Computing power resource allocation method and device, medium and program product

Abstract

The invention relates to the technical field of computing power scheduling, and particularly provides a computing power resource allocation method and device, a medium and a program product, and the method comprises the steps: obtaining a target computing power impact factor of a service device cluster at a current moment, and estimating a computing power parameter weight based on the target computing power impact factor associated with each service device and a computing power parameter weight estimation model; determining a target computing power parameter weight of each service device at the current moment, collecting a target computing power parameter value of each service device at the current moment, and determining a computing power evaluation value of each service device based on the target computing power parameter value and the target computing power parameter weight; and determining a computing power resource allocation result based on the computing power demand information at the current moment and the computing power evaluation values of the plurality of service devices in the service device cluster. According to the method and the device, the accuracy and the reliability of the determined computing power condition of the service equipment can be improved, and the allocation reasonability of the computing power resources is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of computing power scheduling technology, and in particular to a computing power resource allocation method, device, medium and program product. Background Art

[0002] With the continuous development of digital infrastructure, the types and complexity of tasks that the digital infrastructure of data centers needs to handle are constantly increasing. In order to have a clearer understanding of the service capabilities of digital infrastructure equipment, it is necessary to evaluate the computing power of digital infrastructure equipment in order to make a more reasonable allocation of computing resources.

[0003] In the related art, for the service devices commonly used in the basic digital service devices, the computing power of the service devices can be evaluated by obtaining the common computing power parameters of the service devices, where the common computing power parameters may include Million Instructions Per Second (MIPS), Operations Per Second (OPS), or Floating-point Operations Per Second (FLOPS), etc.

[0004] However, the service equipment computing power evaluation scheme provided in the related technology usually uses a single-dimensional indicator to evaluate the service equipment computing power, resulting in low reliability of the service equipment computing power resource evaluation results, which in turn affects the rationality of computing power resource allocation in the business service process. Summary of the invention

[0005] The present disclosure is proposed in view of the above problems. The present disclosure provides a computing power resource allocation method, device, medium and program product, which can improve the accuracy and reliability of the computing power status of the determined service device while improving the rationality of the allocation of computing power resources.

[0006] According to one aspect of the present disclosure, a computing resource allocation method is provided, comprising:

[0007] Obtaining a target computing power impact factor of the service device cluster at the current moment, wherein the computing power impact factor includes device resource information of each service device in the service device cluster, pending task information of the service device cluster, and energy consumption impact information of the service device;

[0008] Determine the target computing power parameter weight of each service device at the current moment based on the target computing power influencing factor associated with each service device and the computing power parameter weight estimation model, wherein the computing power parameter includes at least one of an energy consumption parameter and a task processing duration, and a basic computing power parameter of the device;

[0009] Collecting the target computing power parameter value of each service device at the current moment, and determining the computing power evaluation value of each service device based on the target computing power parameter value and the target computing power parameter weight;

[0010] The computing power resource allocation result is determined based on the computing power demand information at the current moment and the computing power evaluation values ​​of multiple service devices in the service device cluster.

[0011] According to another aspect of the present disclosure, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the above-mentioned computing power resource allocation method.

[0012] According to another aspect of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned computing power resource allocation method is implemented.

[0013] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, which implements the above-mentioned computing power resource allocation method when executed by a processor.

[0014] The computing power resource allocation method, device, medium and program product provided by the present invention can, on the one hand, evaluate the computing power status of the service device based on at least one of the energy consumption parameters and the task processing time, as well as multiple dimensions of the basic computing power parameters of the device, thereby improving the accuracy and reliability of the determined computing power status of the service device, so as to improve the rationality of the allocation of computing power resources; on the other hand, it can combine the current device resource information, the information of the tasks to be processed, and the energy consumption impact information of the service device to determine the weight of the computing power parameters for each dimension, so as to improve the matching of the computing power status of the service device finally determined with the current actual needs or actual environmental conditions, thereby further improving the rationality of the allocation of computing power resources.

[0015] It is to be understood that both the foregoing general description and the following detailed description are exemplary, and are intended to provide further explanation of the technology as claimed. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The above and other purposes, features and advantages of the present disclosure will become more apparent by describing the embodiments of the present disclosure in more detail in conjunction with the accompanying drawings. The accompanying drawings are used to provide a further understanding of the embodiments of the present disclosure and constitute a part of the specification. Together with the embodiments of the present disclosure, they are used to explain the present disclosure and do not constitute a limitation of the present disclosure. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0017] Figure 1 It is a schematic diagram of an implementation scenario of a computing power resource allocation solution according to an embodiment of the present disclosure.

[0018] Figure 2 It is a flowchart illustrating a method for allocating computing resources according to an embodiment of the present disclosure.

