340 results about "Asynchronous processing" patented technology

The invention relates to the technical field of computer communication, provides a distributed accounting asynchronous data processing method and system based on an MEMO state machine, and is used for improving the intelligent adaptability of a distributed accounting network to a dynamic service scene. The method comprises the following steps: acquiring a transaction request set in a distributed accounting network, extracting an asynchronous processing trigger event of the transaction request set, analyzing the asynchronous processing trigger event, generating a state transition instruction sequence containing M state transition paths associated with K transaction data blocks, a parallel prediction module based on an MEMO state machine performs asynchronous processing on the M state transition paths to generate N prediction state vectors and corresponding conflict detection marks; performing feature fusion on the N prediction state vectors according to the conflict detection mark to generate a global state updating matrix; and writing the global state updating matrix into a consensus node set of the distributed accounting network, and generating an asynchronous processing state log matched with the transaction request set.

Provided herein are exemplary systems and methods for an intelligent secure networked system configured by at least one processor to execute instructions stored in memory, the system including a data retention system and an emotional analytics system, the emotional analytics system performing asynchronous processing to determine if interactions with a user's computing device are such that the user is responding from an emotional state of mind or a meditated state of mind.

The invention discloses a hydraulic motor fault diagnosis method and system based on heterogeneous asynchronous data fusion, and the method comprises the steps: collecting heterogeneous asynchronous data of a sensor network, carrying out the preprocessing of the data, inputting the data into a parallel dynamic pruning residual network, extracting features, and carrying out the preliminary fusion, thereby obtaining an initial feature plane; and inputting the initial feature plane into a multi-connection neural network, fusing data, extracting features, obtaining a final feature plane, using the fused feature data as a training set, and building a feature classification model to test and evaluate the performance of the feature classification model. According to the method, the limitation of a traditional fusion model in asynchronous processing of high-frequency vibration signals and low-frequency thermodynamic data of a hydraulic system is effectively overcome, the robustness of key fault features in a strong noise environment is remarkably improved, and rapid virtual-real mapping of bench test simulation data and online monitoring data is realized; typical faults such as plunger pair abrasion and valve plate cavitation of the hydraulic motor can be accurately supported, and safety guarantee is provided for equipment life prediction and safety control.

The invention discloses an interactive large-screen visualization system and method based on multi-dimensional data real-time rendering, relates to the technical field of image rendering, and aims to solve the bottleneck problems of a traditional system in the aspects of data throughput efficiency, rendering delay and interactive response. The method comprises the following steps: dividing space blocks by adopting a non-uniform octree, dynamically adjusting the size of the blocks according to data density, and combining a ZSTD + Hoffman hybrid compression algorithm; a CPU-GPU three-level asynchronous processing assembly line is constructed, and task priorities are dynamically adjusted through a fixation point weight algorithm; the master node dynamically allocates LOD parameters, and the slave nodes achieve special-shaped large-screen sub-pixel-level splicing based on a spherical coordinate system model; a Q-learning algorithm is adopted to predict a user operation sequence, and dual-channel bandwidth adaptive allocation is realized in combination with an ASTP protocol. The system comprises a dynamic partitioning module, a pipeline scheduling module, a multi-node cooperation module, an intelligent preloading module and an exception handling module. Real-time interaction visualization of a large amount of data can be realized, the rendering efficiency is improved, and hardware resource occupation is reduced.

The invention provides an agent memory management method and device based on asynchronous processing and a storage medium. The method comprises the following steps: after an interaction event is written into a first storage area, issuing a processing event associated with an event identifier to an event queue, and reading the corresponding interaction event from the first storage area according to the event identifier; when the interaction event passes the importance judgment, converting the interaction event into structured memory data; determining second storage areas based on the type identification field of the structured memory data, and respectively writing the structured memory data into at least one second storage area; and analyzing the query request to obtain a query condition, determining at least one target second storage area based on the query condition, reading structured memory data related to the query request from the target second storage area, performing aggregation processing, and outputting the aggregated memory data to the agent to generate a response result. According to the method, the interaction response delay can be reduced, the model calling cost is reduced, and the memory processing auditing performance is improved.

