119 results about "Core system" patented technology

The invention discloses a multi-kernel system real-time task scheduling method, system and device based on scene differentiation and a storage medium, and the method comprises the following steps: carrying out the dependency analysis of a task set, and obtaining a task dependency directed graph between tasks; based on the task dependency digraph, determining a regulation and control energy level index corresponding to each task; converting the task dependency directed graph into a task situation graph; determining a scheduling sequence table of the tasks based on the task situation chart and the regulation and control energy level indexes; and allocating real-time tasks to the multi-kernel system based on the scheduling sequence table. The method can be widely applied to the technical field of task scheduling.

The invention discloses a multi-machine multi-card distributed training optimization system and method oriented to a credential heterogeneous environment, and relates to the technical field of distributed training. In order to solve the defects existing in distributed training in a credential and credential heterogeneous environment, the system comprises an application layer for receiving a training task; the core system layer comprises a self-adaptive task scheduling module, a hybrid communication topological optimization module and a dynamic check point and elastic recovery control module, and the scheduling module generates a performance portrait and an optimal scheme based on a training task and adjusts the scheme according to a node fault; the optimization module executes the scheme, firstly detects node topology, constructs a hierarchical path, loads and initializes an optimal communication rear end for target equipment, and synchronizes the path to the target equipment; the control module periodically stores a complete training context and realizes reliable recovery of a training process; the credential hardware abstraction layer is used for providing a uniform interface for realizing underlying resource access and data acquisition; and the hardware resource layer comprises heterogeneous computing nodes and an RoCE network switch.

The application discloses a transaction batch processing method, a server, a computer device and a storage medium, and relates to the technical field of transaction batch processing. The method comprises the following steps: receiving a transaction request, and recording the date of the transaction request according to a natural date during batch processing; if the batch processing is performed before 0 o'clock, performing a first refresh on a core system before 0 o'clock; performing a second refresh on the core system at 0 o'clock, so as to synchronize transaction data between the first refresh time and the second refresh time; recording the date of the transaction request after the second refresh according to the natural date; wherein the natural date is Greenwich Mean Time and cannot be changed. The accuracy of the date record of the transaction of the bank core system can be improved, the situation that the transaction date queried by a customer is inconsistent with the transaction date recorded by the bank system can be avoided, customer complaints can be reduced, and the experience and satisfaction of the customer are improved.

Methods and systems for providing computer implemented services are disclosed. To provide the services, information from a variety of systems may be aggregated. Prior to aggregation, the sources of the information may perform processes for enhancing the provided information. The processes may include removing information that is unlikely to benefit the computer implemented services, and identifying hidden information that is not explicitly noted but that is likely to benefit the computer implemented services. The processes may distribute the workload for preprocessing the aggregated information to reduce the likelihood of bottlenecks or other limits on throughput rate of the computer implemented services from occurring.

The invention discloses a memory access management method for a heterogeneous multi-core system oriented to a deep neural network, belonging to the storage system structure field of a computer system. The invention utilizes CPU-GPU heterogeneous multi-core system to accelerate DNN model training, and designs a memory access controller according to its memory access characteristics. In the DNN training process, the final cache shared by multiple cores is unloaded, preretrieved and released, and the fine-grained data transmission process significantly improves the utilization rate of the final cache. In addition, the memory access controller also designed a delay hiding mechanism, by overlapping the access process of a large number of intermediate data in the feature extraction layer and the calculation process, reducing the calculation performance loss caused by the miss of the final level cache and the need to wait for the memory access response of DRAM during the model calculation process, and optimizing the training efficiency.

The invention relates to a task congestion processing method and device, a control system, a vehicle, a storage medium, a program product and a chip system, and the method comprises the steps that a first processor obtains information generated when a second processor executes a task from a shared memory; the first processor judges whether the task operated by the second processor is congested or not based on the information; and the first processor outputs prompt information of congestion of the second processor under the condition that the tasks operated by the second processor are congested. The method is used for automatically detecting the running state of the multi-core system.

