14results about How to "Reduce scheduling overhead" patented technology

The invention discloses a method and a system for processing a task based on multithreading, the method comprises the following steps of: issuing the task to a task queue corresponding to each thread in parallel through multithreading for arbitration, and accurately updating a control register in combination with configuration queue data in a memory, so that the task processing efficiency is improved. All tasks are serially sent to the execution unit for time-sharing processing, so that the occurrence condition of task gaps caused by untimely interrupt response of the processor can be reduced, the performance loss of the processor caused by null and the like is reduced, the resource utilization rate of arithmetic logic is improved, and the performance of the processor is further improved; the register configuration queue data of the tasks are stored in the memory in advance, so that the execution unit can reach the upper limit of hardware performance under the condition of not increasing hardware resources, and the cost is saved while the performance utilization rate of the execution unit is improved; and the coupling degree of the system and the hardware processor is effectively reduced, and the software scheduling overhead of the processor is reduced.

The invention discloses a process scheduling method and a computer storage medium. The method comprises the following steps: acquiring target interaction feature data, data dependence feature data and running context feature data of a target process, and respectively determining feature dependence intensity indexes of the target interaction feature data, the data dependence feature data and the running context feature data on the target process; determining a comprehensive dependence intensity index corresponding to the target process according to the feature dependence intensity indexes corresponding to the target interaction feature data, the data dependence feature data and the operation context feature data; and obtaining a scheduling overhead proportion of the target process, and adjusting process scheduling parameters of the target process according to the scheduling overhead proportion and the comprehensive dependency intensity index, thereby solving the problems of high scheduling overhead, low resource utilization rate and limited optimization effect of a scheduling method in related technologies, improving the resource utilization rate and the response speed, and improving the user experience. And the overall performance and stability of the system are obviously improved.

The present invention discloses a multi-objective dynamic task scheduling method and system based on an improved ant colony algorithm, the method comprising: splitting a task into a plurality of subtasks, and sorting the subtasks of each task in ascending order according to the priority of each task, to obtain a sequence of tasks to be scheduled; assigning the tasks to be scheduled to each computing node in a sequential scheduling manner, to obtain an initial task-computing node allocation scheme, and to calculate the task execution time and computing cluster load factor; and obtaining an optimal allocation scheme based on an improved ant colony algorithm. The present invention comprehensively considers factors such as the system computing node load capacity, task execution time, and task priority, optimizes the pheromone update strategy and heuristic information in the basic ant colony algorithm to avoid local optimality and achieve rapid convergence, improves the timeliness of task allocation and the utilization rate of computing resources, reduces the cloud computing task scheduling overhead, and realizes global multi-objective task optimization scheduling.

The invention discloses a global nerve drawing method and system based on a programmable rasterization engine, and belongs to the technical field of computer graphics, and the method comprises the steps: at the programmable rasterization engine, analyzing a rasterization descriptor according to a rasterization instruction, and extracting vector microoperation and control parameters; maintaining a task state machine according to the parameters and distributing a control signal, selecting an execution entry from a vector kernel table according to the control signal, and instantiating an operation into a parallel vector thread; in a vector thread execution process, tracking data dependence of a vector register and a synchronization state of a direct memory access unit, executing vector loading / storage operation so as to carry data between the register and an on-chip shared memory according to the data dependence and the synchronization state, and dynamically scheduling vector micro-operation to an execution component so as to complete rasterization calculation; and outputting a result to the neural rendering network to complete global neural rendering. According to the method, the multi-representation neural rendering load can be uniformly and efficiently supported on the AI accelerator, the memory access overhead is remarkably reduced, and the calculation efficiency is improved.

The invention relates to the technical field of video monitoring, in particular to a multi-channel video recording data processing method, electronic equipment and a storage medium, and the method comprises the steps: responding to the received video recording data of each channel, creating a video recording file, and storing the received video recording data in the created video recording file; writing file header information into the beginning part of the video file; writing the received video recording data of each channel into the middle part of the video recording file according to a sequence of writing the frame header information and then writing the frame data; and recording key frame information of the video recording data of each channel at the end part of the video recording file to form a key frame information set. The file header information comprises file verification information and video file information; the frame header information comprises frame header verification information and frame data information, so that the defects of an existing multi-channel video recording data processing method in the aspects of storage, packaging, resource consumption, file management and the like are effectively overcome.

This invention relates to the field of high-performance embedded real-time operating system application technology, and provides a method for task processing and human-machine interface processing on a display console based on VxWorks SMP, including: S1: Building a display console with a computer installed, configuring various control components and a display, and using the VxWorks SMP system to concurrently execute the task processing and human-machine interface processing functions that traditionally require two separate computers on the display console on this computer; S2: Developing and generating system-related project files using VxWorks development tools, and allocating and configuring task processing threads, human-machine interface processing threads, and control response underlying processing tasks, while configuring inter-task cooperation and protection mechanisms; S3: Writing a script program, guided by the VxWorks system, to load the control response underlying processing task binding project, task processing thread project, and human-machine interface processing thread project step by step. This solves the difficulty of concurrent operation of display console task processing and human-machine interface processing in VxWorks SMP.

The application discloses a kind of double host electronic glasses and its image acquisition, processing and display method, belong to image processing technical field, a kind of double host electronic glasses, including: glasses frame, monocular vst perspective unit and the power supply unit compatible with monocular vst perspective unit, the monocular vst perspective unit is equipped with two groups, and two groups monocular vst perspective unit respectively independently handles the image of left and right eyes, the monocular vst perspective unit includes, display component, the display component includes optical module and micro display screen;Sampling component, the sampling component includes corresponding camera module with display component;Soc host unit, the Soc host unit undertakes corresponding monocular image processing task;Realize the frame level synchronous display and the relatively synchronous display of degradable under the premise of not forwarding image data of double eyes, to ensure that electronic glasses can still maintain at least single eye real-time picture output under abnormal condition, high-speed data link is equipped between the Soc host unit.

