Intelligent calculation center system applied to AI large model

Abstract

The invention relates to the technical field of artificial intelligence, and discloses an intelligent calculation center system applied to an AI large model, and the system comprises a cross-data-center calculation power scheduling management module which is used for monitoring and scheduling the calculation resources of different data centers, and scheduling the calculation resources of the different data centers through the scheduling of the calculation resources of the different data centers. Targeted solution is carried out on the element content needing to be calculated at present; the green intelligent calculation center energy efficiency optimization module is used for collecting and analyzing the energy consumption data, performing final calculation and feature recognition on the energy consumption data by using a preset algorithm, generating an energy efficiency optimization strategy and implementing the energy efficiency optimization strategy; and the multi-model cooperative processing module is used for creating a front-end model tree and performing edge calculation on the created front-end model tree. According to the method, the training and reasoning efficiency of the AI large model is improved, the energy consumption is reduced, the resource utilization rate is improved, the value of the method is embodied in multi-dimensional collaborative improvement of technology, economy, society, environment and the like, and the social benefit and the economic benefit of the method are maximized through scientific planning and technical innovation in the future.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial intelligence, and more specifically, to an intelligent computing center system applied to an AI large model. Background Art

[0002] In order to further solidify the industrialization of AI and the AI-ization of industries, the preparation or construction work of intelligent computing centers has been successively carried out in many places in China. With the rapid development of artificial intelligence technology, AI large models have become a key force in promoting technological innovation and industrial upgrading.

[0003] However, the existing intelligent computing center system architectures have deficiencies in aspects such as computing power scheduling, energy efficiency optimization, and multi-model management, and it is difficult to meet the requirements of AI large models for high-performance computing resources. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent computing center system applied to an AI large model, which is used to support the training and inference of AI large models, optimize the scheduling of computing power resources, improve energy efficiency, and achieve the collaborative processing of multiple models, and provide high-performance computing (HPC), distributed storage, and low-latency networks at the technical level, aiming to solve the problem that it is difficult to meet the requirements of AI large models for high-performance computing resources in the prior art.

[0005] The present invention is implemented as follows. An intelligent computing center system applied to an AI large model includes:

[0006] A cross-data center computing power scheduling and management module, which is used to monitor and schedule the computing resources of different data centers, and by scheduling the computing resources of different data centers, achieve targeted solutions to the specific content of the current required computing;

[0007] A green intelligent computing center energy efficiency optimization module, which is used to collect and analyze energy consumption data, use a preset algorithm to calculate and identify the characteristics of the energy consumption data, and generate and implement an energy efficiency optimization strategy;

[0008] A multi-model collaborative processing module, which is used to create a front-end model tree, perform edge computing on the created front-end model tree, optimize the response to user operations, and collaboratively process the front-end models;

[0009] An intelligent computing center operating system intelligent computing OS module, which is used to manage resources, schedule tasks, monitor performance, and ensure security;

[0010] A software ecosystem module, which is used to provide industry large model services, autonomous driving services, metaverse services, and intelligent scientific research services.

[0011] Furthermore, the cross-data center computing power scheduling and management module, which is used to monitor and schedule the computing resources of different data centers, includes:

[0012] Receive a task request, obtain the task analysis path, determine whether resources are sufficient through task analysis, and perform resource allocation or optimization according to the judgment result;

[0013] If resources are sufficient, perform resource allocation according to the task analysis path. After resource allocation is completed, execute the computing resources. After the task execution is completed, recycle the computing resources;

[0014] If resources are insufficient, perform resource optimization according to the task analysis path, re-allocate and analyze the judgment resources, and then execute the computing resources according to the re-allocated analysis and judgment resources. After the task execution is completed, recycle the computing resources.

[0015] Furthermore, the energy efficiency optimization module of the green intelligent computing center is used to collect and analyze energy consumption data, including:

[0016] Receive an energy consumption data collection request to obtain energy consumption data, perform data analysis on the obtained energy consumption data, and determine whether an energy efficiency optimization strategy can be generated from the energy consumption data analysis;

[0017] If the energy consumption data analysis can generate an energy efficiency optimization strategy, then identify the obtained energy efficiency optimization strategy. After the identification is completed, implement the strategy and monitor the effect of the strategy implementation to determine whether the effect of the strategy implementation is satisfactory;

[0018] If the energy consumption data analysis cannot generate an energy efficiency optimization strategy, send a signal to the collection unit to re-obtain the energy consumption data until the energy consumption data analysis can generate an energy efficiency optimization strategy and stop;

[0019] If the effect of the strategy implementation meets the satisfaction preset by the system, it is determined that the strategy implementation has reached the satisfactory value, and the collection and analysis of the energy consumption data are completed. If the effect of the strategy implementation does not meet the satisfaction preset by the system, it is determined that the strategy implementation has not reached the satisfactory value, and return to determine whether the energy consumption data analysis can generate an energy efficiency optimization strategy.

