Patents
Literature
Hiro is an intelligent assistant for R&D personnel, combined with Patent DNA, to facilitate innovative research.
35results about How to "Guaranteed computing power" patented technology
Efficacy Topic
Property
Owner
Technical Advancement
Application Domain
Technology Topic
Technology Field Word
Patent Country/Region
Patent Type
Patent Status
Application Year
Inventor
Bilayer sliced high-efficiency block chain consensus mechanism and working method thereof
ActiveCN109150972AStrong processing ability and stabilityAvoid malicious node attacksFinanceUser identity/authority verificationBlockchainDistributed computing
The invention discloses a bilayer sliced high-efficiency block chain consensus mechanism and a working method thereof. The consensus mechanism is based on a bilayer gossip topology network, and comprises a full node, a local full node, a light node and a micro node, wherein the full node is in charge of node topology and slice maintenance and is elected and updated through a full node maintenancemechanism; and the local full node is elected and updated through a local full node maintenance mechanism and is in charge of transaction validation, transaction consensus, transaction storage and account synchronization. The invention also provides a working method for the consensus mechanism. A credible and incentive node joining and updating mechanism is provided, the stability, the bandwidth capability and the computing capability of the full node and the local full node are ensured, and the impacts on a transaction confirmation speed are reduced. The consensus mechanism proposed by the invention forms a divide-and-conquer distributed account system by means of intra-slice autonomy and inter-slice collaboration to ensure to consider decentralization, scalability and security.
Owner:湖南宸瀚科技有限公司
Cooperative quantum computer architecture structural conceptual
InactiveCN1870015AIncreased instruction parallelismGuaranteed computing powerQuantum computersCoprocessorComputerized system
A system structure of coquantum computer is prepared as extending quantum calculation component on traditional computer system and utilizing traditional computer as control center, using quantum calculation component as coprocessor being used only to carry out quantum calculation and adding command concurrency of quantum computer on quantum calculation coprocessor.
Owner:SUN YAT SEN UNIV
Flexibly configurable neural network computing unit, computing array and construction method thereof
ActiveCN109409512AGuaranteed computing powerIncrease flexibilityNeural architecturesPhysical realisationComputer moduleParallel computing
The invention discloses a flexibly configurable neural network computing unit, a computing array and a construction method thereof. The neural network computing unit comprises a configurable storage module, a configurable control module and a time-division multiplexing multiplication and addition computing module. The configurable storage module comprises a feature map data buffer, a step-size data buffer and a weight data buffer. The configurable control module comprises a counter module and a state machine module; the multiplication and addition calculation module comprises a multiplier andan accumulator. The invention can support any type of convolution calculation and parallel calculation of multi-size convolution kernel, fully exploits the flexibility and data reusability of the convolution neural network calculation unit, greatly reduces the system power consumption caused by data moving, and improves the calculation efficiency of the system.
Owner:XI AN JIAOTONG UNIV
Method for controlling energy consumption of cluster system based on changing control mode
InactiveCN101794169AGuaranteed computing powerReduce power consumptionPower supply for data processingTransmissionComputer clusterSleep state
The invention discloses a method for controlling energy consumption of a cluster system based on changing a control mode, which carries out system level management and control under the operating condition of a computation task on a computation node in the cluster system through a control node and belongs to the field of computer cluster systems. In the method for controlling the energy consumption, the control node respectively chooses fixed, manual or semi-automatic control modes for the change of a computation task amount and the frequency of the change and controls the energy consumption through controlling the state of a computation node and the computation node enters different sleep states according to an idle rule so as to save the energy consumption of a processor. In the invention, the energy consumption of the cluster system is greatly reduced aiming at parallel computation of the cluster system under the premise of not greatly degrading computation performance and the balance between the performance and the energy consumption is obtained.
Owner:BEIHANG UNIV
Indoor intelligent mobile robot and control method
InactiveCN108762255AGood positioning stabilityReduce Matching and Dynamic Feature InterferencePosition/course control in two dimensionsObstacle avoidanceEngineering
The invention provides an indoor intelligent mobile robot and a control method. The global advancing path of the robot includes a plurality of key point positions of the path; when an obstacle existsin the advancing path, an obstacle avoidance point is arranged out of the original path, a local path of the robot advancing to the obstacle avoidance point from the current key point is planned, andafter the robot advances to the obstacle avoidance point along the local path, a local path of the robot advancing to a next key point from the obstacle avoidance point is planned again; and if a newobstacle appears during the process of the robot advancing along the local path, another obstacle avoidance point is set. A three-dimensional lidar is adopted for laser oblique shooting, the scanningpoint of the laser avoids a dynamic obstacle, and thus, laser data reflecting the surrounding environment information are acquired for building a surrounding environment map. Trackless autonomous navigation in an indoor complex environment with human mobility can be realized.
