Browsing by Subject "Parallel Processing"
Now showing 1 - 5 of 5
Results Per Page
Sort Options
Item Open Access Efficient parallel digital signal processing algorithms for hypercube-connected multicomputers(1992) Derviş, ArgunIn this thesis, efficient parallelization of Digital Signal Processing (DSP) algorithms, (FFT, FHT and FCT), on multicomputers implementing the hypercube interconnection topology are investigated. The proposed algorithms, maintain perfect load-balance, minimize communication overhead, can overlap communications with computations and achieve regular computational patterns. The proposed parallel algorithms are implemented on Intel’s iPSC/2^ hypercube multicomputer with 32 processors. High efficiency and almost linear speedup values are obtained for even small size problems.Item Open Access Highly parallel execution of production systems: A model, algorithms and architecture(Springer, 1992) Oflazer, K.This paper presents a new parallel processing scheme called DYNAMIC-JOIN for OPS5-like production systems along with associated parallel algorithms, a parallel architecture and simulation results from a number of production systems. The main motivation behind DYNAMIC-JOIN is to reduce the variations in the processing time requirements and improve limited production level parallelism. For this, the model employs some redundancy that allows the processing of a production to be divided into units of small granularity each of which can be processed in parallel. As a consequence in addition to production level parallelism where a set of relevant productions are processed in parallel, a second level of parallelism can be exploited. After a detailed description of the model proposed, the paper presents algorithms for processing productions with DYNAMIC-JOIN, along with a discussion of various issues and possible disadvantages. Subsequently, the paper presents a parallel processor architecture that can implement DYNAMIC-JOIN, along with simulation results from real production systems. © 1992 Ohmsha, Ltd. and Springer.Item Open Access MARS: a tool-based modeling, animation and parallel rendering system(1992) Aktıhanoğlu, MuratA b stract: This thesis describes a system for modeling, animating, previewing and rendering articulated objects. Tl^^ system has a modeler which models objects, consisting of joints and segments. The animatoi- interactively positions the articulated object in its stick, control vertex or rectangular prism representation into the keyframes, interpolates inbetweens and previews the motion in real time. Then the data representing the motion and the models is sent to a multicomputer {iPSC/2 Ilypercube^). The frames are rendered in parallel by distributed processing techniques, exploiting the coherence between successive frames, thus cutting down the rendering time significantly. The main aim of this research has been to make a detailed study on rendering of a sequence of 3D scenes. The results show that due to an inherent correlation between the 3D scenes, a much more efficient rendering than the conventional sequential one can be done.Item Open Access Parallel mapping and circuit partitioning heuristics based on mean field annealing(1992) Bultan, TevfikMoan Field Annealinp; (MFA) aJgoritlim, receñí,ly proposc'd for solving com binatorial optimization problems, combines the characteristics of nenral networks and simulated annealing. In this thesis, MFA is formulated for tlie mapping i)roblcm and the circuit partitioning problem. EHicient implementation schemes, which decrease the complexity of the proposed algorithms by asymptotical factors, are also given. Perlormances of the proposed MFA algorithms are evaluated in comparison with two well-known heuristics: simulated annealing and Kernighan-Lin. Results of the experiments indicate that MFA can be used as an alternative heuristic for the mapping problem and the circuit partitioning problem. Inherent parallelism of the MFA is exploited by designing efficient parallel algorithms for the proposed MFA heuristics. Parallel MFA algorithms proposed for solving the circuit partitioning problem are implemented on an iPS(J/2’ hypercube multicompute.r. Experimental results show that the proposed heuristics can be efficiently parallelized, which is crucial for algorithms that solve such computationally hard problems.Item Open Access Parallelization of an interior point algorithm for linear programming(1994) Simitçi, HüseyinIn this study, we present the parallelization of Mehrotra’s predictor-corrector interior point algorithm, which is a Karmarkar-type optimization method for linear programming. Computation types needed by the algorithm are identified and parallel algorithms for each type are presented. The repeated solution of large symmetric sets of linear equations, which constitutes the major computational effort in Karmarkar-type algorithms, is studied in detail. Several forward and backward solution algorithms are tested, and buffered backward solution algorithm is developed. Heurustic bin-packing algorithms are used to schedule sparse matrix-vector product and factorization operations. Algorithms having the best performance results are used to implement a system to solve linear programs in parallel on multicomputers. Design considerations and implementation details of the system are discussed, and performance results are presented from a number of real problems.