149results about How to "Reduce the number of calls" patented technology

The invention discloses a thread for high-performance computer NUMA perception and a memory resource optimizing method and system. The system comprises a runtime environment detection module used for detecting hardware resources and the number of parallel processes of a calculation node, a calculation resource distribution and management module used for distributing calculation resources for parallel processes and building the mapping between the parallel processes and the thread and a processor core and physical memory, a parallel programming interface, and a thread binding module which is used for providing the parallel programming interface, obtaining a binding position mask of the thread according to mapping relations and binding the executing thread to a corresponding CPU core. The invention further discloses a multi-thread memory manager for NUMA perception and a multi-thread memory management method of the multi-thread memory manager. The manager comprises a DSM memory management module and an SMP module memory pool which manage SMP modules which the MPI processes belong to and memory distributing and releasing in the single SMP module respectively, the system calling frequency of the memory operation can be reduced, the memory management performance is improved, remote site memory access behaviors of application programs are reduced, and the performance of the application programs is improved.

The invention discloses a carrier channel selection method for MF-TDMA satellite system. there are two conditions for channel selection and allocation in the method by judging if terminal A is first time for starting a time slot resource application. Condition 1: if terminal A is first time for starting time slot resource application and has not prearrange carrier channel, then finding the smallest carrier channel for allocation by finding idle carrier channel and computing present carrier channel load function; condition 2: if terminal A is not first time for starting time slot resource application and a foregoing connection is allocated to predetermined carrier channel, then predetermined carrier channel is target carrier channel for allocation, or all connection of terminal A is tuned to other usable carrier channel for allocation. The inventive method makes user carrier channel tunable, solves time slot resource waste problem arising from uneven load of carrier channel for multi-user multi-connection and promotes time slot resource utilization.

A positioning method and device, a server and a computer storage medium. A positioning sharing instruction is sent to terminals in a preset area, a Bluetooth coverage list is obtained according to Bluetooth search lists fed back by the terminals in the preset area in response to the positioning sharing instruction, and then at least one terminal is selected from terminals comprised in the Bluetooth coverage list to serve as a positioning shared terminal, so that other terminals in the Bluetooth coverage list share position information of the selected positioning shared terminal.

The invention discloses a method and system for a multi-interface switching test of a smart television. The method comprises the steps that a test script is obtained; the test script is analyzed so that a control order can be generated; the current control order is sent to the smart television; relevant information of displayed content on an interface of the smart television is obtained after the control order is executed; whether the obtained relevant information of the displayed content accords with a display rule of a target interface corresponding to the current control order or not is judged, and if yes, it is judged that the interface is successfully switched. According to the method and system for the multi-interface switching test of the smart television, the interface switch can be automatically tested and the testing efficiency is improved.

The invention discloses a JavaScript code optimization method based on a front-end byte code technology. According to the method, a conversion tool chain from JavaScript to WebAssembly is provided, acode segment merging strategy is designed, and then a JavaScript performance optimization scheme is provided; and variable type information of each execution path of the program is obtained through dynamic symbolic execution, and a code statement according with a conversion condition is obtained through a screening rule, a C code is generated from a JavaScript code by designing a JavaScript 2C conversion rule, and a WebAssembly code segment is generated through Exscript compiling. And finally, the performance and function consistency before and after optimization are compared through a unit test idea, and the optimization of the code segment is given up if the performance is reduced or the functions are inconsistent. On the other hand, a code segment merging strategy is designed, on the basis of the data dependence relation between the code segments, the code segment merging strategy is provided as the basis of user code reconstruction, the purpose of reducing data interaction performance losses is achieved, and the execution efficiency of the program is further improved.

The invention discloses an air ticket inquiring cache method which includes the steps: inquiry statistics; query cache; cache checking; cache updating; active caching. According to the method, characteristics of air ticket business are considered based on active caching and passive caching, cache updating and adjustment of cache control parameters are combined, cache is timely updated, cache data accuracy is high, cache hit ratio is high, calling frequencies of external interfaces are effectively reduced, cost is controlled, and the method is different from a past method that control parameters and statistic data are adjusted by taking experience values as standards. According to the method, dynamic adjustment is achieved according to actual states, statistic is performed in real time, and timely and effective feedback is achieved.

