Computing power task scheduling method and device, and storage medium

Abstract

The application discloses a computing power task scheduling method and device and a storage medium, relates to the field of resource scheduling, and is used for reducing computing power task accumulation and resource waste. The method comprises the following steps: splitting a first computing power task to obtain at least two computing power task slices; for each computing power task slice, based on resource requirement information corresponding to one computing power task slice, a target matching factor is calculated, and based on the target matching factor, a target computing power task execution center matched with the one computing power task slice is determined; wherein the target matching factor is used for representing required computing power requirement corresponding to one computing power task slice; and N is an integer greater than 1.

Description

Technical Field

[0001] This application relates to the field of resource scheduling, and in particular to a computing power task scheduling method, apparatus and storage medium. Background Technology

[0002] Performing computationally intensive tasks (such as facial recognition, natural language processing, and real-time interactive games) on electronic devices requires more energy and computing resources. However, the limited computing resources on electronic devices cannot meet the needs of computationally intensive tasks. Therefore, computing power networks have emerged, allowing electronic devices to schedule computing resources in the network on demand.

[0003] In related technologies, when executing computing tasks, a computing task execution center is matched based on the resource requirements of the task, and then the task is executed at that center. However, if multiple computing tasks are matched to the same execution center that meets their needs, this can easily lead to task backlog and resource waste.

[0004] For example, the first computing task execution center includes 10 A devices and 10 B devices. Executing computing task 1 requires 6 A devices and 1 B device. If computing task 2 is also matched to the first computing task execution center during the execution of computing task 1, and requires 5 A devices and 5 B devices to execute, then the number of idle A devices in the first computing task execution center cannot meet the requirements for executing computing task 2. Therefore, the first computing task execution center needs to execute computing task 1 before executing computing task 2, which will result in a waste of resources. Summary of the Invention

[0005] This application provides a computing task scheduling method, apparatus, and storage medium to reduce computing task backlog and resource waste.

[0006] To achieve the above objectives, this application adopts the following technical solution:

[0007] Firstly, a computing power task scheduling method is provided, comprising: splitting a first computing power task to obtain at least two computing power task slices; for each computing power task slice, calculating a target matching factor based on the resource requirement information corresponding to a computing power task slice, and determining a target computing power task execution center that matches a computing power task slice based on the target matching factor; wherein, the target matching factor is used to characterize the required computing power requirement corresponding to a computing power task slice; N is an integer greater than 1.

[0008] Based on the above-described computing power task scheduling method, the first computing power task can be split into at least two computing power task slices. This allows for the calculation of a target matching factor (which characterizes the computing power requirement of a computing power task slice) for each slice based on its resource requirement information. Furthermore, based on the target matching factor for each slice, a target computing power task execution center matching each slice is determined. Thus, by splitting the first computing power task and obtaining the target matching factor for each slice, this application effectively reduces computing power task accumulation and resource waste.

[0009] In one possible implementation, the method of "calculating a target matching factor based on the resource requirement information corresponding to a computing power task slice" includes: obtaining the slice weight corresponding to each computing power task slice; calculating a first matching factor based on the slice weight corresponding to each computing power task slice and the computing power requirement information corresponding to each computing power task slice; calculating a second matching factor based on the slice weight corresponding to each computing power task slice and the target parameters of at least one computing power task execution center; and using the sum of the first matching factor and the second matching factor as the target matching factor; wherein the target computing power task execution center is one of at least one computing power task execution center; the slice weight of a computing power task slice includes the weight corresponding to each of the at least one business feature corresponding to the computing power task slice; the first matching factor is used to characterize the computing power requirement required by a computing power task slice in terms of computing equipment dimension; and the second matching factor is used to characterize the computing power requirement required by a computing power task slice in terms of electricity cost and network cost dimension.

[0010] In one possible implementation, the computing power requirement information corresponding to a computing power task slice includes: information on the single-cabinet density requirements of computing devices in the computing power task execution center for a computing power task slice, and information on the heat dissipation scheme requirements of the computing power task execution center for a computing power task slice; and / or, the target parameters include: the unit price of computing power resources for each computing power task execution center, the unit price of task computing power for each computing power task execution center, and the environmental parameters, node status parameters, and load awareness parameters of the multiple computing terminals included in each computing power task execution center.

[0011] In one possible implementation, the method for "obtaining the slice weight corresponding to each computing power task slice" includes: for each computing power task slice, decomposing the resource requirement index corresponding to a computing power task slice to obtain the computing power requirement corresponding to a computing power task slice; and calculating the slice weight corresponding to a computing power task slice based on the computing power requirement corresponding to a computing power task slice.

[0012] In one possible implementation, the method of "determining the target computing power task execution center that matches a computing power task slice based on the target matching factor" includes: obtaining the computing power factor of each computing power task execution center in at least one computing power task execution center, the computing power factor being used to characterize the computing power that the computing power task execution center can provide; and taking the computing power task execution center in at least one computing power task execution center whose computing power factor matches the target matching factor as the target computing power task execution center corresponding to a computing power task slice.