The invention discloses a heterogeneous multi-core system, a vehicle-mounted application operation method, computer equipment and a computer readable storage medium, the heterogeneous multi-core system comprises a plurality of processors, at least one processor is provided with a high computing power system, and at least another processor is provided with a low computing power system. A mathematics arithmetic unit service module is deployed on the high computing power system, mathematics arithmetic unit services are published through a service oriented architecture (SOA), and the mathematics arithmetic unit services comprise a basic mathematics arithmetic service and an enhanced mathematics arithmetic service; and the low computing power system subscribes to the mathematics arithmetic unit service through a service oriented architecture (SOA). Through the computing power deployment, if the application deployed in the low computing power system has a high computing requirement, a plurality of applications do not need to be specially added and deployed on the high computing power system, so that the development workload is greatly reduced; and there is no need to deploy applications and construct inter-core communication protocols in high and low computing power systems, thereby simplifying the software architecture.

The invention discloses an AI virtual employee system based on multi-modal deep learning and an implementation method. The system comprises a multi-modal fusion processing module, which is used for carrying out cross-modal fusion processing on multi-modal input data input by a user by adopting a multi-head attention mechanism of a Transform architecture; the role customization engine constructs AI employee roles according to specific post requirements of the user; the knowledge graph construction module is used for constructing structured knowledge representation of professional knowledge of the construction industry; the task understanding and planning module is used for analyzing user requirements and making an execution plan for an AI employee role according to the fused multi-modal data and the structured knowledge representation of the professional knowledge of the construction industry; the main execution controller is used for the AI employee role to coordinate each function module to complete a specific task according to the execution plan; the enterprise system integration interface is used for carrying out data interaction and function calling with an enterprise core system; and the result output module is used for outputting an execution result of the execution plan.

This invention provides a parallel replay method and system, relating to the field of Internet technology, comprising: acquiring all gateway log files and synchronizing them to a replay system database; parsing the gateway log files according to a specified date using multiple threads to obtain original transaction information and registering it to the replay system database; segmenting all transactions corresponding to the gateway log files by time and replaying them in chronological order; acquiring the original transaction data and sending request messages to the new core system through the replay system to obtain new response messages and response results; configuring a transaction parameter table, assembling comparison expressions based on the transaction parameter table, and comparing the response messages returned by the old and new core systems; configuring and maintaining statistical SQL, executing the statistical SQL in both the old and new core systems, and comparing the statistical results. This invention enables the system to accommodate two core business models through technical means and also allows for the verification of the new core system's functionality through parallel replay, reducing the workload of manual testing.

The application discloses a method and system for real-time synchronization and efficient query of insurance policies. The method comprises the following steps: real-time monitoring of changes in policy-related tables in the core database through the CDC capability built-in DTS to generate an event stream containing operation semantics; sending the event stream to Kafka, and then reading and writing the event stream into a MySQL queue table by a Kafka consumer; performing timed deduplication and merging on the events in the MySQL queue table, giving priority to the deletion operation, and synchronizing the merged policy data to an Elasticsearch index; and providing multi-condition combination query and multi-dimensional aggregation statistical services based on the Elasticsearch index. The technical solution realizes real-time synchronization of policy data without affecting the stability of the core system, and provides high-performance multi-condition query and multi-dimensional aggregation statistical capabilities to meet the requirements of the front-end sales tools for query performance and response speed.

The present disclosure discloses a method, device and related product for performing an operation task on a multi-core system. The device can be included in a combined processing device as a processing device, and the combined processing device can further include an interface device and a computing device. The computing device interacts with the processing device to jointly complete a user-specified computing operation. The combined processing device can further include a storage device connected with the computing device and the processing device respectively, for storing data of the computing device and the processing device. The scheme of the present disclosure provides a method for performing an operation task on a multi-core system, which can effectively reduce data access and improve processing efficiency.

Provided in the present disclosure is a method for storing a coherence directory. A multi-core system comprises N processing cores, N being an integer greater than 1. The method comprises: acquiring access information of a plurality of data blocks, wherein the access information comprises address information of each data block and accessor information of the data block; and on the basis of the access information of the plurality of data blocks, generating one storage entry of a coherence directory, wherein the storage entry comprises: a plurality of address fields, which correspond to the plurality of data blocks on a one-to-one basis and are used for carrying the address information of the plurality of data blocks, and an accessor information field, which is used for representing accessor information of the plurality of data blocks. Further provided in the present disclosure are a multi-core system, an electronic device, a computer-readable medium and a computer program product.

