18results about How to "Solve low resource utilization" patented technology

The invention discloses a new energy output prediction data-based power distribution network scheduling method and system, and relates to the technical field of power grid scheduling, and the method comprises the steps: making a day-ahead scheduling plan according to regional prediction power consumption data and regional prediction new energy power generation data; according to the regional real-time weather data, predicting the generation power change trend of each new energy generator set in a short period; obtaining a power grid system load change trend in a short period; and making an intra-day short-term scheduling plan according to the power generation power change trend of each new energy generator set and the load change trend of the power grid system. According to the method, factors such as the electric power cleaning degree and the start-stop cost of the conventional generator set are considered during day-ahead scheduling, the carbon emission is reduced while the start-stop time of the conventional generator set is reasonably arranged, and the randomness and the intermittency of power generation of the new energy generator set are considered during day-ahead short-term scheduling; and the power generation capacity is reasonably distributed according to the demand of the power grid load.

The invention discloses a multi-layer memory management method and device, a storage medium and electronic equipment, and relates to the technical field of memory management, and the method comprises the steps that a prediction frequency set of storage equipment is obtained through prediction by using a reference information set corresponding to the storage equipment, and the storage equipment comprises a first memory and a second memory; the first delay of the first memory is smaller than the second delay of the second memory; determining a plurality of target sub-data contained in the target data of the storage device, and determining predicted access frequencies corresponding to the plurality of target sub-data based on the predicted frequency set; under the condition that the predicted access frequencies corresponding to the multiple pieces of target sub-data meet a preset splitting condition, splitting the target data into multiple pieces of target sub-data; and storing the target sub-data with the predicted access frequency greater than the frequency threshold in the first memory, and storing the target sub-data with the predicted access frequency less than the frequency threshold in the second memory. The technical problem that a memory management method in the related technology is low in resource utilization rate is solved.

The invention discloses a frequency point distribution method. The method comprises the following steps: a first device sends a first request to a first network element of a first gateway device; wherein the first request comprises ephemeris information of the first equipment, first beam position frequency point information of the first equipment and beam position information of the first terminal; the first network element is used for managing terminals in an overlapping coverage area of the first equipment; the terminal comprises a first terminal; the first network element receives a first request sent by the first device; the first network element determines a first frequency point at least based on ephemeris information, first beam position frequency point information and beam position information; wherein the first terminal accesses the first device based on the first frequency point; the first network element sends the information of the first frequency point to the first device; and the first device receives the information of the first frequency point sent by the first network element. The invention further discloses the first network element of the first network device, the first device and a computer readable storage medium.

The invention relates to a resource balancing method and device of a distributed code storage system and electronic equipment. The method comprises the following steps: acquiring storage resource information and hotspot resource information of each storage server of the distributed code storage system; according to the storage resource information, carrying out storage resource evaluation on the storage server, and according to the hotspot resource information, carrying out hotspot resource evaluation on the storage server; generating a copy moving task for moving a target code warehouse copy stored in the storage server with unbalanced resources to other storage servers under the condition that the storage resources or the hotspot resources of the storage server are unbalanced; and allocating the copy moving task to the corresponding storage server for execution, and performing copy moving on the target code warehouse copy. The problem that the resource utilization rate of a storage system is low due to the fact that resource occupancy among storage servers is more unbalanced along with use and capacity expansion of a distributed code storage system in the related technology is solved.

The present application relates to a sterilization control method for a bioreactor, including: when a first bioreactor is sterilized in air and a second bioreactor is fermented in biological fermentation, the second valve of the first bioreactor and the fourth valve of the second bioreactor are controlled to be opened, and the fourth valve of the first bioreactor and the second valve of the second bioreactor are controlled to be closed; when the first bioreactor is fermented in biological fermentation and the second bioreactor is sterilized in air, the fourth valve of the first bioreactor and the second valve of the second bioreactor are controlled to be opened, and the second valve of the first bioreactor and the fourth valve of the second bioreactor are controlled to be closed. The present invention improves the steam utilization rate, thereby reducing the waste of water resources and thermal resources as a whole, and solves the problem of low resource utilization rate in the current steam air sterilization method of reactors.

The invention discloses a Multi-DNN reasoning-oriented mixed precision FPGA (Field Programmable Gate Array) hardware accelerator and a quantification method, and relates to the technical field of neural network accelerators. The multi-DNN reasoning-oriented mixed precision FPGA hardware accelerator comprises a processor, a plurality of calculation core modules, a DDR (Double Data Rate) module and an intermediate data cache module; each computing core module supports one of high, medium and low computing precision, and a computing unit array module in each computing core module consists of computing units for performing three bit width multiply-accumulate operations; the computing unit is realized by different resources in the FPGA; according to the quantization method, weight data of the neural network are quantized by taking a layer as a unit by adopting three bit widths, namely a high bit width, a middle bit width and a low bit width, and intermediate feature map data of a neural network layer are quantized by adopting one bit width. According to the hardware accelerator and the quantification method provided by the invention, the hardware utilization rate, the throughput and the energy efficiency ratio of the accelerator can be improved, the neural network load of mixed precision quantification can be processed in parallel, and the memory access bandwidth and the energy efficiency of Multi-DNN reasoning are optimized.