[0020] Furthermore, if the energy consumption data analysis can generate an energy efficiency optimization strategy, then identify the obtained energy efficiency optimization strategy, including:

[0021] Obtain multiple energy efficiency strategy identification models, and sort the scores of the multiple energy efficiency strategy identification models in descending order;

[0022] Obtain the content text of the energy efficiency optimization strategy, and input the energy efficiency optimization strategy into the energy efficiency strategy identification model ranked first.

[0023] Furthermore, the multi-model collaborative processing module is used to create a front-end model tree, including:

[0024] Receive the creation request of the front-end model tree and complete the creation of the front-end model tree according to the creation request;

[0025] Respond to the user operation instruction and complete the edge computing adjustment of the created front-end model tree;

[0026] Collaborate with the AI large model to process the front-end model and optimize the management and scheduling of the front-end model.

[0027] Furthermore, the intelligent computing center operating system intelligent computing OS module includes:

[0028] The management resource unit is used to complete the resource allocation of the intelligent computing center system and the resource recovery after the computing resource allocation;

[0029] The scheduling task unit is used to execute the task queuing of the intelligent computing center system and complete the task priority adjustment by responding to the user's instruction;

[0030] The performance monitoring unit is used to complete the performance analysis of the intelligent computing center system during operation, determine whether the best performance matches the current computing resources through performance analysis, and then complete the performance tuning of the computing resources;

[0031] The security monitoring unit is used to execute the security event detection of the intelligent computing center system, respond to the security event to the control unit when detected, and implement the security policy through the control unit.

[0032] Furthermore, the software ecosystem module is used to provide industry large model services, autonomous driving services, metaverse services and intelligent scientific research services, including:

[0033] The industry large model service unit is used to complete the model training required by the intelligent computing center system, perform model inference through user response to obtain the model features required finally;

[0034] The autonomous driving service unit is used to perceive the environment in the autonomous driving scenario of the intelligent computing center system and send the environmental perception status to the control unit to complete the driving decision-making;

[0035] The metaverse service unit is used to construct the virtual world in the metaverse service scenario of the intelligent computing center system and set the interactive experience optimization unit in the virtual world;

[0036] The intelligent scientific research service unit is used to manage the data in the intelligent scientific research service scenario of the intelligent computing center system and perform intelligent analysis through data management to obtain the data management required by the user.

[0037] Furthermore, the energy efficiency optimization logic of the intelligent computing center specifically includes the following steps:

[0038] The intelligent computing center monitors the energy consumption of the signals in response to users, and analyzes the energy consumption data of the monitored data to analyze and judge whether the current energy consumption of the intelligent computing center is the optimal energy consumption loss;

[0039] If an updated energy efficiency optimization strategy is generated through analysis, implement and optimize the generated updated energy efficiency optimization strategy, conduct an effect evaluation after the optimization is completed. If an effective effect is output and the energy efficiency optimization strategy is confirmed, then adjust the optimization strategy;

[0040] If an updated energy efficiency optimization strategy is generated through analysis, implement and optimize the generated updated energy efficiency optimization strategy, conduct an effect evaluation after the optimization is completed. If the output effective effect is not confirmed as the energy efficiency optimization strategy, then return to re-conduct the energy consumption data analysis.

[0041] Furthermore, the multi-model management logic of the intelligent computing center specifically includes the following steps:

[0042] The intelligent computing center registers models for the signals in response to users, and conducts model scheduling analysis on the registered models to analyze and judge whether the registered models in the current intelligent computing center are the optimal registered models;

[0043] If the registered models in the current intelligent computing center are confirmed as the optimal registered models, then perform model execution training and implement and feedback the model optimization results. If the feedback model optimization results are within the optimization result threshold preset by the system, then update the models, complete the collaborative decision-making output, and confirm whether to conduct model scheduling;

[0044] If the feedback model optimization results are not within the optimization result threshold preset by the system, then conduct collaborative optimization of the models, re-conduct model scheduling after the collaborative optimization is completed, and judge whether it is the optimal registered model.

