Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

362results about How to "Improve parallelism" patented technology

Decision model training reasoning method, product, electronic equipment and medium

ActiveCN119647604ANetwork parameter balanceRealize a reasonable splitResource allocationInference methodsDecision modelAlgorithm
The invention discloses a training reasoning method of a decision model, a product, electronic equipment and a medium, and relates to the technical field of artificial intelligence. In the method, when the training reasoning process of the decision model is divided into a plurality of calculation tasks to be distributed to different cluster nodes, the load pressure on the nodes is reduced; and the plurality of data subsets obtained by splitting the input data corresponding to the calculation task are allocated to different calculation nodes, so that the load pressure on the calculation nodes is reduced, the utilization rate of node resources is improved, parallel processing of the data subsets corresponding to the input data is realized, and the efficiency of training reasoning is improved. When the training reasoning process of the decision model is divided into one calculation task, the plurality of data subsets obtained by splitting the input data corresponding to the calculation task are distributed to different calculation nodes, so that the load balance on the calculation nodes is ensured, the load is reasonably distributed to the nodes in the cluster, and the calculation efficiency is improved. And the utilization rate of node resources is improved.
Owner:SHANDONG HAILIANG INFORMATION TECH RES INST

Population mobility OD matrix prediction method

The invention relates to a population mobility OD matrix prediction method, and the method comprises the steps: constructing a plurality of space-time diagrams with different adjacent matrixes according to historical OD flow information and urban region information; performing feature extraction on each time-space diagram by adopting a three-dimensional diagram convolutional network, and performing feature fusion on feature extraction results to obtain time-space extraction results; block embedding is carried out on the time-space extraction result, position coding and time coding are carried out on the result after block embedding processing, and input data are obtained through fusion; a Transform module is adopted to extract long-term time features from input data, and a decoder outputs a final prediction result. Compared with the prior art, the method has the advantage of high prediction accuracy.
Owner:TONGJI UNIV

Multi-rear-end separated memory system and optimization control method thereof

The invention relates to a multi-back-end separated memory system and an optimization control method thereof, the multi-back-end separated memory system comprises an intelligent far memory multi-back-end management control module and a multi-back-end far memory switch, the intelligent far memory multi-back-end management control module analyzes and processes calling of page errors and page switching triggering, and the multi-back-end far memory switch is used for switching the multi-back-end far memory switch. Switching instruction information and parameter regulation and control instruction information are obtained; and the multi-rear-end remote memory exchanger receives and executes the switching instruction information and the parameter regulation and control instruction information, runs applications on the separated memory architecture and the multi-rear-end remote memory software system, and releases resources after running is finished. By supporting a memory exchange strategy and a system architecture of a plurality of heterogeneous far-end separated memory rear ends, parallel access of a plurality of paths of far memories is realized to improve data throughput, application characteristics are analyzed, and intelligent rear end switching and data exchange parameter adjustment are realized; parallel access of a plurality of memory exchange back-end devices, fine-grained parameter configuration of a remote memory data access path, real-time switching of a plurality of memory exchange back-ends and an intelligent control and management method of the remote memory access path can be realized.
Owner:SHANGHAI JIAOTONG UNIV

Federal learning attack prevention method, system and device and storage medium

The invention provides a federal learning attack prevention method, system and device and a storage medium, and belongs to the technical field of privacy computing. The method comprises the following steps: synchronizing clocks of participating nodes through an aggregation node, setting a network delay threshold value, a VDF time parameter and a security parameter, obtaining a public parameter through an initialization function, and issuing related calculation parameters to the participating nodes; randomly generating a cross validation graph, and issuing a model and a front node address to a target node; carrying out local model training by the target node, carrying out VDF calculation by utilizing current model parameters and random numbers generated by the target node and the front node, generating training and calculation result information, and submitting the training and calculation result information to the aggregation node for verification; time verification and VDF verification are carried out on training and calculation result information, and whether model parameters are accepted to participate in aggregation is determined according to verification results; and aggregating the model parameters submitted by the verified nodes by using a weighted average algorithm through the aggregation nodes to form a new global model.
Owner:SHANDONG ZHICHUANG DIGITAL TECH CO LTD