This invention discloses a log collection method, apparatus, device, and medium for a multi-core system. The method is executed by the master core and multiple slave cores in a multi-core system, including: writing log strings into a buffer in the shared memory of the slave cores and generating log descriptors; writing the log descriptors into a lock-free circular queue; monitoring the state count value of the lock-free circular queue in real time by the master core, and obtaining a target log descriptor when the lock-free circular queue is determined to be non-empty; determining a target buffer in the shared memory of the slave cores and transferring the log strings in the target buffer to the master core memory; and determining that the log upload conditions are met based on a timer or master core memory usage, determining at least one target master core memory block in the master core memory and uploading the log strings in the target master core memory block. This technical solution can improve log writing efficiency and log upload real-time performance in high-concurrency scenarios of multi-core systems.

The invention discloses a multi-core system and a debugging management device and method of the multi-core system, and is applied to the technical field of circuits, and the multi-core system comprises a scheduler which is connected with each processor and is used for carrying out enabling control on a tracking debugging function of each processor based on trigger information and sending received tracking debugging information to a filter; based on the trigger information, selecting various filtering rules currently started by the filter; and the filter is connected with the scheduler and used for filtering the received tracking debugging information based on the currently started filtering rules so as to discard the tracking debugging information which does not pass through the filtering rules in the tracking debugging information and output the tracking debugging information which passes through the filtering rules in the tracking debugging information. By means of the scheme, debugging management of the multi-core system can be effectively achieved, loss of key trace information is avoided, flexibility is high, and rapid transmission and processing of the trace information can be achieved.

Embodiments of the present application provide a multi-core system display method, a multi-core system, an electronic device and a storage medium. In some embodiments, the multi-core system display method can include the following steps: in response to the first core system being started, drawing first image data to a first sub-layer of a first display layer, and drawing mask data to a second sub-layer of the first display layer; in response to establishing a heartbeat connection with the second core system, determining whether the video transmission chip receives second image data output by the second core system; in response to the video transmission chip receiving the second image data, drawing the second image data forwarded by the video transmission chip to a second display layer, and removing the mask data in the second sub-layer; and outputting display data according to the first display layer and the second display layer.

The invention discloses a shared debugging terminal implementation method and a shared debugging terminal implementation system for an asymmetric multi-processing system. The shared debugging terminal implementation method and the shared debugging terminal implementation system have the following beneficial effects: (1) a plurality of cores share a single physical serial port through a shared memory area and a central routing agent; (2) each sub-region in the shared memory region comprises a core identifier region, so that a core identifier can be automatically added to each piece of debugging information to ensure that the source of the debugging information is clear, and the debugging efficiency is improved; (3) the method does not depend on a complex network protocol stack, so that the method can work reliably even at the initial stage of system starting or when the system crashes, and the debugging reliability is improved; (4) the resource occupation is less, and the debugging requirement of the multi-core system can be met by only needing one physical serial port and other types of interfaces; and (5) the method also has a failover mechanism, and when the main routing agent fails, the main routing agent can be automatically switched to the standby agent, namely, the main routing agent is automatically switched to the new central routing agent, so that the normal work of a debugging function is better ensured.

The application discloses a heterogeneous multi-core system and a running method thereof, an electronic device and a medium, and relates to the technical field of heterogeneous multi-core systems.The method comprises the following steps: packaging real-time core programs and non-real-time core programs into a unified program package; when the heterogeneous multi-core system starts, separating the real-time core programs and the non-real-time core programs according to the unified program package; performing version detection on the real-time core and the non-real-time core; when the version information of the non-real-time core is inconsistent with the version information of the real-time core, writing version update information into a shared memory area through the non-real-time core; performing version update on the real-time core according to the version update information; and when the version information of the non-real-time core is consistent with the version information of the real-time core, loading and running the real-time core programs and the non-real-time core programs respectively.The method can package the real-time core programs and the non-real-time core programs into a unified program package, and facilitate simultaneous management of the real-time core programs and the non-real-time core programs.

The invention belongs to the field of integrated circuit design and computer networks, and discloses a multi-path polling arbitration method and a related device. Input signals of all request source channels are collected in real time, and when any request state signal is detected, an arbitration working state is entered; and locking the request states of all the current request source channels to generate arbitration input information, positioning the latest request source channel based on the information in combination with the current polling pointer, generating an arbitration result, executing access operation, exiting the arbitration state, and updating the polling pointer to wait for next request triggering. By adopting the method, the scheduling flexibility is remarkably improved to adapt to variable loads, the circuit structure is simplified to reduce the implementation complexity, the adaptability to different system bus competition problems is enhanced to relieve conflicts, the problem of shared resource access conflicts in a network-on-chip and a multi-core system is effectively solved, and the efficient access requirement is met.

