Do-Loop 병렬수행 알고리즘의 문제점 분석 및 고찰

송월봉,Song, Worl-Bong 한국컴퓨터산업학회 2008 컴퓨터産業敎育學會論文誌 Vol.9 No.2

병렬 스케줄링의 목적은 다중프로세서 병렬시스템 환경에서 병렬성을 가진 응용 프로그램에 대하여 최소 동기화 오버헤드(Synchronization overhead) 및 병목현상(Bottleneck) 그리고 부하균등(Load balance)을 달성하도록 스케줄링을 수행하는데 있다. 본 논문에서는 기존의 대표적인 병렬수행 스케줄링 알고리즘들을 분석하고 각 방법들로부터 문제점들을 도출한다. 이는 향후 효율적인 알고리즘을 설계하는데 도움이 필 것이다. The purpose of a parallel scheduling is to carry out the scheduling with the minimum synchronization overhead and bottleneck under a multiprocessor environment and to perform load balance for a parallel iteration. In this paper, analyse the conventional parallel scheduling methods and drive the problems from each method in order to achive the minimum scheduling overhead and load balance. These problems will go far toward solving the design of effective algorithm.

재 제조 부품 활성화를 위한 유통관리 정보 시스템

송월봉 한국지식정보기술학회 2014 한국지식정보기술학회 논문지 Vol.9 No.1

For the activating of remanufacturing industry which automobile disjoint parts are raw materials , This paper is proposed Distribution Management Information System that is able to adaptable in the base information management for the reusing of used automobile parts, the automobile disjoint management and the automobile parts distribution management. This system is presented the guide line of automobile parts management and reusable parts disjoint of used automobiles which are connected with history of reusable used parts, quality level and guaranty, and the necessity which build up jointed distribution center between automobile disjoint industry and automobile repair industry and introduce certification system for raising the confidence in the remanufacturing parts. From now on, This proposed system is expected which contribute in the green growth socially and get raising the confidence of remanufacturing parts for user and in the position of enterprise, get raising the commercial competitive power from decreasing over-head in the process of business action and distribution.

성능 제고를 위한 재구조화 프로그램의 프로시저 변환

송월봉 한국지식정보기술학회 2013 한국지식정보기술학회 논문지 Vol.8 No.2

For the improvement in ability of processing speed, the parallel processing system should be considered not only theelements of hardware but also the elements of software. Exactly the capability of parallel computer is impossible withoutsoftware technic for detection the parallelism. In this paper, Using the restructured sequential loop, The proceduretransformation method for the advanced performance is proposed. This is a procedure for the automatic conversion of asequential loop into a nested DOALL loops at the compile time. For this, the source program of Parafrase Ⅱparallelcompiler is analyzed and through this existing Parafrase parallel compiler, a improved procedure transformation method toparallel processing effectively in the nested loop is implemented. This proposed parallel scheme is certificated to obtain amore parallelism effect in the reiterative sentence.

모바일 기기를 이용한 설비유지관리 플랫폼 설계 및 구현

송월봉 한국지식정보기술학회 2016 한국지식정보기술학회 논문지 Vol.11 No.3

Normally, the CMMS(Computerized Maintenance Management System) has been provided as modules associated with ERP because it is closely related with the BOM(Bill Of Material) of ERP. However the CMMS had not used more actively on account of usability and user type such as the lack of awareness as the nuisance of a input task. But in the case of operating efficiently the CMMS, the companies which recognize the benefits that can be obtained are developing on their own or use the solution in the form of a package that has already been developed or investigation introduction. In addition, The majority of the domestic manufactures have a mind that produce a good product and make a profit by selling a lot of their customers and repairs if they are experiencing breakdown. However in such a case some customers are to report breakdown irregularity regardless of the reserve parts, There are no plans to build with BM. In this paper, in the case of equipment maintenance for increased productivity it is necessary to introduce a CBM(Condition Based Maintenance) for utilizing smart devices rather than conventional BM(Break down Maintenance) or CPM(Computerized Preventive Maintenance) so that design and implement the equipment maintenance platform for mobile devices.

Parafrase II를 이용한 병렬 컴파일러 설계

송월봉,Song Worl-Bong 한국컴퓨터산업학회 2006 컴퓨터産業敎育學會論文誌 Vol.7 No.3

본 논문에서는 Parafrase II를 이용한 간단한 병렬화 컴파일러를 제안한다. 이것은 중첩루프에서 효율적인 병렬처리를 하기위하여 병렬성을 추출하는 새로운 기법이다. 이를 위해서 Parafrase II 병렬화 컴파일러의 원시 프로그램을 분석하였으며 Parafrase II를 이용한 새로운 병렬 컴파일러를 구현하였다. 아울러 본 컴파일러는 불변 및 가변 종속 관계를 갖는 모든 형식에 적용이 가능하다. In this paper, a simple parallel compiler using of Parafrase II is presented. This is a new general method the extracting parallelism in order to parallel processing effectively in nested loop. For this, the source program of Parafrase II parallel compiler is analyzed and implemented. Moreover, this method can be applicable where the dependency relation is both uniform and non-uniform in distance.

