kernel mode supported by the target device; the model computing device 101 determines, from the kernel parameters of the target device, at least one kernel parameter related to a kernel-state power consumption management mechanism supported by the target device, based on the information. For another example, if the model computing device 101 has a voice recognition capability, the administrator may input, to the model computing device 101, the identifier of the target device and information of the kernel-mode power consumption management mechanism supported by the target device in a voice manner; the model computing device 101 determines, from the kernel parameters of the target device, at least one kernel parameter related to a kernel-state power consumption management mechanism supported by the target device, based on the information. Of course, in addition to the above information, the administrator may also directly input to the model computing device 101 an identification of at least one core parameter, such as a parameter name, associated with the core-state power consumption management mechanism supported by the target device.

After determining at least one core parameter related to a core-mode power consumption management mechanism supported by the target device, the model computing device 101 may collect performance data and power consumption data of the target device under a partial value combination corresponding to the at least one core parameter as sample data, and perform model training using the sample data to obtain a performance-power consumption estimation model, as shown in fig. 2. The performance data of the target device may vary according to the service or application running on the target device. The performance data mainly refers to some data capable of reflecting the capacity of the target device to undertake the service and guarantee the service. For example, it may be the actual achieved QoS, QPS, TPS of the service on the target device, or response time to the request, etc.

It should be noted that, if the number of at least one kernel parameter is one, one value combination includes one value; if the number of the at least one kernel parameter is multiple, one value combination comprises multiple values corresponding to the multiple kernel parameters one to one, and the multiple values contained in different value combinations are not completely the same.

In this embodiment, as shown in fig. 2, the model computing device 101 may collect, by means of a load testing tool, performance data and power consumption data of the target device under a partial value combination corresponding to at least one core parameter. And installing and running a load testing tool on the target equipment, wherein the load testing tool can simulate the performance data and the power consumption data of the target equipment under different load conditions. In detail, partial value combinations are selected from multiple value combinations corresponding to at least one kernel parameter, and the number of the partial value combinations only needs to meet the number required by model training; aiming at each value combination in part of value combinations, according to load, power consumption and/or performance requirements which need to be met by target equipment, utilizing a load testing tool to obtain performance data and power consumption data of the target equipment when the target equipment meets corresponding requirements under the value combination; and providing the performance data and the power consumption data of the target equipment meeting the corresponding requirements under the partial value combination to the model computing equipment 101. In order to simplify the description, the performance data and the power consumption data of the target device meeting the corresponding requirements under each value combination are referred to as the performance data and the power consumption data of the target device under each value combination.

The process of obtaining the performance data and the power consumption data of the target equipment under the value combination by using the load testing tool aiming at each value combination in part of the value combinations comprises the following steps: aiming at the value combination, firstly, modifying the value of at least one kernel parameter in the source code of the operating system into the value in the value combination, and installing the operating system on the target equipment according to the modified source code of the operating system; after the operating system is successfully installed, a load testing tool is installed on the target equipment, and according to the load, power consumption and/or performance requirements which the target equipment should meet, the corresponding load condition is simulated by using the load testing tool and the performance data and the power consumption data of the target equipment under the corresponding load condition are obtained.

In this embodiment, the number and type of the load testing tools are not limited, and may be flexibly selected according to the application requirements, the type of the operating system, and the like. For example, this example lists several load testing tools: the Stream testing tool is used for testing the performance of the memory; a Specjbb test tool for testing cpu performance; the Speccpu testing tool is used for testing the performance of the CPU; the Fio test tool is used for testing the IO performance of the disk; the Sysbench test tool is used for testing mysql database performance, and the like.

Alternatively, during the test process of the above-mentioned several load test tools, the power consumption data of the physical device during the test process can be collected by means of a power consumption collecting tool, such as an electricity meter. The relevant performance parameters tested by the load testing tool and the power consumption data collected by the power consumption collecting tool in the testing process are performance data and power consumption data when the physical device 103 meets corresponding requirements under a certain value combination.

After performance data and power consumption data of the target device under the partial value combination corresponding to the at least one kernel parameter are obtained, the model computing device 101 performs model training by using the data as sample data. In this embodiment, the process of model training performed by the model computing device 101 is not limited, and for example, the process may be a model training process based on a deep neural network, or a model training process based on regression analysis, and the model training method that can analyze the association relationship between the value combination corresponding to the at least one kernel parameter and the performance data and the power consumption data of the target device is suitable for the embodiment of the present application.

Regression analysis is a statistical analysis method for determining the interdependent quantitative relationship between two or more variables, and is a predictive modeling technique. In an alternative embodiment, the model computing device 101 may employ a modeling approach based on regression analysis. Based on this, the process of performing model training by the model computing device 101 is actually a process of performing regression analysis on the performance data and the power consumption data of the target device under the partial value combination corresponding to the at least one kernel parameter, and an association relationship between the value combination corresponding to the at least one kernel parameter and the performance data and the power consumption data of the target device, that is, a performance-power consumption estimation model, can be obtained through the regression analysis.

The regression analysis includes a plurality of linear regression analysis, logistic regression analysis, and the like. In an alternative embodiment of the present application, linear regression analysis is preferred for modeling. Based on this, the process of model computing device 101 performing model training includes: and performing linear regression analysis by taking partial value combinations corresponding to at least one kernel parameter as independent variables and taking performance data and power consumption data of the target equipment under the partial value combinations as dependent variables to obtain a performance-power consumption estimation model. The performance-power consumption prediction model obtained in this alternative embodiment is a linear regression model.

In an alternative embodiment, after obtaining the performance-power consumption estimation model, the model computing device 101 may estimate performance data and power consumption data of the target device under other value combinations corresponding to at least one kernel parameter by using the performance-power consumption estimation model, as shown in fig. 2. Compared with a mode of testing by using a load testing tool, the performance data and the power consumption data of the target equipment under other value combinations are estimated by using the performance-power consumption estimation model, the speed is much higher, and a large amount of time cost can be saved. The performance data and the power consumption data of the target equipment under various value combinations corresponding to at least one kernel parameter can be obtained by combining the performance data and the power consumption data of the target equipment under the other value combinations corresponding to at least one kernel parameter, which are estimated by the performance-power consumption estimation model, and the performance data and the power consumption data of the target equipment under the various value combinations corresponding to at least one kernel parameter, which are tested by adopting the load testing tool. The "multiple value combinations" herein may be all value combinations of the at least one kernel parameter, or may be a part of all value combinations of the at least one kernel parameter. Whether the 'multiple value combinations' are all value combinations or partial value combinations of at least one kernel parameter, the 'multiple value combinations' include the 'partial value combinations' and the 'other value combinations' in the above. Alternatively, the first and second electrodes may be,

in another optional embodiment, after obtaining the performance-power consumption estimation model, the model computing device 101 may estimate performance data and power consumption data of the target device under a plurality of value combinations corresponding to at least one kernel parameter by using the performance-power consumption estimation model. Similarly, the "multiple value combinations" herein may be all value combinations of at least one kernel parameter, or may be a part of all value combinations of at least one kernel parameter. Whether the 'multiple value combinations' are all value combinations of at least one kernel parameter or partial value combinations, the 'multiple value combinations' comprise the 'partial value combinations' and the 'rest value combinations' in the above.

After obtaining the performance data and the power consumption data of the target device under the multiple value combinations corresponding to the at least one kernel parameter, the model computing device 101 may actively send the performance data and the power consumption data of the target device under the multiple value combinations corresponding to the at least one kernel parameter to the target device based on the communication connection between the model computing device and the target device, or may also send the performance data and the power consumption data of the target device under the multiple value combinations corresponding to the at least one kernel parameter to the target device according to a request of the target device, or the target device may also actively download the performance data and the power consumption data of the target device under the multiple value combinations corresponding to the at least one kernel parameter to the model computing device 101; or related personnel may send the performance data and the power consumption data of the target device under the multiple value combinations corresponding to the at least one kernel parameter to the target device through the mobile storage device and copy the data from the model computing device 101 to the target device. The embodiment of the application does not limit the specific manner in which the target device obtains the performance data and the power consumption data of the target device under the multiple value combinations corresponding to the at least one kernel parameter.

For the target device, performance data and power consumption data of the target device under various value combinations corresponding to at least one kernel parameter can be acquired, in the parameter setting process, the performance data and the power consumption data of the target device under various value combinations corresponding to at least one kernel parameter are taken as the basis, a target value combination corresponding to at least one kernel parameter is determined, and at least one kernel parameter is set according to the target value combination. The process of setting at least one kernel parameter according to the target value combination comprises the following steps: and modifying the value of at least one kernel parameter related to the power consumption management mechanism in the source code of the operating system into the value in the target value combination, and then installing the operating system on the physical equipment according to the modified source code of the operating system. Wherein installing the operating system on the physical device according to the modified operating system source code comprises: and compiling the modified operating system source code to obtain an installation file of the operating system, and operating the installation file to complete the installation of the operating system.

In the embodiment of the present application, a specific implementation manner of determining a target value combination corresponding to at least one core parameter based on performance data and power consumption data of a target device under multiple value combinations corresponding to the at least one core parameter is not limited. Several alternative embodiments are exemplified below:

in the optional embodiment a, the target value combination can be directly determined from the multiple value combinations according to the performance data and the power consumption data of the target device under the multiple value combinations.

In optional embodiment B, power consumption data and/or performance data that the target device actually needs to meet may be obtained; matching power consumption data and/or performance data which are actually required to be met by target equipment with performance data and power consumption data of the target equipment under various value combinations; and selecting a value combination meeting the requirement of the matching degree from the multiple value combinations as a target value combination according to the matching degree between the power consumption data and/or the performance data actually required to be met by the target equipment and the performance data and the power consumption data of the target equipment under the multiple value combinations.

In optional embodiment C, at least one candidate value combination may be selected from the multiple value combinations according to performance data and power consumption data of the target device under the multiple value combinations; testing performance data and power consumption data of the target equipment under at least one candidate value combination by using a load testing tool; and further, determining a target value combination from at least one candidate value combination according to performance data and power consumption data of the target equipment under the at least one candidate value combination tested by the load testing tool. In the embodiment C, in addition to performance data and power consumption data of the target device under various value combinations, the accuracy of the finally selected target value combination is improved by combining a load testing tool, and the accuracy of parameter setting is improved.

Further, in the embodiment C, in the process of selecting at least one candidate value combination from the multiple value combinations according to the performance data and the power consumption data of the target device under the multiple value combinations, the at least one candidate value combination may be directly selected from the multiple value combinations according to the performance data and the power consumption data of the target device under the multiple value combinations. Or, at least one candidate value combination may be selected from the multiple value combinations by combining power consumption data that actually needs to be satisfied by the target device, or combining performance data that actually needs to be satisfied by the target device, or combining both power consumption data and performance data that actually needs to be satisfied by the target device.

