40results about How to "Reduce the number of threads" patented technology

The invention relates to a transmission method for a publish / subscribe system based on NIO and IO, comprising the following steps: a thread that uniquely receives socket connection is started to use by using NIO at a server terminal, wherein the receiving thread comprises Selector objects, thread reading groups and thread writing groups; the Selector objects capture various events from a client, and the thread reading groups and thread writing groups are used for receiving socket connection from the client and writing data to the client; each thread reading or writing thread corresponds to a socket connection respectively and packages the received socket every time to form an object to read and write data and then registers the object to the unique Selector object. When the client is shutdown, the connection is logged out from the reading and writing threads, and after logout, the reading and writing threads can bind novel connections. The invention has the advantages that thread number is reduced by using NIO at the server terminal based on the mixed use of IO and NIO of java, thereby reducing the resource consumption; and the server terminal is connected at the client by using IO so that the speed for sending send messages can be improved greatly.

A thread realization and dispatch method on data process operational platform relates to avoiding or reducing kernel thread switch, fully utilizing the processor efficiency and reducing system resources cost with the following stops: finishing design of thread controller, thread creation of thread bonding, starting operational thread activation, receiving items trigger transfer into operational state thread operation, finishing thread state switch, putting in order and operational array management thread dispatch and finishing clearance of thread data, releasing thread symbol, exit of thread of adjusting system thread numbers.

The provided method for reliable transmitting data in Ad Hoc network comprises: with syn, data-ack and fin message, building connection between the transmitting and receiving sides to make the receiving side obtain data without waiting for three time handshake; communicating between two sides by data-ack means to send message; once transmitting the last package, sending close-connection message to make the receiving side start the close procedure. This invention reduces message quantity and time delay, and improves system executive efficiency.

The invention relates to a realization method and system of an Rpc client in a performance test platform based on Netty, a server and a storage medium, According to the implementation method, a request message sent by a performance pressure measurement platform is sent to an Rpc client; Message coding is carried out through the client and then the message is sent to the server. and the server returns a response message to the client for decoding and then returns the response message to the performance pressure measurement platform, the packet sticking problem of the Tpc network is solved through fixed-length coding and decoding, the request timeout problem of the Rpc client is solved through a time wheel, the resource utilization rate is improved through an asynchronous mechanism, the thread number is reduced, and thus pressure resources are saved.

The invention discloses a control and routing method and system for requests, a storage and electronic equipment. The control and routing system for the requests at least comprises a filtering layer and a service connector layer. The control and routing method for the requests comprises the following steps of: by the filtering layer, acquiring service requests and an identity mark from a client; carrying out verification on the identity mark of the client, if verification is successful, detecting a current service request frequency of the client, and if the service request frequency is greaterthan a first threshold, limiting a service request amount of the client; detecting a current service request error rate of the client, and if the service request error rate is greater than a second threshold, blocking all requests of the client; and if the service request frequency is smaller than or equal to the first threshold and the service request error rate is smaller than or equal to the second threshold, carrying out parsing on the service requests, and if parsing is successful, routing the service requests to corresponding service interfaces in the service connector layer.

The invention discloses a server pressure testing method and device. The method comprises the steps that testing requests of at least two clients are sent to a first thread pool; Wherein the first thread pool is pre-configured with at least two threads; and processing the test requests of the at least two clients through the threads in the first thread pool so as to test the tested server. Throughthe server pressure testing method provided by the invention, the number of threads can be reduced, so that the system resource cost is reduced.

The invention provides a communication method for a client and a data terminal server, and the method comprises the steps: building a synchronous non-blocking event thread pool based on an event-driven communication mode of Netty; acquiring configuration information of the data terminal servers, and registering all the data terminal server information to the client startup class; for the data terminal server with connection failure, starting a timer, performing reconnecting until the connection is successful, and when the data terminal server sends data to the system, sequentially triggering an offline reconnection processor, an idle state processor, an idle processor, a decoding processor and an IO reading processor, so that a data stream forms a data packet in a json format; and for thedata terminal server with connection failure, carrying out reconnecting until the connection is successful. According to the method, the data communication efficiency can be improved, the number of threads started by the client in the communication process is reduced, and the operation pressure of communication data is relieved.