[0013] In one possible implementation, the aforementioned computing power factor is determined based on at least one of the following: the density of computing devices in a single cabinet within the computing power task execution center;

[0014] The heat dissipation solution adopted by the computing task execution center;

[0015] Environmental parameters of computing equipment in the computing task execution center;

[0016] Node status parameters of computing devices in the computing task execution center;

[0017] Load perception parameters of computing devices in the computing task execution center; unit price of computing resources and unit price of computing power for tasks in the computing task execution center.

[0018] In one possible implementation, after the above-mentioned "determining the target computing task execution center that matches a computing task slice based on the target matching factor", the computing task scheduling method further includes: if a first computing task execution center with a computing power factor higher than the target computing task execution center is found, the target computing task execution center that matches a computing task slice is changed to the first computing task execution center.

[0019] Secondly, a computing power task scheduling device is provided, which can be used to implement the method described in the first aspect or any possible design of the first aspect. This computing power task scheduling device can implement the functions performed by the computing power task scheduling device in the above aspects or possible designs, and the functions can be implemented by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above functions. For example, the computing power task scheduling device includes a splitting unit, a calculation unit, and a determination unit.

[0020] The splitting unit is used to split the first computing power task into at least two computing power task slices.

[0021] The computing unit is used to calculate the target matching factor for each computing task slice based on the resource requirement information corresponding to the computing task slice.

[0022] The determination unit is used to determine the target computing power task execution center that matches a computing power task slice based on the target matching factor calculated by the computing unit.

[0023] The target matching factor is used to characterize the computing power requirement corresponding to a computing power task slice; N is an integer greater than 1.

[0024] In one possible implementation, the computing power task scheduling device further includes an acquisition unit. The acquisition unit is used to acquire the slice weight corresponding to each computing power task slice. Specifically, the calculation unit is used to calculate a first matching factor based on the slice weight corresponding to each computing power task slice acquired by the acquisition unit and the computing power demand information corresponding to each computing power task slice; and to calculate a second matching factor based on the slice weight corresponding to each computing power task slice acquired by the acquisition unit and at least one target parameter of a computing power task execution center; and to use the sum of the first matching factor and the second matching factor as the target matching factor.

[0025] The target computing power task execution center is one of at least one computing power task execution center;

[0026] The slice weight of a computing power task slice includes the weight of each business feature in at least one business feature corresponding to the computing power task slice;

[0027] The first matching factor is used to characterize the computing power requirement of a computing task slice in terms of computing device dimension;

[0028] The second matching factor is used to characterize the computing power requirements of a computing power task slice in terms of electricity cost and network cost.

[0029] In one possible implementation, the computing power requirement information corresponding to a computing power task slice includes: information on the single-cabinet density requirements of computing devices in the computing power task execution center for a computing power task slice, and information on the heat dissipation scheme requirements of the computing power task execution center for a computing power task slice; and / or, the target parameters include: the unit price of computing power resources for each computing power task execution center, the unit price of task computing power for each computing power task execution center, and the environmental parameters, node status parameters, and load awareness parameters of the multiple computing terminals included in each computing power task execution center.

[0030] In one possible implementation, the aforementioned acquisition unit is specifically used to decompose the resource requirement indicators corresponding to a computing power task slice for each computing power task slice to obtain the computing power requirement corresponding to a computing power task slice; and to calculate the slice weight corresponding to a computing power task slice based on the computing power requirement corresponding to a computing power task slice.

[0031] In one possible implementation, the computing power task scheduling device further includes an acquisition unit. The acquisition unit is configured to acquire the computing power factor of each of the at least one computing power task execution centers, the computing power factor being used to characterize the computing power that the computing power task execution center can provide.

[0032] The determining unit is specifically used to identify, among at least one computing power task execution center, the computing power factor that matches the target matching factor, as the target computing power task execution center corresponding to a computing power task slice.

[0033] In one possible implementation, the aforementioned computing power factor is determined based on at least one of the following:

[0034] Density of computing equipment within a single cabinet in a computing task execution center;

[0035] The heat dissipation solution adopted by the computing task execution center;

[0036] Environmental parameters of computing equipment in the computing task execution center;

[0037] Node status parameters of computing devices in the computing task execution center;

[0038] Load perception parameters of computing devices in the computing task execution center;

[0039] The unit price of computing resources and the unit price of computing power for the computing power task execution center.

[0040] In one possible implementation, the computing power task scheduling device further includes a processing unit. The processing unit is configured to, after the determining unit determines a target computing power task execution center matching a computing power task slice based on a target matching factor, and if a first computing power task execution center with a computing power factor higher than the target computing power task execution center is found, change the target computing power task execution center matching a computing power task slice to the first computing power task execution center.

[0041] Thirdly, a computing power task scheduling device is provided. This device can be a computing power task scheduling device or a chip or system-on-a-chip within it. The computing power task scheduling device can implement the functions performed by the device in the above-mentioned aspects or possible designs. These functions can be implemented in hardware. For example, in one possible design, the computing power task scheduling device may include a processor and a communication interface. The processor can be used to support the computing power task scheduling device in implementing the functions involved in the first aspect or any possible design of the first aspect. For example, the processor determines a target computing power task execution center that matches the computing power task slice based on the target matching factor.

[0042] In another possible implementation, the computing task scheduling device may further include a memory for storing necessary computer execution instructions and data. When the computing task scheduling device is running, the processor executes the computer execution instructions stored in the memory to cause the computing task scheduling device to perform the computing task scheduling method described in the first aspect or any possible design of the first aspect.

