ELSA project concerns massively parallelarchitectures on silicon dedicated especially to low-level image processing. Real-time low-level image processing demands a huge amount of computing power. Fortunately, the alg...
详细信息
parallelprogramming has to date remained inaccessible to the average scientific programmer. parallelprogramming languages are generally foreign to most scientific applications programmers who only speak Fortran. Aut...
详细信息
parallelprogramming has to date remained inaccessible to the average scientific programmer. parallelprogramming languages are generally foreign to most scientific applications programmers who only speak Fortran. Automatic parallelization techniques have so far proved unsuccessful in extracting large amounts of parallelism from sequential codes and do not encourage development of new, inherently parallel algorithms. In addition, there is a lack of consistency of programmer interface across architectures which requires programmers to invest a lot of effort in porting code from one parallel machine to another. This paper discusses the object oriented Fortran language and support routines developed at Mississippi State in support of parallelizing complex field simulations. This interface is based on Fortran to ease its acceptance by scientific programmers and is implemented on top of the Unix operating system for portability.< >
The proceedings contain 45 papers. The special focus in this conference is on Applied Algebra. The topics include: algorithms for the shape of semialgebraic sets. a new approach;on the parameters of algebraic geometri...
ISBN:
(纸本)9783540545224
The proceedings contain 45 papers. The special focus in this conference is on Applied Algebra. The topics include: algorithms for the shape of semialgebraic sets. a new approach;on the parameters of algebraic geometric cedes;on wiedemann's method of solving sparse linear systems;fast algorithms for decoding orthogonal and related codes;jacobian matrices and constructions in algebra;homogenity, pseudo-homogenity, and grobner basis computations;arithmetic on non-supersingular elliptic curves;implementing some algorithms of kantor;computing roadmaps of general semi-algebraic sets;an improved sign determination algorithm;buchberger algorithm and integer programming;new systolic architectures for cyclic code encoding;algebraic constructions of efficient broadcast networks;error-correction for WIMs and WUMs;some constructions in rings of differential polynomials;an algorithm for the computation of the radical of an ideal in the ring of polynomials;polynomial-time construction of spherical codes;algorithms for a multiple algebraic extension ii;on the orphans and covering radius of the reed-muller codes;a joint authentication and encryption scheme based on algebraic coding theory;some results on linear unequal-error-protection codes specified by their generator matrix;an ackermannian polynomial idealsome undecidability results for weakly confluent monadic string-rewriting systems;calculating multidimensional symmetric functions using jacobi's formula;multivariate sturm theory;binary spherical geometric codes;an algebraic construction of generalized beenkefs cedes;improving the time complexity of the computation of irreducible and primitive polynomials in finite fields;completely transitive codes and distance transitive graphs.
Details the implementation of the Pascal programming language for a dataflow architecture. The reasons for choosing this particular architectural model and language was to achieve maximum parallelism with minimal spec...
详细信息
Details the implementation of the Pascal programming language for a dataflow architecture. The reasons for choosing this particular architectural model and language was to achieve maximum parallelism with minimal specification by the programmer. Issues that are discussed involve the analysis and transformation of code to maximise parallelism, and the generation of an applicative intermediate code form (IF1) to be later translated into machine code for a wide range of parallelarchitectures. Specific emphasis is given to those features of Pascal and other conventional languages that are omitted from parallel functional languages for dataflow architectures, such as global variables, function side-effects, variable aliasing, and pointers. Loop and general code optimizations are also introduced in order to maximise parallelism. Some simulation results are presented which highlight the extent of parallelism available in a conventional language.< >
Message Driven Computation (MDC) is a model of computation with which they have been experimenting at the Illinois Institute of Technology. The authors aim to prove the viability of MDC in practice for the expression ...
详细信息
Message Driven Computation (MDC) is a model of computation with which they have been experimenting at the Illinois Institute of Technology. The authors aim to prove the viability of MDC in practice for the expression of parallel algorithms and the implementation of functional and dataflow programming languages. In the paper they discuss their implementation of the Lucid programming language in MDC. The discussion presents a subset of Lucid which illustrates the principles of Lucid, Message Driven Computing, and the translation into and the interpretation of dataflow graphs.< >
The proceedings contain 41 papers. The special focus in this conference is on algorithms. The topics include: Decision-making with incomplete information;maximum independent set of a permutation graph in k tracks;algo...
ISBN:
(纸本)9783540549451
The proceedings contain 41 papers. The special focus in this conference is on algorithms. The topics include: Decision-making with incomplete information;maximum independent set of a permutation graph in k tracks;algorithms for square roots of graphs;distributed k-mutual exclusion problem and k-coteries;weighted random assignments with application to hashing;scheduling file transfers under port and channel constraints;substitution decomposition on chordal graphs and applications;mixed-searching and proper-path-width;short wire routing in convex grids;a new approach to knock-knee channel routing;an average case analysis of monien and speckenmeyer's mechanical theorem proving algorithm;an on-line algorithm for navigating in unknown terrain;on maintaining the width and diameter of a planar point-set online;optimal triangulations by retriangulating;an incremental algorithm for constructing shortest watchman routes;on hitting grid points in a convex polygon with straight lines;on the complexity of some hamiltonian and eulerian problems in edge-colored complete graphs;dynamic programming on intervals;combinatorial optimization through order statistics;combinatorics and algorithms of geometric arrangements;an analysis of randomized shear sort on the mesh computer;efficient parallel divide-and-conquer for a class of lnterconnection topologies;optimal specified root embedding of full binary trees in faulty hypercubes;a tight lower bound for the worst case of bottom-up heapsort;historical searching and sorting;comparison-efficient and write-optimal searching and sorting;nearest neighbors revisited;competitiveness and response time in on-line algorithms;a linear time optimal via assignment algorithm for three-dimensional channel routing;symmetry of information and one-way functions.
The authors present four new parallel algorithms for triangulation of points in a plane. The first two, for one-way iterative arrays and two-way cellular arrays, requiring O(n) time and O(n) processors. Next they give...
详细信息
The authors present four new parallel algorithms for triangulation of points in a plane. The first two, for one-way iterative arrays and two-way cellular arrays, requiring O(n) time and O(n) processors. Next they give an algorithm for d-dimensional mesh arrays which requires O(n/sup 1/d/) time and a hypercube algorithm with a worst case running time of O(log/sup 3/ n) and an expected running time of O(log/sup 2/ n), both using O(n) processors. The linear array algorithms are the first known of for these architectures which compute Delaunay and greedy triangulations. The mesh array and hypercube algorithms appear to be the first which directly compute a triangulation on these architectures. These algorithms can be modified to compute the Voronoi diagram of points in the plane and the convex hull of points in 3-space.< >
Time scheduling is one of the important problems faced in mapping algorithms onto parallelarchitectures, particularly in real-time fault-tolerant computing. The paper shows how linear schedules can be used to exploit...
详细信息
Time scheduling is one of the important problems faced in mapping algorithms onto parallelarchitectures, particularly in real-time fault-tolerant computing. The paper shows how linear schedules can be used to exploit parallelism available in nested-loop programs. It relates linear schedules to known loop transformation techniques for modeling the parallel execution and extracting parallelism of nested loop structures. These techniques are the Doaccross technique used to model the simultaneous execution of loop computations that belong to different iterations and a technique for parallelism exploitation called selective cycle shrinking. It is shown how selective shrinking is related to linear scheduling of nested loops and how to find the selective shrinking with minimum total execution time by applying techniques of finding optimal linear schedules. The execution of these schedules can be modeled using the Doaccross technique.< >
暂无评论