Open Access. Powered by Scholars. Published by Universities.®
Articles 1 - 4 of 4
Full-Text Articles in Entire DC Network
Efficient Data Structures And Algorithms For Scientific Computations., Soon Cheol Park
Efficient Data Structures And Algorithms For Scientific Computations., Soon Cheol Park
LSU Historical Dissertations and Theses
Large-scale numerically intensive scientific applications can require tremendous amounts of computer time and space. Two general methods are presented for reducing the computer resources required in scientific computing. The first is a numerical database system which is built on a space and time optimal data structure called a weighted search tree and that allows for the storage and retrieval of valuable intermediate information so costly redundant calculations can be avoided. The second is a matrix algorithm based on a new space optimal representation of sparse matrices that for typical scientific applications can be expected to dramatically decrease the cost of …
Parallel And Distributed Algorithms For A Class Of Graph-Related Computational Problems., Rajanarayanan Subbiah
Parallel And Distributed Algorithms For A Class Of Graph-Related Computational Problems., Rajanarayanan Subbiah
LSU Historical Dissertations and Theses
There exist at least two models of parallel computing, namely, shared-memory and message-passing. This research addresses problems in both these types of systems, and proposes efficient parallel (Shared-Memory Model) and distributed (message-passing) algorithms for a variety of graph related computational problems. In part I, we design algorithms for three generic problems in distributed systems: set manipulation, network structure recognition and facility placement. We present optimal distributed algorithms for recognizing rectangular-mesh networks. The time and message complexity of our algorithm is linear in the number of nodes in the network. We also lay the foundation for the recognition of 2-reducible, outer-planar …
A Theoretical Approach Involving Recurrence Resolution, Dependence Cycle Statement Ordering And Subroutine Transformation For The Exploitation Of Parallelism In Sequential Code., Chih-Ping Chu
LSU Historical Dissertations and Theses
To exploit parallelism in Fortran code, this dissertation consists of a study of the following three issues: (1) recurrence resolution in Do-loops for vector processing, (2) dependence cycle statement ordering in Do-loops for parallel processing, and (3) sub-routine parallelization. For recurrence resolution, the major findings include: (1) the node splitting algorithm cannot be used directly to break an essential antidependence link, of which the source variable that results in antidependence is itself the sink variable of another true dependence so a correction method is proposed, (2) a sink variable renaming technique is capable of breaking an antidependence and/or output-dependence link, …
Parallel Computation On Hypercube-Like Machines., Kyung Hee Kwon
Parallel Computation On Hypercube-Like Machines., Kyung Hee Kwon
LSU Historical Dissertations and Theses
The hypercube interconnection network has been recognized to be very suitable for a parallel computing architecture due to its attractive topological properties. Recently, several modified hypercubes have been propose to improve the performance of a hypercube. This dissertation deals with two modified hypercubes, the X-hypercube and the Z-cube. The X-hypercube is a variant of the hypercube, with the same amount of hardware but a diameter of only $\lceil$(n + 1)/2$\rceil$ in a hypercube of dimension n. The Z-cube has only 75 percent of the edges of a hypercube with the same number vertices and the same diameter as the hypercube. …