Cloud mobile phone GPU computing power resource optimization method and device
By allocating and managing the internal network IP ports of cloud phone GPU computing resources through the management module, and combining this with real-time analysis by the task management module, the problem of inflexible allocation of cloud phone GPU computing resources has been solved, achieving efficient utilization and flexible allocation of resources.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA MOBILE (SUZHOU) SOFTWARE TECH CO LTD
- Filing Date
- 2025-12-29
- Publication Date
- 2026-04-21
AI Technical Summary
The allocation of GPU computing resources in cloud phones in existing cloud computing is inflexible, resulting in resource waste and low utilization, which fails to meet the dynamic needs of users.
The GPU computing power resource intranet IP port allocation module and management module are used for resource configuration and management, and the task management module is used for real-time demand analysis and allocation to realize the virtualization and dynamic scheduling of computing power resources.
It improves the utilization rate of computing resources and enables flexible allocation and dynamic distribution of GPU computing resources in cloud phones, meeting the diverse needs of users.
Smart Images
Figure CN121907876A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of cloud computing technology, and in particular to a method and apparatus for optimizing GPU computing resources in cloud mobile phones. Background Technology
[0002] In existing technologies, the computing model of cloud phones is similar to that of conventional cloud computing. The phone's application is placed on a remote server for image capture, rendering, and encoding. The user terminal accesses the cloud phone via a web application or by clicking the connection button in a local application and using a high-speed internet connection. Commands are issued from the user terminal, the server executes the corresponding rendering tasks, and the rendered image is then transmitted back to the user terminal for display. Currently, for cloud phone business scenarios, cloud computing vendors pre-deploy computing resources. This means that based on the product design, they define the specifications supported by the current cloud phone service, allocate server resources, and deploy the corresponding services on the allocated servers. Users can only use the specifications pre-defined by the cloud computing vendor during actual use.
[0003] Currently, cloud computing vendors allocate resources for various services but wait for users to subscribe. Only after subscription are these resources actually used, while idle resources are wasted, resulting in low utilization of computing power. Furthermore, when building cloud phone resource pools, vendors pre-plan the specifications of resources for different services. However, users may need the resources of an entire server. In this case, the pre-allocated resources are clearly insufficient to meet user needs, making the allocation of computing power resources inflexible. Therefore, how to flexibly allocate the computing power resources of cloud phone GPUs has become a problem to be solved. Summary of the Invention
[0004] This application provides a method and apparatus for optimizing GPU computing resources in cloud phones. It utilizes a GPU computing resource intranet IP port allocation module to allocate intranet IP ports to underlying GPU computing resources. A GPU computing resource management module then configures and manages these resources based on the intranet IP ports, creating a GPU computing pool. Finally, a cloud phone task management module analyzes resource requirements based on phone tasks and processes the data stream to be rendered using the corresponding GPU computing resources to obtain the output data stream. Through virtualized management of computing resources and automatic allocation of underlying computing power for business purposes, the utilization rate of computing resources can be improved. Furthermore, unified scheduling and management across various modules, along with real-time allocation and statistics of computing resources based on user-provided computing power demands, makes the allocation of GPU computing resources in cloud phones more flexible.
[0005] In a first aspect, embodiments of this application provide a method for optimizing GPU computing resources in cloud phones, the method comprising: The GPU computing power resource intranet IP port allocation module allocates intranet IP ports to the underlying GPU computing power resources; The GPU computing power resource management module configures and manages the underlying GPU computing power resources based on the internal network IP port to obtain the GPU computing power pool; The cloud phone task management module analyzes resource requirements based on the phone task to obtain the data stream to be rendered; The cloud phone task management module obtains the corresponding GPU computing resources from the GPU computing power pool based on the data stream to be rendered, processes the data stream to be rendered based on the corresponding GPU computing power resources, and obtains the output data stream.
[0006] Furthermore, the method also includes: The module deletes the underlying GPU computing node when the mobile task is not using the underlying GPU computing node.
[0007] Furthermore, the method also includes: The GPU computing power resource intranet IP port allocation module powers on the underlying GPU computing power resources, connects the underlying GPU computing power resources to the upstream intranet switch via a physical network cable, and allocates the corresponding intranet IP and corresponding port to the underlying GPU computing power resources.
