89results about How to "Ease of parallel processing" patented technology

A method and apparatus is provided for breaking down computing tasks within a larger application and distributing such tasks across a network of heterogeneous computers for simultaneous execution. The heterogeneous computers may be connected across a wide or local area network. The invention supports mobile agents that are self-migrating and can transport state information and stack trace information as they move from one host to another, continuing execution where the mobile agents may have left off. The invention includes a server component for providing an execution environment for the agents, in addition to sub-components which handle real-time collaboration between the mobile agents as well as facilities monitoring during execution. Additional components provide realistic thread migration for the mobile agents. Real-time stack trace information is stored as the computing tasks are executed, and if over-utilization of the computing host occurs, execution of the computing task can be halted and the computing task can be transferred to another computing hosts where execution can be seamlessly resumed using the stored, real-time state information and stack trace information.

Improved method and apparatus for parallel processing. One embodiment provides a multiprocessor computer system that includes a first and second node controller, a number of processors being connected to each node controller, a memory connected to each controller, a first input / output system connected to the first node controller, and a communications network connected between the node controllers. The first node controller includes: a crossbar unit to which are connected a memory port, an input / output port, a network port, and a plurality of independent processor ports. A first and a second processor port connected between the crossbar unit and a first subset and a second subset, respectively, of the processors. In some embodiments of the system, the first node controller is fabricated onto a single integrated-circuit chip. Optionally, the memory is packaged on plugable memory / directory cards wherein each card includes a plurality of memory chips including a first subset dedicated to holding memory data and a second subset dedicated to holding directory data. Further, the memory port includes a memory data port including a memory data bus and a memory address bus coupled to the first subset of memory chips, and a directory data port including a directory data bus and a directory address bus coupled to the second subset of memory chips. In some such embodiments, the ratio of (memory data space) to (directory data space) on each card is set to a value that is based on a size of the multiprocessor computer system.

Improved method and apparatus for parallel processing. One embodiment provides a multiprocessor computer system that includes a first and second node controller, a number of processors being connected to each node controller, a memory connected to each controller, a first input / output system connected to the first node controller, and a communications network connected between the node controllers. The first node controller includes: a crossbar unit to which are connected a memory port, an input / output port, a network port, and a plurality of independent processor ports. A first and a second processor port connected between the crossbar unit and a first subset and a second subset, respectively, of the processors. In some embodiments of the system, the first node controller is fabricated onto a single integrated-circuit chip. Optionally, the memory is packaged on plugable memory / directory cards wherein each card includes a plurality of memory chips including a first subset dedicated to holding memory data and a second subset dedicated to holding directory data. Further, the memory port includes a memory data port including a memory data bus and a memory address bus coupled to the first subset of memory chips, and a directory data port including a directory data bus and a directory address bus coupled to the second subset of memory chips. In some such embodiments, the ratio of (memory data space) to (directory data space) on each card is set to a value that is based on a size of the multiprocessor computer system.

One embodiment of the present invention provides a switch that includes a transmission mechanism configured to transmit frames stored in a queue, and a queue management mechanism configured to store frames associated with the queue in a number of sub-queues which allow frames in different sub-queues to be retrieved independently, thereby facilitating parallel processing of the frames stored in the sub-queues.

A signal processor is configured to encode a signal in a hierarchy including multiple levels of quality. The signal processor produces a rendition of the signal for at least a first level of quality. The signal processor generates sets of reconstruction data specifying how to convert the rendition of the signal at the first level of quality into a rendition of the signal at a second (higher) level of quality in the hierarchy, potentially leveraging on available reference signals. According to one arrangement, the signal processor utilizes an entropy encoder to encode the reconstruction data. Based on probability distribution information for one or more symbols in each set of reconstruction data and based on probability distribution information and / or other encoding parameters inherited from previous levels of quality, the entropy encoder encodes the reconstruction data into an encoded value or bit string. Using the probability distribution information, an entropy decoder converts the encoded value or bit string back into the reconstruction data.