[0045] Compared with the prior art, an intelligent computing center system applied to an AI large model provided by the present invention has the following beneficial effects:

[0046] 1. It is used to support the training and inference of AI large models, optimize the scheduling of computing power resources, improve energy efficiency, and achieve the collaborative processing of multiple models. At the technical level, it provides high-performance computing (HPC), distributed storage, and low-latency networks. At the economic level, it reduces R & D costs and accelerates product commercialization. At the social level, it promotes the construction of a smart society and improves life efficiency. And with the popularization of AI, 5G, and the Internet of Things, the intelligent computing center will become a key infrastructure for future technological competition, and its application scenarios will continue to expand to more fields.

[0047] 2. By improving the training and inference efficiency of large AI models, reducing energy consumption, and enhancing resource utilization, the intelligent computing center is not only the "foundation" of the technological revolution, but also the "accelerator" of economic transformation and the "intelligent brain" of social governance. Its value lies in the coordinated improvement in multiple dimensions such as technology, economy, society, and environment. In the future, scientific planning and technological innovation are needed to maximize its social and economic benefits.

[0048] The above-mentioned intelligent computing center system applied to large AI models is used in promoting the digital economy and emerging industries. Description of the Drawings

[0049] Figure 1 It is a schematic structural diagram of an intelligent computing center system applied to large AI models proposed by the present invention;

[0050] Figure 2 It is a cross-data center computing power scheduling flowchart of the cross-data center computing power scheduling management module in an intelligent computing center system applied to large AI models proposed by the present invention;

[0051] Figure 3 It is a green intelligent computing center energy efficiency optimization flowchart of the green intelligent computing center energy efficiency optimization module in an intelligent computing center system applied to large AI models proposed by the present invention;

[0052] Figure 4 It is a multi-model collaborative processing flowchart of the multi-model collaborative processing module in an intelligent computing center system applied to large AI models proposed by the present invention;

[0053] Figure 5 It is a schematic structural diagram of the intelligent computing center operating system intelligent computing OS module in an intelligent computing center system applied to large AI models proposed by the present invention;

[0054] Figure 6 It is a schematic structural diagram of the software ecosystem module in an intelligent computing center system applied to large AI models proposed by the present invention;

[0055] Figure 7 It is a logical structure diagram of the intelligent computing center energy efficiency optimization in an intelligent computing center system applied to large AI models proposed by the present invention;

[0056] Figure 8 It is a logical structure diagram of the intelligent computing center multi-model management in an intelligent computing center system applied to large AI models proposed by the present invention. Detailed Embodiments

[0057] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0058] The implementation of the present invention will be described in detail below in conjunction with specific embodiments.

[0059] In the accompanying drawings of this embodiment, the same or similar reference numerals correspond to the same or similar components; in the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "left", "right", etc. indicating the orientation or positional relationship, they are based on the orientation or positional relationship shown in the accompanying drawings. This is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, the terms describing the positional relationship in the accompanying drawings are only for illustrative purposes and should not be construed as a limitation of the present invention. For those of ordinary skill in the art, the specific meanings of the above terms can be understood according to specific circumstances.

[0060] Referring to Figure 1-8 As shown, an intelligent computing center system applied to an AI large model includes:

[0061] A cross-data center computing power scheduling and management module, which is used to monitor and schedule the computing resources of different data centers. By scheduling the computing resources of different data centers, it can specifically answer the content of the computing tools required currently.

[0062] Among them, the cross-data center computing power scheduling and management module, which is used to monitor and schedule the computing resources of different data centers, includes:

[0063] Receive a task request, obtain the task analysis path, judge whether the resources are sufficient through task analysis, and allocate or optimize resources according to the judgment result.

[0064] If the resources are sufficient, allocate resources according to the task analysis path, execute the computing resources after the resource allocation is completed, and recycle the computing resources after the task execution is completed.

[0065] If the resources are insufficient, optimize the resources according to the task analysis path, re-allocate and analyze the judgment of resources, and then execute the computing resources according to the re-allocated analysis and judgment of resources. Recycle the computing resources after the task execution is completed.