Owner:NANJING UNIV OF SCI & TECH
Method for enhancing sharpness of color image with color constancy
ActiveCN105931206AAvoid distortionGuaranteed computing powerImage enhancementImage analysisColor imageSelf adaptive
The invention discloses a method for enhancing sharpness of color image with color constancy. The method includes the following steps: based on the red, green and blue three wave bands of an image, calculating a brightness image; setting a neighborhood radius, calculating mean value and standard difference of the brightness image in each pixel neighborhood, obtaining an neighborhood mean value image and a neighborhood standard difference image; based on the neighborhood mean value image and the neighborhood standard difference image, conducting adaptive calculation on weight of each pixel point; based on the neighborhood mean value image and the weights, recalculating a new brightness image; in accordance with the theory of color constancy, based on the new brightness image and an original brightness image, conducting same ratio enhancement on the three wave bands of the original image, obtaining the image with enhanced sharpness with color constancy. According to the invention, the method can greatly increase image sharpness, reduce color distortion as much as possible, maintain color constancy, and effectively increase visual effects of the image.
Owner:深圳市和天创科技股份有限公司
Cooperative computing unloading method based on splittable task in vehicle-mounted edge computing environment
ActiveCN110868700AFrequent interactionGuaranteed computing powerResource allocationRegistering/indicating working of vehiclesIn vehicleEdge computing
The invention provides a cooperative computing unloading method based on a splittable task in a vehicle-mounted edge computing environment, and the method comprises the steps: sending a to-be-processed task unloading request when the calculation amount of a to-be-processed task cannot be borne; acquiring relevant information of callable computing resources according to feedback of the to-be-processed task unloading request; and obtaining an optimal splitting ratio for task segmentation according to the related information of the callable computing resources. Therefore, the optimal unloading decision is selected by comprehensively considering the computing power of the vehicle, and the minimum task completion time delay can be conveniently obtained. After the request vehicle receives the relevant information of the callable computing resources, a task splitting model is constructed to perform construction, the optimal splitting ratio is obtained, a distributed task splitting and unloading scheme is practiced, it is guaranteed that each request vehicle autonomously decides splitting and unloading decisions of a task only based on received information, and frequent information interaction in a centralized scheme is avoided.
Owner:SHENZHEN UNIV
Coprocessor, and method and system for matrix operation acceleration
ActiveCN108845828AGuaranteed computing powerImprove data throughputConcurrent instruction executionArchitecture with single central processing unitGeneral purposeCoprocessor
The invention is applicable to the technical field of computers, and provides a coprocessor, and a method and a system for matrix operation acceleration. The coprocessor comprises a controller, a register bank, a multiplier-adder MAC array, and a memory. The controller parses a vector expansion instruction, and calls and sends data to be calculated to the register bank according to a parsing result. The register bank transmits the data to be calculated to the MAC array and transmits an operation result to the memory, the operation result being obtained by matrix operation of the MAC array on the data to be calculated, and the data to be calculated and / or the operation result are stored in the memory. In the process, the MAC array performs matrix operation according to the vector expansioninstruction, and the vector expansion instruction is generated by a main processor according to a preset function, ensuring computational power of a general microprocessor. The register bank transmitsand stores all kinds of data, and a data storage channel is independent of a receiving channel of the vector expansion instruction, improving data throughput of a general-purpose microprocessor in anoperation process.
