CN114416352B - Computing power resource allocation method and device, electronic equipment and storage medium - Google Patents
Computing power resource allocation method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN114416352B CN114416352B CN202111645466.3A CN202111645466A CN114416352B CN 114416352 B CN114416352 B CN 114416352B CN 202111645466 A CN202111645466 A CN 202111645466A CN 114416352 B CN114416352 B CN 114416352B
- Authority
- CN
- China
- Prior art keywords
- sub
- resource
- algorithm
- computing power
- resource pool
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000013468 resource allocation Methods 0.000 title claims abstract description 21
- 238000003860 storage Methods 0.000 title claims abstract description 12
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 273
- 238000004364 calculation method Methods 0.000 claims abstract description 44
- 238000013439 planning Methods 0.000 claims abstract description 19
- 230000015654 memory Effects 0.000 claims description 47
- 238000012544 monitoring process Methods 0.000 claims description 43
- 238000004891 communication Methods 0.000 claims description 18
- 230000004044 response Effects 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 12
- 238000013459 approach Methods 0.000 claims description 6
- 238000004064 recycling Methods 0.000 claims description 3
- 238000007726 management method Methods 0.000 description 12
- 238000010586 diagram Methods 0.000 description 10
- 230000008569 process Effects 0.000 description 10
- 238000012545 processing Methods 0.000 description 9
- 238000009826 distribution Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 3
- 238000000137 annealing Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000013011 mating Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2209/00—Indexing scheme relating to G06F9/00
- G06F2209/50—Indexing scheme relating to G06F9/50
- G06F2209/5011—Pool
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Power Sources (AREA)
Abstract
The application provides a computing power resource allocation method, a computing power resource allocation device, electronic equipment and a storage medium. The method comprises the following steps: according to the respective algorithm attributes of the algorithms, the respective calculation force demand characteristics of the algorithms are determined, and according to the respective calculation force demand characteristics of the algorithms, resource planning is carried out on the resource pool of the calculation force platform, so that a plurality of sub resource pools are obtained. Therefore, the computational power resources of the computational power platform can be planned more finely, and the planned computational power resources are matched with the actual use environment more. And acquiring software image files of each of the algorithms, deploying the software image files in sub-resource pools matched with each of the algorithms, acquiring algorithm application requirements of a user, and determining a target sub-resource pool matched with the algorithm application requirements from the sub-resource pools. The application can reasonably manage and divide the computational power resource pool according to the demand of the algorithm on specific computational power resources. And as the computing power resource planning of the platform is finer, the computing power resource allocation can be faster and more accurate.
Description
Technical Field
The present invention relates to the field of cloud computing technologies, and in particular, to a computing power resource allocation method, a computing power resource allocation device, an electronic device, and a storage medium.
Background
The computing resources are the sum of callable computing resources, including software and hardware, local and remote resources. The hardware includes the physical computing devices such as PCs, workstations and servers, intelligent instruments (such as oscilloscopes), and its mating accessories. The software comprises an operating system (Windows series, linux, etc.), a development environment (IDE, compiler, etc.), industry software (Matlab, CAD, etc.), office and other auxiliary class tools. The allocation of the computing power resources is realized by configuration in the computing power server. And performing deployment of various intelligent analysis algorithms in the calculation force server, and executing corresponding calculation tasks. In the existing computing force server, the binding mode of the algorithm and the computing force server is a tight coupling mode.
In the related art, the algorithm and the computing power server are in a tight coupling mode, so that the computing power resources are unevenly used, the computing power resources are severely fragmented, and the computing power resources cannot be fully utilized.
Disclosure of Invention
The embodiment of the invention provides a computing power resource allocation method, a computing power resource allocation device, electronic equipment and a storage medium, and aims to solve the problems in the background technology.
In order to solve the technical problems, the invention is realized as follows:
in a first aspect, an embodiment of the present invention provides a method for allocating computing power resources, where the method includes:
Determining respective calculation force demand characteristics of a plurality of algorithms according to respective algorithm attributes of the plurality of algorithms;
according to the respective calculation force demand characteristics of the algorithms, resource planning is carried out on a resource pool of the calculation force platform, and a plurality of sub-resource pools are obtained; the computing power resource of one sub-resource pool is matched with the computing power demand characteristics of one class of algorithms;
acquiring software image files of the algorithms, and deploying the software image files in sub-resource pools matched with the algorithms, wherein the software image files comprise various application software;
and acquiring algorithm application requirements of a user, and determining a target sub-resource pool matched with the algorithm application requirements from the plurality of sub-resource pools.
Optionally, the computing platform includes a plurality of service areas, each of the service areas includes a plurality of sub-resource pools, and each of the sub-resource pools matches an algorithm.
Optionally, the step of determining a target sub-resource pool matching the algorithm application requirement from the plurality of sub-resource pools comprises:
acquiring algorithm type requirements and positioning information of a user in algorithm application requirements of the user;
Determining a target service area of the application requirement of the algorithm on the computing platform according to the positioning information of the user;
And determining a target sub-resource pool matched with the algorithm type in the target service area according to the algorithm type indicated by the algorithm type requirement.
Optionally, after determining the target sub-resource pool of algorithm application requirements, the method further comprises:
determining target software from the plurality of application software based on the algorithm type in the target sub-resource pool;
Executing a calculation task on the target software, and generating a calculation result corresponding to the calculation task;
And generating response information from the calculation result and feeding back the response information to the user.
Optionally, the method further comprises:
monitoring the occupation condition of computing power resources in the plurality of sub-resource pools;
Determining a first sub-resource pool with the computing power resource occupation smaller than a preset threshold value in the plurality of sub-resource pools, and continuously monitoring the first sub-resource pool;
And if the computing power resources of the first sub-resource pool are still smaller than a preset threshold value in the continuous monitoring period, recycling the computing power resources of the first sub-resource pool to a standby computing power resource pool.
