121results about How to "Improve resource usage efficiency" patented technology

The invention discloses an accelerated operation method and server for a convolutional neural network, and a storage medium. A map to be split is split into a preset quantity of sub-maps, and the position information of each sub-map is obtained; each sub-map is independently subjected to cross-layer operation; an intermediate result of each-layer operation is obtained and is stored into an internal memory to extract the intermediate result in the internal memory so as to participate in next-layer operation; therefore, the internal memory is fully utilized, resources are reused, interaction time with an external memory is shortened, the operation speed and the resource use efficiency of the CNN (Convolutional Neural Network) are greatly improved, and therefore, the CNN can efficiently operate at a high speed in an embedded terminal.

The invention discloses an arbitration server based cluster split-brain prevent method and device, and belongs to high-availability cluster split-brain prevention technology in the field of computer cluster technology, in order to solve the problem that services cannot be taken over or the services run on two nodes at the same time due to the fact that states of other nodes and running services thereof cannot be accurately distinguished when a cluster heartbeat network is interrupted. The scheme includes that when the heartbeat network is interrupted, cluster nodes not running services can take over the services only by acquiring corresponding service locks through an arbitration server so as to avoid the problem of split-brain; after the services cease, the arbitration server recovers the service locks and allows other cluster nodes to preempt the same again; in the process that multiple nodes preempt one service lock, only one node succeeds in preemption and the services can be started to prevent occurrence of the split-brain.

The present invention provides a neural network accelerator for bit width partitioning and an implementation method of the neural network accelerator. The neural network accelerator includes a plurality of computing and processing units with different bit widths, input buffers, weight buffers, output buffers, data shifters and an off-chip memory; each of the computing and processing units obtains data from the corresponding input buffering area and weight buffer, and performs parallel processing on data of a neural network layer having a bit width consistent with the bit width of the corresponding computing and processing unit; the data shifters are used for converting the bit width of data outputted by the current computing and processing unit into a bit width consistent with the bit width of a next computing and processing unit corresponding to the current computing and processing unit; and the off-chip memory is used for storing data which have not been processed and have been processed by the computing and processing units. With the neural network accelerator for bit width partitioning and the implementation method of the neural network accelerator of the invention adopted, multiply-accumulate operation can be performed on a plurality of short-bit width data, so that the utilization rate of a DSP can be increased; and the computing and processing units (CP) with different bit widths are adopted to perform parallel computation of each layer of a neural network, and therefore, the computing throughput of the accelerator can be improved.

The present invention embodiment discloses a communication resource management method, device, system and computer storage medium, said method comprising: a first UE receives control information; said first UE, in accordance with the received control information, allocates device to device (D2D) communication resources to a second UE.

The invention discloses a user interface protocol stack, applied to radio evoluted networkin two-layer node structure, comprising IP access gateway (IAGW) user interface protocol stack, edge radio station (ERS) user interface protocol stack and uer end (UE) user interface protocol stack, where the IAGW user interface protocol stack comprises: layer L1, layer L2, IP layer and higher layer; the ERS user interface protocol stack comprises equivalent layer L1, layer L2, IP layer and higher layer on E-I interface side between it and IAGW; the radio interface side comprises: radio layer L1, lower layer and head compression layer; the UE user interface protocol stack comprises equivalent radio layer L1, lower layer and head compression layer to those on ERS radio interface side. And the invention also discloses a head compressing method in the two-layer node structure, implementing head compression of air interface data packet between ERS and UE.

The invention discloses a docking workflow engine operation scheduling method and device based on Kubernetes system. The methods include: receiving the executable unit submitted by the workflow engine; generating a configuration file for the Kubernetes system to create a container based on the parameters of the executable unit; sending the configuration file to the Kubernetes system so that the Kubernetes system creates the container according to the configuration file and executes the executable unit; receiving the first result of the executable unit returned by the Kubernetes system; translating the first result information to generate the second result information that can be recognized by the workflow engine and sending the second result information to the workflow engine; and if the second result information is successful, receiving the next executable unit that the workflow engine submits according to the execution sequence. The invention can realize the synergistic work of the workflow engine and Kubernetes system, and effectively improve the efficiency of operation execution and the resource utilization.

