http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
장유숙,이상헌,박두순 순천향대학교 부설 산업기술연구소 2000 순천향 산업기술연구소논문집 Vol.6 No.2
Since most program execution time is spent in the loop structure, the researches of transforming sequential programs into parallel programs have been focused on the method of a loop structure transformation. But, most programs have implicit parallelism of interprocedure. Most parallelization of loop with procedure calls just focus on the uniform code. This paper suggestions a way of procedure analysis (an interprocedural analysis and an intraprocedural analysis). We extend the previous loop transformation algorithms for procedure transformation analysis. By using the suggested transformation analysis, we show the examples of parallelization using uniform code, nonuniform code, and complex code, and then evaluate the performance of a sequential and a parallel code using CRAY-T3E.
장유숙,박두순,Jang, Yu-Suk,Park, Du-Sun 한국정보처리학회 2002 정보처리학회논문지 A Vol.9 No.1
기존의 순차 프로그램에서 병렬성을 추출하는 연구들은 하나의 프로시저 내 변환에 치중되고 있다. 그러나 대부분의 프로그램들은 프로시저간 잠재된 병렬성을 가지고 있다. 본 논문에서는 자료 종속성 제거방법을 이용하여 프로시저 호출을 가진 루프에서 병렬성 추출 방식을 제안한다. 프로시저 호출을 포함하는 루프의 병렬화는 대부분 자료종석거리가 uniform 형태의 코드에서만 연구되었다. 본 논문에서는 자료종속거리가 uniform 코드와 nonuniform 코드에 대해 모두 적용 가능한 프로시저 간 변환 방법을 제시하였으며, 제시된 알고리즘의 성능평가를 위하여 CRAY T3E에서 성능평가하였고, 제시된 방법이 효과적임을 보였다. Most researches of transforming sequential programs into parallel programs have been based on the loop structure transformation method. However, most programs have implicit interprocedure parallelism. This paper suggests a way of extracting parallelism from the loops with procedure calls using the data dependency elimination method. Most parallelization of the loop with procedure calls have been conducted for extracting parallelism from the uniform code. In this paper, we propose interprocedural transformation, which can be apply to both uniform and nonuniform code. We show the examples of uniform, nonuniform, and complex code parallelization. We then evaluated the performance of the various transformation methods using the CRAY-T3E system. The comparison results show that the proposed algorithm out-performs other conventional methods.
장유숙,박두순 순천향대학교 부설 산업기술연구소 2001 순천향 산업기술연구소논문집 Vol.7 No.2
In this paper the proposal of parallelism on sequential loop since most program execution time is consumed in loop structure. This procedure is about automatic transformation to nested DOALL loop and nested loops to reduce data dependency efficiently. And the study to extract the parallelism is about in only one procedure transition. But almost all programs have potential parallelism in between procedures. A procedure call, which has only one loop structure or procedure call, is almost all are concentrated on uniform code, which is in the distance of data dependency in uniform.For the parallelism on loops or procedure that has loops on paralleisms are concentrated on uniform code.We presents algorithms which can be applied with uniform code, nonuniform code and complex code. The proposed algorithm evaluate using CRAY-T3E. The result of performance evaluation is that proposed algorithm is an effect.
송월봉,박두순,장유숙 순천향대학교 부설 산업기술연구소 1998 순천향 산업기술연구소논문집 Vol.3 No.3
In this paper we present a new theory of linear loop transformations called Elimination of Data Dependency(EDD). The current framework of linear loop transformation cannot identify a significant fraction of parallelism. For this reason, we presents a method to extract the maximum loop parallelism in perfect nested loops. Our technique is applicable to general loop nests where the dependences include both distances and directions. This theory provides the foundation for solving an open question in complication for parallel machines: which loop transformations, and in what order, should be applied to aceive a particular goal, such as maximizing parallelism. This paper presents an efficient loop transformation algorithm based on this theory to maximize the degree of parallelism in a loop nest. We present the results of experiments that demonstrate that EDD can result in substantial performance improvements.
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.
장유숙 ( Yu-suk Chang ),박두순 ( Doo-soon Park ) 한국인터넷정보학회 2001 인터넷정보학회논문지 Vol.2 No.4
프로그램 수행시간의 대부분이 루프 구조에서 소비되고 있기 때문에 루프 구조를 가진 순차 프로그램에서 병렬성을 추출하는 연구들이 많이 행해지고 있고 그 연구들은 하나의 프로시저 내 루프 구조의 변환에 치중되고 있다. 그러나 대부분의 프로그램들은 프로시저 간 잠재된 병렬성을 가지고 있다. 본 논문에서는 프로시저 호출을 가진 루프에서 병렬성 추출 방식을 제안한다. 프로시저 호출을 포함하는 루프의 병렬화는 대부분 자료종속거리가 uniform 형태의 코드에서만 집중되었다. 본 논문에서는 자료종속거리가 uniform 코드, nonuniform 코드 그리고 복합된(complex) 코드를 가진 프로그램에서 적용 가능한 알고리즘을 제시하였으며, 제안된 알고리즘과 loop extraction, loop embedding 그리고 procedure cloning변환 방법을 CRAY-T3E로 성능 평가하였다. 성능평가 결과는 제안된 알고리즘이 효율적이라는 것을 보여준다. Since roost of the program execution time is spent in the loop structure, the problem of extracting parallelism from sequential loop has been one of the most important research issues. However, roost programs have implicit interprocedure parallelism. This paper presents a generalized method extracting parallelism in loops having the procedure calls. Most parallelization of loops having procedure calls focus on the uniform code where data dependency distance is constant. We present algorithms which can be applied to uniform code, nonuniform code, and complex code. The performance of the proposed algorithm, loop extraction, loop embedding and procedure cloning transformation methods have been evaluated using CRAY-T3E. The result shows the effective of the proposed algorithm.