A problem in pattern recognition is to find the maximum sum over ail rectangular subregions of a given (n X n) matrix of real numbers. The problem has one-dimensional (1D) and two-dimensional (2D) versions. For the 1D...
详细信息
A problem in pattern recognition is to find the maximum sum over ail rectangular subregions of a given (n X n) matrix of real numbers. The problem has one-dimensional (1D) and two-dimensional (2D) versions. For the 1D version, it is to find the maximum sum over all contiguous subvectors of a given vector of n real numbers. We give an algorithm for the 1D version running in O(log n) time using O((n)/(log n)) processors on the EREW PRAM, and an algorithm for the 2D version which takes O(log n) time using O((n(3))/(log n)) processors on the EREW PRAM.
There is a growing interest in designing parallel algorithms on restricted classes of graphs. Some optimal parallel algorithms are presented on circular-arc graphs - specifically, optimal parallel algorithms for find...
详细信息
There is a growing interest in designing parallel algorithms on restricted classes of graphs. Some optimal parallel algorithms are presented on circular-arc graphs - specifically, optimal parallel algorithms for finding (unweighted) maximum independent set, minimum clique cover, and minimum dominating set in a unified way. Given the sorted array of endpoints of the arcs in the intersection model of the circular-arc graph, all these problems can be solved sequentially in O(n) time where n is the number of arcs. All the parallel algorithms run in O(log n) time on an (n/log n)-processor EREW PRAM (parallel random access memory). Maximum cliques, maximum independent sets, and minimum dominating sets are finding applications in project scheduling, cluster analysis, parallel processing, and facilities location.
Orbital-free kinetic energy density functionals provide the means to accurately simulate the behavior of simple sp-band metals, on a length scale currently unparalleled by other ab initio methods. We present some tech...
详细信息
Orbital-free kinetic energy density functionals provide the means to accurately simulate the behavior of simple sp-band metals, on a length scale currently unparalleled by other ab initio methods. We present some technical aspects of the efficient parallel implementation of kinetic energy functionals, including a functional with a recently developed density-dependent response kernel that provides a good description of metal surfaces. We further illustrate the ability of this new functional to treat finite metallic systems by examining the metal-insulator transition in a 2-dimensional array of metal quantum dots. (C) 2000 Elsevier Science B.V. All rights reserved.
In this paper, we propose efficient parallel algorithms on the EREW PRAM for optimally locating in a tree network a path-shaped facility and a tree-shaped facility of a specified length. Edges in the tree network have...
详细信息
In this paper, we propose efficient parallel algorithms on the EREW PRAM for optimally locating in a tree network a path-shaped facility and a tree-shaped facility of a specified length. Edges in the tree network have arbitrary positive lengths. Two optimization criteria are considered: minimum eccentricity and minimum distancesum. Let n be the number of vertices in the tree network. Our algorithm for finding a minimum eccentricity location of a path-shaped facility takes O(log n) time using O(n) work. Our algorithm far finding a minimum distancesum location of a path-shaped facility takes O(log n) time using O(n(2)) work. Both of our algorithms for finding the minimum eccentricity location and a minimum distancesum location of a tree-shaped facility take O(log n log log n) time using O(n) work. In the sequential case, all the proposed algorithms are faster than those previously proposed by Minieka. Recently, Peng and Lo have proposed parallel algorithms for all the four problems considered in this paper. They assumed that each edge in the tree network is of length 1. Thus, as compared with their algorithms ours are more general. Besides, our algorithms for the problems of finding a minimum eccentricity location of a path-shaped facility, the minimum eccentricity location of a tree-shaped facility, and a minimum distancesum location of a tree-shaped facility are more efficient from the aspect of work. Their algorithms far these three problems use O(n log n) work. Ours use O(n) work. (C) 2000 Academic Press.
We present 0(log n) time algorithms in the EREW PRAM model, using n /log n processors, to find cut vertices, bridges, and blocks (often called biconnected components) of an interval graph having n vertices. It is assu...