Owner:SHENZHEN STATE MICROELECTRONICS CO LTD
Power supply voltage control method and device of digital currency mining machine and digital currency mining machine
ActiveCN111966202AReduce output voltageAvoid wasting powerVolume/mass flow measurementPower supply for data processingPower consumptionIndustrial engineering
The embodiment of the invention provides a power supply voltage control method and device of a digital currency mining machine and the digital currency mining machine. The method comprises the steps of determining a computing power ratio of the digital currency mining machine based on actual computing power and theoretical computing power of the digital currency mining machine; generating a powersupply control instruction based on a comparison result of the calculation force ratio and a preset threshold value; and controlling the output voltage of the mining machine power supply based on thepower supply control instruction. According to the embodiment of the invention, the output voltage of the power supply is controlled according to the computing power ratio, and a good compromise can be obtained between the power consumption loss and the computing power of the mining machine.
Owner:SHENZHEN MICROBT ELECTRONICS TECH CO LTD
Power-stage semi-physical simulation system of robot joint controller
ActiveCN108132608AHigh simulationThe simulation equipment is well simulated by the robot joints in real timeSimulator controlLoop controlClosed loop
Provided in the invention is a power-stage semi-physical simulation system of a robot joint controller. The system is composed of an upper computer, a real-time robot joint simulation system, a CAN communication board card, a to-be-tested motor driver, and a motor power simulation device. The upper computer calculates a robot joint position given signal and outputs the signal to the to-be-tested motor driver through the CAN communication board; the to-be-tested motor driver calculates a control voltage, receives the robot joint position given signal and an actual position feedback signal, anddetermines closed-loop control parameter; the real-time robot joint simulation system carries out real-time simulation of a robot joint control motor and a load to obtain an actual position feedback signal of the robot joint, an adjustment current and a current load; and the motor power simulation device obtains an adjustment voltage, superposes the adjustment voltage with the control voltage of the to-be-tested motor driver, and then outputs the superposed voltages to the real-time robot joint simulation system. Therefore, the variable load performance testing and simulation of the servo system are completed, so that the good simulation of the semi-physical simulation accuracy and dynamic performance of the robot is realized.
Owner:BEIJING RES INST OF PRECISE MECHATRONICS CONTROLS
Water environment analysis method and system based on geographic information system and Delft3D-FLOW calculation engine
ActiveCN104199887AGuaranteed computing powerOptimize interaction friendlinessDesign optimisation/simulationSpecial data processing applicationsGeographic information systemEnvironmental geology
The invention relates to a water environment analysis method based on a geographic information system and a Delft3D-FLOW calculation engine. The water environment analysis method comprises the steps that a software system framework is built through a GIS platform; a Delft3D mesh file format is analyzed; a configuration file MDF file format of the Delft3D-FLOW and meanings of multiple parameters in the configuration file MDF file format are analyzed, and a human-computer interaction configuration interface is developed for configuring the multiple parameters under the GIS platform; a model is built, configuration tasks of a plurality of environmental elements are completed through GIS operation; a configuration result is output as a file format which can be identified by the mathematical calculation engine, and the calculation engine is called to conduct calculation; a calculation result file format is analyzed, a file is read under the GIS platform and processed into a raster image with space coordinates through the image processing technology, and the raster image is loaded into the GIS platform for display. The water environment analysis method combines the mathematical model with spatial and geographic information, thereby improving analysis accuracy. The invention further provides a water environment analysis system based on the geographic information system and the Delft3D-FLOW calculation engine.
Owner:BEIJING SANYI SICHUAN TECH CO LTD
High-performance industrial robot controller
InactiveCN110666820AGuaranteed computing powerLow costProgramme-controlled manipulatorIndustrial roboticsMachine
The invention discloses a high-performance industrial robot controller which comprises a processor board, a machine visual unit, a demonstrator, an external IO unit, a grabber unit, an external sensorand a motor driver. The processor board comprises an ARM processor unit, an FPGA unit, a power module, an Ethernet interface A, an Ethernet interface B, an Ethernet interface C, an Ethernet interfaceD, an IO interface and a CAN interface. The robot controller is small in size, cost is reduced, the computation capacity of the controller can be ensured, and the controller has high expandability and safety; meanwhile, the robot controller is convenient to expand, meanwhile has three-dimensional machine vision and can obtain information (X, Y, Z) of a target, the target is more precisely grabbed, and the robot controller can adapt to various complicated application scenes; and a robot movement control algorithm is evenly allocated in an ARM and an FPGA, and the computation capacity of the robot controller is effectively improved.