[0043] Fourthly, a computing power task scheduling device is provided. This device can be a computing power task scheduling device or a chip or system-on-a-chip within a computing power task scheduling device. This computing power task scheduling device can implement the functions performed by the computing power task scheduling device in the above-mentioned aspects or possible designs. These functions can be implemented in hardware. For example, in one possible design, the computing power task scheduling device may include a processor and a communication interface. The processor can be used to support the computing power task scheduling device in implementing the functions involved in the first aspect or any possible design of the first aspect. For example, the processor determines a target computing power task execution center that matches the computing power task slice based on the target matching factor.

[0044] In another possible design, the computing task scheduling device may further include a memory for storing necessary computer execution instructions and data. When the computing task scheduling device is running, the processor executes the computer execution instructions stored in the memory to cause the computing task scheduling device to perform the computing task scheduling method described in the first aspect or any possible design of the first aspect.

[0045] Fifthly, a computer-readable storage medium is provided, which may be a readable non-volatile storage medium storing computer instructions or programs that, when run on a computer, enable the computer to execute the computing task scheduling method described in the first aspect or any possible design of the above aspects.

[0046] In a sixth aspect, a computer program product containing instructions is provided, which, when run on a computer, enables the computer to execute the computing task scheduling method described in the first aspect or any possible design of the above aspects.

[0047] In a seventh aspect, a computing power task scheduling device is provided. This device can be a computing power task scheduling device or a chip or system-on-a-chip within a computing power task scheduling device. The computing power task scheduling device includes one or more processors and one or more memories. The one or more memories are coupled to the one or more processors and are used to store computer program code, which includes computer instructions. When the one or more processors execute the computer instructions, the computing power task scheduling device performs the computing power task scheduling method as described in the first aspect or any possible design of the first aspect.

[0048] Eighthly, a chip system is provided, comprising a processor and a communication interface. This chip system can be used to implement the functions performed by the computing task scheduling device in the first aspect or any possible design of the first aspect. For example, the processor calculates a target matching factor for each computing task slice based on resource requirement information corresponding to the computing task slice, and determines a target computing task execution center matching a computing task slice based on the target matching factor. In one possible design, the chip system further includes a memory for storing program instructions and / or data. The chip system can be composed of chips or may include chips and other discrete devices, without limitation.

[0049] The technical effects of any of the design methods in aspects two through eight can be found in the technical effects of aspect one mentioned above, and will not be repeated here. Attached Figure Description

[0050] Figure 1 A flowchart illustrating a computing task scheduling method provided in an embodiment of this application;

[0051] Figure 2 A flowchart illustrating another computing power task scheduling method provided in this application embodiment;

[0052] Figure 3 A flowchart illustrating another computing power task scheduling method provided in this application embodiment;

[0053] Figure 4 A schematic diagram of a computing power task scheduling device provided in an embodiment of this application;

[0054] Figure 5 A schematic diagram of another computing power task scheduling device provided in the embodiments of this application;

[0055] Figure 6 This is a schematic diagram of another computing task scheduling device provided in the embodiments of this application. Detailed Implementation

[0056] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0057] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0058] It should also be understood that the term "comprising" indicates the presence of the described feature, whole, step, operation, element and / or component, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements and / or components.

[0059] In existing technologies, cross-regional scheduling employs the following strategy: Step 1: Periodically acquire and store regional cluster resource information. This step must collect necessary resource attributes. Step 2: Configure the weight ratio of resource attributes and the priority factor of regional sub-centers. This configuration is based on the analysis of the computing task model and the importance and access frequency of regional sub-centers. Step 3: Based on the priority factor of the regional sub-centers, compare the computing task resource requirements with the latest records of regional cluster resources. During matching, acquire the computing task resource configuration requirements and, combined with the weight ratio of the corresponding resources in the matching regional cluster, calculate the comprehensive weight value of the computing task in each region. Step 4: Based on the comprehensive weight value, determine the optimal matching region for the computing task. If regions with the same comprehensive weight value exist, they are randomly assigned.

[0060] Existing computing task scheduling strategies are based on static metrics, such as the number of memory cores, storage space, network bandwidth, and latency. Furthermore, computing tasks are distributed as a whole to the execution area for execution. This can easily lead to problems such as task accumulation in the resource pool and resource waste.

[0061] It is foreseeable that future computing power tasks will require increasingly sophisticated and customized solutions, and these needs may change at any time. Static resource scheduling strategies alone will struggle to meet the demands of future computing power tasks. This invention retains the analysis of static resource requirement indicators for computing power tasks and matches user computing power task needs from multiple perspectives through a multi-dimensional integrated scheduling strategy. Simultaneously, it dynamically manages the resources required for computing power tasks by slicing them, dynamically matching computing power tasks to specific regions based on real-time data. This ensures computational efficiency while reducing service costs.

[0062] In view of this, embodiments of this application provide a computing power task scheduling method to reduce computing power task backlog and resource waste during computing power task execution. The method includes: splitting a first computing power task to obtain at least two computing power task slices; for each computing power task slice, calculating a target matching factor based on the resource requirement information corresponding to the computing power task slice, and determining a target computing power task execution center that matches the computing power task slice based on the target matching factor.

