37 results about "Parallel software" patented technology

A method is described for dynamic stitching of a new module of executable code in a parallel processing environment, where access to a data object is shared by the new module and another module previously loaded. A new data object is created for shared access by the new module and by the other module; a data freshness indicator is updated in accordance therewith. A pointer value for the data pointer associated with the other module is modified, thereby terminating reference to an old data object previously accessed and substituting reference to the new data object. A second data freshness indicator is updated in accordance with access by the other module to the new data object. The old data object is deleted when a comparison between freshness indicators shows that access to the old data object is no longer required.

This invention teaches a way of implementing formally verified massively parallel programs, which run efficiently in distributed and shared-memory multi-core chips. It allows programs to be developed from an initial abstract statement of interactions among parallel software components, called cells, and progressively refine them to their final implementation. At each stage of refinement a formal description of patterns of events in computations is derived automatically from implementations. This formal description is used for two purposes: One is to prove correctness, timings, progress, mutual exclusion, and freedom from deadlocks/livelocks, etc. The second is to automatically incorporate into each application a Self-Monitoring System (SMS) that constantly monitors the application in parallel, with no interference with its timings, to identify and report errors in performance, pending errors, and patterns of critical behavior. This invention also teaches a way of organizing shared-memory for multi-processors that minimizes memory interference, protects data and increases execution efficiency.

A passive tracking system is provided with a plurality of ultrawideband (UWB) receivers that is asynchronous with respect to a UWB transmitter. A geometry of the tracking system may utilize a plurality of clusters with each cluster comprising a plurality of antennas. Time Difference of Arrival (TDOA) may be determined for the antennas in each cluster and utilized to determine Angle of Arrival (AOA) based on a far field assumption regarding the geometry. Parallel software communication sockets may be established with each of the plurality of UWB receivers. Transfer of waveform data may be processed by alternately receiving packets of waveform data from each UWB receiver. Cross Correlation Peak Detection (CCPD) is utilized to estimate TDOA information to reduce errors in a noisy, multipath environment.

A method and associated systems for dynamic container deployment with parallel layers. A processor creates an image file by processing each instruction of an image-container file in sequential order, where each instruction adds a layer to the image file that identifies either an unconditional or a parallel software layer of a container. Each unconditional layer is assigned a unique level but parallel layers may share a same level. The processor then deploys an instance of a container by deploying layers of the image file to the container in the order in which they were created by the image-creation file. If an image-file layer is marked as parallel, it is deployed only if an associated condition is satisfied, such that no two deployed layers have the same level.

A non-intrusive debugging framework for parallel software based on a super multi-core framework is composed of a plurality of core clusters. Each of the core clusters includes a plurality of core processors and a debug node. Each of the core processors includes a DCP. The DCPs and the debug node are interconnected via at least one channel to constitute a communication network inside each of the core clusters. The core clusters are interconnected via a ring network. In this way, the memory inside each of the debug nodes constitutes a non-uniform debug memory space for debugging without affecting execution of the parallel program, such that it is applicable to current diversified dynamic debugging methods under the super multi-core system.

The invention discloses a parallel programming model system of DAG oriented data driving type application and a realization method. The parallel programming model system of DAG oriented data driving type application comprises a DAG mode chart module, a user application program module and a DAG runtime system module, wherein the DAG mode chart module comprises a DAG mode chart bank; the user application program module is used for user initialization setup and confirmation of specific parallelization algorithm; and the DAG runtime system module comprises a main thread and a thread pool. The main thread is used for analyzing and updating the DAG mode chart, issuing and dispatching data block, and controlling fault-tolerance; the thread pool comprises a thread pool queue buffer zone and a work thread, wherein the thread pool queue buffer zone is a data interface for the main thread and the work thread, the work thread is used for constantly acquiring computing tasks from the queue buffer zone and carrying out computing. Compared with the prior art, the invention decreases difficulty in designing and developing a parallel computing application program for non-computer professionals, shortens development period of parallel software, and enables the completed parallel computing application program to have more reasonable structure and more optimized performance.