Optionally, the method further comprises:
monitoring the occupation condition of computing power resources in the plurality of sub-resource pools;
Determining a second sub-resource pool with the computing power resource occupation larger than a preset threshold value in the plurality of sub-resource pools, and continuously monitoring the second sub-resource pool;
And if the computing power resources of the second sub-resource pool are still larger than a preset threshold value in the continuous monitoring period, invoking the computing power resources in the standby computing power resource pool and expanding the computing power resources of the second sub-resource pool.
Optionally, the deployment mode adopted by the deployment of the software image files of the algorithms in the sub-resource pools matched with the algorithms is as follows: a container deployment approach or a virtual machine deployment approach.
A second aspect of an embodiment of the present invention proposes a computing power resource allocation apparatus, the apparatus comprising:
The demand determining unit is used for determining the computational power demand characteristics of each of the algorithms according to the algorithm attribute of each of the algorithms;
The planning unit is used for carrying out resource planning on the resource pool of the computing platform according to the computing power demand characteristics of each of the algorithms to obtain a plurality of sub-resource pools; the computing power resource of one sub-resource pool is matched with the computing power demand characteristics of one class of algorithms;
the acquisition unit is used for acquiring the software image files of the algorithms and deploying the software image files in the sub-resource pools matched with the algorithms, wherein the software image files comprise various application software;
The allocation unit is used for acquiring the algorithm application requirements of the user, and determining a target sub-resource pool matched with the algorithm application requirements from the plurality of sub-resource pools.
A third aspect of the embodiment of the invention provides an electronic device, which includes a processor, a communication interface, a memory and a communication bus, where the processor, the communication interface, and the memory complete communication with each other through the communication bus;
a memory for storing a computer program;
And the processor is used for realizing the method steps provided by the first aspect of the embodiment of the invention when executing the program stored in the memory.
A fourth aspect of the embodiments of the present invention proposes a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method as proposed in the first aspect of the embodiments of the present invention.
The embodiment of the application has the following advantages: according to the respective algorithm attributes of the algorithms, the respective calculation force demand characteristics of the algorithms are determined, and according to the respective calculation force demand characteristics of the algorithms, resource planning is carried out on the resource pool of the calculation force platform, so that a plurality of sub resource pools are obtained. Therefore, the computational power resources of the computational power platform can be planned more finely, and the planned computational power resources are matched with the actual use environment more. And acquiring software image files of each of the algorithms, deploying the software image files in sub-resource pools matched with each of the algorithms, acquiring algorithm application requirements of a user, and determining a target sub-resource pool matched with the algorithm application requirements from the sub-resource pools. Compared with the prior art, the method and the device can reasonably manage and divide the computational power resource pool according to the requirement of the algorithm on specific computational power resources. After receiving the algorithm application request of the user, the service area and the computing power resource can be rapidly allocated, the application scene suitability of the client algorithm is improved, and the computing efficiency of the platform is improved. And the application has finer calculation resource planning to the platform, and can be faster and more accurate when the calculation resource is distributed, thereby improving the response efficiency of intelligent analysis.
Drawings
In order to more clearly illustrate the embodiments of the application or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a system architecture diagram of a computing resource allocation system in an embodiment of the invention;
FIG. 2 is a flow chart of steps of a method for computing resource allocation in accordance with an embodiment of the present invention;
FIG. 3 is a schematic block diagram of a computing resource allocation device according to an embodiment of the present invention;
fig. 4 is a schematic diagram of functional modules of an electronic device according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
In the related art, an existing computing platform receives a computing power resource request of a user, executes a corresponding computing task in a computing power server of the computing power platform according to the computing power resource request, and feeds back a corresponding computing result to the user through a network. In the existing computing platform, a binding mode of tight coupling is adopted with a pre-stored algorithm. In the existing computing platform, taking one computing server as an example, corresponding CPU (central processing unit ) resources A, GPU (graphics processing unit, image processor) resources B and MEM (Memory) resources C are configured evenly, and other computing resources such as a video Memory are naturally included, so that a plurality of sub-servers are obtained, and each sub-server is deployed with algorithms of various types. When a user requests the service of the computing platform, the resource allocation principle of the computing platform is that the computing platform is allocated first to first, and when a certain type of algorithm occupies a certain computing resource (namely, a certain resource A, B, C), the user can only wait. And the expansion of the computing power resource needs to be built and expanded on the physical resource of the platform, so that the cycle is long and the cost is high.
Based on this, the inventors have proposed the inventive concept of the present application: and reasonably planning demand characteristics of the algorithm power resources according to the algorithm, and rapidly distributing corresponding algorithm resources according to the algorithm demands of the user when the client acquires different algorithm demands from the algorithm power platform.
The system comprises an algorithm management module, an algorithm platform and a resource management module, wherein the algorithm management module is used for managing the algorithm power demand characteristics, the software images and the algorithm attributes of the input algorithm. The resource management module comprises a resource deployment sub-module, a resource monitoring sub-module and a resource monitoring sub-module, wherein the resource deployment sub-module is used for carrying out resource matching planning according to the calculation power demand characteristics of the pre-input algorithm and acquiring the software mirror image of the pre-input algorithm from the algorithm management module. The resource monitoring sub-module is used for implementing monitoring of the occupation condition of the computing power resources in each sub-resource pool, and when the occupation condition of the computing power resources in the sub-resource pool is larger than a preset threshold value or smaller than the preset threshold value, the computing power resources in the sub-resource pool are expanded or recovered through the resource deployment sub-module. The computing platform comprises a plurality of edge computing servers, wherein each edge computing server comprises a plurality of sub-resource pools, and each sub-resource pool is matched with a corresponding matching requirement algorithm and is used for executing a corresponding target algorithm.