Static random access memory in-memory computing system and method

The embodiment of the invention discloses a static random access memory in-memory computing system and method, and the system comprises a feature memory which is used for storing an input signal of a neural network model; the SRAM storage array is used for storing each weight of the neural network model and performing logical operation on an input signal and each weight to obtain a plurality of logical operation results, each weight is divided into a plurality of binary blocks after CSD coding, complementary mode blocks are reserved after processing of eliminating all-zero mode blocks, and each binary block is divided into a plurality of binary blocks; after information is extracted from each complementary mode block, an index, a symbol and a value are obtained, and each weight is stored in a complementary mode block mode; the metadata storage is used for storing indexes and symbols; the adder tree is used for performing decoding and additive operation on the plurality of logical operation results according to the indexes and the symbols; the shift accumulation unit is used for performing shift accumulation on the addition result; and the post-processing unit is used for post-processing the shift accumulation result to obtain an output signal of the neural network model.
Owner:BEIHANG UNIV

FPGA implementation method of DVB high-speed demodulator

The invention discloses an FPGA (Field Programmable Gate Array) implementation method of a DVB (Digital Video Broadcasting) high-speed demodulator. A digital signal enters a parallel DDC (Direct Digital Control) module and is re-adapted to multiple symbol rates; the parallel AGC module adjusts the power of the digital signal to a fixed power value; the method comprises the following steps: passing through Oamp; m algorithm: completing symbol synchronization; related peaks are extracted, and frame synchronization is completed; correcting the frequency to enable the frequency to reach participation frequency deviation; performing frequency compensation on the digital signal after loop filtering; after error information is calculated and filtered, the digital signal is compensated; and restoring the original information of the digital signal. According to the invention, by adopting the parallel processing module, the real-time performance and the parallelism degree of digital signal processing can be greatly improved, and the performance of the whole demodulator is enhanced. Each module is designed to be adaptive to the change of a digital signal, for example, the AGC module can dynamically adjust the power of the digital signal to ensure that the output digital signal is always kept at a fixed power level so as to adapt to different receiving conditions.
Owner:CHENGDU HAITE INFORMATION TECH CO LTD

Parallel simulation regression method and system based on cluster server

The invention relates to the technical field of simulation regression systems, and provides a parallel simulation regression method and system based on cluster servers. The method comprises the following steps: classifying simulation tasks according to task types, and determining a dependency relationship among a plurality of simulation tasks; generating a parallel simulation task list through a cluster server management tool according to the dependency relationship among the plurality of simulation tasks; and delivering the parallel simulation task list to a cluster server, and running simulation tasks by the cluster server to generate a simulation result. The parallel simulation regression system is conveniently constructed by utilizing the scheduling function of the cluster server, the dependency relationship of various types of tasks can be processed, the degree of parallelism of the simulation regression system is increased to the maximum extent, and the method has universality and can be widely applied to different scenes.
Owner:奕行智能科技(广州)有限公司

Task unloading method in vehicle-mounted edge computing environment based on deep reinforcement learning

The invention relates to the technical field of intelligent Internet of Vehicles, in particular to a task unloading method in a vehicle-mounted edge computing environment based on deep reinforcement learning, which comprises the following steps: acquiring a task vehicle to generate a to-be-executed task, uploading the task to an RSU, generating a scheduling decision according to the current state of the task and available resources, and sending the scheduling decision to the RSU. The task is allocated to a service vehicle or locally executed; performing priority division on the tasks by utilizing an analytic hierarchy process, and obtaining the current state of each to-be-executed task; obtaining service vehicle numbers and task vehicle numbers of the tasks based on task states and priority level utilization model scheduling, and calculating unloading and scheduling strategies; obtaining available computing resources of the current service vehicle; and training the sequence to the sequence model by using an Actor-Cr-it ic algorithm to obtain an optimal task part unloading and scheduling strategy. The computing resources of the service vehicle and the edge server can be fully utilized, the execution time delay of all tasks in the whole time period is smaller, and the task execution success rate is higher.
Owner:SUZHOU VOCATIONAL UNIVERSITY (SUZHOU OPEN UNIVERSITY)