详细信息
We present 0(log n) time algorithms in the EREW PRAM model, using n /log n processors, to find cut vertices, bridges, and blocks (often called biconnected components) of an interval graph having n vertices. It is assumed the interval graph is represented by an interval model, with ends presorted. If the ends are not presorted, our algorithms, preceded by an optimal sort, form an 0(log n) time algorithm using n processors, which is shown to be optimal. The algorithms rely heavily on the parallel prefix algorithm.
The Bethe-Salpeter eigenvalue problem is a dense structured eigenvalue problem arising from discretized Bethe-Salpeter equation in the context of computing exciton energies and states. A computational challenge is tha...
详细信息
The Bethe-Salpeter eigenvalue problem is a dense structured eigenvalue problem arising from discretized Bethe-Salpeter equation in the context of computing exciton energies and states. A computational challenge is that at least half of the eigenvalues and the associated eigenvectors are desired in practice. We establish the equivalence between Bethe-Salpeter eigenvalue problems and real Hamiltonian eigenvalue problems. Based on theoretical analysis, structure preserving algorithms for a class of Bethe-Salpeter eigenvalue problems are proposed. We also show that for this class of problems all eigenvalues obtained from the Tamm-Dancoff approximation are overestimated. In order to solve large scale problems of practical interest, we discuss parallel implementations of our algorithms targeting distributed memory systems. Several numerical examples are presented to demonstrate the efficiency and accuracy of our algorithms. (C) 2015 Elsevier Inc. All rights reserved.
A string is square-free if it does not contain a nonempty subword of the form ww. We give an algorithm testing square-freeness of strings in log n time with n processors of a CRCW PRAM. The input alphabet is not bound...
详细信息
A string is square-free if it does not contain a nonempty subword of the form ww. We give an algorithm testing square-freeness of strings in log n time with n processors of a CRCW PRAM. The input alphabet is not bounded. The best sequential time algorithm for this problem takes O(n log n) time. Hence the total number of operations in our parallel algorithm matches that of the best sequential algorithm. The algorithm relies on an efficient parallel computation of a factorization of words used in text compression.
The anomalous structure in the magnetic field of the Ural Region has been studied in the segment bounded by 52A degrees-64A degrees N and 54A degrees-66A degrees E. Analytical apparatus for upward continuation of airb...
详细信息
The anomalous structure in the magnetic field of the Ural Region has been studied in the segment bounded by 52A degrees-64A degrees N and 54A degrees-66A degrees E. Analytical apparatus for upward continuation of airborne magnetic data to different heights was applied. To recalculate magnetic field, parallel algorithms and software for multiprocessor computers were used. Maps of magnetic anomalies for different ranges of wave lengths showing the distribution of magnetization in the layers of the Earth's crust were built.
The article presents a software centric approach, in which ease of programming is a first priority for both uniprocessors and multiprocessors. Here, The author outline two concrete reasons and one general reason why p...
详细信息
The article presents a software centric approach, in which ease of programming is a first priority for both uniprocessors and multiprocessors. Here, The author outline two concrete reasons and one general reason why parallel programs could give a gain in performance over serial code on uniprocessors, especially with the current trends in uniprocessor architecture. Recent research suggests that parallelism in computer program is a resource that can be used to improve performance on modern uniprocessor computers as well as on parallel computers. Furthermore, exploration of enhancements to uniprocessor organization methods to take advantage of parallel programming will contribute to the more difficult problem of making a transition from the current situation, in which serial programming and single processor computer systems predominate, to having ubiquitous parallel computer systems and, of course, parallel programming. Performance enhancement is a fundamental concern in computer science and engineering.
This correspondence introduces scalable data parallel algorithms for image processing. Focusing on Gibbs and Markov random field model representation for textures, we present parallel algorithms for texture synthesis,...
详细信息
This correspondence introduces scalable data parallel algorithms for image processing. Focusing on Gibbs and Markov random field model representation for textures, we present parallel algorithms for texture synthesis, compression, and maximum likelihood parameter estimation, currently implemented on Thinking Machines CM-2 and CM-5. Use of fine-grained, data parcel processing techniques yields real-time algorithms for texture synthesis and compression that are substantially faster than the previously known sequential implementations. Although current implementations are on Connection Machines, the methodology presented here enables machine-independent scalable algorithms for a number of problems in image processing and analysis.
暂无评论