[0063] Based on the above scheme, the first computing task can be split into at least two computing task slices. This allows for the calculation of a target matching factor (which characterizes the required computing power for a given computing task slice) based on the resource requirement information corresponding to each slice. Furthermore, based on the target matching factor for each slice, a target computing task execution center matching each slice is determined. Thus, by splitting the first computing task and obtaining the target matching factor for each slice, this application effectively reduces computing task backlog and resource waste.

[0064] The methods provided in the embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0065] The execution subject of the computing power task scheduling method provided in this application embodiment is a computing power task scheduling device. This application embodiment uses the execution of the computing power task scheduling method by the computing power task scheduling device as an example to illustrate the technical solution provided in this application embodiment.

[0066] For example, the aforementioned computing power task scheduling device can be the scheduling center in a computing network integrated system, or the computing device in the scheduling center.

[0067] This application provides a method for scheduling computing tasks. For example... Figure 1 As shown, the method may include S101 and S102:

[0068] S101. Split the first computing power task to obtain at least two computing power task slices.

[0069] In this embodiment of the application, after the computing power task scheduling device obtains the first computing power task, it can perform business slicing on the first computing power task to obtain at least two computing power task slices.

[0070] S102. For each computing power task slice, calculate the target matching factor based on the resource requirement information corresponding to a computing power task slice, and determine the target computing power task execution center that matches a computing power task slice based on the target matching factor.

[0071] In this embodiment of the application, the target matching factor is used to characterize the required computing power demand corresponding to a computing power task slice; N is an integer greater than 1.

[0072] In this embodiment of the application, the target computing power task execution center is one of at least one computing power task execution center.

[0073] In this embodiment, each computing power task slice corresponds to a resource requirement information. The computing power task scheduling device can calculate the target matching factor corresponding to each computing power task slice based on the resource requirement information corresponding to each of the at least two computing power task slices. Then, based on the target matching factor corresponding to each computing power task slice, the target computing power task execution center matching each computing power task slice is determined.

[0074] In one possible implementation, the resource requirement information corresponding to a computing task slice may include at least one of the following: central processing unit (CPU) size, memory size, disk throughput, network throughput, etc.

[0075] It can be understood that the target computing task execution center that matches a computing task slice is the optimal computing task execution center for executing that computing task slice.

[0076] In one possible implementation, the step of "calculating the target matching factor based on the resource requirement information corresponding to a computing power task slice" in S102 above can be specifically implemented through the following S102a to S102d:

[0077] S102a. Obtain the slice weight corresponding to each computing power task slice.

[0078] In this embodiment of the application, the slice weight of a computing power task slice includes the weight of each business feature in at least one business feature corresponding to the computing power task slice.

[0079] It is understandable that, for each computing power task slice, the computing power task scheduling device can obtain the weight of each business feature in at least one business feature (e.g., 4G memory, 8G video memory, etc.) corresponding to a computing power task slice, thereby obtaining the slice weight of a computing power task slice.

[0080] For example, the slice weights of a computing task slice are: image processing = 0.15, storage space = 0.1, network latency = 0.6, etc.

[0081] In one possible implementation, S102a can be specifically implemented through the following S102a1 and S102a2:

[0082] S102a1. For each computing power task slice, the resource requirement indicators corresponding to a computing power task slice are decomposed to obtain the computing power requirement corresponding to a computing power task slice.

[0083] In this embodiment of the application, after the computing power task scheduling device obtains at least two computing power task slices, it can obtain the resource requirement index corresponding to each computing power task slice, and decompose the resource requirement index corresponding to each computing power task slice into computing power requirements, so as to obtain the computing power requirements corresponding to each computing power task slice.

[0084] In one possible implementation, the aforementioned resource requirement indicators may include at least one of the following: CPU size, memory size, disk throughput, network throughput, etc.

[0085] In one possible implementation, the computing power requirement corresponding to each computing power task slice is used to determine the size of different types of computing power resources required for each computing power task slice.

[0086] In one possible implementation, the aforementioned computing power requirements may include computing power requirements for computation metrics, computing power requirements for storage metrics, and computing power requirements for input / output (I / O) metrics.

[0087] In one possible implementation, the computing power requirements of the aforementioned computational metrics are used to determine the CPU type, CPU clock frequency, etc.

[0088] In one possible implementation, the computing power requirements of the aforementioned storage metric class are used to determine the storage type, storage capacity, etc.

[0089] In one possible implementation, the computing power requirements of the aforementioned I / O metrics are used to determine bandwidth, latency, etc.

[0090] S102a2. Based on the computing power requirement corresponding to a computing power task slice, calculate the slice weight corresponding to a computing power task slice.

[0091] In this embodiment of the application, after the computing power task scheduling device obtains the computing power requirement corresponding to each computing power task slice, it can calculate the slice weight corresponding to each computing power task slice based on the computing power requirement corresponding to each computing power task slice.

[0092] Specifically, the computing power task scheduling device can assign weights to the computing power requirements corresponding to each computing power task slice based on the computing power requirements corresponding to each computing power task slice, so as to obtain the slice weight corresponding to each computing power task slice.

[0093] S102b: Calculate the first matching factor based on the slice weight corresponding to each computing power task slice and the computing power demand information corresponding to each computing power task slice.

[0094] In this embodiment of the application, the first matching factor is used to characterize the computing power requirement of a computing power task slice in the dimension of computing device.