[0008] Furthermore, the method also includes: The GPU computing resource management module transmits relevant information about the underlying GPU computing resources to the GPU computing pool using the SNMP protocol and the internal network IP port.
[0009] Furthermore, the method also includes: Information related to the underlying GPU computing resources includes the type of GPU computing resources and the performance configuration information of the GPU computing resources.
[0010] Furthermore, the method also includes: The cloud phone task management module sends a virtualization request to the baseboard management controller in the GPU computing power resources based on the intranet IP port; The baseboard management controller receives a virtualization request, obtains a virtualization command based on the virtualization request, and sends it to the GPU computing power virtualization module in the GPU computing power resources; The GPU computing power virtualization module obtains multiple virtual GPUs based on virtualization commands and connected GPU hardware, which serve as corresponding GPU computing power resources.
[0011] Furthermore, the method also includes: The GPU computing power virtualization module connects to each GPU hardware unit via the SNMP protocol.
[0012] Secondly, embodiments of this application provide a cloud phone GPU computing power resource optimization device, the device comprising: The GPU computing power resource intranet IP port allocation module is used to allocate intranet IP ports to the underlying GPU computing power resources. The GPU computing power resource management module is used to configure and manage the underlying GPU computing power resources based on the internal network IP port to obtain the GPU computing power pool. The cloud phone task management module is used to analyze resource requirements based on mobile phone tasks to obtain the data stream to be rendered; it then obtains the corresponding GPU computing resources from the GPU computing power pool based on the data stream to be rendered, processes the data stream to be rendered based on the corresponding GPU computing resources, and obtains the output data stream.
[0013] Thirdly, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it performs the steps of a cloud phone GPU computing power resource optimization method as described in any of the above embodiments.
[0014] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of a cloud phone GPU computing power resource optimization method as described in any of the above embodiments.
[0015] In summary, compared with the prior art, the beneficial effects of the technical solution provided in this application include at least the following: This application provides a method for optimizing GPU computing resources in cloud phones. It employs a GPU computing resource intranet IP port allocation module to allocate intranet IP ports to underlying GPU computing resources. A GPU computing resource management module then configures and manages these resources based on the intranet IP ports, creating a GPU computing pool. Finally, a cloud phone task management module analyzes resource requirements based on phone tasks and processes the data stream to be rendered using the corresponding GPU computing resources to obtain the output data stream. By virtualizing computing resources and automatically allocating underlying computing power for business purposes, the utilization rate of computing resources can be improved. Furthermore, unified scheduling and management across various modules, along with real-time allocation and statistics of computing resources based on user-provided computing power needs, makes the allocation of GPU computing resources in cloud phones more flexible. Attached Figure Description
[0016] Figure 1 A flowchart illustrating a method for optimizing GPU computing resources in a cloud phone, provided as an exemplary embodiment of this application.
[0017] Figure 2This is a structural diagram of a cloud phone GPU computing power resource optimization device provided as an exemplary embodiment of this application. Detailed Implementation
[0018] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.
[0019] Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0020] Please see Figure 1 This application provides a method for optimizing GPU computing resources in cloud phones, which specifically includes the following steps: Step S1: The GPU computing power resource intranet IP port allocation module allocates intranet IP ports to the underlying GPU computing power resources.
[0021] In some embodiments, the method further includes: The GPU computing power resource intranet IP port allocation module powers on the underlying GPU computing power resources, connects the underlying GPU computing power resources to the upstream intranet switch via a physical network cable, and allocates the corresponding intranet IP and corresponding port to the underlying GPU computing power resources.
[0022] Specifically, if the underlying GPU computing resources are powered on, then the power supply of the underlying GPU computing resources will be restarted; if the underlying GPU computing resources are not powered on, then the power supply of the underlying GPU computing resources will be directly started.
[0023] In one feasible implementation, an internal network IP port is allocated to the underlying GPU computing resources; specifically, this step aims to allocate an internal network IP port to the underlying GPU computing resources so as to realize the management and configuration of the GPU computing resources through the internal network IP port.
[0024] In this application, the GPU hardware and related modules are integrated into a single chassis, which is referred to as a GPU computing node.