The power consumption data and/or the performance data which are actually required to be met by the target equipment can be analyzed according to the QoS of the service which is actually required to be carried by the target equipment. Generally, each service has its own set of criteria to measure its service performance, such as how long it is acceptable to send a request to read the database, etc. Alternatively, the power consumption data and/or the performance data required by the service deployment user may also be obtained as the power consumption data and/or the performance data that the target device actually needs to meet. The service deployment user refers to a user who needs to deploy a service on the target device, and the service deployment user has certain requirements on the deployed service, for example, the required performance meets a certain standard, or the required power consumption cannot exceed a certain power threshold, and the like, and can acquire power consumption data that the target device actually needs to meet from the requirements on the power consumption, or acquire performance data that the target device actually needs to meet from the requirements on the performance.

For example, the target device may obtain power consumption data that actually needs to be satisfied, and match the power consumption data that actually needs to be satisfied by the target device with the performance data and the power consumption data of the target device under various value combinations corresponding to at least one kernel parameter; and acquiring a value combination which satisfies the matching degree requirement with the matching degree of the power consumption data actually required to be satisfied by the target equipment from the multiple value combinations as a candidate value combination. In this embodiment, according to the application requirement, the power consumption requirement of the target device may be prioritized, and the power consumption requirement may be satisfied by the target device by setting the kernel parameter related to the power consumption management mechanism.

For another example, the target device may obtain performance data that actually needs to be satisfied, and match the performance data that actually needs to be satisfied by the target device with the performance data and the power consumption data of the target device under various value combinations corresponding to at least one kernel parameter; and acquiring a value combination which meets the matching degree requirement with the matching degree of the performance data actually required to be met by the target equipment from the multiple value combinations as a candidate value combination. In this embodiment, the performance requirement of the target device may be prioritized according to the application requirement, and the target device may meet the power consumption requirement by exchanging performance with energy by setting the kernel parameter related to the power consumption management mechanism.

For another example, the target device may obtain power consumption data and performance data that actually need to be satisfied, and match the performance data and power consumption data that the target device actually needs to satisfy with the performance data and power consumption data of the target device under various value combinations corresponding to at least one kernel parameter; and acquiring a value combination, which meets the matching degree requirement with the matching degree of the power consumption data and the performance data actually required to be met by the target equipment, from the multiple value combinations as a candidate value combination. In this embodiment, according to the application requirements, the power consumption and performance requirements of the target device may be considered at the same time, and the power consumption and performance may be considered by setting the kernel parameter related to the power consumption management mechanism.

Further, in embodiment C, the process of testing the performance data and the power consumption data of the target device under at least one candidate value combination by using the load testing tool is a process of testing the performance data and the power consumption data of the target device under each candidate value combination by using the load testing tool. Taking the first candidate value combination as an example, the process of testing the performance data and the power consumption data of the target device under the first candidate value combination by using the load testing tool includes: modifying the value of at least one kernel parameter in the source code of the operating system into the value in the first candidate value combination, and installing the operating system on the target equipment according to the modified source code of the operating system; and running a load testing tool on the target equipment to test the performance data and the power consumption data of the target equipment under the first candidate value combination. The first candidate value combination is any one of at least one candidate value combination.

Further, as shown in fig. 2, in embodiment C, after the target value combination is selected, the performance-power consumption estimation model may be modified by using performance data and power consumption data of the target device under the target value combination, which are tested by the load testing tool. The performance data and the power consumption data of the target equipment tested by the load testing tool under the target value combination conform to actual requirements better, and the performance data and the power consumption data serve as sample data to modify the performance-power consumption estimation model, so that the accuracy of the performance-power consumption estimation model is improved, and the estimated performance data and the estimated power consumption data conform to the actual requirements better.

In an exemplary embodiment, taking the target device as a device such as a server or a computer device using the Linux operating system as an example, the kernel-mode power consumption management mechanism supported by the target device may include: DVFS, then the at least one kernel parameter associated with DVFS includes: at least one of a lowest operating frequency at which the CPU can operate, a highest operating frequency at which the CPU can operate, and an adjustment mode of the operating frequency of the CPU. Taking the example of simultaneously including the 3 parameters, the model computing device 101 may utilize a load testing tool to test performance data and power consumption data of the target device under partial value combinations corresponding to the 3 parameters, perform model training by utilizing the performance data and power consumption data of the target device under partial value combinations corresponding to the 3 parameters to obtain a performance-power consumption estimation model, and further estimate the performance data and power consumption data of the target device under other value combinations corresponding to the 3 parameters based on the performance-power consumption estimation model to obtain the performance data and power consumption data of the target device under various value combinations corresponding to the 3 parameters; the target device sets the 3 parameters based on performance data and power consumption data of the target device under various value combinations corresponding to the 3 parameters.

Further, taking as an example that the target device is a device such as a server or a computer device using a Linux operating system, the kernel-mode power consumption management mechanism supported by the target device may include, in addition to the DVFS: c-state, C-state contains 3-11 levels of C-mode, each level of C-mode has its own corresponding entry time threshold. Based on this, at least one kernel parameter includes, in addition to the above-mentioned 3 parameters related to the DVFS, an entry time threshold corresponding to the C mode of each level at the C-state. If the C-state includes C modes of 6 levels, which is denoted as C1-C6, the at least one kernel parameter includes an entry time threshold corresponding to the C modes of 6 levels, and the 3 parameters required by DVFS, which are 9 parameters in total, specifically: scaling _ min _ freq, scaling _ max _ freq, scaling _ golden, target _ responsiveness: c1, target _ residual: c2, target _ redundancy: c3, target _ redundancy: c4, target _ residual: c5 and Target _ redundancy: C6. taking the example of simultaneously including the 9 parameters, the model computing device 101 may use a load testing tool to test performance data and power consumption data of the target device under the partial value combinations corresponding to the 9 parameters, perform model training by using the performance data and power consumption data of the target device under the partial value combinations corresponding to the 9 parameters to obtain a performance-power consumption estimation model, and further estimate the performance data and power consumption data of the target device under other value combinations corresponding to the 9 parameters based on the performance-power consumption estimation model to obtain the performance data and power consumption data of the target device under various value combinations corresponding to the 9 parameters, as shown in fig. 3. Then, the target device may set the 9 parameters based on the performance data and the power consumption data of the target device under various value combinations corresponding to the 9 parameters.

Of course, the core-mode power consumption management mechanism supported by the target device may include DVFS alone, DVFS and C-state simultaneously, or C-state alone. In the case where the core-state power management mechanism supported by the target device solely includes a C-state, the at least one core parameter associated with the power management mechanism includes: and the entry time threshold corresponding to the C mode of each level. If the C-state includes 11 levels of C-modes, the at least one kernel parameter includes 11 parameters, which are entry time thresholds corresponding to the 11 levels of C-modes. The model computing device 101 may use a load testing tool to test performance data and power consumption data of the target device under partial value combinations corresponding to the 11 parameters, perform model training by using the performance data and the power consumption data of the target device under partial value combinations corresponding to the 11 parameters to obtain a performance-power consumption estimation model, and further estimate the performance data and the power consumption data of the target device under other value combinations corresponding to the 11 parameters based on the performance-power consumption estimation model to obtain the performance data and the power consumption data of the target device under various value combinations corresponding to the 11 parameters; the target device sets the 11 parameters according to performance data and power consumption data of the target device under various value combinations corresponding to the 11 parameters.

In this embodiment of the application, for a power consumption management mechanism in a kernel mode, the model computing device 101 may combine setting operation of kernel parameters related to the power consumption management mechanism with artificial intelligence, train a performance-power consumption estimation model based on the artificial intelligence, and further obtain performance data and power consumption data of the physical device under various value combinations corresponding to the related kernel parameters based on the performance-power consumption estimation model, so that the physical device may set the related kernel parameters based on the performance data and the power consumption data under the various value combinations, may set appropriate values for the related kernel parameters, and takes into account the power consumption and performance of the physical device, and may be combined with the artificial intelligence, thereby improving parameter setting efficiency and reducing cost.

It should be noted that, in the above embodiment, in the setting operation of the kernel parameter related to the power consumption management mechanism, the performance-power consumption estimation model is trained by comprehensively considering both the power consumption and the performance, but the invention is not limited thereto.

For example, in some application scenarios, the performance of the device may be focused, the requirement on the power consumption of the device is not high or no, and a reasonable value needs to be set for a kernel parameter related to the performance of the device; therefore, the model computing device 101 may also train a performance estimation model in a similar manner, and provide the physical device 103 with performance data of the physical device under various value combinations corresponding to the kernel parameters related to the device performance, so that the physical device 103 sets the kernel parameters related to the device performance accordingly.

For the model computing device 101, at least one kernel parameter related to device performance may be determined from kernel parameters of the physical device; testing performance data of the target equipment under partial value combination corresponding to at least one kernel parameter by using a load testing tool; performing model training according to performance data of the target equipment under partial value combination to obtain a performance estimation model; and estimating the performance data of the target equipment under other value combinations corresponding to at least one kernel parameter by using the performance estimation model. For the target device, at least one kernel parameter related to the device performance can be determined from the kernel parameters of the target device; acquiring performance data of the model computing equipment 101 under various value combinations corresponding to at least one kernel parameter from the model computing equipment; determining a target value combination corresponding to at least one kernel parameter according to the performance data of the kernel under various value combinations; and setting at least one kernel parameter according to the target value combination so as to operate according to the value in the target value combination. For details and optional embodiments of the related description, reference may be made to the foregoing examples, which are not repeated herein.

For another example, in some application scenarios, the power consumption of the device may be focused, the requirement on the performance of the device is not high or no, and a reasonable value needs to be set for a kernel parameter related to the power consumption of the device; therefore, the model computing device 101 may also train a power consumption estimation model in a similar manner, and provide the physical device 103 with power consumption data of the model under various value combinations corresponding to the kernel parameters related to the device power consumption, so that the physical device 103 may set the kernel parameters related to the device performance accordingly.

For the model computing device 101, at least one kernel parameter related to device power consumption can be determined from kernel parameters of the physical device; testing power consumption data of the target equipment under the value combination of the part corresponding to the at least one kernel parameter by using a load testing tool; performing model training according to power consumption data of the target equipment under partial value combination to obtain a power consumption estimation model; and estimating power consumption data of the target equipment under other value combinations corresponding to the at least one kernel parameter by using the power consumption estimation model. For the target device, at least one core parameter related to the power consumption of the device can be determined from the core parameters of the target device; acquiring power consumption data of the model computing equipment 101 under various value combinations corresponding to the at least one kernel parameter from the model computing equipment; determining a target value combination corresponding to at least one kernel parameter according to the power consumption data of the core under various value combinations; and setting at least one kernel parameter according to the target value combination so as to operate according to the value in the target value combination. For details and optional embodiments of the related description, reference may be made to the foregoing embodiments, which are not repeated herein.

In the foregoing embodiments of the present application, a data center system is taken as an example for description, but the solution that combines the setting operation of the kernel parameter and the artificial intelligence provided in the embodiments of the present application is not limited to the data center system. The scheme combining the setting operation of the kernel parameters and the artificial intelligence provided by the embodiment of the application can be expanded to any system or equipment needing to set the kernel parameters of the equipment.

