Computing power operation task disassembling method and device of intelligent computing center

Abstract

The invention provides a computing power operation task disassembling method and device of an intelligent computing center, and relates to the technical field of intelligent computing centers, intelligent computing centers and computing power infrastructure, and the computing power operation task disassembling method of the intelligent computing center comprises the following steps: S1, receiving a target request which is used for requesting to process a computing power operation task; s2, disassembling the computing power operation task to obtain a plurality of computing power operation sub-tasks; s3, the computing power operation sub-tasks are distributed to first computing nodes, computing power resources of the first computing nodes are scheduled to process the computing power operation sub-tasks, and the first computing nodes correspond to the computing power operation sub-tasks in a one-to-one mode. In this way, the processing efficiency of the multiple computing power operation sub-tasks can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent computing centers, smart computing centers and computing power infrastructure, and in particular to a method and device for disassembling computing power operation tasks of an intelligent computing center. Background Art

[0002] With the rapid development of artificial intelligence technology, "intelligent computing centers" and "intelligent computing centers" have emerged.

[0003] An "Intelligent Computing Center" is a facility that uses large-scale heterogeneous computing resources, including general-purpose and intelligent computing power, to provide the computing power, data, and algorithms required for AI applications (such as AI deep learning model development, model training, and model inference). The Intelligent Computing Center encompasses facilities, hardware, and software, and provides a full stack of capabilities, from bottom-level computing power to top-level application enablement.

[0004] “Intelligent Computing Center” includes but is not limited to “Smart Computing Center”.

[0005] "Intelligent Computing Center" refers to an artificial intelligence computing center. It is a type of computing power infrastructure that is based on artificial intelligence theory, adopts artificial intelligence computing architecture, and provides computing power services, data services, and algorithm services required for artificial intelligence applications.

[0006] "Computing power" is the core of "intelligent computing center" and "intelligent computing center". It is the ability of computer equipment or computing / data center to process information. It is the ability of computer hardware and software to work together to perform certain computing needs. It is the computing power to achieve target result output by processing information data. It is a new type of productivity that integrates information computing power, network carrying capacity, and data storage capacity. It mainly provides services to society through computing power infrastructure.

[0007] When the current intelligent computing center receives a request for processing computing power running tasks, it usually needs to process the above computing power running tasks according to the request. However, the processing efficiency of the current intelligent computing center in processing the above computing power running tasks is usually very low. It can be seen that since the emergence of the intelligent computing center, the low processing efficiency of computing power running tasks is a problem that needs to be solved urgently. Summary of the Invention

[0008] The present invention provides a method and device for disassembling computing power operation tasks of an intelligent computing center, which are used to solve the problem of low processing efficiency of computing power operation tasks.

[0009] In order to solve the above-mentioned technical problems, the present invention is achieved as follows:

[0010] In a first aspect, the present invention provides a method for decomposing computing power operation tasks in an intelligent computing center, comprising:

[0011] Step S1: Receive a target request, where the target request is used to request processing computing power to run a task;

[0012] Step S2: Decompose the computing power operation task to obtain multiple computing power operation subtasks;

[0013] Step S3: Allocate the computing power operation subtask to the first computing node, and schedule the computing power resources of the first computing node to process the computing power operation subtask, and the first computing node corresponds to the computing power operation subtask one by one.

[0014] Optionally, step S2 includes:

[0015] Step S21: Obtain task disassembly parameters;

[0016] Step S22: Decompose the computing power operation task according to the task decomposition parameters to obtain multiple computing power operation subtasks.

[0017] Optionally, step S21 includes:

[0018] Step S211: obtaining the current computing power parameters of the intelligent computing center and the expected processing completion time parameters of the computing power operation task;

[0019] Step S212: determining the task decomposition parameters according to the computing power parameter and the expected processing completion time parameter, wherein the task decomposition parameters include at least one of the following: a central processing unit (CPU) parameter, a graphics processing unit (GPU) parameter, and a memory parameter;

[0020] Step S213: Obtain the task disassembly parameters.