Micro-area variable-angle grazing incidence wide-angle scattering test method

The invention provides a micro-area variable-angle grazing incidence wide-angle scattering test method based on a single-crystal X-ray diffractometer, which comprises the following steps: preparing a sample from a test material deposited on a substrate, and placing the sample on an improved sample rack which is provided with a specific structure and a driver for controlling an incident angle and the like; selecting an X-ray light source, aligning the X-ray light source with the sample, and switching and matching a copper target and a molybdenum target; z-scan scanning and tilting scanning along a light path direction and a direction vertical to the light path are carried out, and related parameters are set; the scattered X-rays are collected through a two-dimensional surface detector; and finally obtaining a correlation graph through software integration. Specific requirements for the table size, scanning parameters and the like of the improved sample holder are met, a height reference position, an adjusting structure and the like are further arranged, and a connecting part is installed on an equipment sample table through screw joint. The problems of inaccurate incident angle and height control and the like in the prior art are solved, and the test accuracy and reliability are improved.
Owner:WESTLAKE UNIV

Multi-process combined injection molding equipment and injection molding method

ActiveCN119658955AGuaranteed parallel movementavoid mobilityInjection productInjection moulding
The invention relates to multi-process combined injection molding equipment and an injection molding method.The injection molding equipment comprises a mold closing assembly, a range finder, a first injection assembly, a gas injection assembly, a second injection assembly, a manipulator and an infrared heater, the mold closing assembly comprises a fixed mold plate and a movable mold plate which are oppositely arranged, and a mold is arranged between the fixed mold plate and the movable mold plate; the range finder is arranged on the mold; the first injection assembly is used for injecting a foaming material of a product main body into the mold, and the first injection assembly communicates with the gas injection assembly; the second injection assembly is used for injecting a second component material into the mold; the manipulator is used for clamping the thermoplastic organic sheet, and the infrared heater is used for heating the thermoplastic organic sheet. Compared with the prior art, integrated lightweight product machining is completed on the same equipment, multiple parts are directly connected when formed in the mold, product transfer and recombination caused by multiple independent procedures are avoided, and the production efficiency and the production quality are effectively improved.
Owner:KRAUSSMAFFEI MACHINERY ZHEJIANG CO LTD

Data processing method and data processing chip

The embodiment of the invention provides a data processing method and a data processing chip, and the method comprises the steps: determining a first operation object corresponding to a first operation unit under the condition that the state of the first operation unit is an idle state, the first operation unit is any one of at least two operation units, and the first operation object corresponds to the first operation unit; the first operation object comprises one of a first node in the first Merkel tree and a target sequence chain of a second node in the first Merkel tree; and operating the first operation object by using the first operation unit to obtain an operation result. Thus, when any operation unit is in the idle state, the nodes or the sequence chains in the Merkle tree serve as the operation objects of the operation unit, the storage space can be reduced, high-parallel operation under the small storage space is achieved, all the operation units are fully utilized, and the operation efficiency is improved. And the parallel effect is improved while the utilization rate of the operation unit is improved, so that the operation efficiency is improved.
Owner:OPEN SECURITY RES INC

Heterogeneous computing architecture test method of ISP module

The invention discloses a heterogeneous computing architecture testing method of an ISP module, and belongs to the technical field of heterogeneous computing. According to the technical scheme, a high-level synthesis tool is used for generating and verifying an IP to be tested; a target algorithm IP application system is constructed by using a microprocessor IP hardcore and an AXI bus connector of a heterogeneous computing platform, a target algorithm IP core is operated by using software to carry out an application test, and actually measured data is stored in a nonvolatile storage medium; transplanting an operation system and an OpenCV library to the heterogeneous computing platform for the development board, and storing a target algorithm software operation result to a data file after a test program is operated; and evaluating the function correctness and the performance index of the target algorithm IP by using the actually measured data and the software operation result. The method has the advantages that effective testing and optimization of the heterogeneous computing architecture of the ISP module are achieved through the software and hardware co-simulation and IP core verification technology, middle-level optimization, a comprehensive testing framework and other methods, and powerful support is provided for technology development, application and popularization in related fields.
Owner:DALIAN NATIONALITIES UNIVERSITY

