Modeling Algorithm Execution Time on Processor Arrays

Adams,; Crockett,

doi:10.1109/mc.1984.1659184

Cited by 14 publications

(3 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Note that the use of an increasingly larger amount of processors could initially decrease the t ime to solve a problem. However, after some point, the use of too many processors could even take longer time than that required by a smaller number of processors (1,5,6,8,10). This led Flatt and Kennedy to examine the problem of finding the number of processors that would result in the smallest execution time.…”

Section: Introductionmentioning

confidence: 99%

Influence of overhead on processor allocation for multiple loops

Gubitoso

Song

1995

Anais Do VII Simpósio De Arquitetura De Computadores E Processamento De Alto Desempenho (SBAC-PAD 1995)

View full text Add to dashboard Cite

We consider two consecutive and independent forall loops and the strategy to allocate processors for their execution. One strategy is to execute each of the two loops consecutively, each time with all the available processors. Another strategy is to execute both loops simultaneously, each with a fraction of the available processors. We verify that the presence of overhead can influence this strategy, since the second strategy implies the use of a smaller number of processors for each individual loop, reducing thus the effect of the overhead. We establish conditions under which the second strategy is better. Finally we consider the special case when there is a single forall loop. We show conditions under which it is more advantageous to split it into two smaller loops and execute them simultaneously, each with a fraction of the available processors.

show abstract

Section: Introductionmentioning

confidence: 99%

Influence of overhead on processor allocation for multiple loops

Gubitoso

Song

1995

Anais Do VII Simpósio De Arquitetura De Computadores E Processamento De Alto Desempenho (SBAC-PAD 1995)

View full text Add to dashboard Cite

show abstract

“…Furthermore, computations (e.g., [2][3][4]) showed that while the use of an increasingly larger number of processors could initially decrease the time required for solution of a problem, the use of "too many" processors would result in a larger execution time than that required by a smaller number of processors (see Figure 1). This led to investigations concerning the number of processors that would lead to the minimum execution time.…”

Section: Introductionmentioning

confidence: 99%

Further results using the overhead model for parallel systems

Flatt¹

1991

IBM J. Res. & Dev.

View full text Add to dashboard Cite

A performance model that takes into consideration tlie overtiead incurred in the use of a paraliei system is used to show that the maximum value of the speedup achieved by the parallel system for a fixed problem may be much smaller than the number of processors required to achieve that value. It is also shown that under certain conditions, the problem size may be varied so as to achieve a speedup closely approximating the number of processors used. 721IBM J. RES. DEVELOP. VOL. 35 NO. 5/6 SEPTEMBER/NOVEMBER 1991 H. P. FLATT

show abstract

“…Both the time taken to communicate between processors and the synchronization of the computational tasks need to be considered. Adams and Crockett [1984] present methods to quantify these timing costs using a linear system example. More generally, on parallel machines there are two ways by which speedup can be enhanced.…”

Section: Introductionmentioning

confidence: 99%

Supercomputers and their use in modeling subsurface solute transport

1991

International Journal of Rock Mechanics and Mining Sciences &am

View full text Add to dashboard Cite

Abstract. Supercomputers offer a wide range of possibilities for advancing knowledge and solving difficult, real-world problems. Solute transport-related theory and applications have long been an area of specialization within hydrology. In this review, recent applications of supercomputers to solute transport problems are presented. These applications are based mostly on the assumption of a dilute solute, allowing the governing equations to be decoupled. An application of supercomputers is the simulation of solute transport in large heterogeneous spatial domains by both direct solution of the governing equations and by particle-tracking methods. An impetus for this research has been the rapid increase in predictive dispersion models based on analytical stochastic theory. Up until now, relatively few applications have incorporated algorithms designed with a targeted supercomputer in mind, with the result that the machine's potential cannot be exploited fully. Since many numerical problems are reduced to solutions of algebraic systems when they are implemented on computers, algorithms for linear algebraic systems of equations suitable for vector and parallel machines are discussed in some detail. Data analysis, necessary both for evaluating emerging theories and for more "routine" applications like determining model parameters, is identified as another general area that is ably handled by supercomputers. In particular, supercomputers offer a way to become less dependent on parametric assumptions and constraints, substituting, instead, raw computational power. Visualization techniques and associated facilities have flourished alongside the development of supercomputer centers. These offer the possibility for real-time viewing of solute transport processes, both real and simulated, although little has been produced thus far. It is expected that many more researchers involved with subsurface solute transport will make use of supercomputers in the future.

show abstract

Modeling Algorithm Execution Time on Processor Arrays

Cited by 14 publications

References 2 publications

Influence of overhead on processor allocation for multiple loops

Influence of overhead on processor allocation for multiple loops

Further results using the overhead model for parallel systems

Supercomputers and their use in modeling subsurface solute transport

Contact Info

Product

Resources

About