The application provides a concurrent memory database splitting method and relates to the technical field of databases. The ART index is improved to obtain an IART index, and an IARTC algorithm is proposed based on the IART index. The IARTC algorithm uses the IART index as a splitting index to drive database splitting. Meanwhile, a new database splitting concurrent algorithm PIARTC is obtained based on the IARTC algorithm. The PIARTC algorithm controls locking and unlocking of nodes on the IART index through an optimistic lock coupling concurrent strategy to realize concurrent splitting and index updating, so that small locking granularity is maintained in the concurrent execution process. Compared with existing algorithms, the two adaptive indexes proposed by the application have good query performance in single-core and multi-core systems in the case of facing dynamic environments and large-scale data set queries, and still maintain the lightweight characteristics of the original database splitting algorithm.

The invention discloses a many-core system performance verification method based on a portable excitation standard, and belongs to the technical field of integrated circuit design verification, the implementation of the method comprises the following steps: designing PSS actions including atomic actions, composite actions and application-level actions, and realizing precise modeling of hardware resource and data dependence; the method comprises the following steps: optimizing the use of an L1 memory by adopting a double-buffer technology, dividing the L1 memory into independent buffer regions which comprise at least two input buffer regions and two output buffer regions, and realizing data circulation parallelization; quantizing component execution time based on a cost function, and providing a data basis for scheduling optimization; the action priority is determined based on a cost function, and scheduling is optimized for a single-core system and a multi-core system; and cross-stage verification is realized based on the portability of the PSS. According to the method, accurate positioning and optimization of the performance bottleneck of the many-core system are realized, rapid iteration of the AI algorithm and hardware can be adapted without reconstructing a verification framework, and the verification efficiency is effectively improved.

The present invention belongs to the technical field of applications of heterogeneous multi-core systems, in particular relates to a task scheduling method capable of realizing energy consumption sensing under reliability constraints. The method comprises: constructing a heterogeneous multi-core system model, comprising a task model, an energy consumption model and a reliability model; on the basis of the heterogeneous multi-core system model, determining the execution priority of a task; performing first-stage task duplication on the task, ensuring that the reliability value of each subtask is higher than a system reliability requirement, and obtaining the initial number of copies of the task; if an actual reliability value of a system is less than the system reliability requirement, performing second-stage task duplication on the task, so as to obtain the final number of copies of the task; and on the basis of the execution priority of the task, assigning all the copies of the task to corresponding processor cores, and then on the basis of DPM technique use conditions, switching the processor cores from an idle mode to a sleep mode, so as to reduce power consumption. The present invention realizes the joint optimization scheduling for a heterogeneous multi-core system with respect to reliability and low power consumption, and has wide application prospects in heterogeneous multi-core systems.

The invention relates to a software product dynamic configuration and instantiation method and device based on version snapshots. The method comprises the steps of integrating a configuration source and a solidified version, establishing an original hit strategy set of a main body inheritance chain, a resource tree, an environment snapshot and an operation set dynamic tetrad, packaging an effective permission set into a configuration list through collection along the inheritance chain, conflict processing and filtering strategies, and then analyzing the configuration list to complete container layout and rendering. The device comprises a configuration source integration module, an original hit strategy set generation module, a strategy collection module, a conflict processing module, a filtering module, a configuration list analysis module and a container layout and rendering module. According to the method, one-time development and infinite configuration are achieved on the architecture level, configuration is kept in a core system code library, independent system examples in functions, UI, permission, brands and the like are efficiently and flexibly generated, and the life cycle of cost-saving control software in the current market changeable environment is met.

The invention discloses an AI data handling optimization method based on a multi-core system. The data handling optimization method comprises the steps that S1, initialization is carried out; s2, asynchronous request; s3, asynchronous processing; and S4, parallel execution. The processor A for executing the AI operation is always in the user state and does not need to fall into the kernel state, so that the system calling overhead is avoided; in the process that the processor A executes the AI operation, the processor B schedules DMA hardware to complete transmission processing of next data required by the AI reasoning operation, the DMA hardware and the AI reasoning operation are completely parallel, and when DMA transmission time is shorter than AI reasoning operation time, time overhead of DMA data transmission is completely masked; based on the software and hardware collaboration framework, the utilization rate of the computing part of the processor A can be close to 100%.