[0021] Optionally, step S3 includes:

[0022] Step S31: allocating the computing power running subtask to a first computing node, and calculating the compatibility between the first computing node and the computing power running subtask;

[0023] Step S32: When the compatibility between the first computing node and the computing power running subtask is greater than a preset compatibility, the computing power resources of the first computing node are scheduled to process the computing power running subtask.

[0024] Optionally, after step S3, the method further includes:

[0025] Step S4: Obtaining the processing result of the computing power running subtask output by the first computing node;

[0026] Step S5: When the processing result matches the expected result corresponding to the computing power operation subtask, it is determined that the processing of the computing power operation subtask is completed. The expected result is the result obtained when the computing power operation task is disassembled, and the expected result corresponds one-to-one to the computing power operation subtask.

[0027] Optionally, after step S3, the method further includes:

[0028] Step S6: if it is determined that a failure occurs in processing the computing power running subtask by the first computing node, a second computing node is determined, and computing power resources of the second computing node are in an idle state;

[0029] Step S7: Scheduling the computing resources of the second computing node to process the computing operation subtask.

[0030] In a second aspect, the present invention provides a computing power operation task disassembly device for an intelligent computing center, comprising:

[0031] A receiving module, configured to receive a target request, wherein the target request is used to request processing computing power to run a task;

[0032] A disassembly module, configured to disassemble the computing power operation task into multiple computing power operation subtasks;

[0033] The scheduling module is used to allocate the computing power operation subtask to the first computing node and schedule the computing power resources of the first computing node to process the computing power operation subtask, and the first computing node corresponds to the computing power operation subtask one by one.

[0034] In a third aspect, the present invention provides an electronic device comprising: a processor, a memory, and a program stored in the memory and executable on the processor, wherein when the program is executed by the processor, the steps of the method for decomposing computing power operation tasks of the intelligent computing center as described in the first aspect above are implemented.

[0035] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the method for decomposing computing power operation tasks of an intelligent computing center as described in the first aspect above are implemented.

[0036] In a fifth aspect, the present invention provides a computer program product comprising computer instructions, which, when executed by a processor, implement the steps of the method for decomposing computing power operation tasks of an intelligent computing center as described in the first aspect above.

[0037] In the present invention, a target request is received, and the target request is used to request the processing of a computing power operation task; the computing power operation task is disassembled to obtain multiple computing power operation subtasks; the computing power operation subtasks are assigned to a first computing node, and the computing power resources of the first computing node are scheduled to process the computing power operation subtasks, and the first computing node corresponds one to one with the computing power operation subtasks.

[0038] In this way, the computing power operation task can be disassembled to obtain multiple computing power operation sub-tasks, and each computing power operation sub-task can be assigned to the corresponding first computing node, and the computing power resources of the first computing node can be called to process the computing power operation sub-task, so that the above-mentioned multiple computing power operation sub-tasks can be executed in parallel, thereby improving the processing efficiency of multiple computing power operation sub-tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present invention. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0040] Figure 1 A flowchart of a method for decomposing computing power operation tasks in an intelligent computing center provided by the present invention;

[0041] Figure 2 A schematic diagram of the structure of a computing power operation task disassembly device for an intelligent computing center provided by the present invention;

[0042] Figure 3 This is a schematic structural diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0043] The following is a clear and complete description of the technical solutions of the present invention, in conjunction with the accompanying drawings. Obviously, the description is only a portion of the present invention, not all of it. All other contents derived by persons of ordinary skill in the art based on the contents of the present invention without inventive effort are within the scope of protection of the present invention.

[0044] The "computing power" mentioned in the present invention refers to: the ability of computer equipment or computing / data centers to process information, the ability of computer hardware and software to work together to execute certain computing requirements, and the computing power to achieve target result output by processing information data. It is a new type of productivity that integrates information computing power, network carrying capacity, and data storage capacity, and mainly provides services to society through computing power infrastructure.

