31results about How to "Increase visits" patented technology

The invention provides an operation device and method of an accelerating chip which accelerates a depth neural network algorithm. The device comprises a vector addition processor module which carries out addition or subtraction of vectors and / or the vector operation of a pooling layer algorithm in the depth neural network algorithm, a vector function value operator module which carries out the vector operation of the nonlinear evaluation in the depth neural network algorithm, and a vector multiply adder module which carries out the multiply addition operation of the vectors. The three modules execute a programmable instruction and interact with each other to calculate a neural network output result and a synaptic weight change amount which represents the neuron action intensity between neural layers. The three modules are provided with middle value storage areas and carry out reading and writing operations on a master memory. Thus, the reading and writing frequencies of the middle value of the master memory can be reduced, the energy consumption of the accelerator chip can be reduced, and the data loss and replacement problems in a data processing process can be avoided.

A method for processing multi-channel audio signals includes receiving L channel audio signals from participating conference terminals, decoding the L channel audio signals, and determining N conference terminals whose audio signals are to be mixed from the participating conference terminals according to the data obtained through decoding; selecting an encoder as a common encoder from L−N encoders of the participating conference terminals except the N conference terminals whose audio signals are to be mixed, encoding decoded audio signals of the N conference terminals whose audio signals are to be mixed, and sending encoded data to conference terminals corresponding to the L−N encoders (indicating that the number of encoders is equal to the value of L minus N); and updating encoding status information of L−N−1 encoders except the common encoder to synchronize with encoding status information of the common encoder.

Disclosed is an electronic commerce system using a jackpot discount rate and providing a price discount service to a customer. To this end, provided is an electronic commerce system using a jackpot discount rate, the electronic commerce system comprising a web server, which analyzes an electronic commerce earnings rate on the basis of purchase information transmitted from a plurality of purchaser terminals, extracts a jackpot discount rate exceeding earnings expected by sales of a bidding product according to an excess earning rate when the earnings rate exceeds a predetermined earnings rate, determines a jackpot price to which the jackpot discount rate is applied, and provides information on the jacket price to the purchaser terminals. According to the present invention, provided is an opportunity to purchase the bidding product at the price for which the jackpot discount rate is applied according to the generation of earnings, such that a customer's sense of expectation is stimulated and the number of shopping mall visits increases.

Embodiments of the present invention provide a method and an apparatus for constructing a memory access model, and relate to the field of computers. The method includes: obtaining a page table corresponding to a process referencing a memory block, and clearing a Present bit included in each page table entry stored in the page table; and constructing a memory access model of the memory block according to the number of access times of each page in the memory block and time obtained through timing, where the memory access model at least includes the number of access times and an access frequency of each page in the memory block. The apparatus includes: a first obtaining module, a first monitoring module, a first increasing module, and a second obtaining module. The present invention can reduce the memory consumption and an impact on the system performance, and avoid a system breakdown.

There is provided a storage system including one or more LDEVs, one or more processors, a local memory or memories corresponding to the processor or processors, and a shared memory, which is shared by the processors, wherein control information on I / O processing or application processing is stored in the shared memory, and the processor caches a part of the control information in different storage areas on a type-by-type basis in the local memory or memories corresponding to the processor or processors in referring to the control information stored in the shared memory.

The present invention provides a downlink resource scheduling method and a transmission end for a multiple-input multiple-output beam-forming (MIMO+BF) system. The transmission end performs the following steps of: dividing subordinate users into user groups according to respective MIMO+BF weights of the subordinate users; generating a scheduling priority coefficient for each of the user groups according to scheduling priority coefficients of respective users in the user group; allocating group resources to the user groups according to the respective scheduling priority coefficients of the user groups; and allocating sub-channel resources of the group resources to each user in the user groups according to the respective scheduling priority coefficients of users in the user groups.

The invention discloses a caching method and caching equipment and belongs to the field of Internet communication. The method comprises the following steps of: receiving a retrieval request which carries retrieval strings, and searching a search result page which is stored in a file system according to the retrieval strings; if the corresponding search result page is found, increasing an access frequency of the search result page; and if the corresponding search result page is not found, searching search results which correspond to the retrieval strings by a search engine, assembling the search results into the search result page, and storing in the file system. The equipment comprises a searching module, a gradual-increase module and a storage module. By the invention, the consumption oftime and resources can be reduced.

An embodiment of the present invention provides a data processing method, device, system, and application server. The method includes: the first application server determines the access log of each access data based on the user's access log within a preset time period stored by the second application server. times; and determine the weight value corresponding to each access data; for each access data, from the data set to be processed, determine the data associated with the access data, which is the associated data of the access data; for each access data, based on the access data corresponding set the weight value of the associated data corresponding to the access data; send the weight value of the access data and / or the weight value of the associated data to the second application server, so that the second application server, for each received data, based on the The weight value of the data, which sets the cache duration of the data. Realize the intelligent control of the cache duration of each data, to a certain extent reduce the waste of storage resources and reduce the burden of server operation and maintenance.

The invention discloses a method and a device for accessing a survey point of a cluster database, which are applied to a Treap tree established in memory. A node memory index block of the Treap comprises the survey point identification range and access times; the node access times are sequentially reduced according to a preset node searching sequence; and preset number index blocks read from a disk is saved in the memory according to the high-low sequence of the access times. The method comprises the following steps: receiving a survey point access request in a preset cycle; searching a target node from a tree according to the target survey point identification and a preset node searching sequence in the request; increasing the access times of the target node; searching whether an index block corresponding to the target node exists in the memory or not, and determining the location of the survey point corresponding to the target survey point identification in the index block if the index block exists; accessing the survey point according to the survey point location; regulating the position of the target node in the Treap tree after the preset cycle is finished to sequentially reduce the access times of each node according to the node searching sequence.

The embodiment of the invention provides a network voting processing method and device and a server. The method comprises the steps that a client displays the numbers of currently available votes and initial votes of a user and sends voting information to the server by responding voting operation of the user, wherein the voting information comprises the number of the votes; the sever receives the voting information, the number of the currently available votes and the number of the votes in the voting information are compared, and the number of the currently available votes is reduced by M votes according to a comparison result, wherein the M represents the number of the votes in the voting information or the number of the currently available votes; after the server receives the voting information, when the number of the currently available votes is smaller than that of the initial votes, timing is conducted, and when the time reaches a preset value, the number of the available votes is controlled to be increased by at least one vote. By means of the network voting processing method, dwell time and access frequency of the user on a voting page can be increased, and use viscosity and voting frequency of the user on voting are increased.

The present invention provides a method for precise delivery of information, comprising: DMP pre-assigns an authorization code for DSP; the DSP uses the authorization code to access the DMP; the DMP checks the access request of the DSP; The DSP proposes to the DMP the requirement to create an information publishing task; the DMP creates an information publishing task according to the requirements of the DSP; the DMP receives the query request sent by the DSP; the DMP matches the possibility corresponding to each task created by the DSP A list of deployed devices; the DMP returns a list of successfully matched task names to the DSP. The present invention can realize precise delivery of information and increase the number of visits of mobile users to specific targets.