The invention relates to a SQL resisting injection technology using a regular expression, namely a mechanism which uses InforGuard to carry out rule verification for SQL sentences of an operation network databank. Firstly, processes are initialized and then are validated, and a validating process adopts the regular expression mechanism. The invention solves the defects of simple function and poor expansibility of the current validation mechanism and has good expansibility, stable and high-efficiency validation algorithm and strong compatibility.

The present invention discloses a network data packet sending method and device. The method comprises: a Linux operating system kernel space receives a large network data packet sent by a user process, wherein the size of the large network data packet is larger than or equal to the size of network data packet capable of being sent by the kernel space; the kernel space splits the large network data packet into a plurality of small network data packets, wherein the size of each small network data packet is smaller than or equal to the size of network data packet capable of being sent by the kernel space; the kernel space uses a kernel space timer to send the plurality of small network data packets. The present invention reduces system CPU energy consumption and improves packet-sending performance as a whole.

The invention discloses an application screen capture method and device for a mobile terminal, and belongs to the field of mobile terminals. The method comprises the steps of receiving a switching order which is used for switching a designated application of the mobile terminal to the background for operation, capturing a designated area of the application screen of the designated application to obtain a captured picture of the designated area, and storing the captured picture of the designated area, and switching the designated application to the background for operation. According to the method and device, before application screen Activity is turned off and when an Activity window can still be seen, the application screen is subjected to screen capture, the last real screen of the Activity is obtained, the problem of high memory usage due to the fact that screen capture is carried out after the application screen is turned off is avoided, screen capture efficiency is improved, and memory usage and calling frequency are lowered.

The invention belongs to the technical field of smart phones, and particularly relates to a method for reducing standby power consumption of a mobile phone. The method comprises the following steps: recognizing redundant timers: automatically recognizing timers, which are required to be set strictly according to given time, the time of which cannot be changed, which are redundant and too frequent, and the frequencies of which are required to be reduced by adjusting the time of the timers; adjusting the time of the timers, wherein the adjusted timers can ensure normal operation of an application program, without errors; adjusting the interval time of the redundant timers in the standby state of the mobile phone, wherein the frequency of the redundant timers for executing tasks is reduced to achieve the purpose of saving the standby power consumption. By adopting the method, the functions of the mobile phone can be ensured to be completely consistent with the functions of an ordinary mobile phone, in addition, the standby power consumption of the mobile phone can be reduced by 30 to 80%, the usage time of the mobile phone charged for one time is greatly prolonged, and very high social and market values are achieved.

The embodiment of the invention discloses a protocol stack data reading method and device based on transmission control protocol / Internet Protocol (TCP / IP). The method includes: acquiring identification which designates to read a plurality of User Datagram Protocol (UDP) packets in the TCP / IP protocol stack; acquiring length which designates the total length of the UDP packets to be read; acquiring data, reading the corresponding number of UDP packets form the protocol stack based on the designated total length of the UDP packets; copying these UDP packets to a user state. a plurality of UDP packets are destined to read by identification, the number of UDP packets can be read is destined by the length, so that a plurality of UDP packets can be read at one time; the system invoking number is diminished and the occupation time of system invoking is reduced while reading the same number of UDP packets, thereby enhancing the network input and output capacity.

The invention discloses an OPU instruction set defining method for CNN acceleration, and relates to the field of instructions of CNN acceleration processors, and the method comprises the steps of defining a conditional instruction, defining an unconditional instruction and setting the instruction granularity, wherein the unconditional instruction provides configuration parameters for the conditional instruction, the conditional instruction sets a trigger condition, the trigger condition is hard-written in hardware, the conditional instruction sets a trigger condition register corresponding tothe conditional instruction, the conditional instruction is executed after the trigger condition is met, the unconditional instruction is directly executed after being read, and the content of the parameter register is replaced. According to the CNN network and acceleration requirements, the calculation modes of the parallel input and output channels are selected, and the instruction granularity is set. The instruction set provided by the invention avoids the problem that the order of the instructions cannot be predicted due to large uncertainty of the operation period. The instruction set andthe corresponding processor OPU can be realized by an FPGA or an ASIC. The OPU can accelerate different target CNN networks, and hardware reconstruction is avoided.