[0045] The "computing power" (Computational Power, CP) mentioned in the present invention refers to: the ability of a data center server to process data and output results. It is a comprehensive indicator to measure the computing power of a data center, including general computing power, super computing power and intelligent computing power. The commonly used unit of measurement is the number of floating-point operations performed per second (FLOPS, 1EFLOPS=10^18FLOPS). The larger the value, the stronger the comprehensive computing power. According to calculations, 1EFLOPS is approximately the computing power output of 5 Tianhe-2A or 500,000 mainstream server CPUs or 2 million mainstream notebooks. The calculation formula is: CP=CP 通用 +CP 智能 +CP 超级 .

[0046] The "carrying capacity" (Network Power, NP) mentioned in the present invention refers to: it is the performance of the data transmission capability of the computing power facility, which includes comprehensive capabilities such as network architecture, network bandwidth, transmission latency, intelligent management and scheduling, etc. It involves network transmission within and between data centers, and is a comprehensive indicator for measuring network transmission scheduling capabilities.

[0047] The "Storage Power" (SP) described in this invention refers to the comprehensive capabilities of a data center in terms of data storage capacity, performance, security and reliability, and environmental friendliness. It is a comprehensive indicator for measuring a data center's data storage capacity, encompassing both external storage devices such as storage arrays and internal server storage. Storage capacity is commonly measured in exabytes (EB, 1EB = 2^60 bytes), while performance is commonly measured in IOPS / TB (Input / Output Operations Per Second / TB). Disaster recovery ratio is a key indicator of security and reliability.

[0048] The "computing power infrastructure" mentioned in the present invention refers to a new type of information infrastructure that integrates information computing power, network carrying capacity, and data storage capacity, and can realize the centralized calculation, storage, transmission and application of information.

[0049] The "new information infrastructure" mentioned in the present invention refers to: mainly including network infrastructure such as 5G networks, fiber-optic broadband networks, backbone networks, international communication networks, satellite Internet, computing power infrastructure such as data centers, general computing power centers, intelligent computing centers, supercomputing centers, and new technology facilities such as artificial intelligence, blockchain, and quantum computing.

[0050] The "computing power" mentioned in the present invention includes: general computing power, intelligent computing power and super computing power.

[0051] The "general computing power" mentioned in the present invention refers to the computing power provided by servers based on CPU (Central Processing Unit) chips, which is used to support basic general computing such as cloud computing and edge computing.

[0052] The "intelligent computing power" mentioned in this invention refers to: a computing platform based on specialized chips such as GPU (Graphics Processing Unit), FPGA (Field Programmable Gate Array), and ASIC (Application Specific Integrated Circuit) for various innovative artificial intelligence applications, such as natural language processing and machine vision.

[0053] The "supercomputing power" mentioned in the present invention refers to the computing power provided by high-performance computing clusters such as supercomputers. It utilizes the centralized computing resources of multiple computer systems working in parallel and uses a dedicated operating system to handle extremely complex or data-intensive problems. It is mainly used for calculations in cutting-edge scientific fields, such as planetary simulation, drug molecule design, genetic analysis, etc.

[0054] The "intelligent computing center" described in this article refers to a facility that provides the computing power, data, and algorithms required for artificial intelligence applications (such as AI deep learning model development, model training, and model inference) by utilizing large-scale heterogeneous computing resources, including general-purpose computing power (CPU) and intelligent computing power (GPU, FPGA, ASIC, etc.). The intelligent computing center encompasses facilities, hardware, and software, and can provide a full stack of capabilities, from bottom-level computing power to top-level application enablement.

[0055] The "intelligent computing center" mentioned in the present invention includes but is not limited to the "intelligent computing center".

[0056] The "intelligent computing center" mentioned in the present invention is an artificial intelligence computing center, which is a type of computing power infrastructure based on artificial intelligence theory, adopts artificial intelligence computing architecture, and provides computing power services, data services and algorithm services required for artificial intelligence applications.

[0057] The "computing power center" mentioned in the present invention refers to: a facility that is mainly composed of infrastructure such as wind, fire, water, electricity, and IT hardware and software equipment, and has computing power, transportation capacity, and storage capacity, including general data centers, intelligent computing centers, supercomputing centers, etc.