Fig. 4a is a schematic structural diagram of a device management system according to an exemplary embodiment of the present application. As shown in fig. 4a, the device management system 400 includes: at least one physical device 401 and at least one model computing device 402. The number of physical devices 401 is not limited in this embodiment, and may be one or more. Similarly, the number of the model calculation devices 402 is not limited in this embodiment, and may be one or multiple.

In addition, the present embodiment also does not limit the device forms of the physical device 401 and the model calculation device 402. The physical device 401 may include, but is not limited to, at least one of the following device modalities: server devices, computer devices, desktop computers, notebook computers, smart phones, tablet computers, network switching devices, and the like. The server device may be any device including, but not limited to: a regular server, an array of servers or a cloud server, etc. Model computing device 402 may be a server-side device such as a conventional server, a cloud server, or an array of servers. The device modalities and the number of the physical devices 401 and the model calculation devices 402 shown in fig. 4a are merely examples, and are not limited thereto.

In this embodiment, the at least one physical device 401 is a physical device that is installed with an operating system and supports a kernel-mode power consumption management mechanism. It should be noted that, the device management system 400 may include, in addition to the physical device 401 installed with the operating system and supporting the kernel-mode power consumption management mechanism, some devices that do not need the operating system and some devices that are installed with the operating system but do not support the kernel-mode power consumption management mechanism, which is not limited herein. In this embodiment as well as in other embodiments, the focus is on the physical device 401 that is installed with an operating system and supports a kernel-mode power consumption management mechanism.

The kernel-mode power consumption management mechanism supported by the physical device 401 is related to some kernel parameters, and these kernel parameters generally have various values. Different values of kernel parameters cause different energy-saving effects of a power consumption management mechanism. In order to generate a better or optimal energy-saving effect and meet performance requirements under different application scenes and application requirements, in the embodiment, the setting operation of the kernel parameters is combined with artificial intelligence, performance data and power consumption data of the physical device 401 under various value combinations corresponding to the relevant kernel parameters are obtained on the basis of an artificial intelligence model, and further, the setting of the kernel parameters is performed on the basis of the performance data and the power consumption data of the physical device 401 under various value combinations corresponding to the relevant kernel parameters, so that not only can appropriate values be set for the kernel parameters, but also the requirements of the physical device 401 on power consumption and performance can be considered at the same time, and the combination with the artificial intelligence can improve the parameter setting efficiency and reduce the cost.

In this embodiment, the at least one model computing device 402 is mainly configured to: for each physical device 401, at least one core parameter related to a core-state power consumption management mechanism supported by the physical device 401 is determined from the core parameters of the physical device 401, and performance data and power consumption data of the physical device 401 under various value combinations corresponding to the at least one core parameter are provided for the physical device 401 based on an artificial intelligence model, so that the physical device 401 can complete setting of the related core parameters accordingly.

For convenience of description, the present embodiment takes any one physical device 401 as an example, and the detailed operation principle of the system is explained by taking any one physical device 401 as a target device.

The model computing device 402 may determine, from the core parameters of the target device, at least one core parameter associated with a core-state power consumption management mechanism supported by the target device. Then, the model computing device 402 may collect performance data and power consumption data of the target device under a partial value combination corresponding to at least one kernel parameter as sample data, and perform model training using the sample data to obtain a performance-power consumption estimation model.

Optionally, the model computing device 402 may collect, by means of a load testing tool, performance data and power consumption data of the target device under a partial value combination corresponding to the at least one core parameter. And installing and running a load testing tool on the target equipment, wherein the load testing tool can simulate the performance data and the power consumption data of the target equipment under different load conditions. In detail, partial value combinations are selected from multiple value combinations corresponding to at least one kernel parameter, and the number of the partial value combinations only needs to meet the number required by model training; aiming at each value combination in the partial value combinations, according to load, power consumption and/or performance requirements which need to be met by the target equipment, utilizing a load test tool to obtain performance data and power consumption data when the target equipment meets corresponding requirements under the value combination; the performance data and power consumption data of the target device meeting the corresponding requirements under the partial value combinations are provided to the model computing device 402. In order to simplify the description, the performance data and the power consumption data of the target device meeting the corresponding requirements under each value combination are referred to as the performance data and the power consumption data of the target device under each value combination for short.

The process of obtaining the performance data and the power consumption data of the target equipment under the value combination by using the load test tool aiming at each value combination in the partial value combinations comprises the following steps: aiming at the value combination, firstly, modifying the value of at least one kernel parameter in the source code of the operating system into the value in the value combination, and installing the operating system on the target equipment according to the modified source code of the operating system; after the operating system is successfully installed, a load testing tool is installed on the target equipment, and according to the load, power consumption and/or performance requirements which the target equipment should meet, the corresponding load condition is simulated by using the load testing tool and the performance data and the power consumption data of the target equipment under the corresponding load condition are obtained.

After obtaining performance data and power consumption data of the target device under the partial value combination corresponding to the at least one kernel parameter, the model computing device 402 performs model training using the data as sample data. In this embodiment, the process of model training performed by the model computing device 402 is not limited, and for example, the process may be a model training process based on a deep neural network, or a model training process based on regression analysis, and all the model training methods that can analyze the association relationship between the value combination corresponding to the at least one kernel parameter and the performance data and the power consumption data of the target device are suitable for the embodiments of the present application.

In an optional embodiment, the model computing device 402 may adopt a modeling method based on regression analysis, that is, a regression analysis is performed on the performance data and the power consumption data of the target device under the partial value combination corresponding to at least one kernel parameter, and an association relationship between the value combination corresponding to the at least one kernel parameter and the performance data and the power consumption data of the target device, that is, a performance-power consumption estimation model, is obtained through the regression analysis.

Further optionally, the modeling is selected using linear regression analysis. Based on this, the process of model computing device 402 performing model training includes: and performing linear regression analysis by taking partial value combinations corresponding to at least one kernel parameter as independent variables and taking performance data and power consumption data of the target equipment under the partial value combinations as dependent variables to obtain a performance-power consumption estimation model.

In an alternative embodiment, after obtaining the performance-power consumption estimation model, the model computing device 402 may estimate performance data and power consumption data of the target device under other value combinations corresponding to the at least one kernel parameter by using the performance-power consumption estimation model. Alternatively, after obtaining the performance-power consumption estimation model, the model computing device 101 may estimate performance data and power consumption data of the target device under a plurality of value combinations corresponding to at least one kernel parameter by using the performance-power consumption estimation model.

For the target device, performance data and power consumption data of the target device under various value combinations corresponding to at least one kernel parameter can be acquired, in the parameter setting process, the performance data and the power consumption data of the target device under various value combinations corresponding to at least one kernel parameter are taken as the basis, a target value combination corresponding to at least one kernel parameter is determined, and at least one kernel parameter is set according to the target value combination. The process of setting at least one kernel parameter according to the target value combination comprises the following steps: and modifying the value of at least one kernel parameter related to the power consumption management mechanism in the operating system source code into the value in the target value combination, and then installing the operating system on the physical equipment according to the modified operating system source code. Wherein installing the operating system on the physical device according to the modified operating system source code comprises: and compiling the modified operating system source code to obtain an installation file of the operating system, and operating the installation file to complete the installation of the operating system.

Optionally, one embodiment of the determining the target value combination corresponding to the at least one kernel parameter includes: and directly determining the target value combination from the multiple value combinations according to the performance data and the power consumption data of the target equipment under the multiple value combinations. Or, another embodiment of the determining a target value combination corresponding to at least one kernel parameter includes: acquiring power consumption data and/or performance data which are actually required to be met by target equipment; matching power consumption data and/or performance data which are actually required to be met by target equipment with performance data and power consumption data of the target equipment under various value combinations; and selecting a value combination meeting the requirement of the matching degree from the multiple value combinations as a target value combination according to the matching degree between the power consumption data and/or the performance data actually required to be met by the target equipment and the performance data and the power consumption data of the target equipment under the multiple value combinations. Or, another embodiment of the determining the target value combination corresponding to the at least one kernel parameter includes: selecting at least one candidate value combination from the multiple value combinations according to the performance data and the power consumption data of the target equipment under the multiple value combinations; testing performance data and power consumption data of the target equipment under at least one candidate value combination by using a load testing tool; and further, determining a target value combination from at least one candidate value combination according to performance data and power consumption data of the target equipment under the at least one candidate value combination tested by the load testing tool. In this embodiment, in addition to performance data and power consumption data of the target device under various value combinations, the accuracy of the finally selected target value combination is improved by combining a load testing tool, and the accuracy of parameter setting is improved.

Further, in the process of selecting at least one candidate value combination, power consumption data actually required to be met by the target device may be combined, or performance data actually required to be met by the target device may be combined, or both the power consumption data and the performance data actually required to be met by the target device may be combined. For example, power consumption data that actually needs to be satisfied by the target device may be obtained, and the power consumption data and/or performance data that actually needs to be satisfied by the target device may be matched with the performance data and the power consumption data of the target device under a plurality of value combinations corresponding to at least one kernel parameter; and acquiring a value combination which meets the matching degree requirement with the matching degree of the power consumption data and/or the performance data actually required to be met by the target equipment from the plurality of value combinations as a candidate value combination.

Further, in the above embodiment, after the target value combination is selected, the performance-power consumption estimation model may be modified by using performance data and power consumption data of the target device under the target value combination, which are tested by the load testing tool. The performance data and the power consumption data of the target equipment tested by the load testing tool under the target value combination conform to actual requirements better, and the performance data and the power consumption data serve as sample data to modify the performance-power consumption estimation model, so that the accuracy of the performance-power consumption estimation model is improved, and the estimated performance data and the estimated power consumption data conform to the actual requirements better.

For the related description of the model computing device 402, reference may be made to the related description of the model computing device 101 in the foregoing embodiment, and similarly, for the related description of the target device, reference may also be made to the corresponding description in the foregoing embodiment, which is not repeated herein.

It should be noted that the device management system 400 of this embodiment may be implemented as a system of any form or any nature, for example, the system may be a data center, or a cluster, or a computer room system, or an edge cloud network system, or a center cloud network system, and the like, which is not limited in this embodiment.

For example, the device management system 400 is implemented as an edge cloud network system. As shown in fig. 4b, the edge cloud network system includes: the system comprises edge computing nodes and a server deployed in a cloud or a client room. The server is communicated with the edge computing nodes through a network, and can respond to the requests of the edge computing nodes and provide related cloud services for the edge computing nodes; in addition, the server can also perform management and control, operation and maintenance and the like on the edge computing node. The edge computing node comprises a hardware infrastructure, a driver of the hardware infrastructure, an operating system, a relevant application program and the like. Hardware infrastructures include, but are not limited to: CPU, network card and memory. The edge computing node may include: the method comprises the steps of adding a base station, a home gateway, a personal computer, a smart phone, a street lamp, a traffic light and/or an electronic monitoring device on a building and the like in the edge cloud network.

In the edge cloud network system shown in fig. 4b, a server may have the function of the model computing device 402 in fig. 4a, and of course, a server dedicated to model training (for implementing the function of the model computing device 402 in fig. 4 a) may be additionally deployed, and an edge computing node may serve as the physical device 401 in fig. 4 a. Suppose that the edge cloud node supports two mechanisms of DFVS and C-state, and the C-state has 6 levels of C modes, wherein, the kernel parameters related to DFVS and C-state include: scaling _ min _ freq, scaling _ max _ freq, scaling _ golden, and target _ responsiveness corresponding to C1-C6 levels, for 9 parameters.