The embodiment of the application provides a computing power resource allocation method, which is applied based on the computing power resource management system. Referring to fig. 2, fig. 2 shows a flowchart of steps of a method for allocating computing power resources according to an embodiment of the present application, where the method includes:
Step S201, according to the respective algorithm attributes of the algorithms, determining the respective calculation force demand characteristics of the algorithms.
In this embodiment, in the process of the algorithm management module performing the algorithm entry, the algorithm attribute, the software image, and the calculation power demand feature of the target algorithm are extracted. The algorithm attribute is used for identifying the algorithm, namely the algorithm management module determines what type of algorithm the attribute is according to the algorithm attribute in the process of inputting the algorithm. As an example, when the algorithm management module recognizes that the algorithm attribute of the algorithm 1 is a, it is determined that the algorithm type corresponding to the algorithm 1 is a particle cluster algorithm, when the algorithm management module recognizes that the algorithm attribute of the algorithm 2 is B, it is determined that the algorithm type corresponding to the algorithm 2 is an annealing algorithm … …, and when the algorithm management module recognizes that the algorithm attribute of the algorithm n is M, it is determined that the algorithm type corresponding to the algorithm n is an M algorithm. For different algorithms, the respective computational power demand characteristics of the algorithms are different, and the computational power demand characteristics refer to the application demands of the algorithms on CPU resources, GPU resources and MEM resources. As an example, algorithm 1 requires 8 cores of 16 threads for CPU resources, 0 for GPU resources, and 20GB for MEM resources; the requirement of the algorithm 2 on CPU resources is 0, the requirement on GPU resources is a Geforce RTX 2080Ti accelerator card and 10GB video memory, and the requirement on MEM resources is 10GB; the CPU resource requirement of algorithm 3 is 4 cores and 4 threads, the Geforce GTX970 accelerator card and 2GB video memory, and the MEM resource requirement is 100GB. The demand of the algorithm 4 on CPU resources is 9-core 18 threads, the demand on GPU resources is 0, and the demand on MEM resources is 30GB; the requirement of the algorithm 5 on CPU resources is 0, the requirement on GPU resources is a GeForce RTX 2080 accelerator card and 12GB video memory, and the requirement on MEM resources is 15GB; the requirement of the algorithm 6 on CPU resources is 4 cores and 4 threads, a GeForce GTX980 accelerator card and 4GB video memory, and the requirement on MEM resources is 100GB. From the above, it can be seen that the requirements of algorithm 1 and algorithm 4 for CPU resources are higher, the requirements of algorithm 2 and algorithm 5 for GPU resources are higher, and the requirements of algorithm 3 and algorithm 6 for MEM resources are higher, i.e. the requirements of different algorithm categories for computational power resources are different.
Step S202, carrying out resource planning on a resource pool of the computing platform according to respective computing power demand characteristics of the algorithms to obtain a plurality of sub-resource pools; the computing power resources of one sub-resource pool are matched with computing power demand characteristics of one class of algorithms.
In this embodiment, the computing power resources involved in the edge computing power servers are fixed and limited for different edge computing power servers. As an example, for the edge computing force server a, the resources contained are: CPU resources of 128-core 256 threads, GPU resources of a plurality of GeForce RTX 2080Ti video memories 16G and MEM resources of 20 TB. The computing power resource contained in the edge computing power server A is a resource pool, and after the computing power demand characteristics of each input algorithm are obtained, resource planning is needed to be carried out on the computing power resource pool according to computing power demands of different algorithms. Corresponding to the algorithm 1 and the algorithm 4, the demand on CPU resources is 8 core 16 threads, the demand on GPU resources is 0, the demand on MEM resources is 20GB, the demand on CPU resources is 9 core 18 threads, the demand on GPU resources is 0, and the demand on MEM resources is 30GB, and the computing power resources at least meeting the algorithm 1 and the algorithm 4 are allocated from the computing power resource pool, so that a sub-resource pool 1 is generated, and the computing power resources in the sub-resource pool 1 are matched with the demand computing power resources of the algorithm 1 and the algorithm 4. Corresponding to algorithm 2 and algorithm 5, the requirements for CPU resources are 0, the requirements for GPU resources are Geforce RTX 2080Ti accelerator card and 10GB video memory, the requirements for CPU resources are 0, the requirements for GPU resources are Geforce RTX 2080 accelerator card and 12GB video memory, and the requirements for MEM resources are 10GB. And allocating corresponding computing power resources from the computing power resource pool to the algorithm 2 and the algorithm 5 to generate a sub-resource pool 2, wherein the computing power resources in the sub-resource pool 2 are matched with computing power resources required by the algorithm 2 and the algorithm 5. Corresponding to algorithm 3 and algorithm 6, the demand for CPU resources is 4 cores and 4 threads, the demand for MEM resources is 100GB, the demand for CPU resources is 4 cores and 4 threads, the demand for Geforce GTX 970 accelerator card and 4GB video memory, and the demand for MEM resources is 100GB. And allocating corresponding computing power resources from the computing power resource pool to the algorithm 3 and the algorithm 6 to generate a sub-resource pool 3, wherein the computing power resources in the sub-resource pool 3 are matched with computing power resources required by the algorithm 3 and the algorithm 6. After the m types of all the input algorithms are subjected to resource allocation, n sub-resource pools are generated, and all the computing resources of the edge computing server are obtained after computing resources in the n sub-resource pools are added up. Algorithm 1 and algorithm 4 belong to the same class of algorithms, namely algorithm classes with higher requirements on CPU resources, so that the algorithm can point to a resource pool with high CPU matching ratio and share one sub-resource pool. Algorithm 2 and algorithm 5 belong to the same class of algorithm, namely algorithm class with higher requirements on GPU resources, so that the algorithm can point to a resource pool with high GPU matching and share one sub-resource pool. Algorithm 3 and algorithm 6 belong to the same class of algorithm, namely algorithm class with higher requirement on MEM resources, so that the algorithm can point to a resource pool with high MEM ratio and share one sub-resource pool. According to different emphasis on the demand of the computational power resources of each algorithm, the computational power resources of the computational power resource pool are reasonably divided, so that the computational power resources in the sub-resource pools are matched with the computational power demand characteristics of one type of algorithm.