The invention discloses an S-type fluid path layout optimum design method specific to a liquid cooling radiator, which includes steps of confirming parameter of an integrated circuit board; partitioning the integrated circuit board and giving a valuing scale of a heat exchange coefficient in every zone; building a calculation model of the heat exchange coefficient and solving it; confirming a cooling plate parameter of the liquid cooling radiator; confirming basic layout and corresponding fluid path length of the S-type fluid path; structuring a geological function of the S-type fluid path; building the S-type fluid path layout optimum model and solving it; tiny adjusting the fluid path geological function and then outputting the optimum design scheme. According to the parameter of the integrated circuit board to be radiated, the method calculates the optimal heat exchange coefficient required by every zone of the radiating face of the circuit board and designs the corresponding S-typefluid path layout, thus the optimal radiating effect under the equal fluid path length is reached. Compared with the prior art, the method is featured by high efficiency and easy use.

The invention relates to a hidden caption displaying method and system for a digital TV. The method is that when the display model is a Paint on mode or a Roll up mode, and N characters to be displayed are received and copied to the corresponding position coordinates on a caption layer, a system refreshing screen interface is called to display the caption layer and a menu layer through an OSD by a mixing manner, rather than calling the system refreshing screen interface while receiving each character to be displayed, so that the times of calling the system refreshing screen interface can be greatly decreased, the phenomena of continuous flash and blocking due to frequent screen refreshing can be avoided, a user can watch conveniently, and as a result, the user experience can be increased.

The invention embodiment discloses a processing method and an apparatus based on virtual machine of the charging special field, relating to a computer technology field, solving the problem in the prior art with respect to special treatment in the charging field, if a general script interpreted language and a general virtual machine are adopted, complex processing needs to be carried out. The invention provides a technical scheme comprising, obtaining charging rule codes, compiling the charging rule codes into byte code files, [wherein] the byte code files include a large granule instruction which is an instruction comprising a plurality of operation information; interpretively executing the byte code files through the virtual machine based on the charging special field, calling a logic code segment corresponding to the large granule instruction. The processing method and the apparatus based on the charging special field provided by the invention embodiment can be applied to the charging special field.

Embodiments of the present invention provide a method for processing a system function call request and a browser, and relates to the field of internet applications. Resource occupancy of the browser to a system can be reduced. The specific scheme comprises: receiving the system function call request sent by a Web application; determining call information, wherein the call information comprises at least one piece of system information required for the browser to process the system function call request; when it is determined that cached information comprises the call information, obtaining the call information from the cached information, wherein the cached information comprises at least one piece of system information obtained and cached by performing the call of a system function before the system function call request is received by the browser; and according to the system function call request and the call information, returning a calling result to the Web application. The method provided by the present invention is used for development and design of the browser.

The invention relates to a non-recursive SC decoding likelihood ratio determining method and device, and aims at solving the defects that the number of times of calling a recursive function by the existing recursive SC decoding algorithm is too much, and the space complexity is relatively high. The non-recursive SC decoding likelihood ratio determining method provided by the invention comprises the following steps of sequentially computing f2, f4 to fN when (formula) is computed; and sequentially computing (formula) when (formula) is computed, wherein fj represents that a likelihood ratio of the jth layer is computed by using a formula f, and gj represents that the likelihood ratio of the jth layer is computed by using a formula g. The non-recursive SC decoding likelihood ratio determiningdevice comprises a first computing module and a second computing module, wherein the first computing module is used for sequentially computing f2, f4 to fN when i is equal to 1, namely (formula) needs to be computed; the second computing module is used for sequentially computing (formula) when i is not less than 2 and not greater than N, namely (formula) needs to be computed. According to the method and the device, an SC decoding algorithm is realized by adopting a non-recursive mode, the number of times of calling the recursive function is reduced to zero, and the coding rate is improved. Inaddition, by adopting a time division multiplexing storage mode for the likelihood ratio, the space complexity of the algorithm is reduced to 0(N).

The invention relates to a government affair data synchronization method, device and system, computer equipment and a storage medium. The method comprises the steps of obtaining a synchronization task; wherein the synchronization task is created after it is detected that government affair data updating operation of the source platform is completed; determining target government affair data and a target synchronization mode according to the synchronization task; wherein the target government affair data is changed government affair data on the source platform corresponding to the updating operation; and synchronizing the target government affair data to the target platform according to the target synchronization mode. By adopting the method, the data volume to be synchronized can be reduced; the synchronization task is not easy to lose in the synchronization process, the creation of the synchronization task does not need the comparison operation of total data, the calling frequency of the synchronization interface is reduced, millisecond-level government affair data synchronization can be realized, the real-time accuracy of data synchronization is ensured, and the performance of a synchronization system and the use experience degree of a user are effectively improved.