[0019] Figure 3 It is a block diagram illustrating a computing power resource allocation device according to an embodiment of the present disclosure.

[0020] Figure 4 is a schematic diagram illustrating a computer program product according to an embodiment of the present disclosure.

[0021] Figure 5 is a hardware block diagram illustrating an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present disclosure more obvious, the exemplary embodiments according to the present disclosure will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all the embodiments of the present disclosure, and it should be understood that the present disclosure is not limited to the exemplary embodiments described here.

[0023] In the related art, for service devices commonly used in basic digital service devices, the computing power of the service devices can be evaluated by testing the common computing power parameters (for example, FLOPS) in the service devices; it can be understood that the processors installed on different service devices may be of different types. For example, the processors installed on different service devices may be central processing units (CPU), graphics processing units (GPU), field programmable gate arrays (FPGA) processors, application-specific integrated circuits (ASIC) processors, etc. For service devices equipped with different processors, different test suites can usually be designed to measure computing power parameters.

[0024] However, the solution provided in the related technology for evaluating the computing power of service equipment through a single-dimensional indicator has low reliability in the evaluation results of the computing power resources of the service equipment because the information considered in the process of determining the service equipment is relatively single, which in turn affects the rationality of the computing power resource allocation in the business service process.

[0025] In order to solve the above problems, the present disclosure provides a computing resource allocation solution, such as Figure 1 As shown, Figure 1 FIG. 1 is a schematic diagram showing an implementation scenario of a computing resource allocation solution provided by an exemplary embodiment of the present disclosure. Figure 1 As shown, the implementation scenario 100 includes a service device 101 and a computing resource allocation device 102, wherein the service device 101 is a device that can provide services for task processing in a data center, such as a server, or an electronic device such as a gateway, and the computing resource allocation device 102 can be an electronic device that performs computing power allocation, such as a computer, a notebook or a server.

[0026] Among them, a communication link can be established between the service device 101 and the computing power resource allocation device 102 to implement the computing power resource allocation solution provided in the embodiment of the present disclosure.

[0027] Figure 2 A flowchart of a computing resource allocation method according to an exemplary embodiment of the present disclosure is shown. The method can be applied to a computing resource allocation device, such as Figure 2 As shown, the method of the embodiment of the present disclosure may include:

[0028] Step S201, obtaining the target computing power impact factor of the service device cluster at the current moment;

[0029] The computing power impact factor includes the device resource information of each service device in the service device cluster, the pending task information of the service device cluster, and the energy consumption impact information of the service device;

[0030] Step S202, determining the target computing power parameter weight of each service device at the current moment based on the target computing power impact factor associated with each service device and the computing power parameter weight estimation model;

[0031] The computing power parameters include at least one of energy consumption parameters and task processing time, and basic computing power parameters of the device;

[0032] Step S203, collecting the target computing power parameter value of each service device at the current moment, and determining the computing power evaluation value of each service device based on the target computing power parameter value and the target computing power parameter weight;

[0033] Step S204, determining a computing power resource allocation result based on the computing power demand information at the current moment and the computing power evaluation values ​​of multiple service devices in the service device cluster;

[0034] To summarize, the computing power resource allocation method provided in the embodiments of the present disclosure, on the one hand, evaluates the computing power status of the service device based on at least one of the energy consumption parameters and the task processing time, as well as multiple dimensions of the basic computing power parameters of the device, which can improve the accuracy and reliability of the determined computing power status of the service device, so as to improve the rationality of the allocation of computing power resources; on the other hand, it can combine the current device resource information, the information of the tasks to be processed, and the energy consumption impact information of the service device to determine the weight of the computing power parameters for each dimension, so as to improve the matching between the computing power status of the service device finally determined and the current actual needs or actual environmental conditions, thereby further improving the rationality of the allocation of computing power resources.

[0035] The following Figure 2 The specific implementation methods of each step in the embodiment shown are described in detail:

[0036] In step S201, the computing resource allocation device may obtain the target computing power impact factor of the service device cluster at the current moment.

[0037] In the embodiment of the present disclosure, the computing power impact factor includes the device resource information of each service device in the service device cluster, the pending task information of the service device cluster, and the energy consumption impact information of the service device, wherein the service device cluster can be a cluster composed of service devices in a data center.

[0038] It should be noted that in the embodiments of the present disclosure, the device resource information may include at least one of the hardware resource information and network resource information of the service device, wherein the hardware resource information may include: processor type, remaining memory size, memory core frequency and / or I / O controller frequency, etc.; the network resource information may include: network bandwidth, network latency, throughput and / or packet loss, etc.