The invention discloses an interest recommendation method and system based on a user sequence click behavior, and the method comprises the steps: firstly obtaining the historical interaction project data of a user, and forming a historical interaction project sequence of the user; constructing an interest recommendation model; carrying out session division on the historical interaction item sequence of the user by utilizing an interest recommendation model; extracting interests in each session obtained after division, and performing weighting processing on the interests in each session to obtain a session interest sequence of the user; interacting interests among different sessions to obtain a dynamic interaction model among different sessions; and inputting the session interest sequence of the user into a dynamic interaction model between different sessions, and predicting and obtaining a to-be-recommended target project sequence.

According to embodiments, point cloud data transmission method may include encoding point cloud data, encapsulating a bitstream that includes the encoded point cloud data and signaling data into a file, and transmitting the file, the bitstream is stored either in a single track or in multiple tracks of the file, the signaling data include at least one parameter set, and the encoded point cloud data include a geometry bitstream containing geometry data and an attribute bitstream containing attribute data.

The invention discloses an indoor movable positioning and drawing method. The indoor movable positioning and drawing method comprises the following main steps: inputting acquired point cloud data according to requirements; carrying out point cloud primary treatment which specifically comprises point cloud filtering, point cloud coordinate conversion and nonlinear interpolation, and under-sampling; carrying out attitude prediction of a track by adopting a particle filtering method; generating a sub-map, adding a target and carrying out constraint calculation, specifically including generation of a dynamic sub-map, adding a virtual target and an overall target, partial bundling and overall constraint calculation; carrying out overall optimization by utilizing map optimization and reducing accumulation errors caused by frame-by-frame matching; updating a result including updating sub-map data, track data and overall map data generated according to the track; preserving result data. According to the indoor movable positioning and drawing method, a GPS (Global Positioning System) and an IMU (Inertial Measurement Unit) are not used, the data processing speed is rapid and the parallel processing is easy to realize; the drawing precision is high and centimeter-grade precision is realized.

The invention discloses an unmanned aerial vehicle operating parameter automatic acquisition system and an automatic sensing method. The system comprises an operating parameter acquisition module array, a flight parameter acquisition module and a ground station. The method includes steps: acquiring geographic information of each operating parameter acquisition module; extracting operating parameter geographic information characteristics, programming an unmanned aerial vehicle sensing area to obtain a specific regional planar graphic equation of appropriate automatic sensing operation time; acquiring a flight path; pre-judging an unmanned aerial vehicle flight path to obtain a path equation; obtaining an intersection path on a canopy plane according to the path equation and the planar graphic equation, and taking a midpoint coordinate of the intersection path as a sampling target point; making sampling strategies, determining sampling time, acquiring effective operating parameters in a sampling time range, and sending to the ground station to realize processing. The unmanned aerial vehicle operating parameter automatic acquisition system and the automatic sensing method have advantages that simultaneous data acquisition and automatic determination of sampling time are realized, and errors caused by manual visual judgment under different flight conditions are avoided.

The present invention extends to methods, systems, and computer program products for automated statistical analysis job chunking. A computer system provides an interface for a user to submit job requests which pair a script with a query. The computer system and a batch module interoperate with one another to process the job requests and return the job results to the user. The computer system can query the batch module to understand computational resource capability and availability. The computer system can also partition the larger parent job into smaller job chunks for the purpose of multi-threading and to facilitate concurrent parallel processing.

The invention discloses a vehicle-mounted laser point cloud classification method. The method includes the steps of inputting three-dimensional laser point cloud data; utilizing GPS time and a running track, and performing partitioning processing on the three-dimensional laser point cloud data; then performing preprocessing; performing preliminary classification on preprocessed data blocks; and performing secondary classification on the data blocks after preliminary classification and outputting laser point cloud data after classification. The vehicle-mounted laser point cloud classification method provided by the invention is high in classification accuracy, easy to perform parallel processing and fast in processing speed.