The embodiment of the application discloses a brushing and writing performance optimization method, which comprises the following steps: firstly, a storage system receives a write IO sent by an upperlayer module through a lower layer module; if the number of the write IO reaches a first preset value, a corresponding number of threads on the basis of the existing threads is increased to obtain a first number of threads, and the threads are used for processing the write IO; and then it is judged whether the number of the write IO matches the adjusted number of threads; if the number of write IOs matches the adjusted number of threads, the write IO is executed to achieve data writing. The present application can dynamically adjust the number of threads by judging the number of IO so that thenumber of IO and the number of threads are always kept in an ideal state, thereby improving the efficiency of data writing.

The embodiment of the invention provides a file processing method, system and device and a medium, and relates to the technical field of network application and development, and the file processing method comprises the steps: obtaining an uploaded document corresponding to a business object; storing the uploaded document to a target folder corresponding to the business object according to the declaration information corresponding to the business object; compressing the target folder to obtain a compressed file corresponding to the business object; uploading the compressed file to a cloud server for storage to obtain cloud storage address information of the compressed file; and determining data packaging address information corresponding to the business object according to the cloud storageaddress information. According to the embodiment of the invention, files are automatically classified and compressed while being uploaded, so that the number of threads for downloading the files is reduced, and the running efficiency of a Web server is improved.

The invention discloses a rule execution method and device, and relates to the technical field of computers. A specific embodiment of the method comprises the steps of receiving a service request, andobtaining a rule associated with the service request; and judging the rule type of each rule, determining the time sequence of each rule, dividing the rules which are in the same time sequence and have the same rule type into one group, and merging the rules in the same group into the same thread for execution. According to the implementation mode, for the rules capable of being executed in parallel, through reasonable combination, the number of threads can be effectively reduced, the time-consuming proportion of CPU switching threads is reduced, and therefore the execution efficiency of therule engine is improved, and CPU resources are saved.

The invention relates to a CAN bus-based medical X-ray high voltage generator control system, which comprises a console MCU module, a high voltage generator and a bulb tube connected in sequence, wherein the high voltage generator comprises a power interface module, a rotary anode driving module, a high voltage loop control module, a large-small focal filament power module, an AEC shooting control module, an IBS perspective control module, a high voltage inverter and a transformer; the console MCU module, the power interface module, the rotary anode driving module, the high voltage loop control module, the large-small focal filament power module, the AEC shooting control module and the IBS perspective control module are mutually connected via the CAN bus; and the high voltage inverter and the transformer are respectively connected with the high voltage loop control module and the bulb tube. Compared with the prior art, the system has the advantages of flexibility, reliability, low cost, good system performance and the like.

An embodiment of the present invention discloses a data acquisition method for a sensor, an event distributor, a device, a medium and an unmanned vehicle, wherein, the method comprises the following steps: a sensor-based event distributor is created in advance to listen to the data preparation state of all sensors in real time; a data reading task for the target sensor is created when the data ofthe target sensor is sniffed out is ready; a system schedule is invoked to schedule execution of that data read task. An embodiment of the present invention listens to the data readiness of the sensorthrough an event dispatcher which is a thread, when the data of the target sensor is ready, the task thread for fetching data is created for the target sensor, and the task thread is handed to the scheduler for overall scheduling, so as to realize data fetching, thereby reducing the number of threads and saving system resources as a whole.

The embodiments of the invention disclose a service processing method, device and equipment, and a storage medium. The method comprises the following steps of in a preset service period, when data input or data output of a service thread exists in a service system for the first time, acquiring the service data of the service thread at each moment; determining the QoS shared data of the service thread according to the service data at each moment, wherein the QoS shared data is the total amount of the data input and / or output by the service thread in the preset service period, or the QoS shareddata is the data flow rate of the service thread in the preset service period; and controlling the data flow of the service thread according to the QoS shared data and a preset data amount threshold.

The embodiment of the invention provides a serial number generation method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a serial number generation request initiated by a target business, wherein the serial number generation request comprises a node identification corresponding to the target business; generating a current counting serial number according to a historical counting serial number which is a counting serial number generated in response to the last serial number generation request; and generating a target serial number accordingto the current timestamp, the current counting serial number and the node identifier, wherein the target serial number is used for uniquely identifying the target service. In the scheme, the countingserial number is irrelevant to the timestamp, namely, when the time rolls back, the counting serial number is continuously increased and is not reset to the minimum value, so that the serial number generated by the method can tolerate certain time roll-back.