[0095] In one possible implementation, the computing power requirement information corresponding to a computing power task slice includes: information on the single-cabinet density requirements of computing equipment in the computing power task execution center for a computing power task slice, and information on the heat dissipation scheme adopted by the computing power task execution center for a computing power task slice.

[0096] In one possible implementation, the computing task scheduling device can determine the type of each computing task slice by comparing it with the business model based on the slice weight corresponding to each computing task slice. Then, based on the type of each computing task slice, it can determine the single-cabinet interval density requirements of each computing task slice for computing equipment in the computing task execution center, as well as the requirements of each computing task slice for the heat dissipation scheme adopted by the computing task execution center, that is, determine the computing power demand information corresponding to each computing task slice.

[0097] In one possible implementation, the type of computing task slice can be any of the following: balanced, storage-based, or computation-based.

[0098] For example, after the computing task scheduling device obtains the slice weight corresponding to a computing task slice, it can classify the type of the computing task slice as "balanced" by comparing it with the business model. Based on the type of the computing task slice, it can determine that the single-rack interval density of computing equipment in the computing task execution center is required to be 1.3, and at the same time, it can require the computing task execution center to adopt the traditional rack indirect cooling evaporation scheme, thereby obtaining the computing demand information corresponding to the computing task slice.

[0099] In one possible implementation, for each computing task slice, the computing task scheduling device can record the computing demand information corresponding to a computing task slice as a first matching factor to obtain the first matching factor corresponding to each computing task slice.

[0100] S102c: Based on the slice weight corresponding to each computing power task slice and the target parameters of at least one computing power task execution center, calculate the second matching factor.

[0101] In this embodiment of the application, the second matching factor is used to characterize the computing power requirements of a computing power task slice in terms of electricity cost and network cost.

[0102] In one possible implementation, the target parameters include: the unit price of computing resources for each computing task execution center, the unit price of computing power for each task execution center, and the environmental parameters, node status parameters, and load awareness parameters of the multiple computing terminals included in each computing task execution center.

[0103] In one possible implementation, the computing task scheduling device can periodically record and save the environmental parameters, node status parameters, and load awareness parameters of multiple computing terminals included in the computing task execution center.

[0104] In one possible implementation, the computing task scheduling device can estimate the cost of sending each computing task slice to each computing task execution center for execution based on the slice weight corresponding to each computing task slice and the target parameters of each computing task execution center. Then, the information of the computing task execution center with the lowest electricity cost and network cost when executing each computing task slice is determined as the second matching factor.

[0105] For example, the computing power task scheduling device can estimate, based on the slice weight corresponding to a computing power task slice and the target parameters of each computing power task execution center, that a certain computing power task execution center in the Northwest region has the lowest combined electricity and network costs when executing the computing power task slice. Therefore, the information of a certain computing power task execution center in the Northwest region can be recorded as the second matching factor.

[0106] S102d. The sum between the first matching factor and the second matching factor is taken as the target matching factor.

[0107] In one possible implementation, after the computing power task scheduling device obtains the first matching factor and the second matching factor corresponding to each computing power task slice, it can calculate the first matching factor and the second matching factor corresponding to each computing power task slice according to a preset matching algorithm to obtain the target matching factor corresponding to each computing power task slice.

[0108] In one possible implementation, the step of "determining the target computing power task execution center that matches a computing power task slice based on the target matching factor" in S102 above can be specifically implemented through the following S102e and S102f:

[0109] S102e. Obtain the computing power factor of each computing power task execution center in at least one computing power task execution center, wherein the computing power factor is used to characterize the computing power that the computing power task execution center can provide.

[0110] In one possible implementation, the computing power factor is determined based on at least one of the following:

[0111] Density of computing equipment within a single cabinet in a computing task execution center;

[0112] The heat dissipation solution adopted by the computing task execution center;

[0113] Environmental parameters of computing equipment in the computing task execution center;

[0114] Node status parameters of computing devices in the computing task execution center;

[0115] Load perception parameters of computing devices in the computing task execution center;

[0116] The unit price of computing resources and the unit price of computing power for the computing power task execution center.

[0117] In one possible implementation, the computing power task scheduling device can determine the computing power that each computing power task execution center can provide based on the single-cabinet interval density of the computing equipment in each computing power task execution center, the heat dissipation scheme adopted by each computing power task execution center, target parameters, etc., thereby determining the computing power factor of each computing power task execution center.

[0118] S102f: Among at least one computing power task execution center, the computing power factor that matches the target matching factor is taken as the target computing power task execution center corresponding to a computing power task slice.

[0119] In this embodiment of the application, for each computing power task slice, the computing power task scheduling device can match the target matching factor corresponding to a computing power task slice with the computing power factor of each computing power task execution center, and then take the computing power task execution center corresponding to the computing power factor that matches the target matching factor as the target computing power task execution center corresponding to the computing power task slice, so as to execute the computing power task slice in the target computing power task execution center.

[0120] In one possible implementation, after S102 above, combined with Figure 1 ,like Figure 2As shown, the computing power task scheduling method provided in this application embodiment further includes the following S301:

[0121] S301. If a first computing power task execution center with a computing power factor higher than the target computing power task execution center is found, the target computing power task execution center that matches a computing power task slice is changed to the first computing power task execution center.