The invention discloses a self-adaptive multi-sensor fusion SLAM (simultaneous localization and mapping) system implementation method and system, a hierarchical fusion asynchronous architecture with a front-end sensing layer, a middle fusion layer and a rear-end optimization layer is designed, and the front-end sensing layer comprises a feature extraction and matching and timestamp asynchronous processing mechanism. The middle fusion layer comprises sliding of a time window and alignment of data to a time axis, and the rear-end optimization layer comprises a multi-source factor graph model, state estimation, map updating and calculation optimization; a self-adaptive weight adjustment module is arranged between the front-end sensing layer and the middle fusion layer; multi-dimensional evaluation, self-adaptive adjustment and engineering practicability of SLAM spatial positioning mapping are realized; the front-end sensing layer performs data acquisition and preprocessing on each sensor for spatial positioning mapping, and then independently performs feature extraction and matching and timestamp asynchronous processing mechanism implementation. The problem of fusion time sequence synchronization is effectively solved, the environmental adaptability is improved, and the calculation complexity is reduced.

According to the priority asynchronous processing method and system for network security, multiple data sources are integrated, the threat level of each data packet is evaluated in real time, and the processing priority of each data packet is dynamically adjusted according to the threat level. Different from a traditional static method, the method provided by the invention utilizes an advanced machine learning algorithm, can perform dynamic evaluation in multiple dimensions, and timely identifies and preferentially processes high-risk data packets. The mechanism obviously improves the response capability of the system to novel and complex attacks. Meanwhile, a multi-node cooperative attack response mechanism is introduced, when a certain node detects abnormal traffic, other nodes can be automatically notified, threat situation information can be shared, and therefore rapid cooperative reaction is achieved. The integrated innovation not only enhances the overall security of the network, but also optimizes the data packet processing efficiency, effectively solves the main defects in the traditional network security system, and provides a new thought for constructing a safer and more flexible network environment.

The invention provides an asynchronous-to-synchronous method and device, equipment, a storage medium and a product, and the method comprises the steps: responding to a first call of a first thread to an asynchronous interface, generating a first call identifier corresponding to the first call, responding to successful call of the first call, blocking the first thread, and generating a second call identifier corresponding to the first call; the first thread is used for executing synchronous processing logic; when the asynchronous interface stores a first asynchronous processing result to a buffer storage library, a second thread is triggered to obtain the first asynchronous processing result in the buffer storage library according to the first call identifier, and the first asynchronous processing result comprises an asynchronous processing result of the asynchronous interface on the first call; and the second thread sends the first asynchronous processing result to the first thread to wake up the first thread to continue execution. According to the method provided by the invention, excessive consumption of system resources and system crash risks in a high-concurrency scene can be avoided, and the stability of the system is improved.

The invention provides a rural water supply special-purpose load type IC card intelligent water meter and a remote payment method. The method comprises the following steps: a user confirms a water meter account through a recharging program of an intelligent terminal and inputs a recharging amount through the recharging program, and the recharging program sends a water meter account recharging request after successful payment; the water meter account recharging request is subjected to data caching in the local server; after the water meter account recharging request is sent to the local server, the balance is updated in response to the water meter account recharging request of the user based on asynchronous processing of the message queue; the local server selects a communication network according to a network communication condition, and when a communication signal of the narrowband Internet of Things is unqualified, a 5G network is started as an alternative link; waking up the power supply of the transfer machine according to the communication instruction of the local server, and starting the transfer machine; and the transfer machine receives the payment amount sent by the local server. The system comprises a data caching module, a network communication module and a payment transfer module. According to the invention, efficient, reliable and convenient recharging experience is realized.

The invention discloses an AI-based auxiliary system for clinical teaching of cerebrovascular diseases, particularly relates to the field of disease management, and comprises a multi-modal information summarization module, a feature extraction module, a cross-modal deduction module, a dynamic incremental learning module, a virtual case generation module and a teaching evaluation feedback generation module. According to the AI-based auxiliary system for clinical teaching of cerebrovascular diseases, millisecond-level time domain alignment is carried out through a multi-modal information gathering module, a mapping relation table of image space coordinates and physiological signal timestamps is established, the problem of time sequence deviation caused by asynchronous processing of sub-channels is solved, and time-space alignment of multi-source data is achieved; feature decoupling of images and physiological signals is achieved through a feature extraction module, and relevance between cross-modal features is enhanced; dynamic information of blood flow velocity and blood pressure fluctuation of blood vessels is fused into attention scores through a cross-modal deduction module, and a causal relationship between variables is established, so that time-space consistency of clinical logic during deduction of complex cases is improved.