Step S203: and acquiring software image files of the algorithms, and deploying the software image files in sub-resource pools matched with the algorithms, wherein the software image files comprise various application software.
In this embodiment, after resource planning is performed on the resource pool of the computing platform to obtain a plurality of sub-resource pools, the algorithm management module obtains a software image of the first-entry algorithm. The software image file is similar to the ZIP compression package, and a specific series of files are manufactured into a single file according to a certain format, and various application software is contained. The process of resource planning for the resource pool of the computing platform is a process of hardware deployment, and after the hardware deployment is completed, the process of deploying the software image files of each algorithm to the corresponding sub-resource pool is a process of deploying software. And after the hardware deployment and the software deployment are completed, the deployment of the algorithm application environment is realized.
Step S204: and acquiring algorithm application requirements of a user, and determining a target sub-resource pool matched with the algorithm application requirements from the plurality of sub-resource pools.
In this embodiment, in the application stage, the computing platform receives an application requirement of an algorithm of a user, extracts a relevant parameter in the application requirement of the algorithm, matches a corresponding target sub-resource pool for the user from the sub-resource pools which have been planned and configured based on the relevant parameter, and executes a corresponding computing task by using computing resources in the target sub-resource pool.
In the embodiments corresponding to the steps S201 to S204, the computing power resources of the computing power platform are reasonably managed and divided according to the requirements of the algorithm on the specific computing power resources. After receiving the algorithm application request of the user, the service area and the computing power resources can be rapidly distributed, so that the computing efficiency of the computing power platform is improved. And the computational power resource planning is finer, and the computational power resource allocation is faster and more accurate, so that the response efficiency of intelligent analysis is improved.
In a possible implementation, the computing platform includes a plurality of service areas, each of the service areas having a plurality of sub-resource pools, each of the sub-resource pools matching an algorithm.
In an embodiment, the computing platform comprises a plurality of service areas, each service area corresponding to one edge computing server. By way of example, the computing platform contains 4 edge computing servers, each according to its assigned tasks. The edge computing power server 1 is arranged in Beijing, and the service area of the edge computing power server is an A-piece area; the edge computing force server 2 is arranged in a capital, and the service area is a B-piece area; the edge computing force server 3 is arranged in a Fujian, and the service area is a C-piece area; the edge computing server 2 is arranged in the Shanghai, and the service area is a D-patch. The configuration of the sub-resource pools on the edge computing power servers 1-4 is the same, and different sub-resource pools can be configured according to the difference of the regional requirements on the algorithm. And deploy the corresponding specific algorithm in the sub-resource pool. If the configuration of the sub-resource pools on the edge computing force servers 1-4 are the same, the edge computing force servers 1-4 all comprise N sub-resource pools, the N sub-resource pools are provided with the same m algorithms, the sub-resource pool 1 matches algorithm class A, the sub-resource pool 2 matches algorithm class B … …, and the sub-resource pool N matches algorithm class N. If the configuration of the sub-resource pools on the edge computing force servers 1-4 is different, the edge computing force servers 1 and 2 comprise N sub-resource pools, N algorithm categories are arranged in the N sub-resource pools, the sub-resource pool 1 matches the algorithm category A, the sub-resource pool 2 matches the algorithm category B … …, and the sub-resource pool N matches the algorithm category N; the edge force servers 3 and 4 comprise M sub-resource pools, M class algorithms are deployed in the M sub-resource pools, the sub-resource pool 1 matches algorithm class a, the sub-resource pool 2 matches algorithm class B … …, and the sub-resource pool M matches algorithm class M. And the algorithm class a in the edge computing servers 1 and 2 is different from the class to which the algorithm class a in the edge computing servers 3 and 4 belongs.
In a possible implementation, in step S204, the step of determining a target sub-resource pool of the plurality of sub-resource pools that matches the algorithm application requirement includes the sub-steps of:
Step S204-1: acquiring algorithm type requirements and positioning information of a user in algorithm application requirements of the user;
Step S204-2: determining a target service area of the application requirement of the algorithm on the computing platform according to the positioning information of the user;
Step S204-3: and determining a target sub-resource pool matched with the algorithm type in the target service area according to the algorithm type indicated by the algorithm type requirement.
In the embodiments of steps S204-1 to S204-3, when the user sends an algorithm application requirement to the computing platform, the algorithm application requirement at least includes an algorithm type requirement of the user and positioning information of the user, where the positioning information of the user is used to characterize the information of the area to which the user belongs. As an example, when a user located in a capital sends an algorithm application requirement to the computing platform, the application requirement carries identification information representing that the user belongs to the capital, and when the computing platform detects that the user belongs to the capital, the capital belongs to a B-piece area, so that a computing task corresponding to the algorithm application requirement of the user is sent to the edge computing server 2. In the edge algorithm force server 2, the algorithm type requirements among the algorithm application requirements are retrieved again. Among the n sub-resource pools of the edge computing power server 2, a target sub-resource pool is determined according to the algorithm type requirement. As an example, if the algorithm in the user algorithm type requirement is an annealing algorithm, and in the edge algorithm power server 2, the annealing algorithm is deployed in the sub-resource pool a, the sub-resource pool a is determined as the target sub-resource pool of the current algorithm application requirement. The target sub-resource pool is rapidly distributed to the user through the algorithm in the application requirements of the user algorithm and the positioning information of the user, and the calculation requirements of the user are met by utilizing the target sub-resource pool. The algorithm application requirements can also comprise scene application requirements of a user, wherein the scene application requirements refer to time delay requirements of the user on computing tasks and requirements on concurrency of an edge computing force server, and if the requirements are high in response speed or concurrency, the priority of computing force is required to be higher. So that allocation of the target sub-resource pool is performed based on the priority. The utilization rate of the computational resources is improved, and the response speed is also improved.