[0066] A green intelligent computing center energy efficiency optimization module, which is used to collect and analyze energy consumption data, use a preset algorithm to calculate and identify the characteristics of the energy consumption data, and generate and implement an energy efficiency optimization strategy.

[0067] Among them, the green intelligent computing center energy efficiency optimization module, which is used to collect and analyze energy consumption data, includes:

[0068] Receive an energy consumption data collection request to obtain energy consumption data, perform data analysis on the obtained energy consumption data, and judge whether the energy consumption data analysis can generate an energy efficiency optimization strategy.

[0069] If the energy consumption data analysis can generate an energy efficiency optimization strategy, then identify the obtained energy efficiency optimization strategy. After the identification is completed, implement the strategy and monitor the effect of the strategy implementation to determine whether the effect of the strategy implementation is satisfactory;

[0070] If the energy consumption data analysis cannot generate an energy efficiency optimization strategy, send a signal to the acquisition unit to obtain the energy consumption data again until the energy consumption data analysis can generate an energy efficiency optimization strategy and stop;

[0071] If the effect of the strategy implementation meets the satisfaction preset by the system, it is determined that the strategy implementation has reached the satisfactory value, and the acquisition and analysis of the energy consumption data are completed. If the effect of the strategy implementation does not meet the satisfaction preset by the system, it is determined that the strategy implementation has not reached the satisfactory value, and return to judge whether the energy consumption data analysis can generate an energy efficiency optimization strategy;

[0072] The multi-model collaborative processing module is used to create a front-end model tree, perform edge computing on the created front-end model tree, optimize the response to user operations, and collaboratively process the front-end model;

[0073] Among them, the multi-model collaborative processing module is used to create a front-end model tree, including:

[0074] Receive the creation request of the front-end model tree and complete the creation of the front-end model tree according to the creation request;

[0075] Respond to the user operation instruction and complete the edge computing adjustment of the created front-end model tree;

[0076] Collaboratively process the front-end model through the AI large model to optimize the management and scheduling of the front-end model;

[0077] The intelligent computing center operating system intelligent computing OS module is used to manage resources, schedule tasks, monitor performance and security;

[0078] The software ecosystem module is used to provide industry large model services, autonomous driving services, metaverse services and intelligent scientific research services. This technical solution is used to support the training and inference of the AI large model, optimize the computing power resource scheduling, improve the energy efficiency, and realize the collaborative processing of multiple models. At the technical level, it provides high-performance computing (HPC), distributed storage, and low-latency network. At the economic level, it reduces the R & D cost and accelerates the commercialization of products. At the social level, it promotes the construction of a smart society and improves the living efficiency. And with the popularization of AI, 5G, and the Internet of Things, the intelligent computing center will become the key infrastructure for future technological competition, and its application scenarios will continue to expand to more fields.

[0079] In this embodiment, if the energy consumption data analysis can generate an energy efficiency optimization strategy, then identify the obtained energy efficiency optimization strategy, including:

[0080] Obtain multiple energy efficiency policy recognition models, and sort the scores of the multiple energy efficiency policy recognition models in descending order;

[0081] Obtain the content text of the energy efficiency optimization policy, and input the energy efficiency optimization policy into the energy efficiency policy recognition model ranked first.

[0082] In this embodiment, the intelligent computing center operating system intelligent computing OS module includes:

[0083] A management resource unit, which is used to complete the resource allocation of the intelligent computing center system and the resource recovery after the computing resources are allocated;

[0084] A scheduling task unit, which is used to execute the task queuing of the intelligent computing center system and complete the task priority adjustment by responding to the user's instructions;

[0085] A performance monitoring unit, which is used to complete the performance analysis of the intelligent computing center system during operation, determine whether the best performance matches the current computing resources through the performance analysis, and then complete the performance tuning of the computing resources;

[0086] A security monitoring unit, which is used to execute the security event detection of the intelligent computing center system, respond to the control unit when a security event is detected, and implement the security policy through the control unit.

[0087] In this embodiment, the software ecosystem module is used to provide industry large model services, autonomous driving services, metaverse services and intelligent scientific research services, including:

[0088] An industry large model service unit, which is used to complete the model training required by the intelligent computing center system, perform model inference through user response to obtain the model features required finally;

[0089] An autonomous driving service unit, which is used to sense the environment of the intelligent computing center system in the autonomous driving scenario and send the environmental perception status to the control unit to complete the driving decision-making;

[0090] A metaverse service unit, which is used to construct the virtual world of the intelligent computing center system in the metaverse service scenario and complete the setting of the interactive experience optimization unit in the virtual world;

[0091] An intelligent scientific research service unit, which is used to manage the data of the intelligent computing center system in the intelligent scientific research service scenario and perform intelligent analysis through the data management to obtain the data management required by the user.