The invention provides a model quality evaluation optimization method and device for industrial simulation application and a server, and relates to the technical field of simulation application generalization development, and the method comprises the steps: obtaining a to-be-processed geometric model, and carrying out the format detection processing, format conversion processing and asynchronous processing of the geometric model, and determining a target input model; performing grid division processing and finite element analysis processing on the target input model, determining a finite element calculation result, performing local error estimation analysis processing on the finite element calculation result, locally refining a model grid of the target input model, determining a model key region, performing local encryption processing on the key region, and obtaining a local encryption result of the target input model; determining a model output result; and carrying out optimization processing on the model output result through a model quality evaluation optimization system with a layered architecture, and determining an optimized target model output result. The model quality can be remarkably improved.

The invention discloses a high-concurrency inventory cancel-after-verification method, and the method comprises the following steps: S1, a distributed lock stage: locking an inventory item to be modified; s2, at an optimistic lock stage, a version number is added to an inventory record, when the inventory is updated each time, the inventory number needs to be updated, and whether the version number is consistent with the version number during reading needs to be checked; and S3, a pre-deducted inventory stage: when the user submits an order or an exchange request, immediately subtracting a corresponding quantity from the pre-deducted inventory, keeping the actual inventory, and efficiently and accurately processing the inventory management problem, improving the processing speed, avoiding overselling and improving the system stability through methods of integrating a distributed lock, an optimistic lock mechanism, the pre-deducted inventory, asynchronous processing and the like. The high-concurrency inventory cancel-after-verification method provided by the invention has the effects of improving the processing capability of the system under large flow impact, ensuring the accuracy and consistency of inventory data and optimizing the user experience.

The invention provides an AI detection method and device for video stream asynchronous processing based on timestamp alignment, and relates to the technical field of AI detection, and the method comprises the steps: 1, a video stream reading thread receives a detection result fed back by an AI detection thread for a previous latest image frame, and transmits the current latest image frame and a timestamp association thereof to the AI detection thread; step 2, AI detection is carried out on the current latest image frame by an AI detection thread; step 3, during AI detection, the video stream reading thread aligns the image frame and the detection result based on the timestamp carried by the detection result; step 4, displaying the aligned image frame and the detection result by the playing thread; and step 5, when the AI detection is completed, repeating the step 1 to the step 4 until a preset stop condition is met. According to the method, the problems of relatively low rendering performance, obvious resource occupation and delay in serial synchronization processing in a traditional AI detection technology can be effectively relieved.

The invention discloses a data security isolation and sharing framework implementation method for multiple tenants. The method comprises the following steps of performing multi-tenant data isolation design; the method comprises the steps of S1, role-based access control and dynamic authority management design, S3, data sandbox and sharing mechanism design and S4, performance optimization and resource management. S5, a unified security management and monitoring platform, and relates to the technical field of computers.According to the method, multi-tenant data isolation design is adopted, and effective isolation of tenant data is achieved through combination of logic isolation and physical isolation; controlling access of the tenants to the shared data through access control and dynamic authority management of roles; through a data sandbox and a sharing mechanism, the safe operation of the shared data by the tenant is realized; the resource utilization rate is improved through a cache mechanism, load balancing, a resource recycling mechanism and asynchronous processing, and performance optimization and resource management are achieved.

The invention discloses a micro-service asynchronous exporting method. The method comprises the following steps that a user initiates an exporting request through a front end; the business service receives an export request initiated by the front end and assembles the query statement and the export template parameters to the export request according to the business; the business service forwards the processed export request to an asynchronous export service; after receiving an export request of the business service, the asynchronous export service creates an export task and returns a success state to the business service; the asynchronous export service evaluates the complexity, the dependency relationship and the execution time of the task according to the data volume and the system performance of the exported task, selects a task splitting strategy, splits the task into a plurality of sub-tasks and puts the sub-tasks into a task pool; the task pool triggers scheduling, parallel execution of subtasks is achieved through multiple threads or multiple processes, data are read in batches, and the data are written into temporary files. Through asynchronous processing, the user does not need to wait for export completion, and the user experience is improved.