The invention discloses a cross-engine calling method and a cross-engine calling device. The method comprises the following steps: a first engine acquires a first data set, wherein the first engine adopts a compiled programming language, and the first data set comprises serialization results of one or more type callback events; the first engine transmits the first data set into a second engine, wherein the second engine adopts a script programming language; the first engine receives a second data set from the second engine, wherein the second data set comprises serialization results of operation instructions for game objects. According to the cross-engine calling method and the cross-engine calling device disclosed by the invention, the technical problems of lower cross-language calling speed and lower execution efficiency of a technical scheme, provided by the related art, of embedding a Lua script system into a Unity engine are solved.

Embodiments of the invention provide a business processing method and device. The method comprises the following steps: when a business request is received, creating thread cache for the business request; and, under the business request, when processing aiming at a service request is initiated, storing a processing result of the service request in the thread cache so that the processing result of the service request is called from the thread cache when the processing for the service request is initiated again. The processing result is extracted from the thread cache, an external service system does not need to be called again, the call frequency of the external service system is greatly reduced, and the system resource overhead is lowered; and the processing result does not rely on any other storage and neither occupies a memory space of an application, the method is a lightweight thread cache implementation scheme, business interfaces are prevented from being merged, a single responsibility of each interface is kept, and system expansibility and maintainability are guaranteed.

The invention discloses a risk management method, a risk management device and a risk management system. The method comprises the steps of for aiming at a risk management event, acquiring a pre-calculation index which corresponds with the risk management event; acquiring index data of the pre-calculation index from a data platform; respectively supplying the risk management event and the index data of the pre-calculation index to each risk management platform, and performing risk analysis on the risk management event by each risk management platform according to the required index data of the index, wherein the required index data of the index are acquired from the index data of the pre-calculation index. The risk management method, the risk management device and the risk management system have advantages of reducing number of calling times, saving system resource and improving system performance.

The invention provides a caching method and device for a search engine. The method includes the steps of receiving queries and elements obtained according to the queries, generating key values of the queries on the basis of the queries, finding a list corresponding to the key values in a cache, marking and storing the elements and updating the cache. According to the technical scheme, values and state information of all the elements are cached, so that the number of calling times of the MLR is reduced, and therefore the calculation amount of the engine is reduced and time efficiency of the engine is improved. Compositions of each query are refined and the queries are normalized so that the hit rate of the cache can be increased.

The invention provides an automatic scheduling method and system for shipping space information, a storage medium and electronic equipment. The scheduling method comprises the following steps that: classifying all air lines into popular air lines and unpopular air lines; carrying out statistics on the current-day full cabin rate of each popular air line; according to the average full cabin rate of each popular air line in previous several days, judging whether the current-day full cabin rate is abnormal or not; traversing each selected popular air line, and judging whether the current-day full cabin rate falls into a full cabin rate standard reaching range or not; carrying out up-regulation on the instruction calling frequency of the popular air line of which the current-day full cabin rate is greater than the maximum value of the full cabin rate standard reaching range according to a first ratio in a current day; and carrying out down-regulation on the instruction calling frequency of the popular air line of which the current-day full cabin rate is smaller than the minimum value of the full cabin rate standard reaching range according to a first ratio in the current day. According to air line characteristics and a time cycle, instruction calling frequency is automatically regulated, the instruction calling frequency of the unpopular air line can be properly reduced while the full cabin rate of the popular air line is reduced so as to realize a purpose that operation cost is saved while user booking experience is improved while an integral full cabin rate is lowered.

The invention relates to a distribution route optimization method based on vehicle and unmanned aerial vehicle cooperation. The method comprises the following steps: S1, establishing a mixed integer programming model of vehicle and unmanned aerial vehicle cooperation blood distribution; s2, performing equivalent change on the mixed integer programming by adopting a logarithm-based method, and remarkably reducing the number of binary variables by increasing auxiliary linear constraints; and S3, dividing the mixed integer programming model into a Benders main problem and a Benders sub-problem by adopting Benders re-representation, solving the Benders sub-problem by adopting a branch pricing and cutting algorithm, and optimizing the selection of the cluster center hospitals, the distribution of the non-cluster center hospitals and the vehicle driving route to obtain an optimized cluster center hospital selection and distribution strategy.