[0025] In addition, the above-mentioned allocation of internal network IP ports for GPU computing nodes may include: powering on the GPU computing node and physically connecting the GPU computing node to the internal network connection switch in order to configure the corresponding internal network IP port.
[0026] Specifically, if the GPU computing node is powered on, its power is restarted; if it is not powered on, its power is directly started.
[0027] Furthermore, after the GPU computing node is directly started or restarted, the internal network IP port of the GPU computing node can be configured by logging into the BMC management platform (i.e., the baseboard management controller) of the GPU computing node.
[0028] Specifically, the GPU computing node is configured to boot from PXE via the BMC within the GPU computing node. Furthermore, assigning internal network IP ports to the GPU computing node essentially involves assigning IP ports to the BMC and service network cards within the GPU pool node, and binding them to the ports of the upper-layer switch.
[0029] Once a GPU computing node is assigned an internal network IP port, it can be managed and configured through this port. GPU computing nodes can also be dynamically discovered, eliminating the inconvenience of manual configuration when the scale is large. After obtaining the internal network IP port, the GPU computing node automatically establishes a connection with the cloud management terminal. At this time, the cloud management terminal can obtain real-time information on the usage and status of the GPU computing node.
[0030] In step S2, the GPU computing power resource management module configures and manages the underlying GPU computing power resources based on the internal network IP port to obtain the GPU computing power pool.
[0031] In some embodiments, the method further includes: The GPU computing resource management module transmits relevant information about the underlying GPU computing resources to the GPU computing pool using the SNMP protocol and the internal network IP port.
[0032] The relevant information about the underlying GPU computing resources can include the type of GPU computing resources and the performance configuration information of the GPU computing resources.
[0033] After configuring the internal network IP port of the GPU computing resources, the GPU computing resource management module sends a resource management request to the underlying GPU computing resources through the internal network IP port using the SNMP protocol. The SNMP protocol is used to transmit relevant information about the GPU computing resources, including the type and performance configuration information of the GPU computing resources, so that the GPU computing resource management module can manage the underlying GPU computing resources and incorporate the GPU computing performance configuration information into the overall GPU computing pool for subsequent overall scheduling.
[0034] In some feasible implementations, the GPU computing power resource management module manages GPU computing power nodes, obtains information such as the type and performance configuration of the underlying GPU computing power resources, and incorporates the performance data corresponding to the newly managed GPU computing power resources into the GPU computing power pool for subsequent unified allocation of GPU computing power.
[0035] Specifically, after allocating an internal network IP port to the GPU computing power node and establishing a connection between the GPU computing power node and the cloud management terminal through this internal network IP port, the GPU computing power node reports the corresponding GPU computing power type and computing power performance configuration data through the SNMP protocol.
[0036] Furthermore, managed GPU computing nodes can be selected, and the corresponding computing performance data of the GPU computing nodes can be included in the GPU computing resource pool for unified allocation and scheduling of GPU computing resources in the future.
[0037] Step S3: The cloud phone task management module performs resource requirement analysis based on the phone task to obtain the data stream to be rendered.
[0038] The cloud phone task management module analyzes the resource requirements of the mobile files in the mobile task. Based on this, it allocates the corresponding vGPU computing resources in the GPU computing power pool according to the resources required by the task to the corresponding mobile task. Specifically, it transmits the data stream of the mobile file to the corresponding vGPU through the vGPU's internal network IP port to perform mobile encoding operations.
[0039] In step S4, the cloud phone task management module obtains the corresponding GPU computing resources from the GPU computing power pool based on the data stream to be rendered, processes the data stream to be rendered based on the corresponding GPU computing power resources, and obtains the output data stream.
[0040] In some embodiments, the method further includes: The cloud phone task management module sends a virtualization request to the baseboard management controller in the GPU computing power resources based on the intranet IP port; The baseboard management controller receives a virtualization request, obtains a virtualization command based on the virtualization request, and sends it to the GPU computing power virtualization module in the GPU computing power resources; The GPU computing power virtualization module obtains multiple virtual GPUs based on virtualization commands and connected GPU hardware, which serve as corresponding GPU computing power resources.