The present specification relates to the technical field of virtual machines, and provides a virtual machine configuration method and device based on C86 architecture. The method comprises: generating first kernel cluster distribution information according to system information of a host computer; forming an optimization strategy according to the first kernel cluster distribution information and a preset quantization strategy; and obtaining virtual machine configuration information by using the optimization strategy. By the embodiment of the present specification, the scheduling overhead of the virtual machine can be reduced.

The invention discloses a method and an apparatus in a node used for wireless communication. The first receiver is used for receiving a first DCI (Downlink Control Information) group, and the first DCI group comprises at least one piece of DCI; the first transmitter is used for transmitting a first signal, and the first signal carries a first HARQ-ACK bit sequence; wherein, for any DCI in the first DCI group, a first domain included in the DCI is used for indicating the cumulative number of reference cells scheduled by a first type of DCI ending to a current reference cell and a current time interval in a first resource pool, and the first type of DCI is used for indicating the reception of a PDSCH or the release of an SPS PDSCH; the current reference cell is a cell with a default cell index in cells scheduled by any DCI in the first DCI group, the time domain resource occupied by any DCI in the first DCI group belongs to the current time interval, and the number of cells scheduled by at least one DCI in the first DCI group is greater than 1.

The present invention discloses a method for receiving scheduling information, which includes: obtaining downlink control information DCI; determining the scheduling information corresponding to the PUSCH in the DCI according to the mapping relationship between the transmission configuration resources used by the physical uplink shared channel PUSCH and the scheduling information in the DCI. Compared with the prior art, the present invention determines the scheduling information in the DCI by the mapping relationship between the transmission configuration resources used by the UE to transmit the PUSCH and the scheduling information in the DCI, so that the base station can send only one DCI, and realize the scheduling of all UEs with a mapping relationship between the PUSCH transmission configuration resources and the scheduling information in the DCI, thereby reducing the scheduling overhead, reducing the waste of resources, and significantly improving the efficiency of the communication system in scheduling the terminal.

The invention relates to an asynchronous conservative time management method and device based on adaptive task unlocking. The method comprises the steps that a ready queue and a waiting queue are initialized, when a thread in the tth step is idle, if the ready queue is not empty, a first agent group is sequentially taken out from the ready queue according to the interference index priority and executed, and if the first agent group does not meet the execution dependency condition in the tth step, the waiting queue is executed; if so, adding the message dependency group of the first agent in the (t + 1) th step into a waiting queue; if the ready queue is empty, the second agent group is sequentially taken out from the waiting queue according to the dependency index priority, if the message dependency groups of the second agent group all complete the simulation task of the tth step, the simulation task of the second agent group in the (t + 1) th step is executed, and new message dependency is generated, the waiting queue is updated according to the new message dependency. By adopting the method, the simulation operation efficiency can be improved.

A processor and electronic equipment are applied to the field of data processing. The processor includes a scaling factor processing module including a smoothing pre-processing unit configured to receive the first tensor, the second tensor, and a smoothing control parameter, and a processing unit configured to process the first tensor and the second tensor by a plurality of arithmetic logic units and special function units in the processor. Executing smoothing operation of the first tensor and the second tensor to obtain a third tensor and a fourth tensor; the processing unit is configured to determine scaling factors of the third tensor and the fourth tensor and execute matrix multiplication and accumulation operation using the scaling factors based on the scaling factors of the third tensor and the fourth tensor to obtain a matrix multiplication operation result. The processor provides efficient smoothing preprocessing hardware, the processing does not need to be executed in special hardware devices such as a vector core, hardware-level efficient execution of smoothing preprocessing is achieved, hardware resource scheduling overhead and calculation delay are effectively reduced, and hardware execution efficiency is improved.

The embodiment of the invention provides a coding method, a hardware accelerator and a system on chip, and the method comprises the steps: determining a cardinal number frequency corresponding to each to-be-coded cardinal number obtained by mapping each input symbol through a frequency extraction module in the hardware accelerator; generating a target Huffman tree through a code generation module in the hardware accelerator based on the target sorting result of the cardinal number frequency corresponding to each cardinal number to be coded; and generating a compressed data stream through an encoding module in the hardware accelerator based on a target encoding result obtained by dynamically encoding each symbol to be encoded in the input data stream. And constructing a Huffman tree structure according to the target sorting result, so that the high-frequency cardinal number corresponds to the short code and the low-frequency cardinal number corresponds to the long code. And dynamically coding the input data stream based on the target Huffman tree to generate a compressed data stream, thereby realizing rapid coding conversion. The whole coding process is efficiently and cooperatively completed in a hardware accelerator, so that the data transmission and software scheduling overhead is reduced, and the data compression speed and performance are improved.

The invention discloses an automatic charging method for an unmanned vehicle, and relates to the technical field of automatic charging, and the method comprises the following steps: S1, searching a charging station within the maximum mileage of the driving of residual electric quantity; s2, the target vehicle node sends charging request information to the charging station; s3, collecting information of surrounding vehicle nodes which are not added into the charging network; s4, the charging station selectively adds the target vehicle node into a charging station vehicle list; and S5, performing active charging according to the selection information obtained from the charging station. According to the invention, charging demand selection of each vehicle node in the charging network is completed through cooperation among the vehicle nodes on the basis of vehicle node competition, so that the vehicle nodes can select the optimal charging station as much as possible and the common benefits of the vehicle nodes for competition are taken into consideration; and finally, the local overall benefit optimization is achieved, the load balance of the charging station is reduced, and meanwhile, the path overhead between the vehicle nodes can be reduced.