The invention discloses a rapid data correlation method based on angle cotangent values, which mainly solves the problem of slow correlation speed and low correlation precision rate in the measuration data of the existing passive multisensor. The method of the invention comprises the following steps: building statistic by using the angle cotangent values, and building a candidate correlation set by partitioning pretreatment; simplifying the candidate correlation set by azimuth angle and pitch angle measuring which adopt repeatedly random iterations; and finally picking out correct correlation combine through an indicator function and a cost function. The invention directly adopts the angle information for data correlation, avoids the conversion from angle to distance; the operation efficiency is obviously higher than that of the traditional method; and the invention has excellent engineering application value, can be applied to the fields such as infrared guidance, intelligence fusion war, air traffic control and aerospace, aviation, navigation and the like.

The present invention provides a method for star point extracting of a navigational star in cloud environment in daytime, and the method comprises firstly, respectively performing the following operation on each frame of star map of a plurality of frames of daytime cloud star maps which are consecutively taken: (1) using an improved morphological TopHat operator to filter an original start map to obtain a first star map; (2) using a RobinsonGuard filter to filter the first star map to obtain a second star map; (3) using an adaptive threshold method for threshold cutting of the second star map, and storing coordinates and grayscales of all pixels higher than a threshold; and (4) using a clustering method for four-connected domain star point extracting; secondly using a multiframe comparison technique to eliminate pseudo-star points in the connected domain extracted from each frame of star map; and thirdly, calculating the centre of mass of each successfully-compared connected domain to complete the star point extraction. The method can effectively eliminate the interference of daytime sky clouds and other backgrounds, meanwhile avoids mistaken extraction caused by dust and other air floating objects, and ensures the effective and stable star point extraction.

An image processing apparatus includes a division unit configured to divide an input image into a plurality of subblocks subjected to quantization control a subblock quantization parameter calculation unit configured to calculate a quantization parameter of each of the subblocks, a basic block quantization parameter calculation unit configured to set a basic block including at least two subblocks and to calculate a quantization parameter of the basic block, a difference value calculation unit configured to calculate a difference value between the quantization parameter of the basic block and the quantization parameter of each subblock included in the basic block, and a difference value coding unit configured to code the difference value.

Systems, methods, and computer program products for validating billing data and detecting anomalies in billing data are provided. In one embodiment a method is provided, the method comprising: receiving historical billing data for a customer, the historical billing data organized into a plurality of historical data sets; calculating a plurality of statistical representations of each of the plurality of historical data sets; generating a historical profile for the customer based on the plurality of statistical representations of the historical billing data; receiving current billing data for the customer; generating a current profile for the customer; comparing the current profile to the historical profile, the current profile and the historical profile being associated with the same at least one category attribute; and, based at least in part on the result of the comparison, determining whether one or more anomalies are present in the current billing data for the customer.

A computer-based method and apparatus are provided for sourcing persons, organizations or computers capable of executing the tasks required to produce a product or an outcome, and for scheduling the requisite tasks. The computer-implemented method executes a desired workflow or provides a desired workflow deliverable. The method may include entering into workflow management software a desired workflow or a desired workflow deliverable, the workflow management software then defines two or more requisite tasks, and a task executor for each of the tasks by reference to a database of task executors. A task executor may be a human or a computer. The workflow management software then directs the execution of the two or more requisite tasks by the task executor.

A control system for a switched reluctance (SR) machine is provided. The control system may include a converter circuit that is operatively coupled to the SR machine, and a controller in communication with the converter circuit. The controller may be configured to execute two or more processes in parallel, wherein the processes include generating a torque command based on one or more of bus voltage, machine current, rotor speed and rotor position, determining a first set of current control parameters based on the torque command and the rotor speed, determining a second set of current control parameters based on one or more of the torque command, the rotor speed and the rotor position, selecting one of the first and second sets of current control parameters based on the rotor speed, and operating the gates according to the selected set of current control parameters.