The invention relates to a combinable distributed control system and a realization method thereof. The system comprises a distributed control system host serving as a system core portion, and a slave1-distributed system communication module, a slave 2-distributed system digital IO module and a slave 3-distributed system analog quantity IO module and so on, which can realize each single function and is mutually independent. The host and the slaves are initialized and in a standby state. The host communicates with the each slave by identifying single ID of each slave. The single ID of each slave ensures that the slaves are independent to each other, so that it can be ensured that in the distributable control system, required system slaves and host logics can be built according to differentsystem level functions.

The invention provides a non-blocking distributed scheduled task scheduling method based on a JVM (Java Virtual Machine), which solves the problems that the number of threads is too much and nodes are in a distributed state to run threads on the basis of completely realizing original functions. In order to achieve the purpose, the technical scheme adopted by the invention is as follows: the method is applied to a JVM (Java Virtual Machine) configured with a kotlin language library and a coroutines coroutine library, and comprises a distributed plan task framework and a distributed coordination component, a universal non-blocking addition, deletion, modification and check interface is arranged in the distributed coordination component to serve as a non-blocking asynchronous callback interface of a client where the distributed coordination component is located; packaging a non-blocking asynchronous callback interface of the distributed coordination component through a coroutine, and setting a non-blocking circulating monitor to realize continuous monitoring; and finally, a local node cache is realized again based on the packaged coroutine non-blocking api. Through non-blocking transformation, the performance of task execution is remarkably improved.

The invention discloses a message pushing method and device. The method and device are used for solving the problems that in the video transmission process, when videos are frequently abnormal and an existing message pushing mechanism is adopted for sending notification messages to a front-end video player, the memory use ratio of a system is prone to being lowered, and the system crashes when the situation is serious. The method comprises the steps that in each set reporting period, when it is determined that the videos are abnormal, a notification message is generated; when each set reporting period is ended, all the determined notification messages in the reporting period are sent to the video player through an established thread. According to the message pushing method and device, due to the fact that only when each set reporting period is ended, are all the determined notification messages in the reporting period sent to the video player through the established thread, the number of established threads is reduced, and the use rate of memory resources is improved.

The invention discloses a data processing method and related equipment. The method comprises the following steps: obtaining at least two data writing requests; determining a target data writing request from the at least two data writing requests, writing data corresponding to the target data writing request into a storage medium in sequence, the data corresponding to the target data write-in request is written into the storage medium in sequence, so that the plurality of data writing requests can correspond to a thread for writing data into a storage medium; the number of threads corresponding to the storage medium is reduced, so that the technical problem that the storage medium load is too high when data are written into the storage medium through multiple threads concurrently is solved, and the technical effect of improving the read-write efficiency of the storage medium is achieved.

The invention relates to the technical field of data communication and provides a network management system equipment state obtaining method and apparatus based on a Reactor mode, for solving a series of problems of high concurrence, network environment isomerization, message return obstruction and the like of a network management system by taking the idea of an event driving mode as technical guidance. A equipment state detection flow, a query equipment basic information query flow, a sub-equipment information query flow and an equipment warning synchronization flow are corresponding to a handle container of a multiplexing reactor for event formation, and registration time ready callback is returned; and a synchronization time decomposer of the multiplexing reactor monitors an event handle set in the handle container, when a new event takes place, an event multipath decomposer dispatches corresponding events to an event processing program.

The invention provides a method and system for temperature reduction in chips. The method comprises following steps: detecting temperature and operating threads of each kernel; determining whether the kernel temperature exceeds a temperature threshold value; if the kernel temperature exceeds the temperature threshold value, reducing thread quantity of the kernel. By means of the technical scheme provided by the invention, temperature can be effectively reduced.

The invention provides an application method and system of load balancing in a multi-core chip. The method comprises the following steps that the load and running threads of each core are detected; whether the load of the core exceeds a load threshold value or not is judged; if the load of the core exceeds the load threshold value, the number of the threads of the core is decreased. According to the technical scheme, the application method and system of load balancing in the multi-core chip have the advantage of effectively reducing the load.

