Alliant FX/2812 시스템의 병렬처리 구현에 관한 연구

Abstract

본 논문은 Alliant FX/2812 시스템에서 주어지는 프로그램의 동시성과 벡터화 구현을 효율적으로 처리하기 위한 기법을 고찰하고, 간단한 실험 결과를 통하여 대단위 프로그램의 병렬화 작업의 가능성과 기준을 제시하였다. 이 논문은 또한 Alliant FX/2812 시스템의 기본 프로세서인 인텔의 i860프로세서의 특징인 수퍼 스칼라 기능을 살펴보았다.

Objectives of this paper are to study techniques and methods for effective implementation of concurrency and vectorization for large codes; and to determine the feasibility of parallelizing an existing large scientific code; and to estimate potential speedups attainable.

This paper presents the experimental results in order to study the above objectives on Alliant FX/2812 system. This system is based on the Intel's i860 processors. Also, this paper shows the feasibility of super-scalar processors for the benchmark applications.

Objectives of this paper are to study techniques and methods for effective implementation of concurrency and vectorization for large codes; and to determine the feasibility of parallelizing an existing large scientific code; and to estimate potential speedups attainable.

This paper presents the experimental results in order to study the above objectives on Alliant FX/2812 system. This system is based on the Intel's i860 processors. Also, this paper shows the feasibility of super-scalar processors for the benchmark applications.