We have developed a parallel software-only processing system for creating real-time video effects such as titling and compositing (e.g., picture-in-picture) using compressed Internet video sources. the system organize...
详细信息
We have developed a parallel software-only processing system for creating real-time video effects such as titling and compositing (e.g., picture-in-picture) using compressed Internet video sources. the system organizes processors into a hierarchy of levels. Processes at each level of the hierarchy can exploit different types of parallelism and coordinate the actions of lower levels. To control the effect, control messages must be distributed to processors in the hierarchy while preserving the independence of each level. this requires a control mechanism that supports efficient delivery of messages to groups of processors, tunable reliability semantics, and recoverable state information. We describe a mechanism that meets these requirements that uses IP-Multicast, the Scalable Reliable Multicast protocol, and the Scalable Naming and Announcement Protocol. We also describe an optimization that provides a flexible framework for linking the control of different aspects of one or more related video effects.
the proceedings contain 155 papers. the special focus in this conference is on End-User Applications of HPCN. the topics include: High performance integer optimization for crew scheduling;simulating synthetic polymer ...
ISBN:
(纸本)3540658211
the proceedings contain 155 papers. the special focus in this conference is on End-User Applications of HPCN. the topics include: High performance integer optimization for crew scheduling;simulating synthetic polymer chains in parallel;real-time signal processing in a collision avoidance radar system using parallel computing;the impact of workload on simulation results for distributed transaction processing;computer simulation of ageing with an extended penna model;the scenario management tool SMARTFED for real-time interactive high performance networked simulations;an HPCN architecture for distributed component-based real-time simulations;airport simulation using CORBA and DIS;intelligent routing for global broadband satellite internet;adaptive scheduling strategy optimizer for parallel rolling bearing simulation;MPI-based parallel implementation of a lithography pattern simulation algorithm;parallelizing an high resolution operational ocean model;weather and climate forecasts and analyses at MHPCC;high performance parallel FEM for solid earth;elastic matching of very large digital images on high performance clusters;data intensive distributed computing;utilizing HPC technology in 3D cardiac modeling;a parallel algorithm for 3D reconstruction of angiographic images;a diffraction tomography method for medical imaging implemented on high performance computing environment;heterogeneous distribution of computations while solving linear algebra problems on networks of heterogeneous computers;modeling and improving locality for irregular problems and parallelization of sparse cholesky factorization on an SMP cluster.
Network of workstations (NOW) is a cost-effective alternative to a multiprocessor system. Here we propose a centralized architecture for parallel query processing on network of workstations. We describe a three-level ...
详细信息
We have developed a parallel software-only processing system for creating real-time video effects such as titling and compositing (e.g., picture-in-picture) using compressed Internet video sources. the system organize...
详细信息
ISBN:
(纸本)0201485621
We have developed a parallel software-only processing system for creating real-time video effects such as titling and compositing (e.g., picture-in-picture) using compressed Internet video sources. the system organizes processors into a hierarchy of levels. Processes at each level of the hierarchy can exploit different types of parallelism and coordinate the actions of lower levels. To control the effect, control messages must be distributed to processors in the hierarchy while preserving the independence of each level. this requires a control mechanism that supports efficient delivery of messages to groups of processors, tunable reliability semantics, and recoverable state information. We describe a mechanism that meets these requirements that uses IP-Multicast, the Scalable Reliable Multicast protocol, and the Scalable Naming and Announcement Protocol. We also describe an optimization that provides a flexible framework for linking the control of different aspects of one or more related video effects.
A first approach to the parallel implementation of an automated visual quality inspection (VQI) system for printed ceramic dishes is described. the VQI system is designed to detect defects on the surface drawings of d...
详细信息
A first approach to the parallel implementation of an automated visual quality inspection (VQI) system for printed ceramic dishes is described. the VQI system is designed to detect defects on the surface drawings of dishes printed automatically by industrial series production. Quality inspection must be accomplished online, so a parallel implementation is proposed to speed up processing. the paper briefly introduces the algorithms employed to accomplish the various steps of quality inspection, focusing on the parallel implementation of such algorithms and the digital signal processor (DSP) platform employed. the results of the experimental tests show classification abilities of the proposed algorithms, as well as performance times for the DSP based hardware platform and parallelization efficiency of VQI algorithms.
A new class of applications is described which requires cooperation among diverse users in multiple data and problem instance domains. the hierarchy of parallelism includes heterogeneity within a single instance of th...
详细信息
A new flexible architecture for high-end television is presented, featuring multiple video window display with high flexibility and programmability. the architecture is characterized by an ensemble of coprocessors, op...
详细信息
A new flexible architecture for high-end television is presented, featuring multiple video window display with high flexibility and programmability. the architecture is characterized by an ensemble of coprocessors, optimized for individual TV functions, which communicate in parallel via a full switching network. this network enables programmable signal flow-graphs. the architecture is task-oriented, enabling flexibility for various picture formats and leaving computational power for remaining tasks when primary tasks have been carried out. A major benefit of the architectural concept is to create SW programmability for video signal manipulation at functional levels and to have cost-effective hardware support for pixel-based video processing.
this article describes a process mapping algorithm for a virtual parallel host. the algorithm is a part of a mapping and virtual machine resource watcher tool: the Network-Analyser. this system collects information ab...
详细信息
ISBN:
(纸本)3540664432
this article describes a process mapping algorithm for a virtual parallel host. the algorithm is a part of a mapping and virtual machine resource watcher tool: the Network-Analyser. this system collects information about the states of the hosts (CPU, memory, network traffic, etc), saves some of this information, and gives the ability to use different mapping algorithms. All of this is accessible by a graphical interface and an API [1]. the algorithm uses boththe current and the average load of the hosts and the network, so that it is able to estimate the execution time of the tasks. the model of the processor and the network operation is based on a “round robin” model. Next, a differential equation describes the transient behaviour of the queues. the running time of regular parallel applications can be estimated by the integration of that equation, taking into account the stochastic load of the hosts and the networks. this algorithm is compared to other algorithms, by making event simulations.
the exploitation of hidden information from large datasets by means of data mining techniques suffers from long response times. We address this problem by using the processing power of workstation clusters and have st...
详细信息
the scheduling problem dealt with in this paper consists in determining a feasible schedule when each job is defined with a ready date, a processing time different on each machine and a deadline. there are setup time ...
详细信息
the scheduling problem dealt with in this paper consists in determining a feasible schedule when each job is defined with a ready date, a processing time different on each machine and a deadline. there are setup time sequence dependent. the parallel machines are composed of 2 types. the first one is characterized with `cheap' machines and the second one with `expensive' machines. the objective function is to find a feasible schedule first and to minimize the cost due to assignment and setup times costs second. the method presented in this paper is based on a three-phase heuristic. the first phase is based on an iterative heuristic, the second on a genetic algorithm and the last one on a branch and bound for post optimization.
暂无评论