Before the edge computing node provides service for the customer, an operating system needs to be installed in the edge computing node for the customer, and kernel parameters related to DFVS and C-state in the edge computing node are set, so that the edge computing node works under the optimal parameter combination, and the purpose of considering both performance and power consumption is achieved.

In order to achieve the above purpose, the server may test the performance data and the power consumption data of the edge computing node under the partial value combinations corresponding to the above 9 parameters by using a load testing tool, and perform model training by using the performance data and the power consumption data of the edge computing node under the partial value combinations corresponding to the 9 parameters to obtain a performance-power consumption estimation model; further, performance data and power consumption data of the edge computing node under other value combinations corresponding to the 9 parameters are estimated based on the performance-power consumption estimation model, so that the performance data and the power consumption data of the edge computing node under various value combinations corresponding to the 9 parameters are obtained; and issuing the performance data and the power consumption data of the edge computing node under various value combinations corresponding to the 9 parameters to the edge computing node. The edge computing node receives performance data and power consumption data of the edge computing node under various value combinations corresponding to the 9 parameters, which are sent by the server, and sets the 9 parameters according to the performance data and the power consumption data of the edge computing node under various value combinations corresponding to the 9 parameters, so that the edge computing node can work under the optimal parameter combination after being started, and the power consumption is reduced as much as possible under the condition of ensuring the performance.

Fig. 5a is a schematic flowchart of a data processing method according to an exemplary embodiment of the present application. The data processing method can be used to set kernel parameters of a physical device. As shown in fig. 5a, the method comprises:

501. at least one core parameter related to a power consumption management mechanism of a core state supported by the physical device is determined from the core parameters of the physical device.

502. And acquiring performance data and power consumption data of the physical equipment under various value combinations corresponding to the at least one kernel parameter, wherein the performance data and the power consumption data under at least part of the value combinations are estimated based on a performance-power consumption estimation model.

503. And determining a target value combination corresponding to at least one kernel parameter according to the performance data and the power consumption data of the physical equipment under the various value combinations.

504. And setting the at least one kernel parameter according to the target value combination so that the power consumption management mechanism operates according to the values in the target value combination.

The execution subject of this embodiment may be a physical device, where the physical device runs with an operating system and supports a kernel-mode power consumption management mechanism. The physical device may coordinate or manage the kernel-mode power management mechanism by using some kernel parameters, which are parameters of the kernel-mode power management mechanism. The kernel parameters related to the kernel-mode power consumption management mechanism may have various values, and under different values, the power consumption management mechanism may produce different energy saving effects.

In this embodiment, the setting operation of the kernel parameters is combined with artificial intelligence, and the performance data and the power consumption data of the physical device under various value combinations corresponding to the relevant kernel parameters are obtained based on the artificial intelligence model, so that the physical device can set the relevant kernel parameters based on the performance data and the power consumption data of the physical device under various value combinations corresponding to the relevant kernel parameters, can set appropriate values for the relevant kernel parameters, and considers the power consumption and performance of the physical device, and can improve the parameter setting efficiency and reduce the cost by combining with the artificial intelligence.

In an alternative embodiment, one implementation of step 503 includes: selecting at least one candidate value combination from the multiple value combinations according to performance data and power consumption data of the physical equipment under the multiple value combinations; testing performance data and power consumption data of the physical equipment under at least one candidate value combination by using a load testing tool; and determining a target value combination from the at least one candidate value combination according to performance data and power consumption data of the physical equipment tested by the load testing tool under the at least one candidate value combination.

Further, after determining the target value combination, the method further comprises: and correcting the performance-power consumption estimation model by using performance data and power consumption data of the physical equipment under the target value combination tested by the load testing tool.

Further, the testing performance data and power consumption data of the physical device under at least one candidate value combination by using the load testing tool includes: aiming at the first candidate value combination, modifying the value of at least one kernel parameter in the source code of the operating system into the value in the first candidate value combination, and installing the operating system on the physical equipment according to the modified source code of the operating system; running a load testing tool on the physical equipment to test the performance data and the power consumption data of the physical equipment under the first candidate value combination; the first candidate value combination is any one of at least one candidate value combination.

Further, the selecting at least one candidate value combination from the plurality of value combinations according to the performance data and the power consumption data of the physical device under the plurality of value combinations includes: acquiring power consumption data which is actually required to be met by physical equipment; matching power consumption data which are actually required to be met by the physical equipment with performance data and power consumption data of the physical equipment under various value combinations; and acquiring at least one candidate value combination which satisfies the matching degree requirement with the matching degree of the power consumption data actually required to be satisfied by the physical equipment from the multiple value combinations. Alternatively, another embodiment of step 503 comprises: acquiring performance data actually required to be met by the physical equipment, and matching the performance data actually required to be met by the physical equipment with the performance data and the power consumption data of the physical equipment under various value combinations corresponding to at least one kernel parameter; and acquiring at least one candidate value combination which meets the matching degree requirement with the matching degree of the performance data actually required to be met by the physical equipment from the multiple value combinations. Alternatively, another embodiment of step 503 comprises: acquiring performance data and power consumption data which are actually required to be met by physical equipment, and matching the performance data and the power consumption data which are actually required to be met by the physical equipment with the performance data and the power consumption data of the physical equipment under various value combinations corresponding to at least one kernel parameter; and acquiring at least one candidate value combination which meets the matching degree requirement with the matching degree of the performance data and the power consumption data which are actually required to be met by the physical equipment from the plurality of value combinations.

Further optionally, the obtaining of the power consumption data actually required to be met by the physical device includes: analyzing power consumption data which are actually required to be met by the physical equipment according to the QoS of the service which is actually required to be carried by the physical equipment; or, acquiring power consumption data required by a service deployment user as power consumption data actually required to be met by the physical device.

Further optionally, the acquiring performance data actually required to be met by the physical device includes: analyzing performance data which is actually required to be met by the physical equipment according to the QoS of the service which is actually required to be carried by the physical equipment; or acquiring performance data required by a service deployment user as performance data actually required to be met by the physical equipment.

Further optionally, the obtaining of power consumption data and performance data that the physical device actually needs to meet includes: analyzing power consumption data and performance data which are actually required to be met by the physical equipment according to the QoS of the service which is actually required to be carried by the physical equipment; or, acquiring performance data and performance data required by a service deployment user as power consumption data and performance data actually required to be met by the physical device.

In an alternative embodiment, one implementation of step 502 includes: testing performance data and power consumption data of the physical equipment under partial value combinations in the multiple value combinations by using a load testing tool; performing model training according to performance data and power consumption data of the physical equipment under the partial value combination to obtain a performance-power consumption estimation model; and estimating the performance data and the power consumption data of the physical equipment under other value combinations in various value combinations by using the performance-power consumption estimation model.

Further optionally, performing model training according to performance data and power consumption data of the physical device under partial value combinations to obtain a performance-power consumption estimation model, including: and performing regression analysis on the performance data and the power consumption data of the physical equipment under partial value combination to obtain a performance-power consumption estimation model.

Furthermore, performing regression analysis on the performance data and the power consumption data of the physical equipment under partial value combination to obtain a performance-power consumption estimation model, which comprises the following steps: and performing linear regression analysis by taking partial value combinations as independent variables and taking performance data and power consumption data of the physical equipment under the partial value combinations as dependent variables to obtain a performance-power consumption estimation model.

In another optional embodiment, the obtaining performance data and power consumption data of the physical device under multiple value combinations corresponding to the at least one kernel parameter includes: receiving performance data and power consumption data of the physical equipment under various value combinations corresponding to at least one kernel parameter, which are sent by the model computing equipment; and the performance data and the power consumption data under at least part of value combinations are estimated by the model computing equipment based on a performance-power consumption estimation model. For detailed implementation of obtaining the performance-power consumption estimation model by the model computing device and estimating the performance data and the power consumption data of the physical device under at least part of value combinations based on the performance-power consumption estimation model, reference may be made to the foregoing embodiments, which are not described herein again.

Fig. 5b is a schematic flowchart of another data processing method according to an exemplary embodiment of the present application. As shown in fig. 5b, the method comprises:

51. at least one core parameter related to a power consumption management mechanism of a core state supported by the physical device is determined from the core parameters of the physical device.

52. And testing the performance data and the power consumption data of the physical equipment under the partial value combination corresponding to the at least one kernel parameter by using a load testing tool.

53. And performing model training according to the performance data and the power consumption data of the physical equipment under the partial value combination to obtain a performance-power consumption estimation model.

54. And estimating the performance data and the power consumption data of the physical equipment under other value combinations corresponding to the at least one kernel parameter by using the performance-power consumption estimation model.

In an alternative embodiment, one implementation of step 53 includes: and performing regression analysis on the performance data and the power consumption data of the physical equipment under the value combination of the parts to obtain a performance-power consumption estimation model.

Further, a linear regression analysis method may be used for model training. The model training process for obtaining the performance-power consumption estimation model comprises the following steps: and performing linear regression analysis by taking the partial value combinations as independent variables and taking performance data and power consumption data of the physical equipment under the partial value combinations as dependent variables to obtain a performance-power consumption estimation model.

In this embodiment, in combination with artificial intelligence, only performance data and power consumption data of the physical device under a partial value combination corresponding to the at least one kernel parameter need to be collected, a performance-power consumption estimation model is trained based on the collected data, and performance data and power consumption data of the physical device under other value combinations corresponding to the at least one kernel parameter are estimated based on the performance-power consumption estimation model, which can greatly improve data acquisition efficiency, reduce cost, and provide data conditions for setting the kernel parameter of the physical device.

In addition to the above method embodiments, the present application also provides the following method embodiments.

As shown in fig. 6a, another embodiment of a data processing method includes:

601. at least one kernel parameter related to device performance is determined from kernel parameters of the physical device.

602. And acquiring performance data of the physical equipment under various value combinations corresponding to the at least one kernel parameter, wherein at least part of the performance data under the value combinations is estimated based on a performance estimation model.

603. And determining a target value combination corresponding to at least one kernel parameter according to the performance data of the physical equipment under the various value combinations.

604. And setting the at least one kernel parameter according to the target value combination so that the physical equipment operates according to the values in the target value combination.

In this embodiment, the setting operation of the kernel parameters is combined with artificial intelligence, and the performance data of the physical device under various value combinations corresponding to the relevant kernel parameters is obtained based on the artificial intelligence model, so that the physical device can set the relevant kernel parameters based on the performance data of the physical device under various value combinations corresponding to the relevant kernel parameters, and can set appropriate values for the relevant kernel parameters, thereby being beneficial to ensuring the performance of the physical device.

As shown in fig. 6b, another embodiment of a data processing method includes:

61. at least one kernel parameter related to device performance is determined from kernel parameters of the physical device.

62. And testing the performance data of the physical equipment under the value combination of the part corresponding to the at least one kernel parameter by using a load testing tool.