The invention discloses a user positioning method and device. The method comprises the following steps: acquiring cell data and measurement report data; calculating a neighbor cell and a cell coveragerange of the cell according to the cell data; extracting an information fingerprint according to the measurement report data, the neighbor cells of the cell and the cell coverage area; and calculating the position of the user terminal according to the neighbor cell, the cell coverage area and the information fingerprint of the cell. According to the method, the user positioning accuracy is greatly improved, the resource use efficiency is improved, and the method is a basis for all position-related applications; The method can be applied to the FDD/TDD network, and the application range is wider.

The invention discloses a cloud resource scheduling method. The cloud resource scheduling method and device comprises the steps that a resource scheduling model is generated on the basis of an elastic stretching strategy and a dynamic migration strategy by combining the scale of a cloud platform; transverse and/or longitudinal multi-dimensional scheduling is conducted on virtual resources of which resource allocation exceeds a preset value through the elastic stretching strategy and the dynamic migration strategy in the resource scheduling model. According to the cloud resource scheduling method, by adopting the resource scheduling model which is generated on the basis of the elastic stretching strategy and the dynamic migration strategy by combing the scale of the cloud platform to conduct dynamic scheduling on the virtual resources in the platform, the migration and bad dynamic loading effect problems which are likely to be caused by an instant resource utilization rate peak value can be avoided, the resource using efficiency of the cloud platform is effectively improved, the stability of the cloud platform is improved, the cloud platform is safer and more reliable, and the user experience degree is increased. The invention further discloses a cloud resource scheduling device which also has the above-mentioned advantages.

The invention discloses a streaming media broadcast realizing method based on a peer-to-peer network. The method comprises the following steps: a resource scheduling server in the region of a client end receives a resource scheduling request sent by the client end; the resource scheduling server determines the address of a border server corresponding to the client end which sends the resource scheduling request according to a resource scheduling strategy and sends the determined address of the border server corresponding to the client end and the address of a node managing server to the client end. As the resource scheduling server divides the region of the border server according to the resource scheduling strategy and the region is managed by the node managing server, the demand of flow regionalization of an operator is met, the streaming media broadcast scheduling based on the peer-to-peer network is easy, and the resource use efficiency is high.

The invention discloses a carrier aggregation method for a third-generation (3G) network and a fourth-generation (4G) network. The carrier aggregation method for the 3G network and the 4G network comprises the following steps that: when determining to allocate resources of the 4G network to user equipment (UE), a main radio resource connection (RRC) control anchor point informs an auxiliary RRC control anchor point to allocate the resources of the 4G network to the UE; and the main RRC control anchor point acquires information about the resources of the 4G network, which are allocated to the UE by the auxiliary RRC control anchor point. The invention also discloses a carrier aggregation system for the 3G network and the 4G network. The system implements the method. By the invention, the UE can simultaneously use carrier resources of the 3G network and the 4G network, the carrier aggregation of the 3G network and the 4G network is realized, and the speed rate of the UE and the utilization efficiency of the resources are greatly improved.

The invention provides an application process management method. The method comprises: receiving an application process switching instruction; through a switching process function in startup management service, obtaining foreground switching information of a switching foreground process and a background switching information of a switching background process, the switching information being corresponding to the application process switching instruction; according to the foreground switching information, performing foreground switching operation on the switching foreground process; and according to the background switching information, performing background switching operation on the switching background process. The invention also provides an application process management device. The application process management method and the application process management device perform foreground switching operation and background switching operation on the processes according to the foreground switching information and the background switching information of the processes, so as to prevent frequent close and open of the application processes, and improve application usage efficiency of users and resource use efficiency of a system.

The invention relates to a distributed parallel spatial viewshed analysis method. A module is subjected to parallel dynamic division by using digital elevation model data, parallelization and data scheduling of a serial algorithm are quickly completed, and irregular data partitioning and a redundancy mechanism of a sight line region are applied, so that it is ensured that a slave node can access to the digital elevation model data corresponding to a target grid point region in a distributed calculation environment, wherein an irregular data parallel policy has important guidance meaning for viewshed algorithms of different realization modes; and in combination with a buffer region mechanism, correct operations of different dependence degree algorithms are effectively ensured and resource utilization efficiency of the slave node is improved, so that the method has wide industrial application prospects.