The invention relates to the technical field of data acquisition, in particular to an AOP-based user behavior data acquisition method. The method comprises the following steps: firstly, identifying user behavior related code features by statically analyzing Java source codes of an application program, and constructing a feature template library; during a compiling period, according to a template library identification target method, embedding a point burying instruction by utilizing a byte code enhancement technology, and constructing a monitoring section framework; and during running, dynamically adjusting an execution strategy according to the running state of the application program by virtue of a dynamic rule engine. Meanwhile, an annular buffer area is adopted to asynchronously process user behavior data streams, cache data are managed in a sub-generation mode, and asynchronous processing is triggered when a threshold value is exceeded. The method also constructs an adaptive thread pool model based on a hierarchical thread pool architecture, distributes tasks according to data value weights, monitors the task backlog rate of a high-priority sub-pool, adjusts the sampling frequency or acquisition dimension when the task backlog rate exceeds a threshold value, forms adaptive closed-loop control, and effectively acquires and processes user behavior data.

The invention relates to the field of network control system security, in particular to an asynchronous processing method for resisting zero dynamic attacks in a non-uniform sampling control system. Comprising the following steps: comparing and analyzing an attacked system and an unattacked system, and establishing a dynamic relation model between an output signal difference value and an attack signal; based on the relation between the zero-dynamic attack signal and the system state space, a zero-dynamic attack signal design strategy which is kept invisible at the non-uniform sampling moment is given; active introduction of a time difference into a controller of the sampling control system is considered, and a dynamic relation between a corresponding attack signal and an output difference value is reconstructed; active introduction of a time difference into a controller of a sampling control system is considered, and a dynamic relation between an output signal difference value containing the time difference and an attack signal is reconstructed; the existence of the time difference is determined by constructing a system containing the time difference and combining with a zero dynamic attack definition; on the basis of the given time difference, the condition for detecting the zero dynamic attack in the non-uniform sampling control system is determined, so that the system is protected; the method is used for enhancing the detection capability and defense effect of the non-uniform sampling control system on the zero dynamic attack.

The invention provides a high-precision satellite navigation positioning method based on Android, and relates to the technical field of positioning navigation and dead reckoning, the compatibility and adaptability of a system are improved by constructing an abstraction intermediate communication interface between an application layer and an Android kernel layer; asynchronous processing is carried out through monitoring callback, so that deep dependence on a kernel is reduced; batch acquisition is realized by adopting an annular buffer area, and the acquisition frequency and the task priority are adjusted in real time according to the motion state of terminal equipment in combination with a dynamic sampling frequency and scheduling mechanism; the method comprises the following steps: constructing a multi-source data fusion model, fusing multi-source data, and synchronously aligning with timestamps to generate accurate position information; the CPU occupancy rate and the energy consumption in the fusion process are monitored in real time, and the sampling frequency and the scheduling mechanism are dynamically optimized in combination with terminal feedback and a preset threshold value, so that the positioning precision and the equipment load are balanced, and high efficiency and high reliability of the system are realized.

The invention discloses a label-based digital equipment model dynamic creation method and system, and relates to the field of complex data integration and equipment management technology application in the power industry, and the method comprises the following steps: predefining a general data label model, dynamically configuring a query model, and establishing a database; and constructing a digital equipment data entity and designing a technical process of parallel computing and asynchronous processing. Data standardization is realized, and the query efficiency is improved; through a dynamic configuration mechanism, the system can automatically generate the query model, and the flexibility is improved; through the data entity construction method, different business data can be associated through a standardized model, and the data consistency is improved; according to the method, parallel computing and cache optimization technologies are utilized, and data query and processing are accelerated; according to the method, an incremental updating mechanism is adopted, so that the real-time performance and consistency of data are ensured; according to the method, the creation efficiency, query performance and system flexibility of the equipment model can be remarkably improved, and better technical support is provided for digital transformation of the power industry.

The invention discloses an asynchronous processing method and device for business process data, and belongs to a distributed system data processing technology, and the method comprises the following steps: S101, receiving a business request message, and persistently storing the business request message in a main message queue; s102, acquiring a to-be-processed message from the main message queue, and locking an associated target data record before executing processing; s103, verifying data consistency based on the version number recorded by the target data, if the version number is matched, executing processing and updating the version number, and otherwise, discarding the message; s104, after successful processing, sending a confirmation instruction deletion message to the main message queue; through persistent storage of the main message queue, a version number verification mechanism and management of the dead message queue, the problem of data inconsistency caused by concurrent operation is effectively prevented, reliable transmission of messages is ensured, loss of service data is avoided, dead message processing efficiency is optimized, and the method has the advantages of ensuring data consistency, preventing message loss and improving system reliability.