In a possible embodiment, after determining the target sub-resource pool of algorithm application requirements, the method further comprises the steps of:
Step S204-4: determining target software from the plurality of application software based on the algorithm type in the target sub-resource pool;
In this step, after determining the target sub-resource pool required by the user for the present algorithm application according to steps S204-1 to S204-3, the target sub-resource pool at this time may be understood as a PC, and multiple kinds of software satisfying the PC operation and the algorithm operation are preloaded on the PC according to the algorithm software image. Therefore, when executing the algorithm task, determining the target software for executing the calculation task in the software in the target sub-resource pool.
Step S204-5: executing a calculation task on the target software, and generating a calculation result corresponding to the calculation task;
in the step, the initial parameters of the current calculation task are input into the target software in the target software executing the current calculation task, and the initial parameters are brought into the algorithm model, so that the calculation result of the current calculation task is obtained.
Step S204-6: and generating response information from the calculation result and feeding back the response information to the user.
In the step, the calculation result corresponding to the calculation task executed in the sub-resource pool is packaged into a response data packet, and the response data packet is sent to the user according to the positioning information of the user, so that the closed loop of the whole calculation task is realized.
In a possible embodiment, the method further comprises monitoring the computational power resources of the sub-resource pool, which in particular comprises the steps of:
Step S205-1: monitoring the occupation condition of computing power resources in the plurality of sub-resource pools;
step S205-2: determining a first sub-resource pool with the computing power resource occupation smaller than a preset threshold value in the plurality of sub-resource pools, and continuously monitoring the first sub-resource pool;
step S205-3: and if the computing power resources of the first sub-resource pool are still smaller than a preset threshold value in the continuous monitoring period, recycling the computing power resources of the first sub-resource pool to a standby computing power resource pool.
In the embodiments described in step S205-1 to step S205-3, the resource monitoring sub-module is configured to monitor the occupancy of the computing resources in each sub-resource pool. As an example, when the number of the n sub-resource pools is m, the number of the n sub-resource pools is n, and the demand of the computing task to the computing power is smaller, the computing power resource occupation in the m sub-resource pools is smaller than a preset threshold, where the preset threshold is a threshold lower limit value, the threshold lower limit value may be five percent, and the threshold lower limit value may be adjusted according to the actual situation, which is not limited by the present application. And defining a sub-resource pool with the computing power resource occupation of less than five percent as a first sub-resource pool, and carrying out real-time monitoring on the computing power resource occupation condition of the first sub-resource pool by a resource monitoring sub-module, wherein if the computing power resource occupation is less than five percent in a preset time T, the computing power resources in the first sub-resource pool are in an idle and unused state. Thus, all of the computing resources in the first sub-resource pool may be reclaimed and stored in the spare computing resource pool. The use condition of the computing power resources in each sub-resource pool is monitored, the computing power resources are dynamically adjusted and recovered, and the computing power resources are fully utilized.
Similarly, the resource monitoring sub-module is further configured to monitor and adjust that the occupation of the computing power resources in each sub-resource pool is greater than a preset upper threshold, and specifically includes the following steps:
step S205-4: monitoring the occupation condition of computing power resources in the plurality of sub-resource pools;
Step S205-5: determining a second sub-resource pool with the computing power resource occupation larger than a preset threshold value in the plurality of sub-resource pools, and continuously monitoring the second sub-resource pool;
step S205-6: and if the computing power resources of the second sub-resource pool are still larger than a preset threshold value in the continuous monitoring period, invoking the computing power resources in the standby computing power resource pool and expanding the computing power resources of the second sub-resource pool.
In the embodiments described in steps S205-4 to S205-6, for example, the n sub-resource pools in the edge computing power server 1 are corresponding, and when the number p sub-resource pools in the n sub-resource pools have a larger demand for computing power due to the executed computing task, the computing power resources in the p sub-resource pools occupy more than a preset threshold, where the preset threshold is a threshold upper limit, the threshold upper limit may be ninety five percent, and the threshold upper limit may be adjusted according to the actual situation, which is not limited in the present application. And defining a sub-resource pool with the computing power resource occupation higher than ninety-five percent as a second sub-resource pool, and carrying out real-time monitoring on the computing power resource occupation condition of the second sub-resource pool by a resource monitoring sub-module, wherein if the computing power resource occupation is higher than the preset ninety-five percent in the preset time T, the computing power resources in the second sub-resource pool are in a full-load running state. If continuous operation is performed in a full load operation state, resource exhaustion may occur, and when a user who needs to use the computing power resources in the p sub-resource pool occurs, a situation in which the user needs to wait occurs. Therefore, the computing power resources in the standby computing power resource pool can be called and stored, and the computing power resources are expanded for the p sub-resource pool, so that the high computing power requirement is met. By monitoring the use condition of the computing power resources in each sub-resource pool, when the computing power resources are in a spent state, the computing power resources in the standby computing power resource pool and the algorithm mirror image are called for rapid expansion, and the computing power resources are fully utilized. Thereby improving the rapid deployment capability of the computing platform.