[0039] The information of tasks to be processed may include the task type, the task processing volume requirement per unit time and / or the resource requirement of a single task, etc.; the energy consumption impact information may include: energy consumption data of service equipment, and resource prices, etc.

[0040] In an optional embodiment, the process of the computing resource allocation device obtaining the target computing power impact factor of the service device cluster at the current moment may include: sending a device resource information acquisition request to each service device in the service device cluster, and obtaining the target device resource information at the current moment returned by each service device; and reading the target pending task information at the current moment in the task configuration information, and reading the target energy consumption impact information at the current moment in the energy consumption status information.

[0041] In step S202, the computing power resource allocation device may determine the target computing power parameter weight of each service device at the current moment based on the target computing power influence factor associated with each service device and the computing power parameter weight estimation model.

[0042] In the embodiment of the present disclosure, the computing power parameter includes at least one of an energy consumption parameter and a task processing time, and a basic computing power parameter of the device; the energy consumption parameter is used to characterize the energy consumption of the service device during a single task processing process, the task processing time is used to characterize the time spent on processing a single task, and the basic computing power parameter of the device is used to characterize the performance of the service device when only the hardware performance of the device is considered, wherein the basic computing power parameter of the device may be OPS or FLOPS, etc.

[0043] In an optional embodiment, when the computing power parameters include basic computing power parameters of the device, energy consumption parameters and task processing time, the process of training the computing power parameter weight estimation model of the computing power resource allocation device may include: obtaining a sample data set, wherein each sample in the sample data set includes a sample computing power influence factor, a sample computing power parameter value and a computing power parameter weight label value; and inputting the sample computing power influence factor into the computing power parameter weight estimation model to be trained to obtain a computing power parameter weight estimation value; then, based on the computing power parameter weight estimation value and the sample computing power parameter value, determining the sample computing power evaluation prediction value, and, based on the computing power parameter weight label value and the sample computing power parameter value, determining the sample computing power evaluation value; further, inputting the sample computing power evaluation prediction value and the sample computing power evaluation value into the loss function to obtain the loss function value, and when it is determined that the loss function value is less than the loss function threshold, determining that the computing power parameter weight estimation model converges to obtain the computing power parameter weight estimation model. During the training process of the computing power parameter weight estimation model, the sample computing power evaluation prediction value and the sample computing power evaluation value determined by combining the basic computing power parameters, energy consumption parameters and task processing time can be used to optimize the computing power parameter weight estimation model, thereby realizing the computing power parameter weight estimation model based on the basic computing power parameters, energy consumption parameters and task processing time, and improving the reliability of the trained computing power parameter weight estimation model.

[0044] It should be noted that in the embodiments of the present disclosure, the specific structure of the computing power parameter weight estimation model can be determined based on actual needs, and the embodiments of the present disclosure do not limit this; the specific loss function type and the loss function threshold can also be determined based on actual needs, and the embodiments of the present disclosure do not limit this.

[0045] In an optional embodiment, the computing power resource allocation device determines the target computing power parameter weight of each service device at the current moment based on the target computing power influence factor associated with each service device and the computing power parameter weight estimation model. The process may include: inputting the target computing power influence factor associated with each service device into the computing power parameter weight estimation model respectively to obtain the target computing power parameter weight of each service device at the current moment.

[0046] It can be understood that the target computing power parameter weight may include at least one of the target energy consumption parameter weight and the target task processing time weight, as well as the target device basic computing power parameter weight. Specifically, it can be determined based on actual needs. The embodiment of the present disclosure does not limit this. The sum of at least two target computing power parameter weights is 1.

[0047] In step S203, the computing power resource allocation device may collect the target computing power parameter value of each service device at the current moment, and determine the computing power evaluation value of each service device based on the target computing power parameter value and the target computing power parameter weight.

[0048] In the embodiment of the present disclosure, the target computing power parameter value may include at least one of a target energy consumption parameter value and a target task processing duration, and a target device basic computing power parameter value.

[0049] In an optional implementation, the process of the computing power resource allocation device collecting the target computing power parameter value of each service device at the current moment may include: sending a computing power parameter value acquisition instruction to each service device, and obtaining the target computing power parameter value of each service device at the current moment returned by each service device. The target computing power parameter value of each service device can be obtained at the moment when computing power resource allocation is required to perform computing power resource allocation, thereby ensuring the reliability of computing power resource allocation.

[0050] In an optional implementation, the process of the computing power resource allocation device collecting the target computing power parameter value of each service device at the current moment may include: reading the target computing power parameter value of each service device at the current moment from the computing power parameter value information fed back in real time by each service device to the computing power resource allocation device. Since there is no need for the computing power resource allocation device to interact with each service device to obtain the target computing power parameter value at the current moment, the efficiency of obtaining the target computing power parameter value can be improved, thereby improving the efficiency of determining the computing power resource allocation result.