Square wall washing lamp

The utility model provides a square wall washing lamp which comprises a back plate capable of being installed on a wall face in an attached mode and a shell capable of being installed on the back plate in an attached mode, one or more side walls of the shell are provided with notches, and dimming plates and LED lamps electrically connected with a circuit board in the shell are arranged in the notches. The light emitting surface of the LED lamp is perpendicular to the mounting surface of the shell, the two dimming plates are rotationally mounted on the two sides of the LED lamp, the rotating axis of each dimming plate is perpendicular to the mounting surface of the shell, and the light emitting angle of the LED lamp can be adjusted; and the shell is provided with a toggle switch and / or an inductive switch which are / is electrically connected with the circuit board and are / is used for controlling the LED lamp. The square wall washing lamp is compact in structure, attractive in appearance, small in size, good in using effect and wide in application range.
Owner:NINGBO GAOLI ELECTRONICS TECH CO LTD

Vector processing device, chip, electronic equipment and method

The invention discloses a vector processing device, a chip, electronic equipment and a method, and belongs to the technical field of chips. The vector processing device comprises an instruction buffer, a vector decoding unit, a register unit and a plurality of pipelines, the instruction buffer is used for storing instruction packets to be executed, and each instruction packet comprises a plurality of vector instructions which can be executed in parallel; the vector decoding unit is used for decoding each vector instruction included in the instruction packet to obtain a control signal corresponding to each vector instruction, and broadcasting the control signal corresponding to each vector instruction to the register unit and the plurality of assembly lines in a clock period; a register unit to store data related to the vector instruction; and the assembly line is used for processing data related to the vector instruction according to the control signal corresponding to the vector instruction to obtain a processing result corresponding to the vector instruction. By processing a plurality of vector instructions in one clock period, the processing efficiency of the vector instructions can be remarkably improved.
Owner:TENCENT TECHNOLOGY (SHENZHEN) CO LTD

Instruction processing method and device, processor, storage medium and program product

The invention provides an instruction processing method and device, a processor, a storage medium and a computer program product, the instruction processing method is applied to the processor, the processor comprises a plurality of instruction execution ports, and each instruction execution port transmits an execution instruction in a clock period or is in an idle state. The instruction processing method comprises the following steps: respectively assigning corresponding register units for each execution instruction in a plurality of execution instructions which are transmitted in parallel and need to process the same input parameters; storing input parameters in the register unit; and when the plurality of execution instructions are transmitted in parallel, controlling each execution instruction in the plurality of execution instructions, and respectively acquiring the input parameters from the corresponding register units and processing the input parameters. According to the method and the device, parallel emission of multiple execution instructions can be realized, the concurrency characteristic is better utilized, the instruction parallel performance is improved, the overall consumed time of instruction operation is short, and the efficiency is high.
Owner:JD DIGITS HAIYI INFORMATION TECHNOLOGY CO LTD

Waveguide module and augmented reality equipment

The invention provides a waveguide module and augmented reality equipment. The waveguide module comprises an optical waveguide, the optical waveguide comprises an optical conduction layer and a grating structure, and the grating structure is arranged on the surface of the optical conduction layer; the first protection lenses are stacked and arranged on the side, provided with the grating structure, of the optical waveguide at intervals; the first supporting layer comprises a plurality of first supporting pieces, the plurality of first supporting pieces are arranged on the surface, facing the optical waveguide, of the first protection lens at intervals and connected with the first protection lens, and one end, deviating from the first protection lens, of one part of the plurality of first supporting pieces abuts against the optical conduction layer; one end, deviating from the first protection lens, of the other part of the plurality of first supporting pieces abuts against the grating structure; and the first bonding piece is arranged between the optical waveguide and the first protection lens, surrounds the periphery of the first supporting layer and is used for bonding the optical waveguide and the first protection lens. The waveguide module can improve the Newton ring phenomenon generated under low temperature or external force.
Owner:REALME MOBILE TELECOMM SHENZHEN CO LTD

Multi-node metadata management method for persistent memory file system in NUMA system