[0058] The "supercomputing center" mentioned in the present invention refers to: a supercomputing data center, which is a data center based on a supercomputer or a large-scale computing cluster, which can provide large-scale computing, storage and network services and other functions, and is widely used in application scenarios such as aerospace, national defense, oil exploration, climate modeling and genome sequencing.

[0059] The "computing resources" mentioned in the present invention refer to: technologies and facilities with information computing, transmission, storage and application capabilities required for the development of a digital society, including but not limited to computing resources such as CPUs and GPUs, network resources such as switches and routers, storage resources such as storage arrays and distributed storage, security resources such as firewalls and intrusion detection systems, and supporting and guarantee resources such as wind, fire, water and electricity.

[0060] The “model” mentioned in the present invention includes but is not limited to a “large language model” and a “multimodal large model”.

[0061] The "large language model" mentioned in the present invention refers to a large language model (LLM), which is a language model with a large parameter scale. It is designed to understand and generate human language. It is trained with a large amount of text data and can perform a wide range of tasks including text summarization, translation, sentiment analysis, etc.

[0062] The "Multimodal Large Models" mentioned in the present invention refer to models that combine multimodal information such as text, images, video, and audio for training, including but not limited to multimodal large language models.

[0063] The "computing power operation task" mentioned in the present invention refers to: a specific workload or job executed on computing power resources that requires a certain amount of computing power support, usually involving complex data processing, numerical calculations, model training or simulation scenarios.

[0064] See Figure 1 , Figure 1 This is a flow chart of a method for decomposing computing power operation tasks of an intelligent computing center provided by the present invention. Figure 1 As shown, the following steps are included:

[0065] Step S1: Receive a target request, where the target request is used to request processing computing power to run a task.

[0066] Among them, the specific method of receiving the target request is not limited here. Optionally, the present invention can be applied to electronic devices, and the electronic device can be called a computing power platform of an intelligent computing center. The above-mentioned target request can be a request input by the target user on the computing power platform. It should be noted that the specific method of the target user inputting on the computing power platform is not limited here. For example: the method of the target user inputting on the computing power platform can include touch input, press input or voice input.

[0067] Optionally, the target user can also input the target request through other electronic devices, and the other electronic devices can be electrically connected to the computing power platform. In this way, the target user can input on the other electronic devices, so that the other electronic devices send the target request to the computing power platform.

[0068] Among them, the computing power operation task is not specifically limited here. Optionally, the computing power operation task can be a model training task, and the type of the above-mentioned model is not limited here. For example, the above-mentioned model can be a navigation model, a computing model, a large language model, etc., and after the computing power operation task is disassembled in step S2 to obtain multiple computing power operation subtasks, each computing power operation subtask can be a training subtask of part of the function of the above-mentioned model, or each computing power operation subtask can be a training subtask of part of the network layer of the above-mentioned model, or each computing power operation subtask can be a training subtask of a sub-model of the above-mentioned model.

[0069] It should be noted that, optionally, when each computing power running subtask can be a training subtask of a sub-model of the above-mentioned model, the multiple sub-models obtained by training can be connected to obtain the above-mentioned model; optionally, after multiple sub-models are trained, sample testing and screening can be performed on the multiple sub-models, and the sub-model with the best output effect can be screened from the multiple sub-models, and the other sub-models can be guided for training based on the sub-model with the best output effect, and the model composed of the other sub-models after the guided training and the sub-model with the best output effect is determined as the final trained model.

[0070] Alternatively, the computing power operation task may be a computing task, such as a navigation route computing task, a model training computing task, etc.

[0071] Step S2: Decompose the computing power operation task to obtain multiple computing power operation subtasks.

[0072] The specific method of decomposing the computing power operation task to obtain multiple computing power operation subtasks is not limited here.

[0073] Optionally, the computing power running task can be evenly split to obtain multiple computing power running subtasks. For example, when the computing task amount of the computing power running task is 1, when the computing power running task is split into 10 computing power running subtasks, the computing task amount of each computing power running subtask is 1 / 10.

[0074] Alternatively, the computing power operation task can be randomly decomposed to obtain multiple computing power operation subtasks, thereby increasing the diversity and flexibility of the decomposition method of the computing power operation task.