[0122] In one possible implementation, for each computing task slice, after a computing task slice is issued to the corresponding target computing task execution center, the computing task scheduling device can periodically search for computing task execution centers whose computing power factor is higher than that of the target computing task execution center based on the target matching factor corresponding to the computing task slice. When the first computing task execution center with a computing power factor higher than that of the target computing task execution center is found, the target computing task execution center for executing the computing task slice is changed to the first computing task execution center.

[0123] This application provides a computing power task scheduling method that can split a first computing power task into at least two computing power task slices. Based on the resource requirement information corresponding to each computing power task slice, a target matching factor (which represents the required computing power demand for a computing power task slice) can be calculated for each slice. Furthermore, based on the target matching factor corresponding to each computing power task slice, a target computing power task execution center matching each slice is determined. Thus, by splitting the first computing power task and obtaining the target matching factor for each slice, this application finds a computing power task execution center matching each slice, effectively reducing computing power task accumulation and resource waste.

[0124] One possible implementation is, such as Figure 3 As shown in the embodiments of this application, the computing power task scheduling method is applied to a computing network integrated system, which may include: a computing power task scheduling device and multiple computing power task execution centers.

[0125] Specifically, it includes the following S1 to S8:

[0126] S1. The computing power task scheduling device receives the first computing power task and splits the first computing power task to obtain at least two computing power task slices.

[0127] S2. The computing power task scheduling device obtains the resource requirement indicators corresponding to each computing power task slice, and decomposes the resource requirement indicators corresponding to each computing power task slice into computing power requirements to obtain the computing power requirements corresponding to each computing power task slice.

[0128] S3. The computing power task scheduling device assigns weights to the computing power requirements corresponding to each computing power task slice based on the computing power requirements corresponding to each computing power task slice, so as to obtain the slice weight corresponding to each computing power task slice.

[0129] S4. The computing power task scheduling device calculates the first matching factor based on the slice weight corresponding to each computing power task slice and the computing power demand information corresponding to each computing power task slice.

[0130] S5. The computing power task scheduling device calculates a second matching factor based on the slice weight corresponding to each computing power task slice and the target parameters of at least one computing power task execution center.

[0131] S6. The computing power task scheduling device uses the sum of the first matching factor and the second matching factor as the target matching factor.

[0132] S7. The computing power task scheduling device determines the target computing power task execution center that matches each computing power task slice based on the target matching factor.

[0133] S8. When the computing power task scheduling device finds a first computing power task execution center with a computing power factor higher than the target computing power task execution center, it changes the target computing power task execution center that matches a computing power task slice to the first computing power task execution center.

[0134] The various solutions in the above embodiments of this application can be combined without contradiction.

[0135] This application embodiment can divide the computing power task scheduling device into functional modules or functional units according to the above method examples. For example, each function can be divided into a separate functional module or functional unit, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or in software functional modules or functional units. The module or unit division in this application embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0136] When dividing each function into modules according to its corresponding function. Figure 4 A schematic diagram of a computing power task scheduling device 70 is shown. Figure 4The computing task scheduling device 70 shown may include: a splitting unit 701, a calculation unit 702, and a determination unit 703.

[0137] The splitting unit 701 is used to split the first computing power task to obtain at least two computing power task slices.

[0138] The calculation unit 702 is used to calculate the target matching factor for each computing power task slice based on the resource requirement information corresponding to the computing power task slice.

[0139] The determination unit 703 is used to determine the target computing task execution center that matches a computing task slice based on the target matching factor calculated by the computing unit 702.

[0140] The target matching factor is used to characterize the computing power requirement corresponding to a computing power task slice; N is an integer greater than 1.

[0141] In one possible implementation, combining Figure 4 ,like Figure 5 As shown, the computing power task scheduling device 70 also includes an acquisition unit 704.

[0142] The acquisition unit 704 is used to acquire the slice weight corresponding to each computing power task slice.

[0143] The aforementioned calculation unit 702 is specifically used to calculate a first matching factor based on the slice weight corresponding to each computing power task slice obtained by the acquisition unit 704 and the computing power demand information corresponding to each computing power task slice; and to calculate a second matching factor based on the slice weight corresponding to each computing power task slice obtained by the acquisition unit 704 and the target parameter of at least one computing power task execution center; and to use the sum between the first matching factor and the second matching factor as the target matching factor.

[0144] The target computing power task execution center is one of at least one computing power task execution center;

[0145] The slice weight of a computing power task slice includes the weight of each business feature in at least one business feature corresponding to the computing power task slice;

[0146] The first matching factor is used to characterize the computing power requirement of a computing task slice in terms of computing device dimension;

[0147] The second matching factor is used to characterize the computing power requirements of a computing power task slice in terms of electricity cost and network cost.

[0148] In one possible implementation, the computing power requirement information corresponding to a computing power task slice includes: information on the single-rack density requirements of computing equipment in the computing task execution center for a computing power task slice; information on the heat dissipation requirements of the computing task execution center for a computing power task slice; and / or,

[0149] The target parameters include: the unit price of computing resources for each computing task execution center, the unit price of computing power for each task execution center, and the environmental parameters, node status parameters, and load awareness parameters of the multiple computing terminals included in each computing task execution center.

