Introduction to Parallel Computing: Design and Analysis of AlgorithmsBenjamin/Cummings Publishing Company, 1994 - 597 Seiten Take advantage of the power of parallel computers with this comprehensive introduction to methods for the design, implementation, and analysis of parallel algorithms. You'll examine many important core topics, including sorting and graph algorithms, discrete optimization techniques, and scientific computing applications, as you consider parallel algorithms for realistic machine models. Features: presents parallel algorithms as a small set of basic data communication operations in order to simplify their design and increase understanding; emphasizes practical issues of performance, efficiency, and scalability; provides a self-contained discussion of the basic concepts of parallel computer architectures; covers algorithms for scientific computation, such as dense and sparse matrix computations, linear system solving, finite elements, and FFT; discusses algorithms for combinatorial optimization, including branch-and-bound, heuristic search, and dynamic programming; incorporates illustrative examples of parallel programs for commercially available computers; and contains extensive figures and examples that illustrate the workings of algorithms on different architectures. |
Inhalt
CHAPTER | 15 |
CHAPTER 3 | 65 |
CHAPTER 9 | 107 |
Urheberrecht | |
12 weitere Abschnitte werden nicht angezeigt.
Häufige Begriffe und Wortgruppen
all-to-all architecture array assigned Assume begin bitonic block broadcast called chapter column communication comparators complexity concurrency connected Consider contains corresponding cost cost-optimal Derive described determine diagonal dimension discussed distributed edge efficiency elements equal Equation example execution expression factorization Figure four Gaussian elimination given graph grid Hence hypercube illustrates implementation increasing initial isoefficiency function iteration labels linear mapping matrix memory mesh method multiplication node number of processors one-to-all broadcast operations optimal overhead p-processor parallel algorithm parallel computers parallel formulation parallel run parallel system partitioning performance personalized phase pivot points PRAM presented problem procedure Processing Program received requires result ring routing scalability scheme selection sends sequence sequential shortest paths shown in Figure shows single smaller solution solving sort sparse speedup step stored striping takes tree vector vertex