63. And performing model training according to the performance data of the physical equipment under the partial value combination to obtain a performance estimation model.

64. And estimating the performance data of the physical equipment under other value combinations corresponding to at least one kernel parameter by using the performance estimation model.

In this embodiment, in combination with artificial intelligence, only performance data of the physical device under a partial value combination corresponding to the at least one kernel parameter needs to be collected, a performance estimation model is trained based on the collected data, and performance data of the physical device under other value combinations corresponding to the at least one kernel parameter is estimated based on the performance estimation model.

As shown in fig. 7a, another embodiment of a data processing method includes:

701. at least one core parameter related to the power consumption of the device is determined from the core parameters of the physical device.

702. And acquiring power consumption data of the physical equipment under various value combinations corresponding to the at least one kernel parameter, wherein at least part of the power consumption data under the value combinations is estimated based on a power consumption estimation model.

703. And determining a target value combination corresponding to at least one kernel parameter according to the power consumption data of the physical equipment under the plurality of value combinations.

704. And setting the at least one kernel parameter according to the target value combination so that the physical equipment operates according to the values in the target value combination.

In this embodiment, the setting operation of the kernel parameters is combined with artificial intelligence, and the power consumption data of the physical device under various value combinations corresponding to the relevant kernel parameters is obtained based on the artificial intelligence model, so that the physical device can set the relevant kernel parameters based on the power consumption data of the physical device under various value combinations corresponding to the relevant kernel parameters, and can set appropriate values for the relevant kernel parameters, which is beneficial to reducing the power consumption of the physical device.

As shown in fig. 7b, another embodiment of a data processing method includes:

71. at least one core parameter related to the power consumption of the device is determined from the core parameters of the physical device.

72. And testing the power consumption data of the physical equipment under the value combination of the part corresponding to the at least one kernel parameter by using a load testing tool.

73. And performing model training according to the power consumption data of the physical equipment under partial value combination to obtain a power consumption estimation model.

74. And estimating power consumption data of the physical equipment under other value combinations corresponding to the at least one kernel parameter by using the power consumption estimation model.

In this embodiment, in combination with artificial intelligence, only power consumption data of the physical device under a partial value combination corresponding to the at least one kernel parameter needs to be collected, a power consumption estimation model is trained based on the collected data, and power consumption data of the physical device under other value combinations corresponding to the at least one kernel parameter is estimated based on the power consumption estimation model.

As shown in fig. 7c, another embodiment of a data processing method includes:

711. at least one core parameter related to a power consumption management mechanism of a core state supported by the physical device is determined from the core parameters of the physical device.

712. And testing the performance data and the power consumption data of the physical equipment under the partial value combination corresponding to the at least one kernel parameter by using a load testing tool.

713. And performing model training according to performance data and power consumption data of the physical equipment under partial value combination to obtain a performance-power consumption estimation model.

714. Estimating performance data and power consumption data of the physical equipment under various value combinations corresponding to the at least one kernel parameter by using a performance-power consumption estimation model; wherein, the value combinations comprise the above-mentioned part of value combinations.

Besides the data processing method, the embodiment of the application also provides a task scheduling method. As shown in fig. 7d, the task scheduling method includes:

721. and acquiring the task to be scheduled and the performance requirement of the task to be scheduled.

722. And selecting the resource equipment which meets the performance requirement and the kernel-mode power consumption parameter value meets the set power consumption requirement from at least one resource equipment.

723. And scheduling the task to be scheduled to the resource equipment which meets the performance requirement and the kernel-mode power consumption parameter value meets the set power consumption requirement.

In this embodiment, the core-mode power consumption parameter value refers to a combination of values of at least one core parameter related to a core-mode power consumption management mechanism supported by the resource device. The kernel-mode power consumption management mechanism is a mechanism provided by an operating system for managing the power consumption of physical devices, is an operating system-level power consumption management mechanism, and needs to operate in a kernel mode.

In this embodiment, the type of the operating system is not limited, and may be, for example, a Linux operating system, a Windows operating system, a UNIX operating system, or a MAC operating system. The power management mechanism of kernel mode supported by different operating systems may vary. DVFS and C-state are examples of two kernel-mode power management mechanisms supported by the Linux operating system.

The kernel-state power consumption management mechanism is related to some kernel parameters, the kernel parameters are parameters of the kernel-state power consumption management mechanism, and values of the kernel parameters can affect the energy saving effect of the kernel-state power consumption management mechanism. For any resource device, the premise of normally using the power consumption management mechanism of the kernel mode is to preset the value of the kernel parameter related to the power consumption management mechanism of the kernel mode. This means that, in the device in the normal operating state, at least one core parameter related to the power consumption management mechanism in the core state has a certain value, and the combination of the values is the core-state power consumption parameter value of this embodiment. The kernel-state power consumption parameter value of the resource device can reflect the power consumption condition of the resource device to a certain extent.

Based on the above consideration, in a scenario that a machine room, a data center, an edge cloud network, or the like needs to perform task scheduling, the task scheduling may be performed in combination with a kernel-state power consumption parameter value of the resource device. The resource device refers to various physical devices responsible for executing tasks in a resource scheduling scene, and may be terminal devices such as a notebook computer, a tablet computer, a smart phone, or an edge computing node, or may be server-side devices such as some servers, a server cluster, a server array, or a cloud server. For example, a resource device may be a server or a cluster of servers, but is not so limited.

In a task scheduling scenario, task scheduling devices (or called task schedulers) are generally deployed and are mainly used for scheduling tasks to be scheduled to appropriate resource devices. The task scheduling device can run program codes such as task scheduling services, programs or plug-ins and the like to realize the following task scheduling process.

In the task scheduling process, a task to be scheduled and a performance requirement of the task to be scheduled need to be acquired. Optionally, the task scheduling device may provide a human-computer interaction interface for a user, and the user may provide the task to be scheduled and the performance requirement thereof through the human-computer interaction interface. The human-machine interface may be an application page, a web page or a command window. Or, the task scheduling device may also support voice interaction and recognition technology, and the user may submit the task to be scheduled and its performance requirement through a voice instruction. The performance requirements of different tasks to be scheduled may be different, which is not limited.

Then, according to the performance requirement of the task to be scheduled, selecting resource equipment which meets the performance requirement and the kernel-state power consumption parameter value of which meets the set power consumption requirement from at least one resource equipment; and scheduling the task to be scheduled to the resource equipment which meets the performance requirement and the kernel-mode power consumption parameter value meets the set power consumption requirement.

Optionally, in the process of selecting the resource device that meets the performance requirement and the core-state power consumption parameter value meets the set power consumption requirement, the performance data and the power consumption data of at least one resource device under the respective core-state power consumption parameter value may be used. Namely, according to the performance data and the power consumption data of at least one resource device under the respective kernel-mode power consumption parameter values, the resource device which meets the performance requirement and the kernel-mode power consumption parameter values meet the set power consumption requirement is selected.

In an optional implementation manner, selecting, according to performance data and power consumption data of at least one resource device under respective core-state power consumption parameter values, a resource device that meets the performance requirement and the core-state power consumption parameter values meet a set power consumption requirement, includes: selecting candidate resource equipment meeting the performance requirements according to the performance data of at least one resource equipment under the respective kernel-mode power consumption parameter values; and selecting resource equipment with power consumption data meeting the set power consumption requirement from the candidate resource equipment according to the power consumption data of the candidate resource equipment under the respective kernel-state power consumption parameter values. Wherein, the number of candidate resource devices meeting the performance requirement can be one or more. And the power consumption data of the candidate resource equipment under the kernel-state power consumption parameter value meets the set power consumption requirement, and the kernel-state power consumption parameter value of the candidate resource equipment meets the set power consumption requirement.

In another optional implementation, selecting, according to performance data and power consumption data of at least one resource device under respective core-state power consumption parameter values, a resource device that meets the performance requirement and the core-state power consumption parameter values meet a set power consumption requirement, includes: selecting candidate resource equipment with the kernel-state power consumption parameter value meeting the set power consumption requirement according to the power consumption data of at least one resource equipment under the respective kernel-state power consumption parameter value; and selecting resource equipment with power consumption data meeting the set power consumption requirement from the candidate resource equipment according to the performance data of the candidate resource equipment under the respective kernel-state power consumption parameter values.

In the above embodiments, the "setting of the power consumption requirement" is not limited, and may be flexibly set according to the application requirement. For example, in an optional embodiment, the setting of the power consumption requirement may require that the power consumption brought by the core-mode power consumption parameter value to the resource device is the lowest, and based on this, the resource device with the lowest power consumption data may be selected from the candidate resource devices that satisfy the performance requirement; and then scheduling the task to be scheduled to the resource equipment with the lowest power consumption data. In this alternative embodiment, the resource device with the best core-state power consumption parameter value may be selected with guaranteed performance.

Before using the performance data and the power consumption data of the at least one resource device under the respective core-state power consumption parameter value, the performance data and the power consumption data of the at least one resource device under the respective core-state power consumption parameter value need to be obtained.

In some alternative embodiments, performance data and power consumption data for each resource device at the respective core-state power consumption parameter value may be collected by means of a load testing tool.

In other alternative embodiments, an artificial intelligence model, such as a performance-power consumption estimation model, may be trained in advance, and under the condition that the core-state power consumption parameter value of each resource device is known, the performance-power consumption estimation model may be used to predict the performance data and the power consumption data of each resource device under the respective core-state power consumption parameter value. For the training process and the using process of the performance-power consumption estimation model, reference may be made to the foregoing embodiments, and details are not repeated herein.

In still other embodiments, the core-state power consumption parameter value of each resource device is set by using the method provided in the foregoing embodiment of the present application, and in the process of setting the core-state power consumption parameter value for each resource device, the performance data and the power consumption data of each resource device under the respective core-state power consumption parameter value have been learned, and the performance data and the power consumption data of each resource device under the respective core-state power consumption parameter value, which are learned in advance, may be saved, so that, in the task scheduling process, the performance data and the power consumption data of each resource device under the respective core-state power consumption parameter value may be directly read. It should be noted that the core-state power consumption parameter value of each resource device refers to a target value combination finally set by using the method provided by the foregoing embodiment.

In the task scheduling method provided in the embodiment of the present application, in combination with the kernel-mode power consumption parameter value of the resource device, the resource device with a better kernel-mode power consumption parameter value can be preferentially selected to execute the task under the condition that the performance requirement is met, and the power consumption can be reduced while the performance requirement is taken into consideration.

It should be noted that the execution subjects of the steps of the methods provided in the above embodiments may be the same device, or different devices may be used as the execution subjects of the methods. For example, the execution subject of steps 501 to 504 may be device a; for another example, the execution subject of

steps

501, 503, and 504 may be device a, and the execution subject of step 502 may be device B; and so on.

In addition, in some of the flows described in the above embodiments and the drawings, a plurality of operations are included in a specific order, but it should be clearly understood that the operations may be executed out of the order presented herein or in parallel, and the sequence numbers of the operations, such as 501, 502, etc., are merely used for distinguishing different operations, and the sequence numbers themselves do not represent any execution order. Additionally, the flows may include more or fewer operations, and the operations may be performed sequentially or in parallel. It should be noted that, the descriptions of "first", "second", etc. in this document are used for distinguishing different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit the types of "first" and "second".