The invention relates to the technical field of data management, in particular to a cloud-based postpartum health data management system, which comprises a data flow optimization module, a security backup management module, an index adjustment module and a query efficiency optimization module. By analyzing the real-time flow and the access mode, the system can provide higher data access speed, the key point is that the system can adjust and adapt to the quickly changed data access requirement in real time, the optimization not only improves the data access efficiency, but also optimizes the security backup of the data, and the security backup of the data is improved by selecting the most suitable backup center and backup frequency. The multi-dimensional self-adaptive index is introduced, the time of large-scale health data retrieval is shortened, the method is particularly important for health information urgently needing to be processed, the optimal data query mode is evaluated and selected, the operation cost is further reduced, the economic efficiency of a management system is enhanced, and meanwhile the data management response speed is increased through asynchronous processing.

The invention provides a multi-strategy caching and automatic data generation method and system based on a Node environment, and relates to the technical field of data processing, and the method comprises the following steps: receiving a client request, obtaining a data generation template, constructing a data dependency graph to analyze a field dependency relationship, and determining a field generation sequence by adopting a recursive generation strategy and depth-first search; data validity verification is executed, a data generation thread pool is created by adopting an adaptive load balancing mechanism based on the data dependency graph, and asynchronous processing is realized through double buffer queues, so that the data generation efficiency is improved, the data logic integrity is ensured, and the system resource consumption is reduced.

A cybersecurity data loss prevention service stops users from stealing, or exfiltrating, sensitive data. An endpoint cybersecurity agent coordinates the installation of a browser extension. The browser extension adds content scripts to a web browser that monitor for exfiltration events. The exfiltration events represent a user's browser inputs (such as cut-n-paste or drag-n-drop) that can be used to exfiltrate usernames, passwords, credit card numbers, company secrets, and any other sensitive data. When the browser extension detects any exfiltration event, the browser extension intercepts and synchronously blocks the exfiltration event from the web browser. Moreover, the browser extension sends a duplicate copy of the exfiltration event to the cybersecurity agent for evaluation. If the cybersecurity agent determines that the user's browser inputs should have been allowed, then the browser extension is instructed to trigger the duplicate copy. The web browser thus asynchronously processes the user's browser inputs, albeit slightly delayed.

The invention belongs to the technical field of computer graphic processing, and particularly relates to a computer device for chartlet loading, which comprises a central processing unit provided with a main thread for packaging a chartlet loading request into a task object and pushing the task object into a task queue of the central processing unit; the resource loading thread is used for sequentially pulling the to-be-processed task objects from the task queue in an asynchronous processing mode and reading the mapping files in the task objects through the asynchronous I / O interface; the resource decoding thread is used for decoding the mapping file and uploading texture data obtained by decoding to the graphics processor through an asynchronous interface; a plurality of threads are created in the graphics processor and at least comprise a resource processing thread used for receiving texture data uploaded by a resource decoding thread; and the rendering thread is used for carrying out texture binding on the texture data and carrying out rendering according to an instruction of the main thread after binding so as to draw and generate a corresponding map.

The invention discloses a transaction processing method and device, equipment and a storage medium, and relates to the field of distributed systems, and the method comprises the steps: in a distributed system, determining a target transaction processing period through employing an initial transaction processing period, a system load, the number of to-be-processed transactions, a network condition and a resource utilization condition of distributed nodes, which are determined based on an epoch mechanism; allocating the to-be-processed transaction to a target transaction processing cycle, and controlling the start of the target transaction processing cycle, so that the distributed node performs asynchronous processing on the to-be-processed transaction in the target transaction processing cycle, and a processed transaction is obtained; obtaining a transaction processing state of the distributed node, and determining a to-be-submitted transaction and a to-be-rolled-back transaction from the processed transactions based on the transaction processing state; and when the target transaction processing cycle is ended, submitting the to-be-submitted transaction, and redistributing the to-be-rollback transaction to the target transaction processing cycle for processing. Therefore, the transaction processing efficiency can be improved.