Owner:CHINA HEFEI TAIHE OPTOELECTRONICS TECH
Intelligent driving system and method, and computer readable storage medium
PendingCN110398953AGuaranteed computing powerLow costPosition/course control in two dimensionsVehiclesCommunication unitComputer module
The invention discloses an intelligent driving system, the intelligent driving system comprises a sensing and computing module, a decision making and planning module and an underlying control module,wherein the sensing and computing module comprises a first communication unit and a sensing and computing unit comprising a plurality of artificial intelligence chips; the decision making and planningmodule comprises a second communication unit and a decision making and planning unit; and the underlying control module comprises a third communication unit and an underlying control unit. The invention further discloses an intelligent driving method and a computer readable storage medium. By adopting the intelligent driving system disclosed by the invention, the cost of the intelligent driving system can be reduced, and meanwhile, the computing power of the intelligent driving system is ensured.
Owner:CHANGSHA INTELLIGENT DRIVING INST CORP LTD
Embedded platform-oriented dual-mode target detection method and system
PendingCN113255521AGood balance of speed and precisionSatisfy real-timeImage enhancementImage analysisComputational physicsImaging Feature
The invention provides an embedded platform-oriented dual-mode target detection method and system, and the method comprises the steps: carrying out the segmentation processing on a to-be-recognized image according to a preset segmentation algorithm, and obtaining at least one image box; performing feature extraction on the image box through depth separable convolution in the improved YOLOv4-Tiny network to obtain image features corresponding to the image box; determining an infrared image and a visible light image according to the image features, respectively inputting the same a predefined visible light target detection model and an infrared target detection model after registration, and outputting detection results of the visible light target detection model and the infrared target detection model; and performing decision-level image fusion on the detection results through a pre-established decision-level fusion detection model, so that infrared and visible-light multi-band information complementation is realized, and a dual-mode target detection result is obtained. According to the scheme, real-time detection of an embedded platform for a multi-source image is realized, and the method has more obvious advantages in detection precision compared with a single-band target detection model while the requirements of real-time performance and effectiveness are met.
Owner:青岛以萨数据技术有限公司
Neural network weight discretization method and system
ActiveCN109214502AGuaranteed computing powerReduce consumptionMathematical modelsNeural architecturesComputation complexityNerve network
The invention relates to a neural network weight discretization method and system. The method comprises the following steps: obtaining a weight value range and a number of discrete weight states; acquiring a previous time step weight state and a current time step weight increment; obtaining a state transition direction by using a direction function according to the weight increment of the currenttime step; obtaining a current time step weight state according to the previous time step weight state, the current time step weight increment, the state transition direction, the weight value range,and the number of discrete weight states. The invention ensures that the weight value is always constrained in the same discrete state space without storing additional virtual continuous state implicit weights, and greatly reduces the consumption of storage space and the computational complexity under the premise of ensuring the computational performance of the neural network.
Owner:TSINGHUA UNIV
Medical data calculation method and system based on cloud platform
InactiveCN107016231AGuaranteed computing powerImplement split processingTransmissionSpecial data processing applicationsCryptographic protocolFeature data
The invention discloses a medical data calculation method and system based on a cloud platform. The method comprises the steps that medical feature data of a user is collected on the basis of a terminal application layer, and a calculation task is generated on the basis of the medical feature data; the terminal application layer sends a request to a service cluster of the cloud platform on the basis of a security protocol, and the calculation task is sent to the service cluster; when it is judged that medical service data does not exist in a mechache cache unit, the service cluster analyzes cloud calculation keywords to obtain medical feature data classification, and target service program is called for executing cloud calculation according to the cloud calculation keywords and the medical feature data classification, medical service data matched with the cloud calculation keywords is obtained, and the medical service data is sent to the terminal application layer. Accordingly, the problem of the rapid acquisition process of existing data is solved, and the whole platform experience is improved.
Owner:广州七乐康数字健康医疗科技有限公司
Image processing method, image acquisition device, movable platform and storage medium
InactiveCN112335228ANormal exposureGuaranteed computing powerTelevision system detailsColor television detailsImaging processingImaging quality
The present invention relates to an image processing method, an image acquisition device, a movable platform and a storage medium. The method comprises acquiring an image acquired by a first image acquisition device, and determining the current ambient light brightness according to the image (S101); acquiring the image target brightness corresponding to a second image acquisition device, and determining a target exposure parameter corresponding to the second image acquisition device according to the current ambient light brightness and the image target brightness (S102); and controlling the second image acquisition device to generate a target image of the image target brightness based on the target exposure parameter (S103). The method improves the image quality.