The utility model discloses an impact-resistant and wear-resistant hammer head which comprises a hammer head body and a hammer handle, and a connecting sleeve is arranged on the inner side of the hammer handle. The hammerhead is produced in an embedded pouring mode, the hammerhead pouring liquid is made of a high-chromium cast iron material, the whole hammerhead has efficient wear resistance, high temperature resistance and high stability, the service life of the whole hammerhead body is greatly prolonged in the using process of the hammerhead body, meanwhile, the impact resistance of the whole hammerhead body is improved by prearranging the embedded frame, and the service life of the hammerhead body is prolonged. Under the action of the pre-embedded frame, the wear-resisting plate is convenient to disassemble, assemble and replace, when the outer side of the hammer head body is seriously abraded or damaged, targeted replacement is convenient, and the situation that the resource utilization rate is reduced due to overall replacement of the hammer head body is avoided; and the service life and the stability can be kept for a long time, so that the overall service life of the hammer head is prolonged.

This invention provides a self-powered drive device. The self-powered drive device includes a movable support component, a power supply component, and a motor. The movable support component has multiple wheel components at its bottom. The power supply component is disposed on the movable support component. The motor is detachably disposed on the movable support component. The motor is adapted to be detachably connected to the drive connection part of the photovoltaic container to provide power to the photovoltaic container's unfolding and unfolding transmission component. The motor is electrically connected to the power supply component. This invention allows for flexible relocation using the movable support component, enabling the unfolding and folding of photovoltaic panels without the need for additional power supply equipment or long-distance cable laying. This reduces construction difficulty and shortens the construction period, thus meeting the needs of rapid on-site installation.

The invention relates to a decoding resource allocation method, computer equipment and a readable storage medium. The method comprises the following steps: acquiring a first residual decoding capability corresponding to a target decoding system and a decoding demand resource quantity corresponding to a target code stream; under the condition that the first residual decoding capability does not meet the decoding demand resource quantity, determining a target migration memory matched with the decoding demand resource quantity from a resource chain table; performing resource migration on the target migration memory to obtain a first target decoding memory; determining a second residual decoding capability of a target decoding chip corresponding to the first target decoding memory; based on the first target decoding memory and the second residual decoding capability, allocating decoding resources matched with the decoding demand resource quantity to the target code stream; according to the invention, the reasonable utilization and reallocation of the resources can be realized by dynamically adjusting the resources even if the resources of the target decoding system are insufficient, and the resource utilization rate and the resource allocation flexibility are effectively improved.

This application discloses a workflow processing system and method. It relates to the fields of cloud computing, automated task scheduling, and container orchestration. The system includes: a server, used to interact with users through a graphical interface, create workflow template files, and send the workflow template files to a scheduler; a scheduler, used to receive workflow instances, extract a directed acyclic graph (DAG) from the workflow instances, and control clients to execute workflow instances according to the DAG; and a client, used to execute tasks of N task nodes through N containers, updating the status of each task node to the server. This application solves the problem of low resource utilization in workflow scheduling in cloud-native environments in related technologies.

The present application relates to a kind of resource affinity power scheduling method, device, equipment and medium based on, including the following steps, the load state data of each computing node is collected by power node monitoring module;Then the task of user submission is carried out demand analysis, generates task resource demand vector, and with node load data affinity matching, obtain resource affinity matching value;Then the interdependence of node is analyzed, and resource dependence coefficient is calculated, and weighted with affinity matching value to obtain comprehensive affinity value;Comprehensive affinity value is sorted in descending order, and form node priority sequence;Finally, power mapping scheduling is carried out based on the sequence, and the final scheduling scheme is generated.The method considers resource matching degree and node collaborative relationship comprehensively, improves scheduling efficiency and resource utilization, solves the technical problems of low resource utilization, high task response delay and uneven load between nodes caused by traditional scheduling method.

The invention relates to the field of plate glass, and discloses plate glass based on diffusivity strengthening and a preparation method of the plate glass based on diffusivity strengthening. Comprising the following raw materials in parts by weight: 50-80 parts of cullet, 30-60 parts of sand, 10-20 parts of sodium carbonate, 5-10 parts of limestone, 1-5 parts of zinc oxide, 1-5 parts of titanium dioxide, 1-3 parts of bauxite, 1-3 parts of aluminum oxide, 1-3 parts of barium sulfate and 1-3 parts of magnesium oxide. A certain amount of cullet is used as one of the raw materials of the diffusivity-strengthened plate glass, and the cullet, the sand, the sodium carbonate, the limestone, the zinc oxide, the titanium dioxide, the bauxite, the aluminum oxide, the barium sulfate and the magnesium oxide are prepared according to a certain proportion, so that the utilization rate of the cullet is increased, and the waste of resources is reduced; therefore, the problem of low resource utilization rate during production of traditional plate glass due to low utilization rate of cullet of the traditional plate glass is solved.