[0051] In an optional embodiment, when the target computing power parameter value includes the target device basic computing power parameter value, the target energy consumption parameter value and the target task processing time, and the target computing power parameter weight includes the target device basic computing power parameter weight, the target energy consumption parameter weight and the target task processing time weight; the computing power resource allocation device determines the computing power evaluation value of each service device based on the target computing power parameter value and the target computing power parameter weight, and the process may include: for each service device, determining the product of the target device basic computing power parameter value and the target device basic computing power parameter weight to obtain a weighted target device basic computing power parameter value; and determining the product of the target energy consumption parameter value and the target energy consumption parameter weight to obtain a weighted target energy consumption parameter value; and determining the product of the target task processing time and the target task processing time weight to obtain a weighted target task processing time; further, determining the weighted target computing power parameter value associated with each service device, the sum of the weighted target energy consumption parameter value and the weighted target task processing time, and obtaining the computing power evaluation value of each service device. The computing power evaluation value of each service device can be determined by weighting the target computing power parameter values ​​of the three dimensions of each service device. Computing power evaluation from multiple dimensions not only ensures the accuracy and reliability of the computing power evaluation value of the determined service device, but also weights the computing power parameters of each dimension, thereby improving the matching between the determined computing power evaluation value and the actual situation or actual needs.

[0052] The process of determining the computing power evaluation value of each service device by the computing power resource allocation device based on the target computing power parameter value and the target computing power parameter weight can be implemented based on the target formula, and the target formula can be:

[0053] S = w1*I+w2*(E / T)+w3*T; (Formula 1)

[0054] In Formula 1, S is the computing power evaluation value, w1 is the basic computing power parameter weight of the device, I is the basic computing power parameter value of the device, w2 is the energy consumption parameter weight, E / T is the energy consumption parameter value, where E is the device power, T is the task processing time, and w3 is the task processing time weight.

[0055] For example, for two service devices with CPU and GPU processors, assuming that w1, w2, and w3 are 0.4, 0.4, and 0.2 respectively, the CPU service device has I=1TFLOPS, E=150W, and T=30minutes; the GPU service device has I=125TFLOPS, E=250W, and T=5minutes, then the computing power evaluation value S1 of the CPU service device is:

[0056] S1=0.4×1+0.4×(4500 / 1800s)+0.2×30min=0.4+0.4*2.5+0.2*30=7.4

[0057] The computing power evaluation value S2 of the GPU service device is:

[0058] S2=0.4×125+0.4×(1250 / 300s)+0.2×5min=50+1.67+1=52.67

[0059] It can be understood that when the computing power parameters include energy consumption parameters and basic computing power parameters of the equipment, or when the computing power parameters include task processing time and basic computing power parameters of the equipment, the process of the computing power resource allocation device determining the computing power evaluation value of each service device can refer to the above embodiment, and the present disclosed embodiment will not elaborate on this.

[0060] In step S204, the computing power resource allocation device may determine the computing power resource allocation result based on the computing power demand information at the current moment and the computing power evaluation values ​​of multiple service devices in the service device cluster.

[0061] In the disclosed embodiment, the computing power requirement information may include a computing power requirement value, and the computing power requirement information may also include information such as task type, for example, cloud computing tasks, tasks based on machine learning models, or video processing tasks.

[0062] It should be noted that in the embodiments of the present disclosure, the computing power requirement information may be pre-configured, and the computing power resource allocation device may read the computing power requirement information at the current moment in the computing power requirement configuration information.

[0063] In an optional implementation, when the computing power resource allocation device is not configured with computing power demand information, the computing power demand information at the current moment may be determined based on the historical operation status information of the service device cluster in the historical period, wherein the historical operation status information may include the task type and task processing volume of the service device cluster in the historical period, the device resource information of each service device in the service device cluster in the historical period, and network resource information, etc. The historical period may be the historical time length before the current moment, and the specific historical time length may be determined based on actual needs, which is not limited in the embodiments of the present disclosure.

[0064] The computing power resource allocation device can also: obtain multiple historical operating status information of the service device cluster at multiple unit times in the historical period; input the multiple historical operating status information of multiple unit times into the computing power demand prediction model, determine the computing power demand value at the current moment, and determine the computing power demand value as the computing power demand information. Based on the historical operating status information of the service device cluster in the historical period, the computing power demand value of the service device cluster at the current moment can be evaluated to ensure the reliability of computing power resource allocation in the absence of the computing power demand value.