[0075] Step S3: Allocate the computing power operation subtask to the first computing node, and schedule the computing power resources of the first computing node to process the computing power operation subtask, and the first computing node corresponds to the computing power operation subtask one by one.

[0076] There may be multiple first computing nodes, and optionally, the number of first computing nodes may match the number of computing power running subtasks.

[0077] In the present invention, through steps S1 to S3, the computing power operation task can be disassembled to obtain multiple computing power operation subtasks, and each computing power operation subtask is assigned to the corresponding first computing node, and the computing power resources of the first computing node are called to process the computing power operation subtask, so that the above-mentioned multiple computing power operation subtasks can be executed in parallel, thereby improving the processing efficiency of multiple computing power operation subtasks.

[0078] Optionally, step S2 includes:

[0079] Step S21: Obtain task disassembly parameters;

[0080] Step S22: Decompose the computing power operation task according to the task decomposition parameters to obtain multiple computing power operation subtasks.

[0081] Among them, the specific content of the task decomposition parameters is not limited here. Optionally, the task decomposition parameters may include the number of the first computing nodes input by the user. Optionally, the task decomposition parameters may include the number of computing power running subtasks to be decomposed into the computing power running task; optionally, the task decomposition parameters may include CPU parameters, GPU parameters and memory parameters, and the above-mentioned CPU parameters and GPU parameters may include parameters such as model, quantity, computing power, etc., and the memory parameters may include parameters such as model, quantity and storage space, etc.

[0082] It should be noted that the method for obtaining the above-mentioned task decomposition parameters is not limited here. Optionally, the task decomposition parameters manually input by the user can be obtained, and alternatively, the task decomposition parameters can be automatically detected and obtained.

[0083] In the present invention, the computing power operation task is disassembled according to the task disassembly parameters to obtain multiple computing power operation subtasks. In this way, the accuracy of the multiple computing power operation subtasks obtained by disassembly can be higher.

[0084] Optionally, step S21 includes:

[0085] Step S211: obtaining the current computing power parameters of the intelligent computing center and the expected processing completion time parameters of the computing power operation task;

[0086] Step S212: determining the task decomposition parameters according to the computing power parameter and the expected processing completion time parameter, wherein the task decomposition parameters include at least one of the following: CPU, GPU, and memory parameters;

[0087] Step S213: Obtain the task disassembly parameters.

[0088] The computing power parameters of the intelligent computing center at the current moment may include at least one of the following: the computing power utilization rate of the intelligent computing center at the current moment, and the CPU and GPU that are in an idle state at the current moment.

[0089] The expected processing completion time parameter may be understood as a time parameter at which the processing is expected to be completed.

[0090] In the present invention, the task disassembly parameters are determined based on the computing power parameters and the expected processing completion time parameters, so that the accuracy of the determined task disassembly parameters can be further improved.

[0091] Optionally, step S3 includes:

[0092] Step S31: allocating the computing power running subtask to a first computing node, and calculating the compatibility between the first computing node and the computing power running subtask;

[0093] Step S32: When the compatibility between the first computing node and the computing power running subtask is greater than a preset compatibility, the computing power resources of the first computing node are scheduled to process the computing power running subtask.

[0094] Different first computing nodes are adapted to process different types of computing power subtasks. For example, some first computing nodes are adapted to process computing power subtasks for training the target function of the model, while others are adapted to process computing power subtasks for computation. When a first computing node processes the computing power subtask it is adapted to process, it can ensure high processing efficiency and accuracy of the processing results of the computing power subtask.

[0095] It should be noted that the specific method of calculating the adaptability of the first computing node and the computing power operation subtask is not limited here. Optionally, the first computing node is associated with and stores information about historical subtasks that the first computing node is suitable for processing, obtains the information about the historical subtask, and calculates the similarity between the historical subtask and the computing power operation subtask. When the similarity between the historical subtask and the computing power operation subtask is greater than the preset similarity, it can be determined that the adaptability of the first computing node and the computing power operation subtask is greater than the preset adaptability.