[0092] In this embodiment, the energy efficiency optimization logic of the intelligent computing center specifically includes the following steps:

[0093] The intelligent computing center monitors the energy consumption of the signals responding to users, and analyzes the monitored data for energy consumption data analysis to determine whether the current energy consumption of the intelligent computing center is the optimal energy consumption loss;

[0094] If an updated energy efficiency optimization strategy is generated through analysis, implement and optimize the generated updated energy efficiency optimization strategy, conduct an effect evaluation after the optimization is completed. If an effective effect is output to confirm the energy efficiency optimization strategy, then adjust the optimization strategy;

[0095] If an updated energy efficiency optimization strategy is generated through analysis, implement and optimize the generated updated energy efficiency optimization strategy, conduct an effect evaluation after the optimization is completed. If the output effective effect is not confirmed as the energy efficiency optimization strategy, then return to re - conduct the energy consumption data analysis.

[0096] In this embodiment, the multi - model management logic of the intelligent computing center specifically includes the following steps:

[0097] The intelligent computing center registers models for the signals responding to users, and conducts model scheduling analysis on the registered models to determine whether the currently registered models in the intelligent computing center are the optimal registered models;

[0098] If the registered models in the current intelligent computing center are confirmed as the optimal registered models, then perform model execution training and implement and feedback the model optimization results. If the feedback model optimization results are within the optimization result thresholds preset in the system, then update the models, complete the collaborative decision - making output, and confirm whether to conduct model scheduling;

[0099] If the feedback model optimization results are not within the optimization result thresholds preset in the system, then conduct collaborative optimization of the models. After the collaborative optimization is completed, re - conduct model scheduling to determine whether it is the optimal registered model.

[0100] This technical solution is used to support the training and inference of AI large models, optimize the scheduling of computing power resources, improve energy efficiency, and achieve the collaborative processing of multiple models. At the technical level, it provides high - performance computing (HPC), distributed storage, and low - latency networks. At the economic level, it reduces R & D costs and accelerates product commercialization. At the social level, it promotes the construction of a smart society, improves living efficiency, enhances the training and inference efficiency of AI large models, reduces energy consumption, and improves resource utilization rate.

[0101] To improve the training and inference efficiency of AI large models, reduce energy consumption, and improve resource utilization rate, the intelligent computing center is not only the "foundation" of the technological revolution, but also the "accelerator" of economic transformation and the "intelligent brain" of social governance. Its value is reflected in the collaborative improvement in multiple dimensions such as technology, economy, society, and environment. In the future, through scientific planning and technological innovation, its social and economic benefits should be maximized. The specific steps are as follows:

[0102] S1: The construction of the intelligent computing center system starts with the cross-data center computing power scheduling and management module B, which is responsible for monitoring and scheduling the computing resources of different data centers;

[0103] S2: The cross-data center computing power scheduling and management module B first receives the task request A, then conducts task analysis B, and decides whether resource optimization F is needed according to whether the resources are sufficient D;

[0104] S3: The green intelligent computing center energy efficiency optimization module C generates energy efficiency optimization strategies D through energy consumption data collection B and analysis C, implements these strategies E, and finally monitors the effects F;

[0105] S4: The multi-model collaborative processing module D first creates a front-end model tree B, then responds to user operations C, collaboratively processes the front-end model D, and finally optimizes model management and scheduling E;

[0106] S5: The intelligent computing center operating system intelligent computing OS module E is responsible for resource management B, task scheduling C, performance monitoring D, and security monitoring E, including resource allocation F, resource recovery G, task queuing H, task priority adjustment I, performance analysis J, performance tuning K, security event detection L, and security policy implementation M;

[0107] S6: The software ecosystem module F provides industry large model services B, autonomous driving services C, metaverse services D, and intelligent scientific research services E, including model training F, model inference G, environment perception H, decision making I, virtual world construction J, interaction experience optimization K, data management L, and intelligent analysis M;

[0108] S7: The intelligent computing center collects data through energy consumption monitoring B, conducts energy consumption data analysis C, generates energy efficiency optimization strategies D, implements optimization E, evaluates the effects F, and adjusts the optimization strategies G according to the evaluation results;

[0109] S8: The intelligent computing center processes multi-model management, starting from model registration B, then conducting model scheduling C, model execution D, result feedback E, model update F, collaborative decision making G, and collaborative optimization H;

[0110] In this embodiment, the above-mentioned intelligent computing center system applied to the AI large model is specifically manifested in promoting the application of the digital economy and emerging industries as follows:

[0111] Metaverse and virtual reality: Support 3D rendering and real-time interaction of virtual scenes.

[0112] Autonomous driving: Process sensor data and train autonomous driving algorithms.

[0113] Blockchain: Provide high-performance computing power for encrypted computing and distributed ledgers.

[0114] This technical solution is used to support the training and inference of large AI models, providing high-performance computing (HPC), distributed storage, and low-latency networks at the technical level, reducing R & D costs and accelerating product commercialization at the economic level, promoting the construction of a smart society, improving living efficiency, enhancing the training and inference efficiency of large AI models, reducing energy consumption, and improving resource utilization at the social level. The intelligent computing center is not only the "foundation" of the technological revolution, but also the "accelerator" of economic transformation and the "intelligent brain" of social governance. Its value is reflected in the coordinated improvement in multiple dimensions such as technology, economy, society, and environment. In the future, scientific planning and technological innovation are needed to maximize its social and economic benefits.

[0115] In this embodiment, the entire operation process can be controlled by a computer for signal feedback to achieve the sequential execution of steps. These are all common knowledge in current automated control and will not be elaborated one by one in this embodiment.

[0116] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An intelligent computing center system applied to AI large models, characterized in that: include: The cross-data center computing power scheduling management module is used to monitor and schedule computing resources in different data centers. By scheduling computing resources in different data centers, targeted answers can be given to the content of the tools currently required for computing. The energy efficiency optimization module of the green intelligent computing center is used to collect and analyze energy consumption data, calculate and identify the characteristics of energy consumption data using preset algorithms, generate energy efficiency optimization strategies and implement them; A multi-model collaborative processing module is used to create a front-end model tree, perform edge computing on the created front-end model tree, optimize responses to user operations, and collaboratively process front-end models; The intelligent computing center operating system intelligent computing OS module is used to manage resources, schedule tasks, monitor performance and security; The software ecosystem module is used to provide industry large model services, autonomous driving services, metaverse services and intelligent scientific research services.

2. The intelligent computing center system applied to the AI ​​large model according to claim 1, characterized in that: The cross-data center computing power scheduling management module is used to monitor and schedule computing resources in different data centers, including: Receive task requests, obtain task analysis paths, determine whether resources are sufficient through task analysis, and allocate or optimize resources based on the determination results; If resources are sufficient, resources are allocated according to the task analysis path. After resource allocation, computing resources are executed. After task execution is completed, computing resources are recycled. If resources are insufficient, resource optimization is performed according to the task analysis path, and analysis and judgment resources are reallocated. Then, computing resources are executed based on the reallocated analysis and judgment resources. After the task is completed, computing resources are recycled.

3. The intelligent computing center system applied to the AI ​​large model as claimed in claim 2, characterized in that: The energy efficiency optimization module of the green intelligent computing center is used to collect and analyze energy consumption data, including: Receive energy consumption data collection requests to obtain energy consumption data, perform data analysis on the obtained energy consumption data, and determine whether the energy consumption data analysis can generate an energy efficiency optimization strategy; If the energy consumption data analysis can generate an energy efficiency optimization strategy, the obtained energy efficiency optimization strategy is identified, and after the identification is completed, the strategy is implemented, and the effect of the strategy implementation is monitored to determine whether the effect of the strategy implementation is satisfactory; If the energy consumption data analysis cannot generate an energy efficiency optimization strategy, a signal is sent to the acquisition unit to reacquire the energy consumption data until the energy consumption data analysis can generate an energy efficiency optimization strategy; If the effect of the strategy implementation meets the satisfaction level preset by the system, it is determined that the strategy implementation has reached the satisfaction value, and the collection and analysis of energy consumption data are completed. If the effect of the strategy implementation does not meet the satisfaction level preset by the system, it is determined that the strategy implementation has not reached the satisfaction value, and the process returns to determine whether the energy consumption data analysis can generate an energy efficiency optimization strategy.

4. The intelligent computing center system applied to the AI ​​large model as claimed in claim 3, characterized in that: If the energy consumption data analysis can generate an energy efficiency optimization strategy, the obtained energy efficiency optimization strategy is identified, including: Acquire multiple energy efficiency strategy identification models, and sort the scores of the multiple energy efficiency strategy identification models in descending order; The content text of the energy efficiency optimization strategy is obtained, and the energy efficiency optimization strategy is input into the energy efficiency strategy identification model ranked first.

5. The intelligent computing center system applied to the AI ​​large model as claimed in claim 4, characterized in that: Multi-model collaborative processing module, used to create the front-end model tree, including: Receive a request to create a front-end model tree, and complete the creation of the front-end model tree according to the creation request; Respond to user operation instructions and complete edge computing adjustments to the created front-end model tree; Collaboratively process front-end models through large AI models to optimize the management and scheduling of front-end models.

6. The intelligent computing center system applied to the AI ​​large model as claimed in claim 5, characterized in that: The intelligent computing center operating system intelligent computing OS module includes: The resource management unit is used to complete the resource allocation of the intelligent computing center system and the resource recovery after computing resource allocation; The scheduling task unit is used to execute the task queuing of the intelligent computing center system and complete the task priority adjustment by responding to the user's instructions; The performance monitoring unit is used to complete the performance analysis of the intelligent computing center system during operation, determine whether the optimal performance matches the current computing resources through performance analysis, and then complete the performance tuning of the computing resources; The security monitoring unit is used to perform security event detection on the intelligent computing center system, respond to the control unit when a security event is detected, and implement security policies through the control unit.

7. The intelligent computing center system applied to the AI ​​large model according to claim 6, characterized in that: The software ecosystem module is used to provide industry large model services, autonomous driving services, metaverse services and smart scientific research services, including: The industry large model service unit is used to complete the model training required by the intelligent computing center system and perform model reasoning through user responses to obtain the model features required by the final requirements; The autonomous driving service unit is used to sense the environment of the intelligent computing center system in the autonomous driving scenario and send the sensed environment to the control unit to complete driving decision-making; The Metaverse service unit is used to build the virtual world of the intelligent computing center system in the Metaverse service scenario and complete the setting of the interactive experience optimization unit in the virtual world; The smart scientific research service unit is used for data management of the intelligent computing center system in the smart scientific research service scenario, and performs intelligent analysis through data management to obtain the data management required by users.

8. The intelligent computing center system applied to the AI ​​large model as claimed in claim 7, characterized in that: The energy efficiency optimization logic of the intelligent computing center specifically includes the following steps: The intelligent computing center monitors the energy consumption of the signals responded to by the users, and analyzes the monitored data to determine whether the current energy consumption of the intelligent computing center is the optimal energy consumption; If an updated energy efficiency optimization strategy is generated through analysis, the generated updated energy efficiency optimization strategy is optimized and implemented, and the effect is evaluated after the optimization is completed. If an effective effect is output to confirm the energy efficiency optimization strategy, the optimization strategy is adjusted; If the analysis generates an updated energy efficiency optimization strategy, the generated updated energy efficiency optimization strategy is optimized and implemented, and the effect is evaluated after the optimization is completed. If the output effective effect is not confirmed as the energy efficiency optimization strategy, the energy consumption data analysis is returned and re-performed.

9. The intelligent computing center system applied to the AI ​​large model as claimed in claim 8, characterized in that: The multi-model management logic of the intelligent computing center specifically includes the following steps: The intelligent computing center registers the model in response to the user's signal, and performs model scheduling analysis on the registered model to determine whether the current registered model of the intelligent computing center is the optimal registered model; If the current registration model of the intelligent computing center is confirmed as the optimal registration model, the model execution training is carried out, and the feedback model optimization result is implemented. If the feedback model optimization result is within the optimization result threshold preset by the system, the model is updated, the collaborative decision output is completed, and it is confirmed whether to perform model scheduling; If the feedback model optimization result is not within the optimization result threshold preset by the system, the model is collaboratively optimized. After the collaborative optimization is completed, the model is rescheduled to determine whether it is the optimal registered model.

10. The application of an intelligent computing center system applied to an AI large model as described in any one of claims 1-9 in promoting digital economy and emerging industries.

Images