The invention relates to a non-recursive SC decoding partial sum determination method and device. The method is proposed for solving the defect that the existing recursive SC decoding algorithm is frequent in recursive function calling and high in space complexity. The non-recursive SC decoding partial sum determination method comprises the following steps: outputting partial sum of a M0 node corresponding to the formula as shown in description, wherein the M0 node represents a leaf node of a sub-tree taking the formula as shown in description as a root node, and ki is the maximum recursive depth of the likelihood ratio computation when estimating the formula as shown in description; executing the following operations for 2k-1 times so as to compute the partial sum of the node Mk: formulaas shown in description, wherein i is not less than 0 and less than 2k-1; upwardly computing along the rightmost side until the node as shown in description finally outputs the value of the formula asshown in description. The SC decoding algorithm is realized by using the non-recursive way, the calling frequency of the recursive function is reduced to zero, and the decoding rate is improved. Furthermore, the algorithm space complexity is reduced to O(N) by adopting the time-division multiplexing storage way for the likelihood ratio and the partial sum.

The invention relates to a complex high-dimension system optimization method based on quick mode pursuing sampling, and belongs to the technical field of engineering optimization design. The method comprises the following steps that: firstly, in a design space, constructing the agent model of an engineering system complex analysis model; then, dividing sequence sample points iterated each time into global search sample points and local search sample points; and according to the approximation accuracy and approximation optimal solution convergence situation of the agent model, continuously updating a major design space until the optimal solution of an optimal design problem is obtained. By use of the method which is put forward by the invention, the defects of a standard mode pursuing sampling method that time is consumed for calculation when the complex high-dimension system is optimized are overcome, optimization efficiency is improved, calculation cost is lowered, and therefore, themethod is favorable for shortening an optimization design period.

The invention relates to a human body target tracking method based on deep learning and correlation filtering. The method comprises the steps that (1) a current frame of a to-be-tracked video is read,and a human body position is detected through a human body detector; (2) the next frame of the to-be-tracked video is read, whether the end of the video is reached is judged, if yes, the process is ended, or otherwise the step (3) is executed; (3) the human body position obtained through detection in the previous step is utilized to initialize a correlation filtering tracker to perform human bodytracking, and tracking lasts for set time; and (4) a lightweight human body discriminator is utilized to determine whether a current tracking target is a human body, if yes, the human body position is recorded, and the step (2) is returned to, or otherwise the step (1) is returned to. Compared with the prior art, the method has the advantages of being high in accuracy, good in instantaneity and the like.

The invention relates to the technical field of the computer graph, in particular to the large-scale simulated terrain system drawing optimization. Compared with the traditional alpha mixing method, the merging method provided by the invention has the advantages that the drawn model dose not need to be sequenced so that the efficiency is greatly improved, all the ground surface modifiers of the same type can be drawn by batch by adopting the multiflow so that the GPU calling times is reduced, the GPU waiting is prevented and the drawing efficiency is improved, the imperfect alpha testing method is adopted so that the traditional alpha mixing effect is also achieved to some extent, a channel graph is adopted so that any distortion effect can be completely covered, and the calculating speed of the alpha testing method is higher than the calculating speed of the alpha mixing method.

The invention discloses a method and a device of directed migration. The method of the directed migration comprises confirming the position information of a calling terminal and a called terminal and accessed mobile switching center information according to calling information between the calling terminal and the called terminal, confirming that the calling terminal and the called terminal are in an identical set area according to the position information of the calling terminal and the called terminal, and indicating one mobile switching center of the mobile switching centers respectively accessed by the calling terminal and the called terminals to move the calling terminal or the called terminal to another mobile switching center when the calling terminal and the called terminal are confirmed to access different mobile switching centers of one identical area according to the accessed mobile switching center information. In other words, station to station call between the calling terminal and the called terminal with close positions are converted to office call, and therefore the quantity of the station to station call in a network is reduced, the overhead of network resources is reduced, and the perceptibility of a user is improved.

The embodiment of the invention provides a three-dimensional model monomer implementation method. The method comprises the following steps that an octree structure is used for dividing an inclined model; vector data extraction is carried out on the divided eight molecular nodes through a ray method; and inputting the extracted vector data into a database, and layering and monomerizing the inclined model building according to the information in the database. According to the embodiment of the invention, the tilt model triangular grids are combined to reduce the calling times of the rendering pipeline, so that the problem of low rendering efficiency of the tilt model when the data size of the tilt model is relatively large is solved; the method can solve the time-consuming, labor-consuming and precision problems existing in current tilt model vector data extraction, and realizes tilt model layering monomerization.