[0041] In some embodiments, the method further includes: The GPU computing power virtualization module connects to each GPU hardware unit via the SNMP protocol.
[0042] In some feasible implementations, this application can intelligently analyze the pending mobile phone files for mobile tasks, analyze the amount of GPU computing resources required by the current pending mobile phone task, and allocate a corresponding number of GPU computing resources from the GPU computing pool based on the amount of GPU computing resources required by the mobile phone task. This is mainly manifested in obtaining the IP port of the corresponding GPU computing resources, using the SNMP protocol through the internal network IP port of the GPU computing node to achieve information exchange with the GPU computing node, sending a virtualization request to the GPU computing node, and then enabling the GPU computing node to perform virtualization operations upon receiving this virtualization request, thereby obtaining multiple virtual GPUs.
[0043] The above-mentioned step of sending a virtualization request to the GPU computing node through the internal network IP port after configuration, so that the GPU computing node can pool and obtain multiple virtual GPUs, may include: sending a virtualization request to the BMC in the GPU computing node through the internal network IP port after configuration, so that the BMC can send a virtualization command to the GPU computing virtualization module in the GPU computing node, and use the GPU computing virtualization module to virtualize the connected GPU hardware into multiple virtual GPUs.
[0044] Specifically, the BMC in the GPU computing node communicates with the cloud management terminal using the SNMP protocol. The cloud management terminal creates GPU virtualization tasks through SNMP protocol information exchange and sends virtualization requests to the BMC in the GPU computing node. After receiving the virtualization request from the cloud management terminal, the BMC further issues specific virtualization commands to the GPU computing virtualization module in the GPU computing node. The GPU computing virtualization module then virtualizes the connected GPU hardware into multiple virtual GPUs.
[0045] In one specific implementation, the GPU computing power virtualization module can connect to each GPU hardware via the SNMP protocol. After obtaining the corresponding GPU virtualization resources, the task file to be processed is transmitted to the corresponding GPU virtualization resources via data stream, and the screen is captured and encoded. After the encoding is completed, the corresponding data stream is sent back to the cloud phone task management module for output.
[0046] In some embodiments, the method further includes: The module deletes the underlying GPU computing node when the mobile task is not using the underlying GPU computing node.
[0047] Specifically, when a mobile task no longer uses GPU computing resources, the virtual GPU allocated to that task is deleted. This means that GPU resources can be deleted even when they are no longer in use, achieving dynamic assembly and allocation of cloud phone resources. This not only enables dynamic reuse of cloud phone GPU resources, improving resource utilization, but also allows for resource allocation, breaking the limitations of traditional hardware.
[0048] The cloud phone GPU computing power resource optimization method provided in the above embodiments can allocate internal network IP ports to the underlying GPU computing power resources using a GPU computing power resource internal network IP port allocation module. Then, a GPU computing power resource management module configures and manages the underlying GPU computing power resources according to the internal network IP ports, resulting in a GPU computing power pool. Finally, a cloud phone task management module analyzes resource requirements based on phone tasks and processes the data stream to be rendered according to the corresponding GPU computing power resources to obtain the output data stream. Through virtualized management of computing power resources and automatic allocation of underlying computing power for business purposes, the utilization rate of computing power resources can be improved. Unified scheduling and management of various modules, real-time allocation of computing power and statistics of computing power resources based on user-provided computing power needs, makes the allocation of cloud phone GPU computing power resources more flexible.
[0049] Please see Figure 2 Another embodiment of this application provides a cloud phone GPU computing power resource optimization device, the device comprising: The GPU computing power resource intranet IP port allocation module 101 is used to allocate intranet IP ports to the underlying GPU computing power resources.
[0050] The GPU computing power resource management module 102 is used to configure and manage the underlying GPU computing power resources based on the internal network IP port to obtain the GPU computing power pool.
[0051] The cloud phone task management module 103 is used to analyze resource requirements based on the phone task to obtain the data stream to be rendered; obtain the corresponding GPU computing resources from the GPU computing power pool based on the data stream to be rendered; process the data stream to be rendered based on the corresponding GPU computing resources to obtain the output data stream. The specific limitations of the cloud phone GPU computing power resource optimization device provided in this embodiment can be found in the embodiment of the cloud phone GPU computing power resource optimization method described above, and will not be repeated here. Each module in the above-described cloud phone GPU computing power resource optimization device can be implemented entirely or partially through software, hardware, or a combination thereof. Each module can be embedded in or independent of the processor in the computer device in hardware form, or it can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each module.