[0096] Correspondingly, when the similarity between the historical subtask and the computing power operation subtask is less than or equal to the preset similarity, it can be determined that the fitness of the first computing node and the computing power operation subtask is less than or equal to the preset fitness.

[0097] It should be noted that the method of calculating the similarity between the historical subtask and the computing power operation subtask can be referred to the following expression: extracting the historical vector of the historical subtask and extracting the computing power operation vector of the computing power operation subtask, calculating the similarity between the historical vector and the computing power operation vector, and determining the similarity between the historical vector and the computing power operation vector as the similarity between the historical subtask and the computing power operation subtask.

[0098] In the present invention, the fitness of the first computing node and the computing power operation subtask can be calculated. When the fitness of the first computing node and the computing power operation subtask is greater than the preset fitness, it means that the first computing node is suitable for processing the computing power operation subtask, and the first computing node corresponding to the computing power operation subtask is correctly allocated. In this way, the computing power resources of the first computing node are scheduled to process the computing power operation subtask, thereby improving the processing efficiency and processing accuracy of the computing power operation subtask.

[0099] Optionally, after step S3, the method further includes:

[0100] Step S4: Obtaining the processing result of the computing power running subtask output by the first computing node;

[0101] Step S5: When the processing result matches the expected result corresponding to the computing power operation subtask, it is determined that the processing of the computing power operation subtask is completed. The expected result is the result obtained when the computing power operation task is disassembled, and the expected result corresponds one-to-one to the computing power operation subtask.

[0102] Among them, the expected result corresponding to the computing power operation subtask can be understood as the result that is expected to be achieved by processing the computing power operation subtask. For example, the expected result corresponding to the computing power operation subtask can be to obtain the target sub-model through training before the target time. When the processing result of the computing power operation subtask output by the first computing node is to obtain the target sub-model through training before the target time, it can be determined that the processing result matches the expected result corresponding to the computing power operation subtask; when the processing result of the computing power operation subtask output by the first computing node is to obtain the target sub-model through training after the target time, it can be determined that the processing result does not match the expected result corresponding to the computing power operation subtask.

[0103] For another example: the expected result corresponding to the computing power operation subtask can be a target sub-model obtained by training to realize the navigation function. When the processing result of the computing power operation subtask output by the first computing node is a target sub-model obtained by training to realize the navigation function, it can be determined that the processing result matches the expected result corresponding to the computing power operation subtask; when the processing result of the computing power operation subtask output by the first computing node is a sub-model obtained by training to realize the communication function, it can be determined that the processing result does not match the expected result corresponding to the computing power operation subtask.

[0104] In the present invention, when the processing result matches the expected result corresponding to the computing power operation subtask, it is determined that the computing power operation subtask processing is completed. In this way, the accuracy of the determined computing power operation subtask processing completion can be improved.

[0105] It should be noted that when the processing result does not match the expected result corresponding to the computing power operation subtask, it is determined that the computing power operation subtask processing is not completed. At this time, the computing power resources of the first computing node can be scheduled to continue processing the computing power operation subtask; or, a new first computing node can be replaced, and the new first computing node can be scheduled to reprocess the above computing power operation subtask to ensure the smooth completion of the computing power operation subtask. Alternatively, the expected result of the computing power operation subtask can be adjusted, and the computing power resources of the first computing node can be scheduled to continue processing the computing power operation subtask so that the processing result of the computing power operation subtask matches the expected result of the adjusted computing power operation subtask. At this time, it can be determined that the computing power operation subtask processing is completed.

[0106] Optionally, after step S3, the method further includes:

[0107] Step S6: if it is determined that a failure occurs in processing the computing power running subtask by the first computing node, a second computing node is determined, and computing power resources of the second computing node are in an idle state;

[0108] Step S7: Scheduling the computing resources of the second computing node to process the computing operation subtask.

[0109] Among them, the specific method of determining that the first computing node has failed to process the computing power operation subtask is not limited here. Optionally, when the processing progress of the computing power operation subtask by the first computing node is always maintained at the target progress value, and the time for maintaining the target progress value exceeds the preset time, and the target progress value is less than 100%, it can be determined that the first computing node has failed to process the computing power operation subtask.