Owner:SZ DJI TECH CO LTD
Method and device for controlling computing power board, equipment, medium and program product
PendingCN111859829AImprove computing powerImprove energy efficiency ratioDigital computer detailsCAD circuit designComputer architectureSoftware engineering
The invention provides a method and device for controlling a computing power board, digital certificate processing equipment, a storage medium and a computer program product. The computing power boardcomprises a plurality of voltage domains, wherein each voltage domain comprises a plurality of chips; the plurality of voltage domains adopt series power supply; and the plurality of chips in each voltage domain adopt parallel power supply. The method comprises the following steps: obtaining an effective frequency combination, wherein the effective frequency combination comprises the frequency ofeach chip of each voltage domain during effective operation; and controlling the chip of each voltage domain to operate at the corresponding frequency in the effective frequency combination. According to the method and device for controlling the computing power board and the digital certificate processing equipment, for the computing power board with the voltage domains connected in series, the probability of chip failure caused by insufficient voltage distribution of the voltage domains can be reduced, so that each voltage domain chip operates effectively, and the computing power and the energy efficiency ratio of the computing power board are improved.
Owner:BEIJING BITMAIN TECH LTD
Method for generating an evaluation of the activities of a user on a virtual document
InactiveUS20170300858A1Guaranteed computing powerReduced dimensionDigital data information retrievalEncryption apparatus with shift registers/memoriesComputer hardwareDocumentation procedure
A method for generating an evaluation of a user's activities on a virtual document, includes:(a) providing a server system, having a server memory and server processing unit;(b) providing an electronic device, having a device memory, device processing unit, elements for connecting with the server system, elements for interacting with and allowing a user to create and modify a virtual document;whereinthe device processing unit: (s1) receives information relating the creation or opening of a virtual document; (s2) encrypts the virtual document at subsequent time intervals, to create a plurality of encrypted copies of the virtual documents, each encrypted copy associated with a different time interval; (s3) sends the encrypted copies to the server memory; (s4) interrupts steps (s2) and (s3) when the virtual document is closed;the server processing unit (s4) compares the encrypted copies; and (s5) generates a report containing the evaluation of the user's activities.
Owner:PMTP SA
A working method of the consensus mechanism of an efficient blockchain with double-layer sharding
ActiveCN109150972BImprove stabilityEasy to handleFinanceUser identity/authority verificationConsensusFinancial transaction
The invention discloses a consensus mechanism and working method of a double-layer fragmented high-efficiency block chain. The consensus mechanism is based on a double-layer gossip topology network, including: full nodes, local full nodes, light nodes and micro nodes, full Nodes are responsible for the maintenance of node topology and sharding. They are elected and updated through the full node maintenance mechanism, and local full nodes are elected and updated through the local full node maintenance mechanism. They are responsible for transaction verification, transaction consensus, transaction storage, and ledger synchronization. The invention also proposes How this consensus mechanism works. A credible and incentivized node joining and updating mechanism proposed by the present invention ensures the stability, bandwidth capacity, and computing power of full nodes and partial full nodes, and reduces the impact on transaction confirmation speed. The consensus mechanism proposed by the present invention passes The "intra-chip autonomy, inter-chip collaboration" approach forms a divide-and-conquer distributed ledger system to ensure decentralization, scalability, and security.
Owner:湖南宸瀚科技有限公司
Task processing method, device and system in cloud computing platform
InactiveCN108958913AGuaranteed computing powerImprove performanceProgram initiation/switchingResource allocationQuantum harmonic oscillatorEnergy consumption
The invention discloses a task processing method in a cloud computing platform, the method comprising generating a predetermined number of first scheduling schemes according to scheme generation information sent by a user; conducting iterative processing on the first scheduling schemes based on a multi-scale quantum harmonic oscillator algorithm to obtain a predetermined number of current scheduling schemes; calculating the energy consumption of each current scheduling scheme, and selecting the lowest energy consumption scheduling scheme; processing the tasks in the cloud computing platform according to the lowest energy consumption scheduling scheme. While guaranteeing the computing ability of the cloud computing platform, this method is based on the multi-scale quantum harmonic oscillator algorithm to optimize a predetermined number of scheduling schemes and obtain the lowest energy consumption scheduling scheme, effectively ensuring the cloud computing platform to play a high-performance and low-energy consumption strength enabling the performance of the platform to be fully exerted. The invention also discloses a task processing device and system and a computer-readable storagemedium in a cloud computing platform, all of which have the above beneficial effects.
