A multi-paradigm parallel programming system is described in this paper. the system supports a specification language that can conveniently encode a fork-join graph by which many parallel applications are representabl...
详细信息
ISBN:
(纸本)0818656026
A multi-paradigm parallel programming system is described in this paper. the system supports a specification language that can conveniently encode a fork-join graph by which many parallel applications are representable. the language, called parScript, mandates the separation of parallel control from sequential codes that perform the actual computation, resulting in a program that is concise, portable, and yet efficient. We investigate the feasibility of this approach by building a prototype system on a NUMA multiprocessor. Experimental data shows that the performance of parScript programs can be comparable to that of their C counterparts, both on uniprocessors and on NUMA multiprocessors.
Fortran D extends Fortran to parallel computers via specification of the distribution of array variables across processors. When multidimensional arrays have been linearized for optimal performance on vector processor...
详细信息
ISBN:
(纸本)0818656026
Fortran D extends Fortran to parallel computers via specification of the distribution of array variables across processors. When multidimensional arrays have been linearized for optimal performance on vector processors, Fortran D cannot produce the best parallelization because it is limited to one-dimensional distribution, which is less efficient due to surface-to-volume effects. We propose Fortran D extensions and compiler technology to support natural topology parallelization of such codes. Experimental results are presented to illustrate the improvement in performance of this approach over the performance of currently available methods.
the broad thesis presented in this paper suggests that the serial emulation of a parallel algorithm has the potential advantage of a parallel algorithms has the potential advantage of running on a serial machine faste...
详细信息
ISBN:
(纸本)0818656026
the broad thesis presented in this paper suggests that the serial emulation of a parallel algorithm has the potential advantage of a parallel algorithms has the potential advantage of running on a serial machine faster than a standard serial algorithm for the same problem. It is too early to reach definite conclusions regarding the significance of this thesis. However, using some imagination, validity of the thesis and some arguments supporting it may lead to several far-reaching outcomes: (1) Reliance on 'predictability of reference' in the design of computer systems will increase. (2) parallel algorithms will be taught as part of the standard computer science and engineering undergraduate curriculum irrespective of whether (or when) parallelprocessing will become ubiquitous in the general-purpose computing world.
We apply parallel heuristic solving techniques to the exponentially complex problem of interpretation of pure logic programs. parallel heuristics incorporated within the control strategy employed to solve logic progra...
详细信息
ISBN:
(纸本)0818656026
We apply parallel heuristic solving techniques to the exponentially complex problem of interpretation of pure logic programs. parallel heuristics incorporated within the control strategy employed to solve logic programs are used to guide the search for a solution in parts of the search space most likely to yield a solution. the results show that our strategy when applied to standard benchmarks not only improves the time taken to arrive at a solution by more than a linear factor, but reduces the memory requirements as well.
CC+PLU is an object-oriented parallel programming language that uses parallel composition, atomic functions, and single-assignment variables to express concurrency. We show that this programming paradigm is equivalent...
详细信息
ISBN:
(纸本)0818656026
CC+PLU is an object-oriented parallel programming language that uses parallel composition, atomic functions, and single-assignment variables to express concurrency. We show that this programming paradigm is equivalent to several traditional imperative communication and synchronization models, namely monitors and asynchronous channels. Furthermore, the object-oriented nature of CC+PLU provides an ideal framework for integrating these paradigms. We specify, implement, and formally verify a collection of libraries that integrates these traditional models with CC+PLU.
Symmetry is a parallel programming language intended for specifying scalable computation and communication in (K-1)-dimensions of discrete space and 1-dimension of discrete time. A program specifies causal and geometr...
详细信息
ISBN:
(纸本)0818656026
Symmetry is a parallel programming language intended for specifying scalable computation and communication in (K-1)-dimensions of discrete space and 1-dimension of discrete time. A program specifies causal and geometric relationships between space time regions and indicates when and where each subcomputation takes place. Symmetry unifies the notions of memory and communication in an expanded concept of 'variable'. A Symmetry variable propagates its value in any timelike direction in spacetime, not just in a direction parallel to the time axis as is the case for memory variables in traditional languages. A performance model for Symmetry properly accounts for the all costs of computation including communication delay.
In this paper we present an algorithm for Permutation Routing on a reconfigurable mesh based on the divide and conquer paradigm. the algorithm has a worst case running time of 4n+O(log2n) communication steps and requi...
详细信息
ISBN:
(纸本)0818656026
In this paper we present an algorithm for Permutation Routing on a reconfigurable mesh based on the divide and conquer paradigm. the algorithm has a worst case running time of 4n+O(log2n) communication steps and requires a maximum queue size of 5. the O(log2n) terms is associated with communication of single words which convey information about the distribution of packets. Hence, if the packet length is much larger than a single word, then the constant associated withthis term is negligible.
Explicitly parallel programs have the potential for greater performance than their implicitly parallel counterparts. However, this benefit can be accompanied by additional programming difficulties. this paper addresse...
详细信息
ISBN:
(纸本)0818656026
Explicitly parallel programs have the potential for greater performance than their implicitly parallel counterparts. However, this benefit can be accompanied by additional programming difficulties. this paper addresses one particular problem that has implications for both scalability and portability: the need for programs to accommodate diverse data decompositions. We explain why programs with explicit communication have difficulties in handling changes in data decomposition, and we present our solution to this problem which involves the notions of derivative functions and configuration parameters. We illustrate our technique by using three different data decompositions to solve the Modified Gram-Schmidt method on four parallel machines.
Path planning is an essential step in any mobile robot system. In this paper, we propose a parallel algorithm and architecture for this computationally intensive problem. the parallel implementation is based on the tr...
详细信息
ISBN:
(纸本)0818656026
Path planning is an essential step in any mobile robot system. In this paper, we propose a parallel algorithm and architecture for this computationally intensive problem. the parallel implementation is based on the trulla algorithm proposed earlier by the authors in [5]. the algorithms uses a wavefront like propagation technique and computes the near-optimal paths from every point in the grid space to the specified destination. the algorithm is implemented on a linear systolic array architecture which is simple and regular in structure and can be implemented as a VLSI system. Currently, a prototype VLSI chip with 5 processor cells in being implemented at the University of South Florida.
the growing variety of parallel computers has made it difficult to design portable tools for parallel programs. this paper shows how an interactive visualization tool can be designed to work with a variety of parallel...
详细信息
ISBN:
(纸本)0818656026
the growing variety of parallel computers has made it difficult to design portable tools for parallel programs. this paper shows how an interactive visualization tool can be designed to work with a variety of parallel machines. the design includes a strategy for adapting to differences in the interfaces and capabilities of the low-level debuggers supplied by hardware vendors. Our tool uses these debuggers to perform basic tasks like setting breakpoints and examining variables. By dividing each interaction between the visualization tool and the 'base debugger' into a sequence of customizable steps, we can write code that adapts cleanly and efficiently to differences in the debuggers. this design has been implemented in the Panorama parallel debugger, which runs on several message-passing multicomputers.
暂无评论