The invention discloses a multi-node metadata management method for a persistent memory file system in an NUMA system, and relates to the technical field of novel storage of data centers. File data access is a mixing process of file metadata and file data, for example, in a file creating and deleting process, global information of a file system needs to be accessed so as to update access time, idle data blocks, idle inode and the like of the file system; in the file reading and writing process, a directory structure needs to be traversed, a file metadata inode structure corresponding to a file name is searched, space is distributed, data is copied, file system global information and file metadata inode information are updated, and file reading and writing are completed. In the design of a current NUMA persistent memory file system, the operation of remotely accessing metadata is involved in the file data access process, so that the file read-write performance is reduced. According to the method, a redundant dynamic attribute metadata management mechanism and a centralized directory management mechanism based on the access frequency are adopted, metadata management of the persistent memory file system in the NUMA system is optimized, and the influence of cross-node metadata access on file data access performance is reduced.
Owner:CHONGQING UNIV OF POSTS & TELECOMM

Method, system and device for optimizing ReLee filtering algorithm based on CUDA (Compute Unified Device Architecture) platform

The invention discloses an optimization method, system and device for a ReLee filtering algorithm based on a CUDA platform, and relates to the technical field of signal processing.The optimization method comprises the following steps that a first memory and a second memory are applied in a global memory of a GPU, image data to be filtered in a memory of a CPU is copied to the first memory, and the second memory is copied to the second memory; according to the to-be-filtered image data layout block stored in the first memory, after a CUDA kernel function is started, all threads in the block are allowed to execute the kernel function so as to obtain a filtering result, the filtering result is stored in the second memory, and the filtering result is copied to the memory of the CPU. Through the steps, the optimization method is favorable for improving the calculation efficiency.
Owner:SEA EAGLE DEEP SEA TECH CO LTD +1

High-density near-memory computing and in-memory computing hybrid architecture based on eDRAM (enhanced dynamic random access memory) and computing method

The invention provides an eDRAM-based high-density near-storage calculation and in-storage calculation hybrid architecture and a calculation method. The architecture comprises an eDRAM near-storage and storage calculation hybrid array, a sensitive amplifier, an ADC circuit, a storage calculation result register, a characteristic value updating drive circuit, a controller, a row decoder and a column decoder. According to the calculation method, the architecture is adopted for calculation. By combining a near memory and an in-memory calculation structure, the characteristics of the near memory and the in-memory calculation structure are fully matched to carry out optimization design on an array structure and a basic unit and complement the independent defects of the two architectures, meanwhile, a deep groove capacitor is adopted to replace a plane capacitor, and the advantages of low power consumption, low time delay, high calculation efficiency, high density and the like are achieved; the bandwidth limitation of a traditional near storage circuit and the process limitation of an in-storage computing circuit are effectively overcome, and the method can be applied to a deep learning accelerator computing core and is particularly suitable for edge end hardware deployment with high requirements for the integration level and the power consumption performance.
Owner:HANGZHOU VANGO TECH

A multi-core processing system and its task scheduling method, chip, and storage medium

The embodiment of the present application discloses a multi-core processing system and its task scheduling method, chip, and storage medium. The multi-core processing system includes multiple control cores and multiple processing cores. The multiple control cores are respectively connected to the multiple processing cores. The first control core reads a first instruction and executes the first instruction when it is determined that the first instruction is a control instruction executed by the control core; when the first control core determines that the first instruction is a data processing instruction executed by the processing core, it sends the first instruction to the first processing core for execution. In this way, a one-to-many topological structure is configured between the control core and the processing core according to the processing requirements, so as to improve the parallelism of data processing instruction processing, thereby improving data processing efficiency and reducing chip area.
Owner:GUANGDONG OPPO MOBILE TELECOMMUNICATIONS CORP LTD

Electron gun frame and space traveling wave tube

The present invention relates to the technical field of vacuum electronic devices, and provides an electron gun frame, comprising an insulating porcelain seat, a first anode seat and a second anode seat arranged on the insulating porcelain seat, and a positioning member for fixing the first anode seat and the second anode seat, wherein the insulating porcelain seat comprises an insulating porcelain ring, and at least two positioning block groups evenly distributed along the circumference of the top surface of the insulating porcelain ring, each positioning block group comprises a first positioning block, a second positioning block and a third positioning block, the first positioning block, the second positioning block and the third positioning block are arranged in sequence and equidistantly, the setting heights of the first positioning block, the second positioning block and the third positioning block are successively larger, and each of the first positioning block, the second positioning block and the third positioning block is fixedly connected to the first anode seat, the second anode seat and the positioning member respectively by a clamping method. The present invention solves the technical problem that the miniaturized traveling wave tube in the prior art cannot meet the requirements in terms of weight reduction, processing, assembly and precision.
Owner:山东微波电真空技术有限公司 +1