Fig. 8a is a schematic structural diagram of a physical device according to an exemplary embodiment of the present application. As shown in fig. 8a, the physical device includes: a memory 81a and a processor 82a.

The memory 81a is used for storing a computer program and may be configured to store other various data to support operations on the physical device. Examples of such data include instructions for any application or method operating on the physical device, contact data, phonebook data, messages, pictures, videos, and so forth.

A processor 82a, coupled to the memory 81a, for executing the computer program in the memory 81a for: determining at least one core parameter related to a core-state power consumption management mechanism supported by the physical device from the core parameters of the physical device; acquiring performance data and power consumption data of the physical equipment under various value combinations corresponding to the at least one kernel parameter, wherein the performance data and the power consumption data under at least part of the value combinations are estimated based on a performance-power consumption estimation model; determining a target value combination corresponding to the at least one kernel parameter according to the performance data and the power consumption data of the physical device under the plurality of value combinations; and setting the at least one kernel parameter according to the target value combination so that the power consumption management mechanism operates according to the values in the target value combination.

In an optional embodiment, the processor 82a, when determining the target combination of fetching values corresponding to the at least one kernel parameter, is specifically configured to: selecting at least one candidate value combination from the multiple value combinations according to the performance data and the power consumption data of the physical equipment under the multiple value combinations; testing performance data and power consumption data of the physical equipment under the at least one candidate value combination by using a load testing tool; and determining the target value combination from the at least one candidate value combination according to the performance data and the power consumption data of the physical equipment under the at least one candidate value combination, which are tested by the load testing tool.

Further optionally, the processor 82a is further configured to, after determining the target value combination, modify the performance-power consumption estimation model by using performance data and power consumption data of the physical device under the target value combination, which are tested by the load testing tool.

Further, when the processor 82a tests the performance data and the power consumption data of the physical device under the at least one candidate value combination by using the load testing tool, the processor is specifically configured to: aiming at a first candidate value combination, modifying the value of the at least one kernel parameter in the source code of the operating system into the value in the first candidate value combination, and installing the operating system on the physical equipment according to the modified source code of the operating system; running the load testing tool on the physical equipment to test performance data and power consumption data of the physical equipment under the first candidate value combination; wherein the first candidate value combination is any one of the at least one candidate value combination.

Further, when selecting at least one candidate value combination from the plurality of value combinations, the processor 82a is specifically configured to: acquiring power consumption data which is actually required to be met by physical equipment; matching power consumption data which are actually required to be met by the physical equipment with performance data and power consumption data of the physical equipment under various value combinations; and acquiring at least one candidate value combination which satisfies the matching degree requirement with the matching degree of the power consumption data actually required to be satisfied by the physical equipment from the multiple value combinations. Alternatively, the first and second liquid crystal display panels may be,

in an optional embodiment, when the processor 82a selects at least one candidate value combination from the multiple value combinations, the processor is specifically configured to: the method comprises the steps of acquiring performance data actually required to be met by physical equipment, and matching the performance data actually required to be met by the physical equipment with the performance data and the power consumption data of the physical equipment under various value combinations corresponding to at least one kernel parameter; and acquiring at least one candidate value combination which meets the matching degree requirement with the matching degree of the performance data actually required to be met by the physical equipment from the multiple value combinations. Alternatively, the first and second electrodes may be,

in an optional embodiment, when the processor 82a selects at least one candidate value combination from the multiple value combinations, the processor is specifically configured to: acquiring performance data and power consumption data which are actually required to be met by physical equipment, and matching the performance data and the power consumption data which are actually required to be met by the physical equipment with the performance data and the power consumption data of the physical equipment under various value combinations corresponding to at least one kernel parameter; and acquiring at least one candidate value combination which satisfies the matching degree requirement with the performance data and the power consumption data actually required to be satisfied of the physical equipment from the multiple value combinations.

Further optionally, when acquiring the power consumption data actually required to be satisfied by the physical device, the processor 82a is specifically configured to: analyzing power consumption data which are actually required to be met by the physical equipment according to the QoS of the service which is actually required to be carried by the physical equipment; or, acquiring power consumption data required by a service deployment user as power consumption data actually required to be met by the physical device.

Further optionally, when the processor 82a obtains the performance data actually required to be satisfied by the physical device, it is specifically configured to: analyzing performance data which is actually required to be met by the physical equipment according to the QoS of the service which is actually required to be carried by the physical equipment; or, acquiring the performance data required by the service deployment user as the performance data actually required to be met by the physical device.

Further optionally, when acquiring the power consumption data and the performance data that the physical device actually needs to meet, the processor 82a is specifically configured to: analyzing power consumption data and performance data which are actually required to be met by the physical equipment according to the QoS of the service which is actually required to be carried by the physical equipment; or acquiring performance data and performance data required by a service deployment user as power consumption data and performance data actually required to be met by the physical equipment.

In an optional embodiment, when the processor 82a acquires the performance data and the power consumption data of the physical device under the multiple value combinations corresponding to the at least one kernel parameter, the processor is specifically configured to: testing performance data and power consumption data of the physical equipment under partial value combinations in the multiple value combinations by using a load testing tool; performing model training according to performance data and power consumption data of the physical equipment under the partial value combination to obtain a performance-power consumption estimation model; and estimating the performance data and the power consumption data of the physical equipment under other value combinations in various value combinations by using the performance-power consumption estimation model.

Further optionally, when obtaining the performance-power consumption estimation model, the processor 82a is specifically configured to: and performing regression analysis on the performance data and the power consumption data of the physical equipment under partial value combination to obtain a performance-power consumption estimation model.

Further, the processor 82a is specifically configured to: and performing linear regression analysis by taking partial value combinations as independent variables and taking performance data and power consumption data of the physical equipment under the partial value combinations as dependent variables to obtain a performance-power consumption estimation model.

In an optional embodiment, when the processor 82a acquires the performance data and the power consumption data of the physical device under the multiple value combinations corresponding to the at least one kernel parameter, the processor is specifically configured to: receiving performance data and power consumption data of the physical equipment under various value combinations corresponding to at least one kernel parameter, which are sent by the model computing equipment; and the performance data and the power consumption data under at least part of value combinations are estimated by the model computing equipment based on a performance-power consumption estimation model. For detailed implementation of obtaining the performance-power consumption estimation model by the model computing device and estimating the performance data and the power consumption data of the physical device under at least part of value combinations based on the performance-power consumption estimation model, reference may be made to the foregoing embodiments, which are not described herein again.

Optionally, in addition to the above functions, the processor 82 of the present embodiment may also implement the following functions: determining at least one kernel parameter related to the device performance from kernel parameters of the physical device; acquiring performance data of the physical equipment under various value combinations corresponding to the at least one kernel parameter, wherein at least part of the performance data under the value combinations is estimated based on a performance estimation model; determining a target value combination corresponding to the at least one kernel parameter according to the performance data of the physical device under the plurality of value combinations; and setting the at least one kernel parameter according to the target value combination so that the physical equipment operates according to the values in the target value combination.

Optionally, in addition to the above functions, the processor 82 of the present embodiment may also implement the following functions: determining at least one core parameter related to the power consumption of the device from the core parameters of the physical device; acquiring power data of the physical equipment under various value combinations corresponding to the at least one kernel parameter, wherein at least part of the power data under the value combinations is estimated based on a power estimation model; determining a target value combination corresponding to the at least one kernel parameter according to the power data of the physical device under the plurality of value combinations; and setting the at least one kernel parameter according to the target value combination so that the physical equipment operates according to the values in the target value combination.

Further, as shown in fig. 8a, the physical device further includes: communication component 83a, display 84a, power component 85a, audio component 86a, and the like. Only some of the components are schematically shown in fig. 8a, and it is not meant that the physical device comprises only the components shown in fig. 8 a. In addition, the components within the dashed box in fig. 8a are optional components, not necessary components, according to the implementation form of the physical device. For example, when the physical device is implemented as a terminal device such as a smart phone, a tablet computer, or a desktop computer, the physical device may include components within a dashed box in fig. 8 a; when the physical device is implemented as a server device such as a conventional server, a cloud server, a data center, or a server array, the components within the dashed box in fig. 8a may not be included.

Accordingly, the present application also provides a computer readable storage medium storing a computer program, and when the computer program is executed by a processor, the computer program causes the processor to implement the steps in the method embodiments shown in fig. 5a, fig. 6a or fig. 7 a.

Fig. 8b is a schematic structural diagram of a model computing device according to an exemplary embodiment of the present application. As shown in fig. 8b, the physical device includes: a memory 81b and a processor 82b.

The memory 81b is used for storing a computer program and may be configured to store other various data to support operations on the physical device. Examples of such data include instructions for any application or method operating on the physical device, contact data, phonebook data, messages, pictures, videos, and so forth.

A processor 82b, coupled to the memory 81b, for executing the computer program in the memory 81b for: determining at least one core parameter related to a power consumption management mechanism of a core state supported by a physical device from core parameters of the physical device; testing performance data and power consumption data of the physical equipment under partial value combination corresponding to the at least one kernel parameter by using a load testing tool; performing model training according to the performance data and the power consumption data of the physical equipment under the partial value combination to obtain a performance-power consumption estimation model; and estimating performance data and power consumption data of the physical equipment under other value combinations corresponding to the at least one kernel parameter by using the performance-power consumption estimation model.

In an alternative embodiment, the processor 82b, when obtaining the performance-power consumption prediction model, is specifically configured to: and performing regression analysis on the performance data and the power consumption data of the physical equipment under the partial value combination to obtain a performance-power consumption estimation model.

Further optionally, the processor 82b is specifically configured to: and performing linear regression analysis by taking the partial value combination as an independent variable and taking the performance data and the power consumption data of the physical equipment under the partial value combination as dependent variables to obtain a performance-power consumption estimation model.

Optionally, in addition to the above functions, the processor 82 of the present embodiment may also implement the following functions: determining at least one kernel parameter related to the device performance from the kernel parameters of the physical device; testing the performance data of the physical equipment under the value combination of the part corresponding to the at least one kernel parameter by using a load testing tool; performing model training according to the performance data of the physical equipment under the partial value combination to obtain a performance estimation model; and estimating the performance data of the physical equipment under other value combinations corresponding to the at least one kernel parameter by using the performance estimation model.

Optionally, in addition to the above functions, the processor 82 of the present embodiment may also implement the following functions: determining at least one core parameter related to the power consumption of the device from the core parameters of the physical device; testing power data of the physical equipment under the partial value combination corresponding to the at least one kernel parameter by using a load testing tool; performing model training according to the power data of the physical equipment under the partial value combination to obtain a power estimation model; and estimating power data of the physical equipment under other value combinations corresponding to the at least one kernel parameter by using the power estimation model.