[0065] It should be noted that in the embodiments of the present disclosure, the computing power demand prediction model can be a time series model, which can be specifically determined based on actual needs, and the embodiments of the present disclosure do not limit this; the process in which the computing power resource allocation device determines the computing power demand value as computing power demand information can include: determining the computing power demand value and the task type in the historical operating status information as computing power demand information, or determining the computing power demand value as computing power demand information.

[0066] In an optional implementation, when the service device cluster provides services for a special type of task, the computing power demand information may include a computing power demand value, and the computing power resource allocation is based on the computing power demand information at the current moment, as well as the computing power evaluation values ​​of multiple service devices in the service device cluster. The process of determining the computing power resource allocation result may include: reading the computing power demand value in the computing power demand information; wherein, if the computing power demand value is less than or equal to the total computing power evaluation value of multiple service devices, then determining the proportional relationship value of the computing power evaluation values ​​of the multiple service devices, and determining the proportional relationship value as the proportional relationship value of the task processing amount of the multiple service devices, and obtaining the computing power resource allocation result. When the computing power demand value at the current moment is less than or equal to the total computing power evaluation value of the multiple service devices, it is determined that the computing power resources at the current moment are sufficient, and the reliability of the task processing service can be guaranteed. At the same time, according to the proportional relationship of the computing power evaluation values ​​of the multiple service devices, the task processing amount is allocated to each service device, ensuring that the task processing amount of each service device meets the computing power that the service device can provide, and improving the rationality of the allocation of computing power resources.

[0067] It is understandable that if the computing power demand value is less than or equal to the total computing power assessment value of multiple service devices, a prompt message of insufficient computing power resources is generated to remind the operation and maintenance personnel to expand the capacity of the service device cluster to ensure the reliable implementation of the task service.

[0068] It should be noted that in the embodiments of the present disclosure, the service device cluster can provide services for various types of tasks, and the service device cluster may include service devices equipped with different types of processors. For example, a service device equipped with a CPU processor can be used to process routine tasks, a service device equipped with a GPU processor can be used to process video services, and a service device equipped with an ASIC processor can be used to process tasks that require the participation of a machine learning model.

[0069] In an optional implementation, when a service device cluster provides services for multiple types of tasks, the computing power requirement information may include a computing power requirement value and a task type, and the computing power requirement value may include a computing power requirement sub-value associated with each type. The computing power resource allocation is based on the computing power requirement information at the current moment, and the computing power evaluation values ​​of multiple service devices in the service device cluster. The process of determining the computing power resource allocation result may include: reading the task type in the computing power requirement information, and the computing power requirement sub-value associated with each service type; wherein, for each service type, when the service device cluster contains at least one target service device associated with the task type, and the total computing power evaluation value of at least one target service device is greater than or equal to the computing power requirement sub-value, an association relationship between the target service device and the task type is generated to obtain a computing power resource allocation result. When it is determined that the computing power resources used to provide task processing services for different task types are sufficient, different types of tasks are allocated to the target service device used to process the task, so as to ensure that the service device used to provide services for each task matches the current task type, thereby improving the rationality of computing power resource allocation.

[0070] In an optional implementation, when there are multiple target service devices, after the target service device generates the association relationship between the target service device and the task type, the computing power resource allocation device can also determine the target proportional relationship value of the computing power evaluation values ​​of the multiple target service devices, and determine the target proportional relationship value as the proportional relationship value of the task processing volume of the multiple target service devices. According to the proportional relationship of the computing power evaluation values ​​of the multiple target service devices, the task processing volume of the corresponding task type is allocated to each target service device, ensuring that the task processing volume of each target service device meets the computing power that the target service device can provide, further improving the rationality of the allocation of computing power resources.

[0071] It can be understood that, if the service device cluster does not contain at least one target service device associated with the task type, a prompt message indicating that the target service device is missing is generated to remind the operation and maintenance personnel to deploy the target service device to ensure the reliable implementation of the relevant type of task service; or, if the total computing power evaluation value of at least one target service device is less than the computing power requirement sub-value, a prompt message indicating that the computing power resources of the target service device are insufficient is generated to remind the operation and maintenance personnel to expand the capacity of the target service device to ensure the reliable implementation of the relevant type of task service.

[0072] In an optional implementation, after the computing power resource allocation device obtains the computing power resource allocation result, it can also determine the target computing power impact factor at the current moment as the sample computing power impact factor, and determine the target computing power parameter weight at the current moment as the computing power parameter weight label value to obtain an updated sample data set; then, after a preset time, the computing power parameter weight estimation model is trained based on the updated sample data set to obtain an updated computing power parameter weight estimation model; the preset time can be determined based on actual needs, and the present disclosure embodiment does not limit this. For example, the preset time can be one month or half a month. The computing power parameter weight estimation model can be updated and trained according to the updated sample data to further improve the reliability of the determined computing power parameter weights, thereby further improving the rationality and reliability of the determined computing power resource allocation results.