The invention relates to the technical field of network teaching, and specifically provides a teaching material resource management system and a teaching material resource management method. The teaching material resource management system comprises a teaching material management module, a resource uploading module, a resource indexing module, a resource checking module, a resource center module, a statistical analysis module, a data interface module and a resource retrieval module. The system solves the problem that a teacher needs to retrieve teaching material resources in a simplified, foolish and pertinent manner due to disorder resource storage in the existing technology. Effective management and circulation of a school or regional resources are improved. The resource utilization rate of the teacher is improved.

This invention relates to message initial compression method, which comprises the following steps: defining wireless interface user interface agreement shed; processing compression process between agreement shed internet and core internet mouth. This invention can process effective compression between internet and core interface.

The invention discloses a staged construction method of a multi-bridge-floor-layer suspension bridge for a highway and a rail. The suspension bridge is a three-bridge-floor-layer suspension bridge composed of a double-layer parallel main truss structure, rail transit is located on the uppermost layer, and highway traffic is located on the middle layer and the lower layer. The staged construction method of firstly constructing the highway traffic and then constructing the rail transit is adopted, construction on the rail transit can be carried out in the running process of the highway traffic, the scheme of the three-bridge-floor-layer suspension bridge is adopted, structural efficiency can be improved, the adaptive capacity of the structure on terrain is enhanced, the structure can adapt to the requirement for staged construction of the rail transit and the highway traffic, the recent engineering construction investment is effectively reduced, idle resources are avoided, meanwhile, the requirement that existing traffic circulation is not interrupted in the long-term construction period can be met, and the influence of construction on traffic is reduced to the largest extent.

Embodiments of the invention provide a data processing method and apparatus of a data warehouse. The method comprises the steps of receiving a query condition input by a user, wherein the query condition comprises a keyword of to-be-queried data; determining a dependence relationship between the to-be-queried data and other data in the data warehouse according to the keyword, wherein the dependence relationship is one of the following dependence relationships: no dependence, strong dependence and weak dependence; returning the dependence relationship to the user; receiving a data processing instruction sent by the user according to the dependence relationship; and triggering the data warehouse to execute a data processing instruction for the to-be-queried data. By adopting the method provided by the embodiment of the invention, the resource usage efficiency of the data warehouse can be improved.

The embodiment of the invention provides an application scheduling method and device, and relates to the technical field of communication. The success rate of application deployment can be increased.The method includes the steps that a scheduling device receives an application set from a deployment managing device, wherein the application set comprises corresponding relations between each application to be deployed in the application set and the resource requirement of the application to be deployed; the scheduling device obtains the free resource quantities of multiple resource nodes and a preset deployment condition, wherein the deployment condition includes the preset SLA condition of each resource node; for each application to be deployed, the scheduling device deploys the applicationto be deployed to the resource node of which the free resource quantity meets the resource requirement of the application to be deployed, wherein the service level agreement SLA parameters of the resource node after the application to be deployed is deployed meet the preset SLA condition of the resource node. The method is applied to application scheduling.

The invention relates to an energy storage system, in particular to a solar thermal dynamic energy storage system, belonging to the technical field of energy storage. The solar thermal dynamic energy storage system comprises a solar collecting system, a compressed energy storage system and a thermal power expansion system, wherein a photovoltaic power generation device can generate electricity to drive a high-pressure compressor to work, external electric energy can drive the high-pressure compressor and a low-pressure compressor to work through a power grid of a power station, and an expansion machine can swell to generate electricity to release energy. According to the energy storage system, the valley electric energy of the power grid of the power station and residual electric quantity generated through daytime solar photovoltaic power generation can be stored, the solar photovoltaic power generation and sunshine thermal power stored heat power generation can be fully utilized, thereby improving the utilization efficiency of energy, lowering the power generation cost and increasing the economic benefit.