Optionally, in addition to the above functions, the processor 82 of the model calculation device of the present embodiment may also implement the following functions: determining at least one core parameter related to a core-state power consumption management mechanism supported by a physical device from core parameters of the physical device; testing performance data and power consumption data of the physical equipment under partial value combination corresponding to the at least one kernel parameter by using a load testing tool; performing model training according to the performance data and the power consumption data of the physical equipment under the partial value combination to obtain a performance-power consumption estimation model; estimating performance data and power consumption data of the physical equipment under various value combinations corresponding to the at least one kernel parameter by using the performance-power consumption estimation model; wherein the plurality of value combinations comprises the partial value combinations.

Further, as shown in fig. 8b, the model calculation apparatus further includes: communication component 83b, power component 85b, and the like. Only some of the components are schematically shown in fig. 8b, and it is not intended that the model computing device comprises only the components shown in fig. 8 b. Alternatively, the model computing device may be implemented as a server-side device such as a conventional server, a cloud server, a data center, or a server array.

Accordingly, the present application also provides a computer readable storage medium storing a computer program, and when the computer program is executed by a processor, the processor is caused to implement the steps in the method embodiments shown in fig. 5b, fig. 6b, fig. 7b or fig. 7 c.

Fig. 8c is a schematic structural diagram of a task scheduling apparatus according to an exemplary embodiment of the present application. As shown in fig. 8c, the task scheduling apparatus includes: a memory 81c and a processor 82c.

The memory 81c is used for storing a computer program and may be configured to store other various data to support operations on the task scheduling device. Examples of such data include instructions for any application or method operating on the task scheduling device, task lists, messages, pictures, videos, and so forth.

A processor 82c coupled to the memory 81c for executing the computer program in the memory 81c for: acquiring a task to be scheduled and a performance requirement of the task to be scheduled; selecting resource equipment which meets the performance requirement and the kernel-mode power consumption parameter value meets the set power consumption requirement from at least one resource equipment; scheduling the task to be scheduled to the resource equipment which meets the performance requirement and the kernel-state power consumption parameter value meets the set power consumption requirement; the kernel-mode power consumption parameter value refers to a value combination of at least one kernel parameter related to a kernel-mode power consumption management mechanism supported by the resource device.

Alternatively, the resource device may be a server, or a cluster of servers.

In an optional embodiment, when selecting the resource device that meets the performance requirement and the core-state power consumption parameter value meets the set power consumption requirement, the processor 82c is specifically configured to: and selecting the resource equipment which meets the performance requirement and the kernel-state power consumption parameter value meets the set power consumption requirement according to the performance data and the power consumption data of at least one resource equipment under the respective kernel-state power consumption parameter value.

Further optionally, the processor 82c is specifically configured to: selecting candidate resource equipment meeting the performance requirement according to the performance data of at least one resource equipment under the respective kernel-mode power consumption parameter value; and selecting resource equipment with power consumption data meeting the set power consumption requirement from the candidate resource equipment according to the power consumption data of the candidate resource equipment under the respective kernel-state power consumption parameter values.

Further, the processor 82c is specifically configured to: and selecting the resource equipment with the lowest power consumption data from the candidate resource equipment according to the power consumption data of the candidate resource equipment under the respective kernel-mode power consumption parameter values.

Further, as shown in fig. 8c, the task scheduling apparatus further includes: communications component 83c, display 84c, power component 85c, audio component 86c, and the like. Only some of the components are schematically shown in fig. 8c and it is not meant that the task scheduling apparatus comprises only the components shown in fig. 8 c. In addition, the components within the dashed box in fig. 8c are optional components, not mandatory components, according to the implementation form of the task scheduling apparatus. For example, when the task scheduling device is implemented as a terminal device such as a smartphone, a tablet computer, or a desktop computer, the components in the dashed box in fig. 8c may be included; when the task scheduling device is implemented as a server device such as a conventional server, a cloud server, a data center, or a server array, the components in the dashed box in fig. 8c may not be included.

Accordingly, the present application further provides a computer readable storage medium storing a computer program, and when the computer program is executed by a processor, the processor is caused to implement the steps in the method embodiment shown in fig. 7 d.

The communication components of fig. 8 a-8 c described above are configured to facilitate communication between the device in which the communication component is located and other devices in a wired or wireless manner. The device where the communication component is located can access a wireless network based on a communication standard, such as a WiFi, a 2G, 3G, 4G/LTE, 5G and other mobile communication networks, or a combination thereof. In an exemplary embodiment, the communication component receives a broadcast signal or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component may further include a Near Field Communication (NFC) module, radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, ultra Wideband (UWB) technology, bluetooth (BT) technology, and the like.

The displays in fig. 8 a-8 c described above include screens, which may include Liquid Crystal Displays (LCDs) and Touch Panels (TPs). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive an input signal from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation.

The power supply components of fig. 8 a-8 c described above provide power to the various components of the device in which the power supply components are located. The power components may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device in which the power component is located.

The audio components of fig. 8 a-8 c described above may be configured to output and/or input audio signals. For example, the audio component includes a Microphone (MIC) configured to receive an external audio signal when the device in which the audio component is located is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signal may further be stored in a memory or transmitted via a communication component. In some embodiments, the audio assembly further comprises a speaker for outputting audio signals.

As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include forms of volatile memory in a computer readable medium, random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.

Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrases "comprising one of 8230; \8230;" 8230; "does not exclude the presence of additional like elements in a process, method, article, or apparatus that comprises that element.

The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method of data processing, comprising:

determining at least one kernel parameter related to a kernel-mode power consumption management mechanism supported by a physical device from kernel parameters of the physical device, wherein the kernel parameter refers to a parameter in an operating system source code and is used for limiting a value range or a threshold value of a parameter which can be adjusted by the power consumption management mechanism in a power consumption management process;

acquiring performance data and power consumption data of the physical equipment under various value combinations corresponding to the at least one kernel parameter, wherein the performance data and the power consumption data under at least part of the value combinations are estimated based on a performance-power consumption estimation model;

determining a target value combination corresponding to the at least one kernel parameter according to the performance data and the power consumption data of the physical device under the various value combinations;

before the operating system is installed or reinstalled on the physical device, setting the at least one kernel parameter according to the target value combination, wherein the setting of the at least one kernel parameter refers to a process of modifying the value of the at least one kernel parameter in the source code of the operating system into the value in the target value combination;

installing an operating system on the physical device according to the operating system source code subjected to parameter setting, and starting a power consumption management mechanism to enable the power consumption management mechanism to perform parameter adjustment within a parameter value range or a threshold value limited by values in the target value combination in the power consumption management process so as to realize power consumption management.

2. The method of claim 1, wherein determining a target value combination corresponding to the at least one kernel parameter according to the performance data and the power consumption data of the physical device under the plurality of value combinations comprises:

selecting at least one candidate value combination from the multiple value combinations according to the performance data and the power consumption data of the physical equipment under the multiple value combinations;

testing performance data and power consumption data of the physical equipment under the at least one candidate value combination by using a load testing tool;

and determining the target value combination from the at least one candidate value combination according to the performance data and the power consumption data of the physical equipment under the at least one candidate value combination tested by the load testing tool.

3. The method of claim 2, wherein testing the performance data and the power consumption data of the physical device under the at least one candidate combination of values using a load testing tool comprises:

aiming at a first candidate value combination, modifying the value of the at least one kernel parameter in the source code of the operating system into the value in the first candidate value combination, and installing the operating system on the physical equipment according to the modified source code of the operating system;

running the load testing tool on the physical equipment to test performance data and power consumption data of the physical equipment under the first candidate value combination; wherein the first candidate value combination is any one of the at least one candidate value combination.

4. The method of claim 2, wherein selecting at least one candidate combination of values from the plurality of combinations of values according to the performance data and the power consumption data of the physical device under the plurality of combinations of values comprises:

acquiring power consumption data and/or performance data which are actually required to be met by the physical equipment;

matching the power consumption data and/or the performance data which are actually required to be met by the physical equipment with the performance data and the power consumption data of the physical equipment under the various value combinations;

and acquiring at least one candidate value combination which meets the matching degree requirement with the matching degree of the power consumption data and/or the performance data which is actually required to be met by the physical equipment from the plurality of value combinations.

5. The method of claim 4, wherein obtaining power consumption data and/or performance data that the physical device actually needs to meet comprises:

analyzing power consumption data and/or performance data which are actually required to be met by the physical equipment according to the QoS of the service which is actually required to be carried by the physical equipment; or

And acquiring power consumption data and/or performance data required by a service deployment user as the power consumption data and/or performance data which are actually required to be met by the physical equipment.

6. The method of claim 2, wherein after determining the target combination of values, further comprising:

and correcting the performance-power consumption estimation model by using the performance data and the power consumption data of the physical equipment under the target value combination, which are tested by the load testing tool.

7. The method of claim 1, wherein the power consumption management mechanism comprises: DVFS, then the at least one kernel parameter comprises: at least one of a lowest operating frequency at which the processor is capable of operating, a highest operating frequency at which the processor is capable of operating, and a throttling mode of the processor operating frequency.

8. The method of claim 7, wherein the power consumption management mechanism further comprises: c-state mechanism, then the at least one kernel parameter further comprises: the entry time threshold corresponding to the C mode of each level; wherein the entry time threshold represents at least a time that the physical device needs to remain after entering the respective C-mode.

9. The method of claim 1, wherein the kernel-mode power management mechanism supported by the physical device comprises: c-state mechanism, then the at least one kernel parameter comprises: the entry time threshold corresponding to the C mode of each level; wherein the entry time threshold represents at least a time that the physical device needs to remain after entering a respective C-mode.

10. The method according to any one of claims 1 to 9, wherein obtaining performance data and power consumption data of the physical device under a plurality of value combinations corresponding to the at least one kernel parameter comprises:

testing performance data and power consumption data of the physical equipment under partial value combinations in the multiple value combinations by using a load testing tool;

performing model training according to the performance data and the power consumption data of the physical equipment under the partial value combination to obtain a performance-power consumption estimation model;

and estimating the performance data and the power consumption data of the physical equipment under other value combinations in the multiple value combinations by using the performance-power consumption estimation model.

11. The method of claim 10, wherein performing model training based on the performance data and the power consumption data of the physical device under the partial value combinations to obtain a performance-power consumption estimation model comprises:

and performing regression analysis on the performance data and the power consumption data of the physical equipment under the partial value combination to obtain a performance-power consumption estimation model.

12. The method of claim 11, wherein performing a regression analysis on performance data and power consumption data of the physical device under the partial value combinations to obtain a performance-power consumption estimation model comprises:

and taking the partial value combination as an independent variable, and taking performance data and power consumption data of the physical equipment under the partial value combination as dependent variables to perform linear regression analysis to obtain a performance-power consumption estimation model.

13. The method according to any one of claims 1 to 9, wherein obtaining performance data and power consumption data of the physical device under a plurality of value combinations corresponding to the at least one kernel parameter comprises:

receiving performance data and power consumption data of the physical equipment under various value combinations corresponding to the at least one kernel parameter, which are sent by model computing equipment; and the performance data and the power consumption data under at least part of value combinations are estimated by the model computing equipment based on a performance-power consumption estimation model.