[0150] In one possible implementation, the acquisition unit 704 is specifically used to decompose the resource requirement index corresponding to a computing power task slice for each computing power task slice to obtain the computing power requirement corresponding to a computing power task slice; and to calculate the slice weight corresponding to a computing power task slice based on the computing power requirement corresponding to a computing power task slice.

[0151] In one possible implementation, the computing power task scheduling device 70 further includes an acquisition unit 704.

[0152] The acquisition unit 704 is used to acquire the computing power factor of each computing power task execution center in at least one computing power task execution center. The computing power factor is used to characterize the computing power that the computing power task execution center can provide.

[0153] The aforementioned determining unit 703 is specifically used to identify, among at least one computing power task execution center, the computing power factor that matches the target matching factor, as a target computing power task execution center corresponding to a computing power task slice.

[0154] In one possible implementation, the computing power factor is determined based on at least one of the following:

[0155] Density of computing equipment within a single cabinet in a computing task execution center;

[0156] The heat dissipation solution adopted by the computing task execution center;

[0157] Environmental parameters of computing equipment in the computing task execution center;

[0158] Node status parameters of computing devices in the computing task execution center;

[0159] Load perception parameters of computing devices in the computing task execution center;

[0160] The unit price of computing resources and the unit price of computing power for the computing power task execution center.

[0161] In one possible implementation, combining Figure 4 ,like Figure 6 As shown, the computing power task scheduling device 70 also includes a processing unit 705.

[0162] The processing unit 705 is configured to, after the determining unit 703 determines the target computing power task execution center that matches a computing power task slice based on the target matching factor, change the target computing power task execution center that matches a computing power task slice to the first computing power task execution center if a first computing power task execution center with a computing power factor higher than the target computing power task execution center is found.

[0163] As another feasible approach Figures 4 to 6 The units in the diagram, including the splitting unit 701, the calculation unit 702, the determination unit 703, the acquisition unit 704, and the processing unit 705, can be replaced by a processor, which can be integrated. Figures 5 to 6 The function of the unit in the text.

[0164] This application also provides a computer-readable storage medium. All or part of the processes in the above method embodiments can be executed by a computer program instructing related hardware. This program can be stored in the computer-readable storage medium, and when executed, it can include the processes of the above method embodiments. The computer-readable storage medium can be an internal storage unit of the computing task scheduling device in any of the foregoing embodiments, such as the hard disk or memory of the computing task scheduling device. The computer-readable storage medium can also be an external storage device of the terminal device, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the terminal device. Further, the computer-readable storage medium can include both the internal storage unit of the computing task scheduling device and external storage devices. The computer-readable storage medium is used to store the computer program and other programs and data required by the computing task scheduling device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0165] It should be noted that the terms "first" and "second," etc., in the specification, claims, and drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0166] It should be understood that in this application, "at least one (item)" means one or more, "more than one" means two or more, "at least two (items)" means two or three or more, and "and / or" is used to describe the relationship between related objects, indicating that there can be three relationships. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the related objects before and after are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0167] Through the above description of the embodiments, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0168] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0169] The units described as separate components may or may not be physically separate. A component shown as a unit can be one or more physical units; that is, it can be located in one place or distributed in multiple different locations. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0170] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0171] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solutions of this application embodiment, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0172] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A computing power task scheduling method, characterized in that, The method includes: The first computing power task is split into at least two computing power task slices; For each computing power task slice, a target matching factor is calculated based on the resource requirement information corresponding to a computing power task slice, and a target computing power task execution center that matches the computing power task slice is determined based on the target matching factor. The target matching factor is used to characterize the required computing power demand corresponding to a computing power task slice; The calculation of the target matching factor based on the resource requirement information corresponding to a computing power task slice includes: Obtain the slice weight corresponding to each of the computing power task slices; The first matching factor is calculated based on the slice weight corresponding to each computing power task slice and the computing power demand information corresponding to each computing power task slice; Based on the slice weight corresponding to each computing power task slice and the target parameters of at least one computing power task execution center, a second matching factor is calculated. The sum of the first matching factor and the second matching factor is taken as the target matching factor; Wherein, the target computing power task execution center is one of the at least one computing power task execution centers; The slice weight of a computing power task slice includes the weight of each business feature in at least one business feature corresponding to the computing power task slice; The first matching factor is used to characterize the computing power requirement of the computing task slice in terms of computing device dimension; The second matching factor is used to characterize the computing power requirements of a computing power task slice in terms of electricity cost and network cost.

2. The method according to claim 1, characterized in that, The computing power requirement information corresponding to a computing power task slice includes: information on the single-rack density requirements of computing equipment in the computing task execution center for the computing power task slice; information on the heat dissipation requirements of the computing power task slice for the cooling scheme adopted by the computing task execution center; and / or, The target parameters include: the unit price of computing resources for each computing task execution center, the unit price of computing power for each task execution center, and the environmental parameters, node status parameters, and load awareness parameters of the multiple computing terminals included in each computing task execution center.

3. The method according to claim 1, characterized in that, The step of obtaining the slice weight corresponding to each computing power task slice includes: For each computing power task slice, the resource requirement index corresponding to a computing power task slice is decomposed to obtain the computing power requirement corresponding to the computing power task slice. Based on the computing power requirement corresponding to the computing power task slice, calculate the slice weight corresponding to the computing power task slice.