The invention discloses a large-scale monitoring video stream converging method based on stream-oriented computation. The method comprises the following four steps: I, acquiring information about all cameras, and performing ranking according to importance degrees, required central processing unit (CPU) resources and Internet protocol (IP) addresses; II, acquiring information about Workers, and performing ranking according to CPU computing capabilities and IP addresses; III, selecting a properly-identified Worker computing container to serve as a startup position of a receiver according to a computing resource chain table; and IV, estimating resource computing capabilities of selected Workers. According to the method, the complexity in inputting data sources and the computing capabilities of the Workers are considered fully, and the video stream startup position is computed reasonably, so that data are executed locally as much as possible; the CPU utilization ratio of a system is increased; and the real-time computing capability of the system is enhanced. The method has a high practical value and a good application prospect in the technical field of parallel real-time processing of massive video data.

An image processing apparatus includes a division unit configured to divide an input image into a plurality of subblocks subjected to quantization control a subblock quantization parameter calculation unit configured to calculate a quantization parameter of each of the subblocks, a basic block quantization parameter calculation unit configured to set a basic block including at least two subblocks and to calculate a quantization parameter of the basic block, a difference value calculation unit configured to calculate a difference value between the quantization parameter of the basic block and the quantization parameter of each subblock included in the basic block, and a difference value coding unit configured to code the difference value.

The invention discloses an FCT-FDM forward simulation method based on a dispersion viscosity wave equation. A propagation characteristic of a dispersion viscosity wave is researched from the perspective of the numerical simulation. Calculating steps of FCT-FDM is stated, a fluid saturation model is designed, the method is utilized to simulate propagation of the dispersion viscosity wave, and comparing and analyzing are carried out on a numerical result and a sound wave simulation result. The result indicates that relative to a sound wave, the dispersion viscosity wave is obviously attenuated on the amplitude, and the phase position is notably changed. The attenuation degree of the dispersion viscosity wave mainly depends on a dispersion attenuation coefficient and a value of the dispersion attenuation coefficient. According to the FCT-FDM forward modeling method based on the dispersion viscosity wave equation, an effective tool is provided for the numerical simulation of a seismic wave, in particular to a medium containing fluid. The FCT-FDM forward modeling method based on the dispersion viscosity wave equation can be used for describing the attenuation and the phase-shift characteristic of the seismic wave in the fluid medium. In addition, the method is easy to achieve, and the operability is high.

One embodiment of the present invention provides a switch that includes a transmission mechanism configured to transmit frames stored in a queue, and a queue management mechanism configured to store frames associated with the queue in a number of sub-queues which allow frames in different sub-queues to be retrieved independently, thereby facilitating parallel processing of the frames stored in the sub-queues.

The invention discloses a real-time video stitching algorithm based on optimal stitching line updating, and belongs to the technical field of digital image processing. According to the real-time videostitching scheme provided by the invention, an image registration technology based on feature points is used, and internal and external parameters of each camera are estimated through matching feature points of different view angle images. When image fusion is carried out between different visual angles, an image fusion algorithm based on an optimal suture line is adopted, and a background elimination algorithm based on KNN and a suture line updating algorithm based on dynamic programming are used for avoiding discontinuity and ghosting caused by the fact that a moving object penetrates through the suture line in order to solve the problems of ghosting and discontinuity possibly generated by the moving object. And finally, splicing seams are eliminated by using an image fusion algorithm based on a convolution pyramid. By adopting the video stitching scheme designed by the invention, a good large-view-field image can be provided under the condition of ensuring the stitching real-time performance.