The invention relates to the technical field of parallel computing, in particular to a parallel computing component and method and a corresponding parallel software development method and system. The parallel computing component comprises a parallel computing component interface, a data distribution storage unit, a data parallel communication unit and a numerical parallel computing unit, wherein the parallel computing component interface is used for providing numerical data for other units and feeding back numerical data subjected to numerical computation; the data distribution storage unit is used for storing a data field and the numerical data received from the parallel computing component interface in a private memory space of a progress; the data parallel communication unit is used for copying data fields from private memory spaces of other progresses and storing the data fields to a local progress; the numerical parallel computing unit is used for computing the numerical data of the local progress independently according to a numerical analog computation process of a numerical algorithm. By means of the method and system, the reusability, maintainability and extendibility of a software module can be improved remarkably, and the computing efficiency of application software is improved remarkably.

The invention discloses a Harris corner detecting parallel software system based on the design idea of a Graphics Processing Unit (GPU). The time-consuming image Gaussian convolution smoothing filtering part in calculation is improved to a Single Instruction Multiple Thread (SIMT) mode through a plurality of threads, and the whole process of image corner detecting is finished on a Compute Unified Device Architecture (CUDA) through a shared memory, a constant memory and a lock page internal storage mechanism in the GPU. Corners detected by the software system are distributed evenly, and good effects in the aspects of corner extracting and precise positioning are achieved. Compared with a serial algorithm based on a CPU, a Harris corner detecting parallel algorithm based on the GPU can obtain a speed-up ratio reaching up to 60 times. The executing efficiency of the Harris corner detecting parallel algorithm based on the GPU is improved obviously, and the good real-time processing capacity is shown in the aspect of large-scale data processing.

The invention discloses a communication method suitable for monitoring MPI (Message Passing Interface) parallel software. The method comprises an MPI communication process, an MPI calculation process and host monitoring software, and the MPI communication process comprises an OpenMP command thread and a state thread. The method comprises the following steps: 1, the MPI communication process receives a command of the host monitoring software, and forwards the command to the MPI calculation process in a non-blocking way; 2, the MPI calculation process receives the command transmitted by the MPI communication process, executes the command to control a calculation task, and transmits state information to the MPI communication process regularly or in an event-driving way; 3, the MPI communication process receives the state information of the MPI calculation process, and transmits the state information to the host monitoring software. By adopting the communication method, parallel software consisting of a plurality of MPI processes can be monitored via the host monitoring software, and the aim of hardly influencing the parallel software calculation performance is fulfilled.

A method and a system for identifying and resolving conflicts between design results from a parallel software design. The method includes: receiving a design diagram, wherein the design diagram includes a plurality of nodes and arrows connecting different nodes, with each node indicating a design artifact, and an arrow directed from one node to another node indicating that a design artifact corresponds to the one node depends on a design artifact corresponding to the other node; determining a level of a design artifact in the design diagram, identifying different design artifacts at a given level of the design diagram that depend on a common design artifact, and marking them as isomorphic design artifacts; and outputting a design diagram with the isomorphic design artifacts marked. A conflict between relevant designs are automatically identified in a bottom-up approach according to a software design hierarchy to facilitate conflict resolution.

A method for allocating memory in a parallel processing computing system in which there is first provided a system memory available for parallel processing and first and second threads. The method includes using the first thread to request memory from the system memory; allocating to the first thread a first pool of memory and associating the memory pool with the second thread; using the second thread to request memory from the system memory; allocating to the second thread a second pool of memory and associating the memory pool with the first thread; using the first thread to request further memory from the second thread; and allocating to the first thread a portion of the second pool of memory from the second thread without making a request to the system memory.

The invention provides a method and system for monitoring IB network flow. The method comprises the steps: deploying a server cluster, and configuring an IB network card and a driver for nodes in thecluster; starting an IB networking service, and carrying out IB communication of the cluster server; and running parallel software on the multiple nodes, and executing the monitoring script on the to-be-tested node to obtain the occupation condition of the IB network flow. By running the parallel software on the cluster node, executing the monitoring script on the node and obtaining the real-timeflow values sent and received by the performance counting port, whether the IB network on the current node is occupied or not is judged, the defect that the use condition of the IB network cannot be obtained in the prior art is overcome, and daily maintenance of the cluster network is facilitated.