Polarization code soft input soft output decoding method, decoding device and communication equipment

The invention provides a polarization code soft input soft output decoding method, a decoding device and communication equipment. The decoding method comprises the following steps: (S1) aiming at all special nodes forming a polarization code, fast serial cancellation list decoding under dynamic frozen bit constraint is executed, and the types of the special nodes comprise rate zero, rate one, repetition and single parity check; (S2) after each special node is decoded, calculating the approximation of the posterior probability sum of legal code words which are not accessed under the node; and (S3) utilizing the approximate estimation codebook probability obtained in the step (S2) and calculating bit-level soft information of the output polarization code. Compared with an SO-SCL decoding algorithm, the low-delay soft input and soft output decoding method for the polarization codes, provided by the invention, has the advantages that the degree of parallelism is improved and the decoding delay is remarkably reduced on the premise that the soft output precision is not lost by realizing rapid soft output decoding based on the Rate0, the Rate1, the REP and the SPC nodes.
Owner:SHANGHAI JIAOTONG UNIV

Data processing method of neural network, neural network and chip

PendingCN119272827AReduce high energy consumptionReduce computing energy consumptionArchitecture with single central processing unitPhysical realisationEngineeringTerm memory
The invention relates to a data processing method of a neural network, the neural network and a chip. The data processing method of the neural network comprises the following steps: analyzing a neural network model to an internal graph representation; identifying neural network calculation which needs to be optimized to PIM.PNM; optimizing an external interaction instruction of the PIM.PNM; and generating a calculation instruction of PIM * PNM. The application of PIM and PNM technologies in the neural network has the effects of reducing the data transmission bottleneck, improving the calculation efficiency, optimizing the storage and calculation parallelism, reducing the energy consumption, improving the memory bandwidth utilization rate and better supporting a large-scale model and real-time application.
Owner:WUXI MICRONANO CORE ELECTRONIC TECH CO LTD

Dynamic backlight regulation and control method and system of display device

The invention discloses a dynamic backlight regulation and control method and system for a display device, and the method comprises the steps: forming a grayscale image according to the feature value of each pixel point in a current to-be-displayed original input image; inputting the grayscale image into a preset convolutional neural network encoder model to obtain a compressed feature map; inputting the compressed feature map into a preset coupling diffusion algorithm decoder model to obtain an updated image; adjusting and controlling the backlight of the display based on the gray value of each pixel point of the updated image; according to the method, the neural network method is adopted to extract the compression features of the image backlight, the adaptability of the algorithm to different display scenes is improved, the operation in the expansion stage is not too complex through the coupling of bilinear interpolation and nearest neighbor interpolation in the decoder stage, the relatively reliable and smooth image effect is provided, and the method is suitable for large-scale popularization and application. A ping-pong design and a pipeline architecture design are adopted for video stream processing, the degree of parallelism and the data throughput rate are increased, and the data processing efficiency is improved.
Owner:NANJING UNIV

Data processing method, encryption and decryption algorithm module, chip and electronic equipment

The embodiment of the invention provides a data processing method, an encryption and decryption algorithm module, a chip and electronic equipment, and relates to the technical field of computers, the data processing method is applied to the encryption and decryption algorithm module, and the encryption and decryption algorithm module comprises a transposition distribution sub-module, a transposition reading sub-module, an unpacking and packaging sub-module and a plurality of operation sub-modules. The method comprises the following steps: receiving a plurality of to-be-processed data packets through a transpose distribution sub-module, extracting data from each to-be-processed data packet in sequence, combining the data into one data packet, and sending a plurality of first transpose data packets to an unpacking and grouping sub-module in sequence; splitting the first transposed data packet into a plurality of first sub-data packets through an unpacking and grouping sub-module, and respectively sending the plurality of first sub-data packets to respective corresponding operation sub-modules; performing encryption or decryption processing on the data through the operation sub-module to obtain a processing result corresponding to the to-be-processed data packet; and reading the processing result through the transposition reading sub-module, and generating a processed data packet.
Owner:LOONGSON TECH CORP