In order to reduce the number of data transfers and to increase parallel processing of decryption processing and authentication processing, an encryption processing apparatus is provided that includes an input / output data that processes input / output data to an encryption / decryption processing unit and an authentication processing unit, where the input / output data processing unit calculates a parameter used by the authentication processing unit from input data to the input / output data processing unit and forms input data to the authentication processing unit from the calculated parameter or a parameter calculated from data processed by the encryption / decryption processing unit and the input data to the input / output data processing unit.

The invention discloses an image enhancing method based on frequency domain and belongs to the technical field of image enhancement methods. An image is regarded as a two-dimensional signal. Signal enhancement based on two-dimension Fourier transform is performed on the two-dimensional signal. By adopting the lowpass filtering method, i.e., just allowing low-frequency signals to pass, noise in pictures can be removed. By adopting the highpass filtering method, high frequency signals on edges can be increased, and vague pictures can become clear. The method comprises steps of 1) obtaining an original image; 2) switching images in the space of the original image to frequency domain space by means of the Fourier transformation; 3) using transition function to process images in the frequency domain space; and 4) after images in the frequency domain space is processed via the transition function, enhancing obtained images via inverse transformation. The image enhancing method is used for image processing.

The invention belongs to the technical field of heterogeneous multi-processor computing, and particularly relates to a heterogeneous platform task scheduling method and system based on Q learning, which take all tasks as a Q learning state space, take a processor set as an action space, and wait for an allocated task as a current state. The method comprises the following steps: obtaining a task initial mapping scheme according to execution time required for mapping a task to an action space in Q learning; creating a genetic algorithm model, carrying out fitness evaluation on the task initial mapping scheme, setting individuals copied to the next generation of population in the genetic algorithm model according to fitness, carrying out crossover variation on reserved individuals, and determining new population optimization efficiency and a minimum threshold value; obtaining an approximately optimal solution mapped from a task to a processor in the model according to the genetic algorithm model; converting the model approximate optimal solution into ant colony information initial information distribution, and iteratively searching and outputting an optimal path through an ant colonyalgorithm according to the information distribution to obtain a task scheduling optimal scheme, so as to better improve the performance of the heterogeneous platform.

The invention relates to a data communication circuit arranged among a plurality of processors and adopting a data driving mechanism. The data communication circuit arranged among the plurality of processors and adopting the data driving mechanism comprises a crossbar switch matrix, four data stream memorizers and four match circuits. As shown in the figure, when data transfer is sent, a source CPU sends data to the crossbar switch matrix to achieve data transmission according to a preset protocol through a match circuit which is connected with the source CPU. When data are received, source data are transmitted to a corresponding output port of the crossbar switch matrix after the source data pass through the crossbar switch matrix to carry out path selection, the output data enter a data flow memory (DFM), and function numbers and the data are extracted and are transmitted to corresponding goal CPUs after completeness detection of the DFM is carried out on the data. The data communication circuit arranged among the plurality of processors and adopting the data driving mechanism can well solve the problem of a communication bottleneck which appears in a multi-core structure.

The invention discloses a three-dimensional imaging method for a multi-transmitting and multi-receiving synthetic aperture radar. The method includes the following steps: carrying out two-dimensional imaging on echo data collected by a multi-transmitting and multi-receiving array at all scanning positions; projecting two-dimensional imaging results which are obtained after processing and correspond to all the scanning positions to three-dimensional space; and adding three-dimensional imaging results which are obtained after processing and correspond to all the scanning positions after processing. The three-dimensional imaging method of the invention can avoid calculating Fourier transform of measurement of a multi-transmitting and multi-receiving synthetic aperture radar imaging system along a scanning direction, can perform multi-transmitting and multi-receiving array scanning and imaging processing at the same time, is suitable for the multi-transmitting and multi-receiving synthetic aperture radar imaging system with a multi-transmitting and multi-receiving array scanning track being a straight line or a curve, which means that an equivalent antenna aperture is a plane or a curved surface, reduces the calculation amount and is beneficial to parallel processing.