In a possible implementation manner, a deployment manner adopted by deploying software image files of each of a plurality of algorithms in sub-resource pools matched with each of the plurality of algorithms is as follows: a container deployment approach or a virtual machine deployment approach.
In this embodiment, the deployment method of the virtual machine mainly includes the following advantages: virtual machines can reduce the expenditure on server devices and can utilize one physical server resource to split into multiple independent virtual machines to accomplish a lot of work. All virtual environments can be efficiently managed using the centralized functionality of the virtual machine manager. And the systems are completely independent of each other, which means that different system environments can be installed in different virtual machines. Most importantly, the virtual machine is isolated from the host operating system and is a secure place for experimental and development applications. The container deployment mode mainly comprises the following advantages: the size of the occupied container is much smaller than that of the virtual machine, even can be as small as 10MB, and the memory and CPU utilization rate of the container can be easily limited. The container is very compact and starts up quickly compared to deploying an application requiring the deployment of a virtual machine of the entire operating system. It is possible to quickly expand the containers and add the same containers. Therefore, selecting a container deployment manner or a virtual machine deployment manner may be selected based on the scenario of the application, which is not limited by the present application.
The embodiment of the invention also provides a computing power resource distribution device, and referring to fig. 3, a functional block diagram of the computing power resource distribution device is shown, and the device can comprise the following modules:
A demand determining unit 301, configured to determine respective computational power demand characteristics of a plurality of algorithms according to respective algorithm attributes of the plurality of algorithms;
A planning unit 302, configured to perform resource planning on a resource pool of the computing platform according to respective computing power demand characteristics of the multiple algorithms, so as to obtain multiple sub-resource pools; the computing power resource of one sub-resource pool is matched with the computing power demand characteristics of one class of algorithms;
a deployment unit 303, configured to obtain software image files of the multiple algorithms, and deploy the software image files to sub-resource pools matched with the multiple algorithms, where the software image files include multiple application software;
and the allocation unit 304 is configured to obtain an algorithm application requirement of a user, and determine a target sub-resource pool matched with the algorithm application requirement from the multiple sub-resource pools.
In one possible embodiment, the distribution unit 304 includes:
the acquisition subunit is used for acquiring the algorithm type requirement and the positioning information of the user in the algorithm application requirement of the user;
The determining subunit is used for determining a target service area required by the application of the algorithm on the computing platform according to the positioning information of the user;
And the matching subunit is used for determining a target sub-resource pool matched with the algorithm type in the target service area according to the algorithm type indicated by the algorithm type requirement.
In a possible embodiment, the distribution unit 304 further comprises:
a positioning subunit, configured to determine, in the target sub-resource pool, target software from the plurality of application software based on the algorithm type;
The execution subunit is used for executing the calculation task on the target software and generating a calculation result corresponding to the calculation task;
and the feedback subunit is used for generating response information from the calculation result and feeding back the response information to the user.
In a possible embodiment, the apparatus further comprises: a monitoring and scheduling unit, the monitoring and scheduling unit comprising:
The first monitoring subunit is used for monitoring the occupation condition of the computing power resources in the plurality of sub-resource pools;
A first determining subunit, configured to determine a first sub-resource pool whose computing power resource occupation is smaller than a preset threshold, and continuously monitor the first sub-resource pool;
And the recovery subunit is used for recovering the computing power resources of the first sub-resource pool to the standby computing power resource pool if the computing power resources of the first sub-resource pool are still smaller than a preset threshold value in the continuous monitoring period.
In a possible embodiment, the monitoring and scheduling unit further comprises:
The second monitoring subunit is used for monitoring the occupation condition of the computing power resources in the plurality of sub-resource pools;
The second monitoring subunit is used for determining a second sub-resource pool with the computing power resource occupation larger than a preset threshold value in the plurality of sub-resource pools and continuously monitoring the second sub-resource pool;
And the expansion subunit is used for calling the computing power resources in the standby computing power resource pool to expand the computing power resources of the second sub-resource pool if the computing power resources of the second sub-resource pool are still larger than a preset threshold value in the continuous monitoring period.
In one possible implementation, the deployment unit 303 includes:
the first monitoring subunit is used for deploying the software image files of the algorithms in a virtual machine deployment mode in sub-resource pools matched with the algorithms.
And the second monitoring subunit is used for deploying the software image files of the algorithms in a container deployment mode in sub-resource pools matched with the algorithms.
Based on the same inventive concept, another embodiment of the present invention provides an electronic device, as shown in fig. 4, comprising a processor 41, a communication interface 42, a memory 43 and a communication bus 44, wherein the processor 41, the communication interface 42, the memory 43 complete communication with each other through the communication bus 44,
A memory 43 for storing a computer program;
The processor 41 is configured to implement the steps of the first aspect of the embodiment of the present invention when executing the program stored in the memory 43.
The communication bus mentioned by the above terminal may be a peripheral component interconnect standard (PERIPHERAL COMPONENT INTERCONNECT, abbreviated as PCI) bus or an extended industry standard architecture (Extended Industry Standard Architecture, abbreviated as EISA) bus, etc. The communication bus may be classified as an address bus, a data bus, a control bus, or the like. For ease of illustration, the figures are shown with only one bold line, but not with only one bus or one type of bus.
The communication interface is used for communication between the terminal and other devices.
The memory may include random access memory (Random Access Memory, RAM) or may include non-volatile memory (non-volatile memory), such as at least one disk memory. Optionally, the memory may also be at least one memory device located remotely from the aforementioned processor.