4. The method according to claim 1, characterized in that, The step of determining the target computing power task execution center that matches the computing power task slice based on the target matching factor includes: Obtain the computing power factor of each computing power task execution center in at least one computing power task execution center, wherein the computing power factor is used to characterize the computing power that the computing power task execution center can provide; Among the at least one computing power task execution centers, the computing power task execution center whose computing power factor matches the target matching factor is taken as the target computing power task execution center corresponding to the computing power task slice.

5. The method according to claim 4, characterized in that, The computing power factor is determined based on at least one of the following: The density of single-cabinet intervals of computing devices in the computing power task execution center; The heat dissipation scheme adopted by the computing power task execution center; Environmental parameters of the computing equipment in the computing task execution center; The node status parameters of the computing devices in the computing task execution center; The load perception parameters of the computing devices in the computing task execution center; The unit price of computing resources and the unit price of computing power for the computing power task execution center.

6. The method according to claim 1, characterized in that, After determining the target computing power task execution center that matches the computing power task slice based on the target matching factor, the method further includes: If a first computing power task execution center with a computing power factor higher than the target computing power task execution center is found, the target computing power task execution center that matches the computing power task slice is changed to the first computing power task execution center.

7. A computing power task scheduling device, characterized in that, The device includes: a splitting unit, a calculation unit, and a determination unit; The splitting unit is used to split the first computing power task to obtain at least two computing power task slices; The computing unit is used to calculate the target matching factor for each computing power task slice based on the resource requirement information corresponding to the computing power task slice. The determining unit is used to determine the target computing power task execution center that matches the computing power task slice based on the target matching factor calculated by the computing unit. The target matching factor is used to characterize the required computing power demand corresponding to a computing power task slice; The device further includes: an acquisition unit; The acquisition unit is used to acquire the slice weight corresponding to each computing power task slice; The calculation unit is specifically used to calculate a first matching factor based on the slice weight corresponding to each computing power task slice obtained by the acquisition unit and the computing power demand information corresponding to each computing power task slice; and to calculate a second matching factor based on the slice weight corresponding to each computing power task slice obtained by the acquisition unit and the target parameter of at least one computing power task execution center; and to use the sum between the first matching factor and the second matching factor as the target matching factor. Wherein, the target computing power task execution center is one of the at least one computing power task execution centers; The slice weight of a computing power task slice includes the weight of each business feature in at least one business feature corresponding to the computing power task slice; The first matching factor is used to characterize the computing power requirement of the computing task slice in terms of computing device dimension; The second matching factor is used to characterize the computing power requirements of a computing power task slice in terms of electricity cost and network cost.

8. The apparatus according to claim 7, characterized in that, The computing power requirement information corresponding to a computing power task slice includes: information on the single-rack density requirements of computing equipment in the computing task execution center for the computing power task slice; information on the heat dissipation requirements of the computing power task slice for the cooling scheme adopted by the computing task execution center; and / or, The target parameters include: the unit price of computing resources for each computing task execution center, the unit price of computing power for each task execution center, and the environmental parameters, node status parameters, and load awareness parameters of the multiple computing terminals included in each computing task execution center.

9. The apparatus according to claim 7, characterized in that, The acquisition unit is specifically used to decompose the resource requirement index corresponding to each computing power task slice to obtain the computing power requirement corresponding to the computing power task slice for each computing power task slice; and to calculate the slice weight corresponding to the computing power task slice based on the computing power requirement corresponding to the computing power task slice.

10. The apparatus according to claim 7, characterized in that, The device further includes: an acquisition unit; The acquisition unit is used to acquire the computing power factor of each computing power task execution center in at least one computing power task execution center, and the computing power factor is used to characterize the computing power that the computing power task execution center can provide; The determining unit is specifically used to select the computing power task execution center whose computing power factor matches the target matching factor from the at least one computing power task execution center as the target computing power task execution center corresponding to the computing power task slice.

11. The apparatus according to claim 10, characterized in that, The computing power factor is determined based on at least one of the following: The density of single-cabinet intervals of computing devices in the computing power task execution center; The heat dissipation scheme adopted by the computing power task execution center; Environmental parameters of the computing equipment in the computing task execution center; The node status parameters of the computing devices in the computing task execution center; The load perception parameters of the computing devices in the computing task execution center; The unit price of computing resources and the unit price of computing power for the computing power task execution center.

12. The apparatus according to claim 7, characterized in that, The device further includes: a processing unit; The processing unit is configured to, after the determining unit determines the target computing task execution center that matches the computing task slice based on the target matching factor, change the target computing task execution center that matches the computing task slice to the first computing task execution center if a first computing task execution center with a computing power factor higher than the target computing task execution center is found.

13. A computing power task scheduling device, characterized in that, include: The processor, memory, and communication interface; wherein the communication interface is used for communication by the computing power task scheduling device. The memory is used to store one or more programs, the one or more programs including computer execution instructions. When the computing power task scheduling device is running, the processor executes the computer execution instructions stored in the memory to cause the computing power task scheduling device to perform the method of any one of claims 1 to 6.

14. A computer-readable storage medium, characterized in that, The readable storage medium stores instructions that, when executed, implement the method as described in any one of claims 1 to 6.

Images