[0073] It should be noted that in the embodiments of the present disclosure, the computing power resource allocation device can periodically reallocate computing power resources, wherein the update period of the computing power resource allocation results can be determined based on actual needs, and the embodiments of the present disclosure do not limit this; for example, computing power resources are reallocated once every two days, or once every week, and the current time is the starting time of the update period of each computing power resource allocation result.

[0074] It is understandable that the computing power resource allocation device can provide a relevant user interface for the operation and maintenance task to configure the relevant configuration information required in the computing power resource allocation process, and can also generate various prompt information in the computing power resource allocation process so that the operation and maintenance personnel can promptly perform corresponding maintenance on the service equipment cluster according to the prompt information to ensure the reliability of the task service; it can also display the computing power resource allocation results so that the operation and maintenance personnel can view and change the computing power resource allocation results.

[0075] An exemplary embodiment of the present disclosure provides a computing power resource allocation apparatus, which may be a computing power resource allocation device or a chip applied to a computing power resource allocation device. Figure 3 A schematic block diagram of functional modules of a computing resource allocation device according to an exemplary embodiment of the present disclosure is shown.

[0076] like Figure 3 As shown, the computing resource allocation device 300 includes:

[0077] The acquisition module 301 is configured to acquire a target computing power impact factor of the service device cluster at the current moment, wherein the computing power impact factor includes device resource information of each service device in the service device cluster, pending task information of the service device cluster, and energy consumption impact information of the service device;

[0078] The first determination module 302 is configured to determine the target computing power parameter weight of each service device at the current moment based on the target computing power influencing factor associated with each service device and the computing power parameter weight estimation model, wherein the computing power parameter includes at least one of an energy consumption parameter and a task processing duration, and a basic computing power parameter of the device;

[0079] The processing module 303 is configured to collect the target computing power parameter value of each service device at the current moment, and determine the computing power evaluation value of each service device based on the target computing power parameter value and the target computing power parameter weight;

[0080] The second determination module 304 is configured to determine the computing power resource allocation result based on the computing power demand information at the current moment and the computing power evaluation values ​​of the multiple service devices in the service device cluster.

[0081] Optionally, the target computing power parameter value includes a target device basic computing power parameter value, a target energy consumption parameter value and a target task processing time, and the target computing power parameter weight includes a target device basic computing power parameter weight, a target energy consumption parameter weight and a target task processing time weight.

[0082] The processing module 303 is configured to:

[0083] For each service device, determine the product of the target device basic computing power parameter value and the target device basic computing power parameter weight to obtain a weighted target device basic computing power parameter value;

[0084] Determine the product of the target energy consumption parameter value and the target energy consumption parameter weight to obtain a weighted target energy consumption parameter value;

[0085] Determine the product of the target task processing time and the target task processing time weight to obtain a weighted target task processing time;

[0086] Determine the weighted target computing power parameter value associated with each service device, the sum of the weighted target energy consumption parameter value and the weighted target task processing duration, and obtain the computing power evaluation value of each service device.

[0087] Optionally, the second determining module 304 is configured to:

[0088] Reading the computing power requirement value in the computing power requirement information;

[0089] If the computing power requirement value is less than or equal to the total computing power evaluation value of the multiple service devices, the proportional relationship value of the computing power evaluation values ​​of the multiple service devices is determined, and the proportional relationship value is determined as the proportional relationship value of the task processing volume of the multiple service devices to obtain the computing power resource allocation result.

[0090] Optionally, the second determining module 304 is configured to:

[0091] Read the task type in the computing power requirement information, and the computing power requirement sub-value associated with each service type;

[0092] For each service type, if the service device cluster includes at least one target service device associated with the task type, and the total computing power evaluation value of the at least one target service device is greater than or equal to the computing power requirement sub-value, an association relationship between the target service device and the task type is generated to obtain a computing power resource allocation result.

[0093] Optionally, the computing power parameters include basic computing power parameters of the device, energy consumption parameters and task processing time, and the device further includes a model training module 305 configured to:

[0094] Acquire a sample data set, wherein each sample in the sample data set includes a sample computing power influence factor, a sample computing power parameter value, and a computing power parameter weight label value;

[0095] Inputting the sample computing power influencing factor into the computing power parameter weight estimation model to be trained to obtain a computing power parameter weight estimation value;

[0096] Determine a sample computing power evaluation prediction value based on the computing power parameter weight estimate value and the sample computing power parameter value, and determine a sample computing power evaluation value based on the computing power parameter weight label value and the sample computing power parameter value;

[0097] The sample computing power evaluation prediction value and the sample computing power evaluation value are input into the loss function to obtain the loss function value, and when it is determined that the loss function value is less than the loss function threshold, it is determined that the computing power parameter weight estimation model converges to obtain the computing power parameter weight estimation model.