[0110] Optionally, when the intelligent computing center receives a warning message indicating that a failure has occurred in the first computing node processing the computing power operation subtask, it indicates that a failure has occurred in the first computing node processing the computing power operation subtask, and the above warning message can be issued by the monitoring system.

[0111] In the present invention, when it is determined that a failure occurs in the first computing node processing the computing power operation subtask, a second computing node whose computing power resources are in an idle state is determined, and the computing power resources of the second computing node are scheduled to process the computing power operation subtask, thereby ensuring that the computing power operation subtask is processed smoothly and timely.

[0112] See also Figure 2 , Figure 2 This is a structural diagram of a computing power operation task disassembly device for an intelligent computing center provided by the present invention, such as Figure 2 As shown, the computing power operation task decomposition device 200 of the intelligent computing center includes:

[0113] A receiving module 201 is configured to receive a target request, wherein the target request is used to request processing computing power to run a task;

[0114] A disassembly module 202 is used to disassemble the computing power operation task to obtain multiple computing power operation subtasks;

[0115] The scheduling module 203 is used to allocate the computing power operation subtask to the first computing node and schedule the computing power resources of the first computing node to process the computing power operation subtask. The first computing node corresponds to the computing power operation subtask in a one-to-one manner.

[0116] Optionally, the disassembly module 202 includes:

[0117] Get submodule, used to obtain task disassembly parameters;

[0118] The disassembly submodule is used to disassemble the computing power operation task according to the task disassembly parameters to obtain multiple computing power operation subtasks.

[0119] Optionally, obtain submodules, including:

[0120] A first acquisition unit is used to obtain the computing power parameters of the intelligent computing center at the current moment and the expected processing completion time parameters of the computing power operation task;

[0121] a determining unit, configured to determine the task decomposition parameters according to the computing power parameters and the expected processing completion time parameters, wherein the task decomposition parameters include at least one of the following: parameters of a central processing unit (CPU), parameters of a graphics processing unit (GPU), and parameters of a memory;

[0122] The second acquiring unit is used to acquire the task disassembly parameters.

[0123] Optionally, the scheduling module 203 includes:

[0124] an allocation submodule, configured to allocate the computing power operation subtask to a first computing node and calculate the compatibility between the first computing node and the computing power operation subtask;

[0125] The scheduling submodule is used to schedule the computing power resources of the first computing node to process the computing power operation subtask when the compatibility between the first computing node and the computing power operation subtask is greater than a preset compatibility.

[0126] Optionally, the computing power operation task decomposition device 200 of the intelligent computing center further includes:

[0127] An acquisition module, configured to acquire a processing result of the computing power execution subtask output by the first computing node;

[0128] The first determination module is used to determine that the processing of the computing power operation subtask is completed when the processing result matches the expected result corresponding to the computing power operation subtask, and the expected result is the result obtained when the computing power operation task is disassembled, and the expected result corresponds one-to-one to the computing power operation subtask.

[0129] Optionally, the computing power operation task decomposition device 200 of the intelligent computing center further includes:

[0130] A second determining module is configured to determine a second computing node when it is determined that a failure occurs in processing the computing power running subtask by the first computing node, and the computing power resources of the second computing node are in an idle state;

[0131] A scheduling module is used to schedule the computing power resources of the second computing node to process the computing power execution subtask.

[0132] The computing power operation task disassembly device 200 of the intelligent computing center provided by the present invention can execute each step of the computing power operation task disassembly method of the above-mentioned intelligent computing center, and thus has the same beneficial technical effects as the computing power operation task disassembly method of the above-mentioned intelligent computing center, and the details will not be repeated here.

[0133] Please refer to Figure 3 The present invention also provides an electronic device 30, including a processor 31, a memory 32, and a computer program stored in the memory 32 and executable on the processor 31. When the computer program is executed by the processor 31, the various processes shown in the above-mentioned method for decomposing computing power operation tasks of the intelligent computing center are implemented, and the same technical effects can be achieved. To avoid repetition, they will not be described here.