Underground garage electronic map drawing method and device based on Internet of Vehicles

The invention provides an underground garage electronic map drawing method and device based on the Internet of Vehicles, and relates to the technical field of the Internet of Vehicles. The method comprises the following steps: detecting the position of a vehicle in response to receiving an underground garage drawing task sent by map maintenance equipment; after it is detected that the vehicle enters the underground garage, first certification information is sent to map maintenance equipment, so that after the map maintenance equipment verifies the first certification information, a data collection instruction is sent to the vehicle-mounted terminal; in response to the received data acquisition instruction, collecting and storing monitoring data obtained by monitoring the vehicle by a steering sensor, a speed sensor and a gyroscope; and after it is detected that the vehicle leaves the underground garage, second proof information and the monitoring data are sent to the map maintenance equipment, so that after the map maintenance equipment verifies the second proof information, an electronic map of the underground garage is drawn according to the monitoring data. In this way, the electronic map of the underground garage can be accurately drawn.
Owner:BEIJING ZHI YANG NORTH INTERNAITONAL EDUCATION TECH CO LTD

RISC-V-based AI processor design method for small convolution kernel acceleration

The invention relates to the technical field of RISC-V instruction set processors, in particular to an RISC-V-based AI processor design method for small convolution kernel acceleration, which comprises the following steps: determining the acceleration demand of a 3 * 3 convolution kernel in a convolutional neural network (CNN), selecting an open architecture RV32I of RISC-V as a basis, defining a core architecture of an AI processor, comprising a calculation unit, a register file and a control unit, the processing of the small convolution kernel is optimized; designing and realizing a vector arithmetic logic unit (VALU) and a convolution operation unit (CUU); the VALU has parallel computing capability and can execute addition, subtraction and multiplication vector operations to support instruction set extension; a convolution arithmetic unit CUU processes 3 * 3 convolution operations, including dedicated hardware logic to perform element-by-element multiplication and accumulation operations, thereby accelerating the convolution calculation process.
Owner:CIVIL AVIATION UNIV OF CHINA

Method, device and system for setting transaction operation dependence based on log sequence

The invention relates to a method, a device and a system for setting transaction operation dependence based on a log sequence. The method mainly comprises the steps that after a target end data synchronization service receives a DML operation of a source end, an MD5 value calculated by the source end in the operation is extracted to serve as a KEY, and then a transaction ID in the operation is extracted to serve as a TRXID; constructing a detection record X by using the currently operated KEY, merging the detection record X into the hash table R, performing conflict detection judgment in the merging process, and returning a TRXID value corresponding to a conflicting KEY if the conflicting KEY exists; otherwise, 0 is returned to serve as the corresponding TRXID, and no conflict exists; and setting the TRXID obtained by conflict detection judgment as a conflict TRXID, writing the initial LSN of the first memory chip in the memory linked list corresponding to the current hash table R into the current operation, and setting the initial LSN as a conflict LSN. According to the method, the dependency relationship between transaction operations can be determined, which transactions need to wait can be determined during synchronous operation, and the synchronization parallelism can be improved to the maximum extent.
Owner:WUHAN DAMENG DATABASE

Kernel driving method and device based on multi-core collaboration and electronic equipment

The invention provides a kernel driving method and device based on multi-kernel collaboration and electronic equipment. The method comprises the following steps: receiving a calling task of a user mode by a service driver; the service driver sends the calling task to a task queue of an IP driver of a corresponding type according to the IP type corresponding to the calling task; selecting a calling task from the task queue as a target task by the IP driver; if the IP corresponding to the target task is of a multi-core structure, the IP driver sends the target task to a corresponding multi-core cooperation unit; selecting, by the multi-core cooperation unit, a target core from the plurality of cores of the IP according to a priority policy; the multi-core cooperation unit schedules the target task to the target core; and executing the target task by the target core. In this way, the task processing capacity of the system is greatly improved, and the overall efficiency is improved.
Owner:FUZHOU ROCKCHIP SEMICON