14. A method of data processing, comprising:

testing performance data and power consumption data of the physical equipment under the partial value combination corresponding to the at least one kernel parameter by using a load testing tool;

estimating performance data and power consumption data of the physical equipment under other value combinations corresponding to the at least one kernel parameter by using the performance-power consumption estimation model to obtain the performance data and the power consumption data under various value combinations corresponding to the at least one kernel parameter;

the target value combination corresponding to the at least one kernel parameter is used for the power consumption management mechanism to perform parameter adjustment within a parameter value range or a threshold value defined by values in the target value combination in the power consumption management process so as to achieve power consumption management, the target value combination is determined according to performance data and power consumption data under multiple value combinations corresponding to the at least one kernel parameter, and the value of the at least one kernel parameter is set as a value in the target value combination before an operating system is installed or reinstalled on a physical device.

15. The method of claim 14, wherein performing model training based on the performance data and the power consumption data of the physical device under the partial value combinations to obtain a performance-power consumption estimation model comprises:

16. The method of claim 15, wherein performing regression analysis on performance data and power consumption data of the physical device under the partial value combinations to obtain a performance-power consumption estimation model comprises:

and performing linear regression analysis by taking the partial value combination as an independent variable and taking the performance data and the power consumption data of the physical equipment under the partial value combination as dependent variables to obtain a performance-power consumption estimation model.

17. A data processing method, comprising:

acquiring performance data of the physical equipment under various value combinations corresponding to the at least one kernel parameter, wherein at least part of the performance data under the value combinations is estimated based on a performance estimation model;

determining a target value combination corresponding to the at least one kernel parameter according to the performance data of the physical device under the various value combinations;

installing an operating system on the physical device according to the operating system source code after parameter setting, and starting a power consumption management mechanism to enable the power consumption management mechanism to carry out parameter adjustment in a parameter value range or a threshold value limited by values in the target value combination in the process of carrying out power consumption management so as to realize power consumption management.

18. A data processing method, comprising:

testing the performance data of the physical equipment under the partial value combination corresponding to the at least one kernel parameter by using a load testing tool;

performing model training according to the performance data of the physical equipment under the partial value combination to obtain a performance estimation model;

estimating performance data of the physical equipment under other value combinations corresponding to the at least one kernel parameter by using the performance estimation model to obtain performance data under various value combinations corresponding to the at least one kernel parameter;

the target value combination corresponding to the at least one kernel parameter is used for the power consumption management mechanism to perform parameter adjustment within a parameter value range or a threshold value defined by values in the target value combination in the power consumption management process so as to achieve power consumption management, the target value combination is determined according to performance data under multiple value combinations corresponding to the at least one kernel parameter, and before an operating system is installed or reinstalled on a physical device, the value of the at least one kernel parameter is set to be a value in the target value combination.

19. A data processing method, comprising:

acquiring power data of the physical equipment under various value combinations corresponding to the at least one kernel parameter, wherein at least part of the power data under the value combinations is estimated based on a power estimation model;

determining a target value combination corresponding to the at least one kernel parameter according to the power data of the physical device under the plurality of value combinations;

20. A method of data processing, comprising:

testing power data of the physical equipment under the partial value combination corresponding to the at least one kernel parameter by using a load testing tool;

performing model training according to the power data of the physical equipment under the partial value combination to obtain a power estimation model;

estimating power data of the physical equipment under other value combinations corresponding to the at least one kernel parameter by using the power estimation model to obtain power data under various value combinations corresponding to the at least one kernel parameter;

the target value combination corresponding to the at least one kernel parameter is used for the power consumption management mechanism to perform parameter adjustment within a parameter value range or a threshold value defined by values in the target value combination in a power consumption management process so as to achieve power consumption management, the target value combination is determined according to power data under multiple value combinations corresponding to the at least one kernel parameter, and before an operating system is installed or reinstalled on a physical device, the value of the at least one kernel parameter is set to be a value in the target value combination.

21. A method of data processing, comprising:

testing performance data and power consumption data of the physical equipment under partial value combination corresponding to the at least one kernel parameter by using a load testing tool;

estimating performance data and power consumption data of the physical equipment under various value combinations corresponding to the at least one kernel parameter by using the performance-power consumption estimation model; wherein the plurality of value combinations comprises the partial value combinations;

the target value combination corresponding to the at least one kernel parameter is used for the power consumption management mechanism to perform parameter adjustment in a parameter value range or a threshold value defined by values in the target value combination in a power consumption management process so as to achieve power consumption management, the target value combination is determined according to performance data and power consumption data under multiple value combinations corresponding to the at least one kernel parameter, and the value of the at least one kernel parameter is set as a value in the target value combination before an operating system is installed or reinstalled on a physical device.

22. A device management system, comprising: at least one physical device and at least one model computing device; wherein the at least one physical device supports a power consumption management mechanism of a kernel state respectively;

the at least one model computing device is configured to determine, from kernel parameters of a target device, at least one kernel parameter related to a kernel-state power consumption management mechanism supported by the target device, and obtain performance data and power consumption data of the target device under multiple value combinations corresponding to the at least one kernel parameter based on an artificial intelligence model; the target device is any one of the at least one physical device, and the kernel parameter refers to a parameter in an operating system source code and is used for limiting a value range or a threshold value of a parameter which can be adjusted by the power consumption management mechanism in a power consumption management process;

the target device is configured to determine a target value combination corresponding to the at least one kernel parameter according to performance data and power consumption data of the target device under multiple value combinations corresponding to the at least one kernel parameter, which are obtained by the model computing device, and set the at least one kernel parameter according to the target value combination before the operating system is installed or reinstalled on the target device, where setting the at least one kernel parameter refers to a process of modifying a value of the at least one kernel parameter in a source code of the operating system to a value in the target value combination;

installing an operating system on target equipment according to an operating system source code subjected to parameter setting, and starting a power consumption management mechanism to enable the power consumption management mechanism to carry out parameter adjustment in a parameter value range or threshold value limited by values in the target value combination in the process of carrying out power consumption management so as to realize power consumption management.

23. The system of claim 22, wherein the model computing device is specifically configured to:

testing performance data and power consumption data of the target equipment under the partial value combination corresponding to the at least one kernel parameter by using a load testing tool;

performing model training according to the performance data and the power consumption data of the physical equipment under the partial value combination to obtain a performance-power consumption estimation model; and

and estimating performance data and power consumption data of the physical equipment under other value combinations corresponding to the at least one kernel parameter by using the performance-power consumption estimation model to obtain the performance data and the power consumption data of the target equipment under the various value combinations of the physical equipment.

24. A data center system, comprising: the system comprises model computing equipment and at least one machine room, wherein the at least one machine room comprises at least one piece of physical equipment, and the at least one piece of physical equipment supports a kernel-state power consumption management mechanism respectively;

the model computing device is used for determining at least one kernel parameter related to a kernel-mode power consumption management mechanism supported by a target device from kernel parameters of the target device, and obtaining performance data and power consumption data of the target device under various value combinations corresponding to the at least one kernel parameter based on an artificial intelligence model; the target device is any one of the at least one physical device, and the kernel parameter refers to a parameter in an operating system source code and is used for limiting a value range or a threshold value of a parameter which can be adjusted by the power consumption management mechanism in a power consumption management process;

25. A physical device, comprising: a memory and a processor;

the memory for storing a computer program;

the processor, coupled with the memory, to execute the computer program to:

determining at least one kernel parameter related to a kernel-mode power consumption management mechanism supported by the physical device from kernel parameters of the physical device, wherein the kernel parameter refers to a parameter in an operating system source code and is used for limiting a value range or a threshold value of a parameter which can be adjusted by the power consumption management mechanism in a power consumption management process;

determining a target value combination corresponding to the at least one kernel parameter according to the performance data and the power consumption data of the physical device under the plurality of value combinations;

before installing or reinstalling the operating system on the physical device, setting the at least one kernel parameter according to the target value combination so that the power consumption management mechanism operates according to the value in the target value combination, wherein the setting of the at least one kernel parameter refers to a process of modifying the value of the at least one kernel parameter in the source code of the operating system into the value in the target value combination;

installing an operating system on the physical device according to the operating system source code subjected to parameter setting, and starting a power consumption management mechanism to enable the power consumption management mechanism to carry out parameter adjustment within a parameter value range or a threshold value limited by values in the target value combination in the power consumption management process so as to realize power consumption management.

26. A model computing device, comprising: a memory and a processor;

the memory for storing a computer program;

the processor, coupled with the memory, to execute the computer program to:

27. A computer storage medium having a computer program stored thereon, which, when executed by a processor, causes the processor to carry out the steps of the method of any one of claims 1 to 21.

28. A method for task scheduling, comprising:

acquiring a task to be scheduled and a performance requirement of the task to be scheduled;

selecting resource equipment which meets the performance requirement and the kernel-mode power consumption parameter value meets the set power consumption requirement from at least one resource equipment;

scheduling the task to be scheduled to resource equipment which meets the performance requirement and the kernel-mode power consumption parameter value meets the set power consumption requirement;

the kernel-state power consumption parameter value refers to a value combination of at least one kernel parameter which is set before an operating system is installed or reinstalled on a resource device and is related to a kernel-state power consumption management mechanism supported by the resource device, performance data corresponding to a performance requirement and power consumption data corresponding to a power consumption requirement under the value combination of the at least one kernel parameter are obtained based on a performance-power consumption estimation model trained by artificial intelligence, and the kernel parameter refers to a parameter in an operating system source code and is used for limiting a value range or a threshold of a parameter which can be adjusted by the power consumption management mechanism in a power consumption management process.

29. The method of claim 28, wherein selecting a resource device from the at least one resource device that meets the performance requirement and whose core-state power consumption parameter value meets a set power consumption requirement comprises:

and selecting the resource equipment which meets the performance requirement and the kernel-state power consumption parameter value meets the set power consumption requirement according to the performance data and the power consumption data of the at least one resource equipment under the respective kernel-state power consumption parameter value.

30. The method of claim 29, wherein selecting resource devices that meet the performance requirement and whose core state power consumption parameter values meet a set power consumption requirement based on performance data and power consumption data of the at least one resource device at the respective core state power consumption parameter values comprises:

selecting candidate resource equipment meeting the performance requirement according to the performance data of the at least one resource equipment under the respective kernel-mode power consumption parameter values;

and selecting resource equipment with power consumption data meeting the set power consumption requirement from the candidate resource equipment according to the power consumption data of the candidate resource equipment under the respective kernel-state power consumption parameter values.

31. The method of claim 30, wherein selecting resource devices from the candidate resource devices whose power consumption data meets a set power consumption requirement based on the power consumption data of the candidate resource devices at the respective core-state power consumption parameter values comprises:

and selecting the resource equipment with the lowest power consumption data from the candidate resource equipment according to the power consumption data of the candidate resource equipment under the respective kernel-mode power consumption parameter values.

32. A task scheduling apparatus, comprising: a memory and a processor;

the memory for storing a computer program;

the processor, coupled with the memory, to execute the computer program to:

scheduling the task to be scheduled to the resource equipment which meets the performance requirement and the kernel-mode power consumption parameter value meets the set power consumption requirement;

33. A computer storage medium having a computer program stored thereon, which, when executed by a processor, causes the processor to carry out the steps of the method of any one of claims 28-31.