The processor may be a general-purpose processor, including a central processing unit (Central Processing Unit, abbreviated as CPU), a network processor (Network Processor, abbreviated as NP), etc.; but may also be a digital signal processor (DIGITAL SIGNAL Processing, DSP), application Specific Integrated Circuit (ASIC), field-Programmable gate array (FPGA) or other Programmable logic device, discrete gate or transistor logic device, discrete hardware components.
In yet another embodiment of the present application, a computer readable storage medium is provided, which includes the computing power resource allocation apparatus according to the second aspect of the embodiment of the present application.
It will be apparent to those skilled in the art that embodiments of the present invention may be provided as a method, apparatus, or computer program product. Accordingly, embodiments of the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, embodiments of the invention may take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
Embodiments of the present invention are described with reference to flowchart illustrations and/or block diagrams of methods, terminal devices (apparatus), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Finally, it is further noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. "and/or" means either or both of which may be selected. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or terminal. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or terminal device that comprises the element.
The foregoing describes in detail a method, apparatus, electronic device and storage medium for computing resource allocation, and specific examples are applied to illustrate the principles and embodiments of the present invention, and the above examples are only used to help understand the method and core idea of the present invention; meanwhile, as those skilled in the art will have variations in the specific embodiments and application scope in accordance with the ideas of the present invention, the present description should not be construed as limiting the present invention in view of the above.
Claims (10)
1. A method of computing power resource allocation, the method comprising:
Determining respective calculation force demand characteristics of a plurality of algorithms according to respective algorithm attributes of the plurality of algorithms;
according to the respective calculation force demand characteristics of the algorithms, resource planning is carried out on a resource pool of the calculation force platform, and a plurality of sub-resource pools are obtained; the computing power resource of one sub-resource pool is matched with the computing power demand characteristics of one class of algorithms;
acquiring software image files of the algorithms, and deploying the software image files in sub-resource pools matched with the algorithms, wherein the software image files comprise various application software;
and acquiring algorithm application requirements of a user, and determining a target sub-resource pool matched with the algorithm application requirements from the plurality of sub-resource pools.
2. The method of claim 1, wherein the computing platform comprises a plurality of service areas, each of the service areas comprising a plurality of sub-resource pools, each of the sub-resource pools matching an algorithm.
3. The method of claim 2, wherein the step of determining a target sub-resource pool from the plurality of sub-resource pools that matches the algorithm application requirements comprises:
acquiring algorithm type requirements and positioning information of a user in algorithm application requirements of the user;
Determining a target service area of the application requirement of the algorithm on the computing platform according to the positioning information of the user;
And determining a target sub-resource pool matched with the algorithm type in the target service area according to the algorithm type indicated by the algorithm type requirement.
4. A method according to claim 3, wherein after determining the target sub-resource pool of algorithm application requirements, the method further comprises:
determining target software from the plurality of application software based on the algorithm type in the target sub-resource pool;
Executing a calculation task on the target software, and generating a calculation result corresponding to the calculation task;
And generating response information from the calculation result and feeding back the response information to the user.
5. The method according to claim 1, wherein the method further comprises:
monitoring the occupation condition of computing power resources in the plurality of sub-resource pools;
Determining a first sub-resource pool with the computing power resource occupation smaller than a preset threshold value in the plurality of sub-resource pools, and continuously monitoring the first sub-resource pool;
and if the computing power resources of the first sub-resource pool are still smaller than a preset threshold value in the continuous monitoring period, recycling the computing power resources of the first sub-resource pool to a standby computing power resource pool.
6. The method of claim 5, wherein the method further comprises:
monitoring the occupation condition of computing power resources in the plurality of sub-resource pools;
Determining a second sub-resource pool with the computing power resource occupation larger than a preset threshold value in the plurality of sub-resource pools, and continuously monitoring the second sub-resource pool;
and if the computing power resources of the second sub-resource pool are still larger than a preset threshold value in the continuous monitoring period, invoking the computing power resources in the standby computing power resource pool and expanding the computing power resources of the second sub-resource pool.
7. The method of claim 1, wherein the deployment of the software image files of each of the plurality of algorithms in the sub-resource pools matched by each of the plurality of algorithms is as follows: a container deployment approach or a virtual machine deployment approach.
8. A computing power resource allocation apparatus, the apparatus comprising:
The demand determining unit is used for determining the computational power demand characteristics of each of the algorithms according to the algorithm attribute of each of the algorithms;
The planning unit is used for carrying out resource planning on the resource pool of the computing platform according to the computing power demand characteristics of each of the algorithms to obtain a plurality of sub-resource pools; the computing power resource of one sub-resource pool is matched with the computing power demand characteristics of one class of algorithms;
the acquisition unit is used for acquiring the software image files of the algorithms and deploying the software image files in the sub-resource pools matched with the algorithms, wherein the software image files comprise various application software;
The allocation unit is used for acquiring the algorithm application requirements of the user, and determining a target sub-resource pool matched with the algorithm application requirements from the plurality of sub-resource pools.
9. The electronic equipment is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory are communicated with each other through the communication bus;
a memory for storing a computer program;
a processor for carrying out the method steps of any one of claims 1-7 when executing a program stored on a memory.
10. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the method according to any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111645466.3A CN114416352B (en) | 2021-12-29 | 2021-12-29 | Computing power resource allocation method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111645466.3A CN114416352B (en) | 2021-12-29 | 2021-12-29 | Computing power resource allocation method and device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114416352A CN114416352A (en) | 2022-04-29 |
| CN114416352B true CN114416352B (en) | 2024-10-11 |
Family
ID=81268995
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111645466.3A Active CN114416352B (en) | 2021-12-29 | 2021-12-29 | Computing power resource allocation method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114416352B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114596009B (en) * | 2022-05-09 | 2022-07-22 | 苏州浪潮智能科技有限公司 | Computing resource deployment method, device, equipment and storage medium of intelligent computing center |
| CN115051988B (en) * | 2022-06-07 | 2024-08-16 | 瞳见科技有限公司 | Fusion scheduling system based on distributed computing power |
| CN115048224B (en) * | 2022-08-13 | 2022-11-29 | 北京蔚领时代科技有限公司 | Computing power reuse management method and device based on multiple cloud service providers |
| CN115550370B (en) * | 2022-12-01 | 2023-03-31 | 浩鲸云计算科技股份有限公司 | Computing power resource optimal scheduling allocation method based on multi-factor strategy |
| CN115643263B (en) * | 2022-12-08 | 2023-03-21 | 阿里巴巴（中国）有限公司 | Cloud native platform resource allocation method, storage medium and electronic device |
| CN117421130A (en) * | 2023-12-18 | 2024-01-19 | 成都凌亚科技有限公司 | Cloud computing power distribution system and method |
| CN117611425B (en) * | 2024-01-17 | 2024-06-11 | 之江实验室 | Method, apparatus, computer device and storage medium for configuring computing power of graphic processor |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105912397A (en) * | 2016-03-31 | 2016-08-31 | 乐视控股（北京）有限公司 | Resources management method and device |
| CN107567102A (en) * | 2017-09-22 | 2018-01-09 | 上海华为技术有限公司 | A kind of resource allocation methods and device |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10291548B2 (en) * | 2014-08-08 | 2019-05-14 | Oracle International Corporation | Contribution policy-based resource management and allocation system |
| US20200174844A1 (en) * | 2018-12-04 | 2020-06-04 | Huawei Technologies Canada Co., Ltd. | System and method for resource partitioning in distributed computing |
| CN111580951B (en) * | 2019-02-15 | 2023-10-10 | 杭州海康威视数字技术股份有限公司 | Task allocation method and resource management platform |
| CN111176852B (en) * | 2020-01-15 | 2024-04-16 | 上海依图网络科技有限公司 | Resource allocation method, device, chip and computer readable storage medium |
| CN111679905B (en) * | 2020-05-11 | 2022-03-08 | 天津大学 | Calculation network fusion network model system |
| CN111966485B (en) * | 2020-06-30 | 2024-03-15 | 北京百度网讯科技有限公司 | Scheduling method and device of computing resources, electronic equipment and storage medium |
| CN111880914A (en) * | 2020-07-20 | 2020-11-03 | 北京百度网讯科技有限公司 | Resource scheduling method, resource scheduling apparatus, electronic device, and storage medium |
| CN112783659B (en) * | 2021-02-01 | 2023-08-04 | 北京百度网讯科技有限公司 | Resource allocation method and device, computer equipment and storage medium |
| CN113590282A (en) * | 2021-07-19 | 2021-11-02 | 海宁奕斯伟集成电路设计有限公司 | Calculation force scheduling method, system, electronic equipment and computer readable storage medium |
| CN113791906A (en) * | 2021-08-09 | 2021-12-14 | 戴西(上海)软件有限公司 | Scheduling system and optimization algorithm based on GPU resources in artificial intelligence and engineering fields |
-
2021
- 2021-12-29 CN CN202111645466.3A patent/CN114416352B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105912397A (en) * | 2016-03-31 | 2016-08-31 | 乐视控股（北京）有限公司 | Resources management method and device |
| CN107567102A (en) * | 2017-09-22 | 2018-01-09 | 上海华为技术有限公司 | A kind of resource allocation methods and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114416352A (en) | 2022-04-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114416352B (en) | Computing power resource allocation method and device, electronic equipment and storage medium | |
| CN109783229B (en) | Thread resource allocation method and device | |
| CN108632365B (en) | Service resource adjusting method, related device and equipment | |
| CN111768006B (en) | Training method, device, equipment and storage medium for artificial intelligent model | |
| CN115328663B (en) | Method, device, equipment and storage medium for scheduling resources based on PaaS platform | |
| CN112465146B (en) | Quantum and classical hybrid cloud platform and task execution method | |
| CN108572845B (en) | Upgrading method of distributed micro-service cluster and related system | |
| CN106033373A (en) | A method and a system for scheduling virtual machine resources in a cloud computing platform | |
| CN111464659A (en) | Node scheduling method, node pre-selection processing method, device, equipment and medium | |
| CN113986534A (en) | Task scheduling method and device, computer equipment and computer readable storage medium | |
| CN114168302A (en) | Task scheduling method, device, equipment and storage medium | |
| CN109117244B (en) | Method for implementing virtual machine resource application queuing mechanism | |
| CN113886069A (en) | Resource allocation method and device, electronic equipment and storage medium | |
| CN112486642A (en) | Resource scheduling method and device, electronic equipment and computer readable storage medium | |
| CN118364918B (en) | Reasoning method, device, equipment and storage medium of large language model | |
| CN116069493A (en) | Data processing method, device, equipment and readable storage medium | |
| CN112860442A (en) | Resource quota adjusting method and device, computer equipment and storage medium | |
| CN112003931A (en) | Method and system for deploying scheduling controller and related components | |
| CN116881003A (en) | Resource allocation method, device, service equipment and storage medium | |
| CN115309558A (en) | Resource scheduling management system, method, computer equipment and storage medium | |
| CN114237902A (en) | Service deployment method and device, electronic equipment and computer readable medium | |
| CN113722091A (en) | Simulation task deployment method, system and related device | |
| CN117992239B (en) | Resource management allocation method, intelligent computing cloud operating system and computing platform | |
| CN116954812B (en) | Service request processing method, device, equipment and storage medium | |
| CN118426911B (en) | CAE preprocessing simulation software containerization service method, device, server and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |