318results about How to "Avoid performance bottlenecks" patented technology

The invention provides a big data platform scheduling task and data collaborative smooth migration method and system, and relates to the technical field of big data, and the method comprises the steps: constructing a digital twin environment of a target system, collecting task operation state data of a source system, extracting task features through a deep neural network, and carrying out the clustering and grouping. And deploying the task group to a digital twin environment, analyzing a task association relationship and a resource competition relationship based on a graph neural network, and generating a task stability score in combination with a task operation index. And constructing a migration risk prediction model to perform quantitative evaluation on the migration risk, and determining a migration priority and an execution sequence according to the quantitative evaluation. And making corresponding resource configuration, data synchronization and fault recovery plans for each task group, carrying out migration drilling in a digital twinning environment, and optimizing a migration scheme by utilizing a reinforcement learning algorithm. The migration risk can be effectively reduced, the smoothness and stability of the migration process are guaranteed, and the migration efficiency is improved.

The invention discloses a Transform large model training method based on cloud edge collaboration. The method comprises the following steps: S1, constructing a data set; s2, preprocessing the data set; s3, segmenting the preprocessed data set by using a data parallelism strategy, and performing distributed training by combining data parallelism and model parallelism technologies; s4, monitoring computing resources and network conditions of the edge equipment and the cloud server in real time, dynamically adjusting a data distribution strategy, and optimizing computing load distribution; s5, a multi-agent dominant actor-commentator algorithm is combined with a multi-arm tiger machine strategy, and the layered structure of the Transform model is adjusted in real time; s6, performing partial model reasoning tasks on the edge equipment in combination with the edge reasoning capability; s7, protecting data privacy by adopting a federal learning strategy and utilizing a differential privacy technology; and S8, performing hardware acceleration by utilizing heterogeneous computing resources. According to the method, a cloud edge cooperative computing and intelligent scheduling method is adopted, and large model training and reasoning with efficient privacy protection are realized.

The invention relates to the technical field of computer storage, in particular to a multi-level cache management method, system and equipment for solid-state storage equipment and a medium. The method comprises the following steps that multiple layers of caches are configured in the solid-state storage device and comprise a high-speed cache layer, a medium-speed cache layer and a low-speed cache layer, and each layer of cache has different access speeds and capacities and is used for storing data blocks with different priorities and access frequencies; dynamically scheduling the data blocks to a proper cache layer through a cache allocation algorithm according to the real-time access frequency, size, life cycle and importance of the data blocks; establishing a comprehensive data block priority evaluation model, and calculating a priority score of each data block; according to the multi-level cache management method, by configuring the multi-layer cache, the priority evaluation model and the multi-path scheduling mechanism, dynamic scheduling and load balancing of the data blocks among different cache layers are achieved, and the overall efficiency of data access is improved.

The invention relates to the technical field of big data storage, in particular to a storage control method and device based on big data. The method comprises the following steps: acquiring an original data stream and application demand information; performing data dynamic correlation analysis on the original data flow according to the application demand information to obtain a dynamic update data correlation graph; performing feature vector generation on the dynamic update data association graph to obtain a data feature vector; performing intelligent quality evaluation on the data feature vector to obtain a data quality label; performing timeliness evaluation according to the data feature vector, and performing access popularity evaluation to obtain a timeliness score and an access popularity score; and acquiring system resource load condition information. The storage resource utilization rate, the data access performance and the overall efficiency of the system are improved by solving the limitation of a traditional method in the aspects of static data management and resource allocation.

The invention discloses a personalized recommendation system based on semantic analysis, and relates to the technical field of personalized recommendation systems.The personalized recommendation system comprises a data acquisition module, a dynamic situational semantic modeling module and a recommendation engine module, the data acquisition module transmits acquired user behavior data and situational data to the dynamic situational semantic modeling module, and the dynamic situational semantic modeling module transmits the acquired user behavior data and situational data to the recommendation engine module; the dynamic situational semantic modeling module extracts sequence features of user behavior data through a behavior context analyzer, generates context embedding through a situational semantic embedding network, completes cross-modal semantic alignment of multi-modal data through a multi-modal fusion layer, and generates multi-level semantic embedding features. According to the method, the dynamic situational semantic modeling module is designed, the semantic alignment function of the situation and the behavior is achieved, the problem that a traditional recommendation system is insufficient in situation data utilization is solved, the adaptability of the recommendation result to the dynamic situation is enhanced, and the method is suitable for dynamically changing user demand scenes.

The invention discloses a method and a system for realizing mine communication signal improvement based on edge computing, and relates to the technical field of mine communication signal improvement, and the method comprises the following steps: integrating a real-time monitoring system on each edge node, so as to continuously detect the fluctuation of communication flow and data request quantity, the real-time monitoring system identifies a potential communication peak period by collecting real-time data and generates an early warning signal to notify a network administrator or automatically trigger a load management program. According to the invention, through the real-time monitoring and early warning system on the edge node, active monitoring and peak prediction of mine communication are realized, the overload risk caused by traffic surge is reduced, and the stability of data transmission is ensured. And intelligent scheduling and dynamic resource allocation are combined, so that efficient execution of tasks under high load is ensured, and node overload is prevented. A self-adaptive load balancing and feedback mechanism enables the system to dynamically adjust distribution, quickly respond to a bottleneck and improve self-adaptability and stability.

The invention discloses a block chain multi-node secure communication method for an industrial internet, and the method comprises the steps: constructing a multi-chain cooperation mechanism, and obtaining an inter-chain interaction path table; determining a task partition optimization scheme and inter-chain interaction path update data based on the inter-chain interaction path table; matching an adaptive node resource for each task according to a task partition optimization scheme; determining an encryption mechanism between a cross-chain protocol and a block chain according to the security requirement of the task; constructing a multi-level block chain node hierarchical architecture and an intra-chain and inter-chain communication architecture, and formulating a parallel communication scheduling scheme; formulating a dynamic load balancing strategy based on the node resource use condition and the task execution state; and obtaining the task execution progress and the resource use condition of each block chain node, and adjusting the task scheduling and cross-chain communication path. Through the multi-dimensional collaborative design, the execution efficiency, the resource utilization rate and the communication security of the industrial internet are remarkably improved, and the stability and the reliability in a complex multi-chain collaborative scene are ensured.

The invention provides a railway signal monitoring data generation method and system, and relates to the technical field of railways, and the method comprises the steps: carrying out the time-space correlation fusion processing of obtained carrier frequency characteristics, response coding information, the receiving time of transponder signal data, train operation speed data and train positioning data, and generating a monitoring data record under a unified coordinate system; adopting a rule-based data compression algorithm to adaptively select a compression rule and a compression rate to perform data compression on the monitoring data record; packaging the compressed monitoring data record into a monitoring data file in a standard format; and the data management server decompresses and decodes the monitoring data file, cleans, classifies and indexes the data in the monitoring data file by adopting a MapReduce parallel computing framework, stores the structured monitoring data into a distributed database, and performs fragmentation and redundant backup management on the database.

The invention provides a log analysis adaptive optimization method and system based on swarm intelligence, and relates to the technical field of log analysis, and the method comprises the steps: obtaining a database log, and carrying out the feature extraction to generate a multi-dimensional feature matrix; constructing a hierarchical analysis network and deploying an agent analysis unit cluster with a self-learning capability; the agent analysis unit executes an analysis task through a feature recognition module, a parallel computing engine and a decision optimizer; constructing a multi-target evaluation model based on the performance evaluation indexes; and generating an optimization strategy set by adopting a self-adaptive evolution model and updating the optimization strategy set to the analysis network. According to the invention, the accuracy and efficiency of log analysis are improved, and the self-adaptive capability of the system is enhanced.

The invention provides a three-dimensional scene rendering method and device, equipment and a medium, relates to the field of computers, in particular to the technical field of page development, and can be used for application scenes such as generative search, intelligent document editing, intelligent assistants, virtual assistants and intelligent e-commerce. The method comprises the following steps: determining a plurality of elements in a to-be-rendered three-dimensional scene; based on a preset rule, putting the plurality of elements into a plurality of independent rendering layers; determining respective rendering priorities of the plurality of rendering layers; and based on the corresponding rendering priorities, sequentially rendering the plurality of rendering layers.

The invention relates to the technical field of transaction data verification, in particular to a transaction data verification method and system. The method comprises the following steps: carrying out transaction data acquisition on block chain network nodes to obtain an original transaction data set; performing adaptive heterogeneous fragmentation on the original transaction data set to obtain a fragmented transaction data set; performing verification weight coefficient distribution on the fragment transaction data set to obtain a weighted fragment transaction data set; performing multi-level verification engine deployment on the verification nodes to obtain a layered verification node network; performing asynchronous message queue mechanism configuration on the hierarchical verification node network to obtain a parallel verification network cluster; and performing load balancing scheduling on the weighted fragment transaction data set based on the parallel verification network cluster to obtain a transaction verification task sequence. According to the invention, the throughput and transaction confirmation speed of the block chain system can be effectively improved.

The invention relates to the technical field of data processing, in particular to a cloud computing-based software development application data processing method, which comprises the following steps of: collecting original data generated by software development application from a plurality of heterogeneous data sources in real time through a distributed data acquisition module; cleaning, filtering and formatting the original data, removing noise data and redundant information, and generating a standardized data set; the standardized data set is fragmented according to a preset fragmentation strategy; analyzing the stored data in real time by using the parallel processing capability of cloud computing to generate an analysis result; a result storage and updating step; a data visualization step; a security and authority management step; and an anomaly detection and fault-tolerant processing step. According to the method, the data processing speed and the throughput of the system are improved, and the requirement for real-time feedback in an agile development environment is met.

The invention provides a DETR-based 3D target detection and tracking method. The method comprises the steps of obtaining a point cloud data stream of continuous frames from an LIDER camera; original point cloud data of continuous frames are converted into regular voxel grids, and high-dimensional video features of the regular voxel grids are extracted; a group of 4D Tubelet Query is learned, and a time self-attention module and a space cross attention module are utilized to perform modeling on dynamic space-time characteristics of video clips so as to enhance the characterization capability of the dynamic space-time characteristics; and finally, performing bounding box prediction on each Tubelet Query by the model, and measuring the difference between a true value and a predicted value by using set-to-set loss to realize end-to-end learning of a multi-target tracking task. According to the method, the Tubelet is designed, the information in the space-time dimension is deeply mined and integrated, and a multi-target tracking task is simplified into a detection task, so that the unification of the detection task and the tracking task is realized, and the serious dependence of traditional multi-target tracking on a frame-level detector is reduced.

The invention discloses a notebook computer storage performance optimization method, and relates to the technical field of storage performance optimization. According to the notebook computer storage performance optimization method, by evaluating the state of the target storage area, storage abnormity is found and prompted in time, and the situation that data are lost or cannot be stored is avoided. Meanwhile, the initial storage speed and the removal speed of the data needing to be cleaned are optimized in the storage process, and efficient data writing and management are ensured. In addition, the storage state is dynamically monitored and adjusted, the user is helped to keep the stability and smoothness of the system, the service life of equipment is prolonged, the overall use experience is improved, and the performance bottleneck problem of the notebook computer in the data storage process can be effectively solved.

The invention discloses a quantum key encryption communication method and system based on a virtual private network, relates to the technical field of data encryption transmission, and solves the technical problems that the existing quantum key online distribution is covered by an OKD network, the key providing capability and the key generation rate are far lower than the data transmission speed of a modern network, and the performance bottleneck exists. According to the method, the data of the virtual private network is transmitted and encrypted by taking online distribution of the quantum key as a main part and pre-charging as an auxiliary part; the data is encrypted by the online distributed quantum key or the pre-charged quantum key, so that the data security can be improved, and the performance bottleneck of an online distribution mode can be avoided; according to the application, when the key supply gap is made up in the pre-charging mode, the quantum key surplus in the key demand gap before the time range corresponding to the key supply gap and closest to the key supply gap is preferentially selected, and when pre-charging is carried out through the selected quantum key, due to the fact that the time from generation to use is short, the leakage risk can be greatly reduced.

The invention discloses an animation rendering method and device, a medium and a product, and relates to the field of animation rendering, and the method comprises the steps: obtaining to-be-processed data, and carrying out the preprocessing of the to-be-processed data; the preprocessing comprises data format conversion and data simplification; performing fragmentation storage on the preprocessed data to be processed, and creating an index corresponding to each piece of fragmented data; storing an index corresponding to each piece of fragmented data; the GPU rendering environment is initialized; obtaining corresponding fragmented data according to the index, and performing rendering task scheduling in the initialized GPU rendering environment; and in the real-time interaction process of the user, GPU parallel processing is carried out by adopting a WebGL incremental rendering method and an optimization strategy. According to the method, the performance bottleneck of traditional Canvas paintable rendering under a large data volume can be solved, and the fluency of data rendering and animation is improved.

The invention discloses a micro-service-based enterprise software system construction method and device, and relates to the technical field of enterprise software system construction, and the method comprises the steps: carrying out the micro-service division of an initial software system according to enterprise demand information and micro-service boundary information, obtaining an independent micro-service system, and carrying out the evaluation of the independent micro-service system, determining communication architecture information between the independent micro-service systems according to the micro-service configuration information, evaluating the independent micro-service systems to obtain a micro-service configuration influence index, acquiring the micro-service configuration information, evaluating the independent micro-service systems to obtain the micro-service configuration influence index, and determining the communication architecture information between the independent micro-service systems according to the micro-service configuration information. The micro-service architecture is evaluated and optimized through the micro-service configuration influence index and the micro-service independence index, and through quantitative evaluation and the influence index, the micro-service architecture can be judged and adjusted in time, the system design is optimized, and potential performance bottleneck and system excessive dependence are avoided.

The invention relates to the technical field of data processing, in particular to a privacy data encryption method and system of an information service platform. The method comprises the following steps: acquiring running state data of the information service platform and privacy data of a user; constructing an adaptive homomorphic encryption parameter model according to the running state data and the privacy data; and encrypting the privacy data of the user by using the self-adaptive homomorphic encryption parameter model. According to the embodiment of the invention, the operation state of the information service platform and the privacy estimation value of the user privacy data are combined, and the dynamically adjusted self-adaptive homomorphic encryption model is adopted, so that the encryption strength can be automatically adjusted according to the resource load and the data privacy requirement of the platform. Through the mode, on the premise that the privacy of the user is protected, the encryption efficiency and effect can be improved, and the platform performance bottleneck caused by overhigh load in the encryption process is avoided.

The invention discloses a character string data read optimization method based on a B + tree and a learning index. A learning index generated based on a machine learning (ML) model focuses on integer data of a fixed size, but is lack of consideration for character strings of variable lengths and is poor in performance in character string data. In order to solve the problem, the invention provides a character string slicing method which can convert a character string with a variable length into an integer with a fixed size for prediction and extract a common prefix from a character string key to reduce a model fitting error. According to the method, the characteristics of character string data are fully considered, complex multi-dimensional linear regression is simplified into one-dimensional linear regression, and the model calculation overhead is reduced. Besides, the invention provides a self-adaptive strategy to select a proper index to manage the calculation cost, the method for dynamically selecting the index nodes can utilize the advantages of the two index types to the maximum extent, and the calculation overhead is effectively reduced while the index efficiency is ensured.

The invention discloses a data processing system for processing mass data and a processing method thereof, and relates to the technical field of data processing, the system comprises the following components: a data receiving module for monitoring a data access port, receiving data from an external data source, and performing primary processing on the received data for subsequent processing, according to the invention, the intelligent data routing module utilizes a reinforcement learning model to make an optimal routing decision based on data characteristics and real-time load conditions so as to ensure that data can be efficiently processed by specified nodes in a processed node cluster, and the load balancing module monitors and processes the load conditions of the node cluster in real time, so that the load balancing efficiency of the node cluster is improved. According to the method and the system, the future load peak is predicted, and the data routing strategy is intelligently adjusted according to the analysis result, so that the performance bottleneck and the fault risk caused by overload are effectively avoided, and the intelligent data routing and load balancing mechanism enables the system to keep efficient performance and stable operation when processing a large amount of data.

The invention discloses a self-supervised clustering method for massive single cell data. The method comprises the following steps: 1) downloading single cell RNA sequencing data from a disclosed gene expression database; 2) preprocessing the single-cell RNA-seq data, wherein the preprocessing comprises filtering, standardization and feature selection; (3) carrying out preliminary clustering on the preprocessed single cell RNA sequencing data by utilizing a rapid Louvain algorithm; 4) screening representative cell points from the preliminary clustering result by using an index of clustering representativeness; 5), after representative cell points are screened out from each preliminary cluster, a Transform model is used, and a final clustering result is obtained.According to the method, the rapid Louvain algorithm and the Transform neural network are combined, the efficiency and accuracy of clustering analysis are remarkably improved, and an efficient and accurate solution is provided for analysis of mass single cell data.

The invention relates to the technical field of data storage management, in particular to a data management method and system for an automobile transaction platform. The method comprises the following steps: obtaining a to-be-stored data sequence corresponding to an automobile transaction platform, and carrying out index pair classification extraction and storage space limit quantification calculation to obtain a storage release memory space limit degree corresponding to each type of index field; performing sequence storage division on the corresponding to-be-stored data sequence to obtain a storage data sequence segment corresponding to each type of index field; obtaining a query access rate and a storage interval approach access degree of the corresponding sequence segment, and carrying out storage activity evaluation analysis and index storage priority measurement analysis to obtain an index storage priority measurement of each type of index fields for each sequence segment; and carrying out priority storage management on the corresponding storage data sequence segments. According to the invention, mass data in the automobile transaction platform can be managed efficiently and accurately, so that the storage process of the data is optimized.

The invention relates to the technical field of computer networks, in particular to an implementation method and system for an RDMA network card request queue, and the method comprises the steps: checking whether a residual space exists in an RDMA Cache module or not when a write request is received; if yes, writing the writing information of the writing request into an RDMA Cache module, updating a free bit, and recording a queue pair number and an unfinished pointer; if not, writing the writing information of the writing request into a host memory module; when a read request is received, checking whether the queue pair number and the unfinished pointer carried by the read request are matched with historical records in the RDMA Cache module or not; if yes, reading from the RDMA Cache module; and if not, reading from the host memory module. According to the invention, multiple QPs in the RDMA network card share the internal resources of the Cache, the resource consumption is reduced, and the utilization rate of the cache is improved.

The invention discloses an industrial equipment data intelligent management method based on a big data algorithm. The method comprises the following steps: acquiring operation data of industrial equipment from a plurality of data sources; the missing part in the operation data is filled based on a data filling algorithm; performing format conversion and standardization processing on the multi-source heterogeneous data; positive and negative samples are balanced by adopting a resampling technology in a model training process; and carrying out large-scale data processing by utilizing a distributed computing framework. The industrial equipment data management method based on the big data algorithm is high in accuracy, high in real-time performance, simpler in data processing and high in data processing upper limit.

The invention discloses a science answering method and system based on an aligned thinking chain, and the method comprises the following steps: inputting a science question into a known LLM, generating a natural language reasoning step under a zero sample training condition, and recognizing the native style of the natural language reasoning step as a local style; compiling a few-sample prompt example according to the local style, wherein the few-sample prompt example comprises questions, reasoning steps and answers; modifying the format and reasoning steps of the few-sample prompt example according to the local style, so that the few-sample prompt example is aligned with the local style; optimizing the prompt example retriever based on the few-sample prompt example to obtain an optimized prompt example retriever; and inputting a new science question into the optimized prompt example searcher, and performing answering in combination with the LLM reasoning path to generate an answer. By aligning the native style of the language model, unifying the inference step format and enhancing the retrieval method, the manual workload is reduced, and the performance and reliability of the existing science question answering system are remarkably improved.

The invention relates to a database operation instruction execution method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: receiving a database operation instruction; the database operation instruction is used for executing data operation on a distributed database; the distributed database stores data in a distributed table through a plurality of fragments; splitting the database operation instruction to obtain a plurality of subtasks; and according to the dependency relationship between the sub-tasks and the task attributes corresponding to the sub-tasks, executing the sub-tasks on the fragments corresponding to the sub-tasks. By adopting the method, the operation and maintenance cost can be reduced.

The invention relates to a multi-board division cooperative control system based on blood purification equipment, which comprises a man-machine interaction unit, a liquid balance unit, a blood loop unit, a blood anticoagulation unit and a power supply unit, each unit is an operation unit with independent hardware, is provided with an independent PCB (Printed Circuit Board) and an independent DCDC (Direct Current-Direct Current) circuit, and performs data interaction through a custom protocol on a CAN (Controller Area Network) bus. The man-machine interaction unit receives user input and displays an equipment state; the liquid balance unit adjusts liquid input and output in the purification process; the blood loop unit is used for establishing and maintaining in-vitro blood circulation; the blood anticoagulation unit controls the injection of an anticoagulant; and the power supply unit provides power support for each module. Each unit operates independently, and only a single unit needs to be replaced when a fault occurs, so that the reliability and the maintenance efficiency of the system are improved. According to the system, modularization and efficient cooperation are achieved through distributed design, and it is ensured that equipment operates stably and safely in the blood purification process.

The invention provides a server device control method and device and a medium. The server device control method comprises the steps that a sensor is used for collecting operation data, automatic modeling is carried out to determine a resource allocation scheme, and resource allocation is controlled. Collecting data after resource allocation, dividing a task queue, matching a server, and dividing sub-tasks; acquiring subtask execution information data; controlling the plurality of servers to execute tasks according to a preset speed, and acquiring environment information and data processing temperature; the processing speed is adjusted in real time by combining task execution information data, environment data and data processing temperature; and evaluating the completion degree after task processing is completed, and judging whether processing is continued or not. According to the method, the resource allocation and task processing speed can be optimized, the task processing quality and efficiency of the server are improved, and the running stability and reliability of the server are ensured.

The invention belongs to the technical field of Lustre file management, and particularly relates to a method and system for improving Lustre small file management and performance, and the system comprises a nested file system, a three-level index module, a cache index module and a self-adaptive cache module; the nested file system realizes management and storage of small files; the three-level index module comprises a small file global metadata index, a container file internal index and a cache index and is used for optimizing retrieval and storage management of small files; the cache index module is used for managing small files accessed at high frequency and reducing repeated I / O requests; and according to the self-adaptive caching module, the client dynamically adjusts a caching strategy of a small file according to the file access frequency and mode, so that the reading speed of a high-frequency access file is improved, and network I / O (Input / Output) operation is reduced. The performance bottleneck of a Lustre file system during small file processing is solved through a nested file system, and the access burden of a metadata server and an object storage server is reduced through a three-level index architecture.

The invention discloses a multi-bus decoupling distributed embedded hardware rendering architecture, system and method, and relates to the technical field of ultra-high-definition video hardware processing, and the key points of the technical scheme are that the architecture comprises a multi-bus decoupling module which comprises a plurality of hardware buses which are physically isolated from one another, each hardware bus is an exclusive hardware transmission link bearing a single-type manufacturing task; the modular processing unit is an independent hardware board card and is integrated with an external signal input port, an upstream bus signal input port and a downstream bus signal output port; the plurality of modular processing units are cascaded in sequence through a hardware bus in the multi-bus decoupling module to form a distributed assembly line rendering link; when the modular processing units are newly added, the signal processing capability and the rendering synthesis capability are linearly improved along with the increase of the number of the modular processing units. According to the invention, the problem of performance bottleneck of the traditional architecture in processing 4K / 8K ultra-high-definition and multichannel rendering is fundamentally solved.