The invention relates to the technical field of computer communication and equipment monitoring, in particular to an equipment interface protocol interaction method based on a plug-in type. Through a three-layer architecture of an interface layer, a protocol adaptation layer and a core service layer, and in combination with a plug-in management module and a global exception handling module, a unified method, parameters and a return value format of a standardized interactive interface are defined; various protocol analysis and framing logics such as Socket, SNMP, Modbus and the like are packaged in a plug-in form, and dynamic loading, multi-version coexistence and hot deployment are realized through a plug-in management center; designing a standardized frame model and a general coding and decoding logic, and unifying frame processing rules of different protocols; the global exception handling module realizes unsupported function default response and exception degradation strategies, and the configuration center is combined to realize hot update of parameters such as protocol mapping and plug-in paths. According to the method, the protocol logic and the core system are thoroughly decoupled, the development efficiency is improved, and the time for adapting to a complex equipment environment is shortened.

The embodiment of the invention provides a synchronization control method, system, device and equipment. The method comprises the following steps: firstly, acquiring phase state information and a frequency measurement value of a target nuclear system; determining synchronization deviation information of the target nuclear system according to the phase state information and the frequency measurement value; then, synchronous control information is generated according to the synchronous deviation information; and finally, sending the synchronization control information to the target core system, so that the target core system eliminates the synchronization deviation according to the synchronization control information. Through the method, efficient and stable synchronization control can be kept in a complex environment, and the universality, reliability and synchronization precision of the system are remarkably improved.

The invention provides an intelligent operation and maintenance method and device, electronic equipment and a computer program product, belongs to the field of software operation and maintenance, and aims to solve the problems that an existing method lacks data fusion and automatic decision making, and high requirements of a core system for stability and efficiency are difficult to meet. The method comprises the following steps: in response to a notification sent to an agent by an alarm server, triggering the agent to enter a root cause analysis mode; in the root cause analysis mode, the agent collects multi-dimensional operation and maintenance data related to a fault event from a plurality of servers based on a standardized tool calling protocol; and analyzing and reasoning the multi-dimensional operation and maintenance data through the intelligent agent, and determining a fault root cause.

The invention provides a cache optimization method and device for an RISC-V multi-core processor. The cache optimization method comprises the following steps: identifying a shared variable generated by a transmission side RISC-V processor and read by a receiving side RISC-V processor in a software code; the receiving side RISC-V processor allocates a cache reserved space corresponding to the special address identifier in a local cache according to the configuration information; constructing an address mapping table according to the configuration information and the special address identifier; querying an address mapping table according to the target address; and if the target address is queried to be the special address identifier, transmitting the data from the transmission side register file to the cache reserved space corresponding to the receiving side through a direct cache link. According to the invention, a special direct cache data transmission link is constructed, and a producer-consumer model is adopted on the link to replace a traditional global cache consistency protocol, so that the problem of consistency traffic overload in a multi-core system is greatly relieved, and the overall efficiency of the system is improved.

This enables users without specific technical knowledge to easily perform tasks for business support using APIs. [Solution] An API processing server 30 connected via a network to a user terminal 10, a core system 40 having multiple APIs configured to provide predetermined business support services, and comprising: a message sending and receiving function 31a that exchanges messages with the generation AI 20 and the core system 40 respectively; an API selection support function 33a that sends an API information message including API list information 32a and selection support information 32b to the generation AI 20 in response to an information inquiry from the generation AI 20; and an API execution function 34a that, upon receiving a processing request message from the generation AI 20 including a request to execute a business for a specific API, sends an API execution message to the core system 40 in accordance with the specifications of the specific API.

Embodiments of the present application disclose a resource allocation method and device for an operator graph, a computer device and a storage medium. The method comprises: in a set of operator graphs to be run on a many-core system, obtaining operator graphs with running demand information; dividing a running period into multiple running time periods, and allocating running time periods to each of the operator graphs with running demand information according to the running demand information of each of the operator graphs; and determining target core resources for running each of the operator graphs in each running time period according to a resource allocation type and the running demand information of each of the operator graphs. The running data of the operator graphs with running demand information in the allocated running time periods meets the running demand information in the running period. Embodiments of the present application can reasonably allocate core resources of a many-core system.