[0052] This application provides a computer device that may include a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it causes the processor to perform the steps of a cloud phone GPU computing resource optimization method as described in any of the above embodiments.
[0053] The working process, working details, and technical effects of the computer device provided in this embodiment can be found in the embodiment of a cloud phone GPU computing power resource optimization method described above, and will not be repeated here.
[0054] This application provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements the steps of a cloud phone GPU computing power resource optimization method as described in any of the above embodiments. The computer-readable storage medium refers to a data storage medium, which may include, but is not limited to, floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or memory sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.
[0055] The working process, working details, and technical effects of the computer-readable storage medium provided in this embodiment can be found in the embodiment of a cloud phone GPU computing power resource optimization method described above, and will not be repeated here.
[0056] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
[0057] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0058] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.
Claims
1. A method for optimizing GPU computing resources in cloud phones, characterized in that, The method includes: The GPU computing power resource intranet IP port allocation module allocates intranet IP ports to the underlying GPU computing power resources; The GPU computing power resource management module configures and manages the underlying GPU computing power resources based on the internal network IP port to obtain a GPU computing power pool. The cloud phone task management module analyzes resource requirements based on the phone task to obtain the data stream to be rendered; The cloud phone task management module obtains the corresponding GPU computing resources from the GPU computing power pool based on the data stream to be rendered, processes the data stream to be rendered based on the corresponding GPU computing resources, and obtains the output data stream.
2. The cloud phone GPU computing power resource optimization method according to claim 1, characterized in that, The method further includes: The module deletes the underlying GPU computing node when the mobile task does not use the underlying GPU computing node.
3. The cloud phone GPU computing power resource optimization method according to claim 1, characterized in that, The method further includes: The GPU computing power resource intranet IP port allocation module powers on the underlying GPU computing power resource, connects the underlying GPU computing power resource to the upstream intranet switch via a physical network cable, and allocates the corresponding intranet IP and corresponding port to the underlying GPU computing power resource.
4. The cloud phone GPU computing power resource optimization method according to claim 1, characterized in that, The method further includes: The GPU computing power resource management module transmits relevant information about the underlying GPU computing power resources to the GPU computing power pool according to the SNMP protocol and the internal network IP port.
5. The cloud phone GPU computing power resource optimization method according to claim 4, characterized in that, The method further includes: The relevant information about the underlying GPU computing resources includes the type of GPU computing resources and the performance configuration information of the GPU computing resources.
6. The cloud phone GPU computing power resource optimization method according to claim 1, characterized in that, The method further includes: The cloud phone task management module sends a virtualization request to the baseboard management controller in the GPU computing power resources based on the intranet IP port; The baseboard management controller receives the virtualization request, obtains a virtualization command based on the virtualization request, and sends it to the GPU computing power virtualization module in the GPU computing power resources; The GPU computing power virtualization module obtains multiple virtual GPUs based on the virtualization command and the connected GPU hardware, which serve as corresponding GPU computing power resources.
7. The cloud phone GPU computing power resource optimization method according to claim 6, characterized in that, The method further includes: The GPU computing power virtualization module connects to each GPU hardware via the SNMP protocol.
8. A cloud phone GPU computing power resource optimization device, characterized in that, The device includes: The GPU computing power resource intranet IP port allocation module is used to allocate intranet IP ports to the underlying GPU computing power resources. The GPU computing power resource management module is used to configure and manage the underlying GPU computing power resources according to the internal network IP port to obtain a GPU computing power pool. The cloud phone task management module is used to analyze resource requirements based on mobile phone tasks to obtain a data stream to be rendered; obtain corresponding GPU computing resources from the GPU computing power pool based on the data stream to be rendered; process the data stream to be rendered based on the corresponding GPU computing resources to obtain an output data stream.
9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the cloud phone GPU computing power resource optimization method as described in any one of claims 1 to 7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the cloud mobile phone GPU computing power resource optimization method as described in any one of claims 1 to 7.