The invention discloses a dynamic partition-based tax form intelligent asynchronous processing and real-time query method, which comprises the following steps of: in response to multiple groups of tax form data, analyzing a characteristic field numerical value to generate a unique request identification code; carrying out numerical processing on the feature field to generate a dynamic partition sub-identifier, and aggregating to generate a dynamic partition combination identifier; matching a target data partition corresponding to the dynamic partition combination identifier through a constructed dynamic partition mapping table, and importing structured form data to form a dynamic partition database; constructing an asynchronous task key value pair, writing the asynchronous task key value pair into an asynchronous message queue, regularly scanning the asynchronous message queue, and when a to-be-processed task is detected, identifying whether the task queue contains a priority mark or not, and dynamically adjusting the execution priority of the task queue; and implementing multi-dimensional tax data query from the dynamic partition database based on real-time query conditions and the optimized task queue. The tedious and low-efficiency manual operation is avoided, and the processing efficiency of the form and the tax data is improved.

The invention provides a low-delay cross-chain interaction method and device based on asynchronous processing, and the method comprises the steps: a source chain generates a cross-chain message, screens the cross-chain message through an eICN node, and asynchronously transmits the cross-chain message to a target chain; the target link receives and stores the cross-chain message, and temporarily stores the cross-chain message as an unconfirmed state under the condition that the validity of the cross-chain message is not verified; after the source chain confirms a block corresponding to the cross-chain message, the eICN node sends verification data of the block to the destination chain; and the target chain performs validity verification on the cross-chain message in the unconfirmed state based on the verification data, and executes a corresponding cross-chain service logic or deletes an invalid cross-chain message according to a verification result. According to the method, the combination of a cross-chain interaction process and a block chain execution process is realized, and the cross-chain interaction time delay is reduced through asynchronous driving of message transmission and message confirmation.

The invention discloses a volunteer service work order intelligent processing method and system based on multistage classification and similarity matching, and the method comprises the steps: carrying out the differential cleaning of a work order text and case data, constructing a structured template, and carrying out the label mapping coding; through coarse-grained classification of the second-level classification model, triggering fine-grained division of the third-level classification model when a large class is specific and the confidence coefficient is in a preset interval; retrieving associated historical cases based on a semantic similarity algorithm; and constructing a dynamic prompt template according to matched case elements, and calling a large model API to generate a mediation suggestion. Further, a paging mechanism is adopted to asynchronously process the work order, store the result and integrate an exception handling mechanism. According to the method, the work order processing intelligence level can be improved, and the classification and matching efficiency can be optimized. According to the method, field adaptive learning, hybrid index optimization and dynamic Prompt engineering are fused, an intelligent closed loop from work order analysis to mediation suggestion generation is constructed, and the matching efficiency of community service resources and the scientificity of decision making are remarkably improved.

The invention relates to the technical field of data analysis, and discloses an AI-based cloud data intelligent analysis and service management system and method. The system comprises a cloud data priority division module, a resource demand prediction module, a cloud data scheduling scheme generation module and a service management dynamic adjustment module. The method comprises the steps of firstly collecting multi-source cloud data, performing data preprocessing, and then performing data priority division; secondly, introducing a hierarchical asynchronous processing architecture to construct a hierarchical neural network prediction model, and outputting a predicted value according to a data priority; establishing a resource scheduling objective function according to the predicted value, optimizing cloud data scheduling parameters in a resource scheduling process by using a multi-strategy fusion grey wolf algorithm, and generating a cloud data scheduling scheme; and finally, calculating a service management health degree, and establishing a health degree response mechanism closed-loop dynamic adjustment cloud data scheduling scheme. By analyzing and processing the cloud data, the purposes of intelligent analysis and service management are achieved, and the method is accurate and objective.

The invention provides a graph calculation method and device, electronic equipment and a storage medium, and relates to the technical field of data processing. The graph calculation method comprises the steps of determining a node task corresponding to a graph calculation task and a thread pool corresponding to the node task according to real-time system resources and task information of the graph calculation task, so that the thread pool obtains exchange data of other node tasks and node cache data in a cache module; under the condition that the node cache data is not obtained, searching a preset database in an asynchronous mode, and writing the searched node cache data into a cache module, so that the thread pool obtains the node cache data again based on a synchronous mode; and under the condition that the node cache data is obtained, performing multi-thread synchronous calculation according to the node cache data and the exchange data so as to determine a graph calculation result. According to the method and the device, through combination of synchronous processing, asynchronous processing and utilization of the preset database when the cache is missed, the overall efficiency during large-scale data processing is improved.