[0134] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the various processes of the above-mentioned intelligent computing center computing power operation task decomposition method and can achieve the same technical effect. To avoid repetition, the details are not repeated here. The computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0135] The present invention also provides a computer program product, comprising computer instructions, which, when executed by a processor, implement the above Figure 1 The various processes of the computing power operation task decomposition method of the intelligent computing center shown in the figure can achieve the same technical effect. To avoid repetition, they will not be repeated here.

[0136] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0137] Through the description of the above embodiments, those skilled in the art can clearly understand that the method provided by the above invention can be implemented by means of software plus the necessary general hardware platform, or of course by hardware, but in many cases the former is a better embodiment. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a terminal (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the various methods provided by the present invention.

[0138] The present invention is described above with reference to the accompanying drawings, but the present invention is not limited to the above-mentioned specific embodiments. The above-mentioned specific embodiments are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are protected by the present invention.

Claims

1. A method for decomposing computing power operation tasks in an intelligent computing center, characterized in that: include: Step S1: Receive a target request, where the target request is used to request processing computing power to run a task; Step S2: Decompose the computing power operation task to obtain multiple computing power operation subtasks; Step S3: Allocate the computing power operation subtask to the first computing node, and schedule the computing power resources of the first computing node to process the computing power operation subtask, and the first computing node corresponds to the computing power operation subtask one by one.

2. The method according to claim 1, characterized in that The step S2 comprises: Step S21: Obtain task disassembly parameters; Step S22: Decompose the computing power operation task according to the task decomposition parameters to obtain multiple computing power operation subtasks.

3. The method according to claim 2, characterized in that The step S21 includes: Step S211: obtaining the current computing power parameters of the intelligent computing center and the expected processing completion time parameters of the computing power operation task; Step S212: determining the task decomposition parameters according to the computing power parameters and the expected processing completion time parameters, wherein the task decomposition parameters include at least one of the following: parameters of a central processing unit (CPU), parameters of a graphics processing unit (GPU), and parameters of a memory; Step S213: Obtain the task disassembly parameters.

4. The method according to claim 1, wherein The step S3 comprises: Step S31: allocating the computing power running subtask to a first computing node, and calculating the compatibility between the first computing node and the computing power running subtask; Step S32: When the compatibility between the first computing node and the computing power running subtask is greater than a preset compatibility, the computing power resources of the first computing node are scheduled to process the computing power running subtask.

5. The method according to any one of claims 1 to 4, characterized in that After step S3, the method further includes: Step S4: Obtaining the processing result of the computing power running subtask output by the first computing node; Step S5: When the processing result matches the expected result corresponding to the computing power operation subtask, it is determined that the processing of the computing power operation subtask is completed. The expected result is the result obtained when the computing power operation task is disassembled, and the expected result corresponds one-to-one to the computing power operation subtask.

6. The method according to any one of claims 1 to 4, characterized in that After step S3, the method further includes: Step S6: if it is determined that a failure occurs in processing the computing power running subtask by the first computing node, a second computing node is determined, and computing power resources of the second computing node are in an idle state; Step S7: Scheduling the computing resources of the second computing node to process the computing operation subtask.

7. A computing power operation task disassembly device for an intelligent computing center, characterized in that: include: A receiving module, configured to receive a target request, wherein the target request is used to request processing computing power to run a task; A disassembly module, configured to disassemble the computing power operation task into multiple computing power operation subtasks; The scheduling module is used to allocate the computing power operation subtask to the first computing node and schedule the computing power resources of the first computing node to process the computing power operation subtask, and the first computing node corresponds to the computing power operation subtask one by one.

8. An electronic device, characterized in that: include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein when the program is executed by the processor, the steps of the method for decomposing computing power operation tasks of an intelligent computing center as described in any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for decomposing computing power operation tasks of an intelligent computing center according to any one of claims 1 to 6.

10. A computer program product, characterized in that It includes computer instructions, which, when executed by a processor, implement the steps of the method for decomposing computing power operation tasks of an intelligent computing center as described in any one of claims 1 to 6.

Images