Owner:GUANGDONG UNIV OF TECH
Water environment analysis method and system based on geographic information system and delft3d-flow calculation engine
ActiveCN104199887BGuaranteed computing powerOptimize interaction friendlinessDesign optimisation/simulationSpecial data processing applicationsImaging processingSoftware system
The invention relates to a water environment analysis method based on a geographic information system and a Delft3D-FLOW calculation engine. The water environment analysis method comprises the steps that a software system framework is built through a GIS platform; a Delft3D mesh file format is analyzed; a configuration file MDF file format of the Delft3D-FLOW and meanings of multiple parameters in the configuration file MDF file format are analyzed, and a human-computer interaction configuration interface is developed for configuring the multiple parameters under the GIS platform; a model is built, configuration tasks of a plurality of environmental elements are completed through GIS operation; a configuration result is output as a file format which can be identified by the mathematical calculation engine, and the calculation engine is called to conduct calculation; a calculation result file format is analyzed, a file is read under the GIS platform and processed into a raster image with space coordinates through the image processing technology, and the raster image is loaded into the GIS platform for display. The water environment analysis method combines the mathematical model with spatial and geographic information, thereby improving analysis accuracy. The invention further provides a water environment analysis system based on the geographic information system and the Delft3D-FLOW calculation engine.
Owner:BEIJING SANYI SICHUAN TECH CO LTD
A Color Constant Color Image Definition Enhancement Method
ActiveCN105931206BAvoid distortionGuaranteed computing powerImage enhancementImage analysisColor imageSelf adaptive
A method for enhancing the clarity of color images with constant color, including calculating the brightness image based on the three bands of red, green and blue of the image; setting the neighborhood radius, calculating the mean and standard deviation of the brightness image in each pixel neighborhood, and obtaining the neighborhood Domain mean image and neighborhood standard deviation image; adaptively calculate the weight of each pixel based on the neighborhood mean image and neighborhood standard deviation image; recalculate a new brightness image based on the neighborhood mean image and weight; based on color constancy Based on the principle of nature, based on the new brightness image and the original brightness image, the three bands of the original image are enhanced year-on-year to obtain a sharpness enhancement result image with constant color. The method of the present invention can greatly improve the clarity of the image while reducing color distortion as much as possible, maintaining color constancy, and effectively improving the visual effect of the enhanced image.
Owner:深圳市和天创科技股份有限公司
Coprocessor, matrix operation acceleration method and system
ActiveCN108845828BGuaranteed computing powerImprove data throughputConcurrent instruction executionArchitecture with single central processing unitCoprocessorMaster processor
Owner:SHENZHEN STATE MICROELECTRONICS CO LTD
Computing system for evaluating parameters of comprehensive energy equipment
PendingCN114764545AEasy to calculateFully calculatedDigital data protectionDesign optimisation/simulationData privacy protectionData mining
The invention discloses a calculation system for evaluating parameters of comprehensive energy equipment, which comprises an evaluation system body, the evaluation system body internally comprises a global model, an evaluation model A, an evaluation model B and an evaluation model C, and the global model internally comprises a first algorithm module, a second algorithm module, a third algorithm module and a profit factor formula module. According to the calculation system for evaluating the parameters of the comprehensive energy equipment, the comprehensive energy scene, the profit factor, the calculation process, the required parameter profit factor application scene, the leveling unit energy cost and the effect of the profit factor in production are combined with a learning framework, so that the parameters of the comprehensive energy equipment are conveniently and fully calculated; the three algorithms unify energy output and facilitate calculation, so that operation of workers is facilitated, guidance is provided for follow-up construction of a new comprehensive energy station through collection of a large amount of comprehensive energy equipment profit factor data, data privacy protection is achieved, and the model popularization degree is improved.