[0098] Optionally, the apparatus further includes a prediction module 306 configured to:

[0099] Acquire multiple historical operation status information of the service device cluster at multiple unit times in a historical period;

[0100] The multiple historical operating status information of the multiple unit times are input into the computing power demand prediction model to determine the computing power demand value at the current moment, and the computing power demand value is determined as the computing power demand information.

[0101] Optionally, the model training module 305 is further configured to:

[0102] Determine the target computing power impact factor at the current moment as the sample computing power impact factor, and determine the target computing power parameter weight at the current moment as the computing power parameter weight label value, to obtain an updated sample data set;

[0103] After a preset period of time, the computing power parameter weight estimation model is trained based on the updated sample data set to obtain an updated computing power parameter weight estimation model.

[0104] The exemplary embodiment of the present disclosure also provides an electronic device, comprising: at least one processor; and a memory connected to the at least one processor in communication. The memory stores a computer program that can be executed by the at least one processor, and the computer program is used to cause the electronic device to perform the method according to the embodiment of the present disclosure when executed by the at least one processor.

[0105] The exemplary embodiments of the present disclosure also provide a non-transitory computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor of a computer, is used to cause the computer to perform the method according to the embodiments of the present disclosure.

[0106] like Figure 4 As shown, the exemplary embodiment of the present disclosure further provides a computer program product 400, including a computer program 401, wherein when the computer program is executed by a processor of a computer, it is used to enable the computer to perform the method according to the embodiment of the present disclosure.

[0107] refer to Figure 5 , a block diagram of an electronic device 500 that can be used as a computing resource allocation device of the present disclosure will now be described, which is an example of a hardware device that can be applied to various aspects of the present disclosure. The electronic device is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0108] like Figure 5As shown, the electronic device 500 includes a computing unit 501, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the electronic device 500 can also be stored. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0109] A plurality of components in the electronic device 500 are connected to the I / O interface 505, including: an input unit 506, an output unit 507, a storage unit 508, and a communication unit 509. The input unit 506 may be any type of device capable of inputting information to the electronic device 500, and the input unit 506 may receive input digital or character information, and generate key signal inputs related to user settings and / or function control of the electronic device. The output unit 507 may be any type of device capable of presenting information, and may include, but is not limited to, a display, a speaker, a video / audio output terminal, a vibrator, and / or a printer. The storage unit 508 may include, but is not limited to, a disk, an optical disk. The communication unit 509 allows the electronic device 500 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks, and may include, but is not limited to, a modem, a network card, an infrared communication device, a wireless communication transceiver, and / or a chipset, such as a Bluetooth™ device, a WiFi device, a WiMax device, a cellular communication device, and / or the like.

[0110] The computing unit 501 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 501 performs the various methods and processes described above. For example, in some embodiments, the method of the exemplary embodiment of the present disclosure may be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as a storage unit 508. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 500 via the ROM 502 and / or the communication unit 509. In some embodiments, the computing unit 501 may be configured to perform the method of the exemplary embodiment of the present disclosure by any other appropriate means (e.g., by means of firmware).

[0111] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0112] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0113] As used in this disclosure, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and / or device (e.g., disk, optical disk, memory, programmable logic device (PLD)) for providing machine instructions and / or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor.

[0114] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0115] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0116] A computer system may include clients and servers. Clients and servers are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship to each other.

[0117] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented by software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer programs or instructions. When the computer program or instruction is loaded and executed on a computer, the process or function described in the embodiment of the present disclosure is executed in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, a terminal, a user device or other programmable device. The computer program or instruction may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer program or instruction may be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless means. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server, data center, etc. that integrates one or more available media. The available medium may be a magnetic medium, for example, a floppy disk, a hard disk, a tape; it may also be an optical medium, for example, a digital video disc (DVD); it may also be a semiconductor medium, for example, a solid state drive (SSD).

[0118] Although the present disclosure has been described in conjunction with specific features and embodiments thereof, it is apparent that various modifications and combinations may be made thereto without departing from the spirit and scope of the present disclosure. Accordingly, this specification and the drawings are merely exemplary illustrations of the present disclosure as defined by the appended claims, and are deemed to have covered any and all modifications, variations, combinations or equivalents within the scope of the present disclosure. Obviously, those skilled in the art may make various modifications and variations to the present disclosure without departing from the spirit and scope of the present disclosure. Thus, if these modifications and variations of the present disclosure fall within the scope of the claims of the present disclosure and their equivalents, the present disclosure is also intended to include these modifications and variations.