The present application discloses a resource allocation method, which belongs to the field of computer technology and is used to solve the problem of low resource utilization in the prior art. The resource allocation method disclosed in the embodiment of the present application includes: allocating the resources of the platform according to the predetermined resource configuration weights matching each user, determining the basic resource quantity allocated to each user; and controlling resource transfer according to the basic resource quantity and the resource usage information of the user obtained during the operation of the platform. The resource allocation method disclosed in the embodiment of the present application performs preliminary resource allocation according to historical resource usage information when the platform is started, and then, during the operation of the platform, flexibly allocates resources according to the resource usage of each user, and handles the peaks and troughs of resource usage by transferring resources between users, thereby effectively improving resource utilization.

The invention provides a component editing method which comprises the steps that a game editing scene is displayed in a graphical user interface, and the game editing scene comprises a plurality of virtual components; and in response to a selection operation for a first virtual component and a second virtual component in the plurality of virtual components, displaying an associated display effect of the second virtual component in a target surface area of the first virtual component, the associated display effect being determined based on visual features of the second virtual component. Therefore, the user can directly use the existing component in the scene to generate the sticker effect and preview the sticker effect in real time without the tedious process of screenshot, export and reimport.

The application discloses a MAC protocol implementation method based on clustering management and multi-priority scheduling, and is implemented according to the following steps: first, a node with completed initialization interacts neighbor information by periodically sending a HELLO packet, and then each node has neighbor information within a two-hop range, and a clustering degree value of the node is calculated based on multiple interaction packets; then, if there is no cluster head within the two-hop range, a cluster head election process is completed based on a clustering management mechanism, and the cluster is organized into clusters; if there is a cluster head within the two-hop range, window information broadcasted by the cluster head is listened to, and network access is requested within the window; finally, after clustering is completed, the cluster head periodically completes cluster group maintenance, information aggregation and time slot resource scheduling according to a TDMA time frame structure; and the method divides the cluster into multiple clusters for independent management through the clustering mechanism, and solves the problems of difficult node access, poor flexibility and poor expansibility of the network of the traditional MAC protocol.

The application provides a large model instance scheduling method supporting heterogeneous computing power nodes, comprising: acquiring real-time resource state data of each computing power node in a cluster, wherein the resource state data at least includes processor occupancy, memory occupancy, accelerator occupancy and video memory occupancy; comprehensively evaluating the real-time resource state data according to a preset weight strategy, and calculating the load score of each computing power node, wherein a penalty factor is applied to the computing power node without an accelerator to improve the load score; in response to a start request of a large model instance, selecting a node with the lowest load score from all online computing power nodes as a target deployment node; and sending a model start instruction containing parallel configuration parameters to the target deployment node to run the large model instance on the target deployment node. The problem of low resource utilization under a hybrid architecture is effectively solved, and the flexibility of large model deployment and the overall performance of the cluster are improved.

The present application provides a method, device and terminal for side link communication, the method comprising: determining a maximum AGC change corresponding to each second terminal, the maximum AGC change being the maximum fluctuation value of an AGC gain value within a transmission time interval; determining a configuration quantity of AGC symbols according to the maximum AGC change corresponding to each second terminal, wherein when the maximum AGC change corresponding to the second terminal is less than a preset threshold, determining the configuration quantity of the AGC symbols of the second terminal to be 0, otherwise, determining the configuration quantity of the AGC symbols of the second terminal to be a positive integer N; the first terminal sends the configuration quantity of the AGC symbols to each second terminal through indication information, thereby solving the problem in the prior art that the side link communication method cannot dynamically adjust the quantity of AGC symbols, resulting in low resource utilization.

The present invention discloses a data center job scheduling method and system based on temperature prediction, including: obtaining relevant parameters of data center cabinets, relevant parameters of servers in the cabinets, resource sizes required for jobs to be scheduled in a job queue, and relevant parameters of cooling equipment; preprocessing the acquired data, and performing feature screening on the preprocessed data; predicting the temperature of the cabinets within a future set time period based on a trained machine learning model and the screened features, and selecting the cabinet with the lowest temperature; performing initial scheduling and optimal scheduling of the jobs to be scheduled in several servers in the cabinet with the lowest temperature, and selecting the best mapping scheme between the servers and the jobs to be scheduled through multiple iterations; and scheduling the jobs to be scheduled according to the best mapping scheme.