Owner:ENNEW DIGITAL TECH CO LTD
Power supply voltage control method and device of digital currency mining machine and digital currency mining machine
ActiveCN111966202BAvoid wasting powerGuaranteed computing powerVolume/mass flow measurementPower supply for data processingDigital currencyReliability engineering
Embodiments of the present invention provide a power supply voltage control method and device for a digital currency mining machine, and a digital currency mining machine. The method includes: determining the computing power ratio of the digital currency mining machine based on the actual computing power and theoretical computing power of the digital currency mining machine; generating a power control instruction based on the comparison result of the computing power ratio and a predetermined threshold value; The above power control command controls the output voltage of the power supply of the mining machine. The embodiment of the present invention controls the output voltage of the power supply according to the calculation power ratio, and can achieve a good compromise between the power consumption loss and the calculation power of the mining machine.
Owner:SHENZHEN MICROBT ELECTRONICS TECH CO LTD
Storage space estimation method and device, electronic equipment and readable storage medium
ActiveCN113296689AGuaranteed efficiencyGuaranteed computing powerInput/output to record carriersAlgorithmIndustrial engineering
The embodiment of the invention discloses a storage space estimation method and device, electronic equipment and a readable storage medium. The storage space estimation method comprises the following steps: calculating and estimating a first storage space occupied by result data after data calculation based on sampling sum of data; writing the result data into the estimated first storage space; and writing the residual result data into a second storage space based on the fact that the residual result data is not written after the first storage space is fully written, so that the overall calculation performance is ensured while the space use efficiency is ensured.
Owner:ALIBABA GRP HLDG LTD
Method for controlling energy consumption of cluster system based on changing control mode
InactiveCN101794169BState avoidance changesGuaranteed computing powerPower supply for data processingTransmissionComputer clusterSleep state
The invention discloses a method for controlling energy consumption of a cluster system based on changing a control mode, which carries out system level management and control under the operating condition of a computation task on a computation node in the cluster system through a control node and belongs to the field of computer cluster systems. In the method for controlling the energy consumption, the control node respectively chooses fixed, manual or semi-automatic control modes for the change of a computation task amount and the frequency of the change and controls the energy consumption through controlling the state of a computation node and the computation node enters different sleep states according to an idle rule so as to save the energy consumption of a processor. In the invention, the energy consumption of the cluster system is greatly reduced aiming at parallel computation of the cluster system under the premise of not greatly degrading computation performance and the balance between the performance and the energy consumption is obtained.
Owner:BEIHANG UNIV
A flexible configurable neural network computing unit, computing array and construction method thereof
ActiveCN109409512BGuaranteed computing powerIncrease flexibilityNeural architecturesPhysical realisationComputer hardwareComputer architecture
The invention discloses a flexibly configurable neural network computing unit, a computing array and a construction method thereof. The neural network computing unit includes: a configurable storage module, a configurable control module, and a multiply-add computing module capable of time-division multiplexing; a configurable storage module The modules include: feature map data buffer, step size data buffer and weight data buffer; the configurable control module includes: counter module and state machine module; the multiplication and addition calculation module includes: multiplier and accumulator. The invention can support any type of convolution calculation, and supports parallel calculation of multi-size convolution kernels, fully explores the flexibility and data reusability of convolutional neural network calculation units, greatly reduces system power consumption caused by data migration, and improves The computational efficiency of the system.
Owner:XI AN JIAOTONG UNIV
A CNN processing method and device
ActiveCN105678379BGuaranteed computing powerLower deployment costsProgram controlNeural architecturesApplication-specific integrated circuitProcessing
The invention discloses a CNN processing method and device, which are used for realizing the module operation in the CNN model on the ASIC, and reducing the operation cost while ensuring the calculation performance of the CNN. An embodiment of the present invention provides a CNN processing method, including: obtaining the intensive type of the first module in the CNN model; if the intensive type of the first module is computationally intensive, deploying the first module to a dedicated integration on the circuit ASIC, and obtain the multiple computing resources of the ASIC occupied by the first module; merge the same computing resources among the multiple computing resources of the ASIC occupied by the first module, and obtain the deployment The first module on the ASIC after completing the combination of computing resources; running the first module deployed on the ASIC after completing the combining of computing resources.
Owner:TENCENT TECH (SHENZHEN) CO LTD
Features
- R&D
- Intellectual Property
- Life Sciences
- Materials
- Tech Scout
Why Patsnap Eureka
- Unparalleled Data Quality
- Higher Quality Content
- 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com