The invention provides a method and a system for applying current limiting in a multi-core chip. The method comprises the following steps of detecting the current and running thread of each inner core; judging whether the current of each inner core exceeds a current threshold value or not; when the current of each inner core exceeds the current threshold value, reducing the number of the threads of the corresponding inner core. By adopting the technical scheme, the method has the advantage that the current is effectively reduced.

The invention discloses an ETL process execution method based on an asynchronous mode, relates to the field of medical information system data processing, solves the technical problems that pause, recovery and stop of a thread in an existing Java language consume resources, and similar spark needs a lot of resources for operation. The method comprises the following steps: step 1, carrying out configuration analysis; step 2, initializing a task; step 3, running the task; and step 4, after the task is finished, summarizing task results and sending the task results to a task result processing module. Asynchronous communication is carried out between the Actors, and even if the Actors send messages, other things can be handled without blocking or waiting. Each Actor has a mailbox for message storage, an Actor scheduler carries out Actor event calling after retrieving that the Actor has a message, the Actor carries out event processing after receiving the message, and a result message is sent to other Actors after processing is completed.

The invention relates to a structurally-modified chemical fiber paper bobbin comprising a bobbin body. The outer peripheral surface of the top end of the bobbin body is provided with a thread catcher groove comprising a guide groove and a clamp groove. Two sides of an opening of the guide groove are provided with an upper slope and a lower slope respectively. The lower wall of the guide groove intersects the lower slope to form a lower edge. The lower wall of the guide groove and the lower slope incline, recess, and fracture in the lengthwise elongation direction of the lower edge to form a thread hanger part. Through the provision of the upper slope, the lower slope and the thread hanger part, success rate of thread catching is increased, thread guiding frequency is reduced, and efficiency is high.

The invention provides a power distribution method and system in an electronic chip. The method comprises the following steps of detecting the power and operational threads of each inner core; judging whether the power of the inner core exceeds a power threshold value or not; if the power of the inner core exceeds the power threshold value, decreasing the number of threads of the inner core. The power distribution method and device have the advantage of effectively lowering power.

The invention discloses a database connection method, comprising the steps of grouping a plurality of database embodiments, among which, each database embodiment comprises at least one database, and configuring a connection pool for sharing a database of all database embodiments; setting a specified thread of the connection pool to run operation of idle connection inspection and returning an idle connection to the client terminal based on the request and operation result of the idle connection inspection while receiving a request for application for connection with the connection pool from the client terminal. The invention further discloses distributed database connection device. The distributed database connection method and device realizes one connection pool shared by a plurality of databases, reduces thread quantity of connection pool and saves system resources.

The invention relates to a transmission method for a publish / subscribe system based on NIO and IO, comprising the following steps: a thread that uniquely receives socket connection is started to use by using NIO at a server terminal, wherein the receiving thread comprises Selector objects, thread reading groups and thread writing groups; the Selector objects capture various events from a client, and the thread reading groups and thread writing groups are used for receiving socket connection from the client and writing data to the client; each thread reading or writing thread corresponds to a socket connection respectively and packages the received socket every time to form an object to read and write data and then registers the object to the unique Selector object. When the client is shutdown, the connection is logged out from the reading and writing threads, and after logout, the reading and writing threads can bind novel connections. The invention has the advantages that thread number is reduced by using NIO at the server terminal based on the mixed use of IO and NIO of java, thereby reducing the resource consumption; and the server terminal is connected at the client by using IO so that the speed for sending send messages can be improved greatly.

The embodiment of the present invention discloses a data acquisition method, event distributor, equipment, medium and unmanned vehicle for sensors, wherein the method includes: a pre-created sensor-based event distributor, which monitors the data preparation status of all sensors in real time ; When it is detected that the data of the target sensor is ready, create a data reading task for the target sensor; call the system scheduler to schedule and execute the data reading task. The embodiment of the present invention monitors the data preparation status of the sensor through the thread of the event distributor, avoiding data scanning for each sensor corresponding to its own thread. Task threads for fetching data are handed over to the scheduler for overall scheduling to realize data fetching, thereby reducing the number of threads and saving system resources as a whole.