Claims

1. A computing resource allocation method, characterized in that: include: Obtaining a target computing power impact factor of the service device cluster at the current moment, wherein the computing power impact factor includes device resource information of each service device in the service device cluster, pending task information of the service device cluster, and energy consumption impact information of the service device; Determine the target computing power parameter weight of each service device at the current moment based on the target computing power influencing factor associated with each service device and the computing power parameter weight estimation model, wherein the computing power parameter includes at least one of an energy consumption parameter and a task processing duration, and a basic computing power parameter of the device; Collecting the target computing power parameter value of each service device at the current moment, and determining the computing power evaluation value of each service device based on the target computing power parameter value and the target computing power parameter weight; The computing power resource allocation result is determined based on the computing power demand information at the current moment and the computing power evaluation values ​​of multiple service devices in the service device cluster.

2. The computing resource allocation method according to claim 1, characterized in that: The target computing power parameter value includes the target device basic computing power parameter value, the target energy consumption parameter value and the target task processing time, and the target computing power parameter weight includes the target device basic computing power parameter weight, the target energy consumption parameter weight and the target task processing time weight. The determining the computing power evaluation value of each service device based on the target computing power parameter value and the target computing power parameter weight includes: For each service device, determine the product of the target device basic computing power parameter value and the target device basic computing power parameter weight to obtain a weighted target device basic computing power parameter value; Determine the product of the target energy consumption parameter value and the target energy consumption parameter weight to obtain a weighted target energy consumption parameter value; Determine the product of the target task processing time and the target task processing time weight to obtain a weighted target task processing time; Determine the weighted target computing power parameter value associated with each service device, the sum of the weighted target energy consumption parameter value and the weighted target task processing duration, and obtain the computing power evaluation value of each service device.

3. The computing resource allocation method according to claim 1, characterized in that: The determining of the computing power resource allocation result based on the computing power demand information at the current moment and the computing power evaluation values ​​of the plurality of service devices in the service device cluster includes: Reading the computing power requirement value in the computing power requirement information; If the computing power requirement value is less than or equal to the total computing power evaluation value of the multiple service devices, the proportional relationship value of the computing power evaluation values ​​of the multiple service devices is determined, and the proportional relationship value is determined as the proportional relationship value of the task processing volume of the multiple service devices to obtain the computing power resource allocation result.

4. The computing resource allocation method according to claim 1, characterized in that: The determining of the computing power resource allocation result based on the computing power demand information at the current moment and the computing power evaluation values ​​of the plurality of service devices in the service device cluster includes: Read the task type in the computing power requirement information, and the computing power requirement sub-value associated with each service type; For each service type, if the service device cluster includes at least one target service device associated with the task type, and the total computing power evaluation value of the at least one target service device is greater than or equal to the computing power requirement sub-value, an association relationship between the target service device and the task type is generated to obtain a computing power resource allocation result.

5. The computing resource allocation method according to claim 1, characterized in that: The computing power parameters include basic computing power parameters of the device, energy consumption parameters and task processing time. The method further includes: Acquire a sample data set, wherein each sample in the sample data set includes a sample computing power influence factor, a sample computing power parameter value, and a computing power parameter weight label value; Inputting the sample computing power influencing factor into the computing power parameter weight estimation model to be trained to obtain a computing power parameter weight estimation value; Determine a sample computing power evaluation prediction value based on the computing power parameter weight estimate value and the sample computing power parameter value, and determine a sample computing power evaluation value based on the computing power parameter weight label value and the sample computing power parameter value; The sample computing power evaluation prediction value and the sample computing power evaluation value are input into the loss function to obtain the loss function value, and when it is determined that the loss function value is less than the loss function threshold, it is determined that the computing power parameter weight estimation model converges to obtain the computing power parameter weight estimation model.

6. The computing resource allocation method according to claim 1, characterized in that: The method further comprises: Acquire multiple historical operation status information of the service device cluster at multiple unit times in a historical period; The multiple historical operating status information of the multiple unit times are input into the computing power demand prediction model to determine the computing power demand value at the current moment, and the computing power demand value is determined as the computing power demand information.

7. The computing resource allocation method according to claim 1, characterized in that: The method further comprises: Determine the target computing power impact factor at the current moment as the sample computing power impact factor, and determine the target computing power parameter weight at the current moment as the computing power parameter weight label value, to obtain an updated sample data set; After a preset period of time, the computing power parameter weight estimation model is trained based on the updated sample data set to obtain an updated computing power parameter weight estimation model.

8. An electronic device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the computing resource allocation method described in any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the computing power resource allocation method described in any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the computing power resource allocation method described in any one of claims 1 to 7 is implemented.

Images