http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
순차프로그램의 병렬수행을 위한 프로그램 재구조화 기법에 관한 연구
김병수,박두순 순천향대학교 1991 논문집 Vol.14 No.1
In this paper we discuss new techniques for transforming serial programs into a parallel form suitable for execution on parallel processor systems. These schemes can be implemented in a restructuring compiler with minimal effort, and for certain types of serial loops they can obtain significant improvement. More specially, we discuss in detail dependency removing, a compiler transformation that can partially parallelize certain types of serial loops. Four versions of this transformation are presented.
공용해,박두순 순천향대학교 부설 산업기술연구소 1996 순천향 산업기술연구소논문집 Vol.2 No.2
Neural network pattern recognition requires a large amount of computation and often results in local optimum solutions. For these problems, an improvement is attempted to train neural networks more effectively and efficiently. The performance experiments are conducted on a real character recognition problem. Among many neural networks, a multilayer perceptron(MLP) is used. due to its good generalization ability. The initial weight distribution of MLP influences the training process significantly. Therefore, the random initial weights of MLPs are globally optimized by a genetic algorithm. For this, the MLP weight sets and error function are transformed to genetic forms. For the speedup of the optimization process, the partition of patterns and the premature saturation problem are considered. Real character recognition result of the method shows an improved recognition performance and a faster training process.
Data flow Language: introduction and Analysis
朴斗淳 순천향대학교 1986 논문집 Vol.9 No.2
L.S.I의 발달로 제5세대 computer에서는 high speed로 data를 처리할 수 있는 super system이 요구되었는데 이것이 concurrent system이다. current system이 연구됨에 따라 이 system에 맞는 language가 필요하게 되었고 이 language가 data flow language이다. 이 paper에서는 data flow model과 architecture를 서술하였으며 이런 model과 architecture에서 발생하는 문제점도 서술하였다. 그리고 data flow language의 특징과 design방법을 소개하였고 현재 개발되어 있는 language인 VAL과 ID에 대해서 소개하였다.
박두순 순천향대학교 1994 논문집 Vol.17 No.1
This paper presents new method for parallelization of sequential programs which can be applied in parallel processing systems. There exist data dependences between array variables of a loop. The data dependence relations between statements which form variable or constant dependence distances are especially complex. Therefore, extracting parallelism for those statements at compile time is very difficult. In this paper, a general method for the extracting parallelism in nested loop for MIMD machine is presented.
다양한 토폴로지를 이용한 트랜스퓨터에서의 병렬처리에 관한 연구
김영희,박두순 순천향대학교 부설 산업기술연구소 1998 순천향 산업기술연구소논문집 Vol.4 No.1
The parallel computing has come to an age of popularity, and is moving into the mainstream of the computing world. Also, Parallel computers can provide optimal performance, only when it fulfills both the hardware and software specific requirements of the machine at the same time. For our study, we experimented various methodology on a transputer system and we present detailed study on the effect of parallel processing on the transputer system. Our results show that increases in the number of processors do not always provide the increases in the execution performance. For most applications, we found the performance depends on many factors such as, the topology, data distribution model, number of processors and the software techniques being used.
한개의 단순 순환에서의 자료의존성 검사를 위한 개선된 알고리즘
박두순 순천향대학교 1989 논문집 Vol.12 No.2
In translating the sequential programs into parallel programs for multi-processor systems, many algortithms are presented for the automatic testing on data dependency. [13, 19] In these algorithms, the methods by Banerjee and Wolfe are popular. In this paper, we present an improved algorithm which applies for single loop.
PARAFRASE Ⅱ 원시코드 분석을 통한 병렬 컴파일러 설계에 관한 연구
이상헌,박두순 순천향대학교 부설 산업기술연구소 1998 순천향 산업기술연구소논문집 Vol.4 No.1
More powerful computer is required to use large and variable data quickly in information society. To fill up these requirements, parallel computer has appeared. But it is not easy for user to do parallel programming. In addition, it is more difficult to convert existing former sequential programs into parallel programs. The way to solve this problem is to use parallel compilers. Making parallel compilers need many periods, much endeavor. This paper analyses the source code of Parafrase-2, one of parallel compilers, and then designs, develops a simple parallel compiler.
장유숙,박두순 순천향대학교 부설 산업기술연구소 1997 순천향 산업기술연구소논문집 Vol.3 No.2
The parallelization of loops can be made formal by basing it on an algebraic theory of loop transformation. In this theory, the concept of unimodularity and non-unimodularity arises. We discuss the linear loop transformation framework which is the formal basis for state of the art optimization techniques in restructuring compilers for parallel machines. The framework unifies most existing transformations. We discuss in detail the techniques that generates the transformed loop and derives the desired linear transformation. We present method on loop and data transformations and motivate a unified framework.