Non-uniform에서 종속성 제거 방법을 이용한 프로시저 변환 알고리즘

송월봉 한국지식정보기술학회 2011 한국지식정보기술학회 논문지 Vol.6 No.2

Generally, In a application program the core part for parallel processing is a loop. Because the loop exists data dependencies between the array index variable of a loop. Specially, the data dependence relations between statements which of variable or constant dependence distance are very complex. In this paper propose procedure transformation Algorithm which can be apply to non-uniform code. A performance test results show the proposal scheme is superior to conventional method.

루프인터체인지 병렬컴파일러 구현

송월봉,Song, Worl-Bong 한국컴퓨터산업학회 2007 컴퓨터産業敎育學會論文誌 Vol.8 No.3

Generally, In a application program the core part for parallel processing is a loop. therefore in this paper, loop interchange parallel compiler is proposed. this is a procedure for the automatic conversion of a loop interchange. According to execution to the outside CDOALL statements of cedar fortran, loop interchange is more effectively method the extracting parallelism in order to parallel processing in iterations. This method will be expected to effectively execution result with mixed into linear conversion and go far toward solving the effectively implementation of the non-unimodular nested loop. 일반적으로 응용프로그램에서 병렬성 추출에 대한 핵심 부분은 루프이다 따라서 본 논문에서는 Loop Interchange를 자동으로 처리할 수 있는 병렬컴파일러를 구현하고자한다. Loop Interchange는 반복문에서 cedar Fortran의 CDOALL문장을 바깥쪽으로 보냄으로서 특히 더 많은 병렬화 효과를 얻을 수 있기 때문이다. 이는 향후 선형변환과 혼합하여 더욱 효과적인 실행 결과를 기대하고 불완전 중첩루프에 적용하는 방법의 연구에 기여할 것으로 판단된다.

루프구조의 병렬화 컴파일러 설계 및 구현

송월봉 한국컴퓨터산업학회 2002 컴퓨터産業敎育學會論文誌 Vol.3 No.8

본 논문에서는 순차루프를 이용한 간단한 병렬화 컴파일러를 제안한다. 이것은 컴파일 시간에 중첩 병렬 DOALL루프로 바꾸어주는 순차루프의 자동 변환에 관한 절차이다. 이를 위해서, Parafrase II 병렬화 컴파일러의 원시 프로그램을 분석하였으며 중첩루프에서 효율적인 병렬처리를 위한 새로운 병렬성 추출 방법을 구현하였다. In this paper, a simple parallel compiler of a sequential loop is presented. This is a procedure for the automatic conversion of a sequential loop into a nested parallel DOALL loops at compile time. For this. the source program of Parafrase II parallel compiler is analyzed and a new general method the extracting parallelism in order to parallel processing effectively in nested loop is implemented.

병렬처리를 위한 프로그램 재구조화

송월봉 한국컴퓨터산업학회 2003 컴퓨터産業敎育學會論文誌 Vol.4 No.4

본 논문에서는 자료종속성제거와 관련해서 선형루프변환의 새로운 이론을 제안한다. 이는 선형루프변환이 병렬성 추출의 주요한 대상이 되기 때문이다. 이러한 이유 때문에 완전한 중첩루프에서 최대의 루프 병렬성을 추출하는 방법이 제안되었다. 제안된 이론은 거리와 방향에 관계없이 종속성을 갖는 일반적인 중첩루프에 적용할 수 있다. In this paper A new theory of linear loop transformation called Elimination of Data Dependency(BDD) is presented. The current framework of linear loop transformation cannot identify a significant fraction of parallelism. For this reason, a method to extract the maximum loop parallelism in perfect nested loops is presented. This technique is applicable to general loop nests where the dependence include both distance and directions.

A Design and Implementation of The RF Sensing Node Using The ATmega Platform

송월봉 한국지식정보기술학회 2015 한국지식정보기술학회 논문지 Vol.10 No.2

As the nervous system which create the new information space from connecting the network that make intellectual space, The sensor network technology for core element which realize ubiquitous computing is in the spotlight recently and the research on sensor network which is composed many different sensor node is proceeding actively. As well the RTOS is used for raising the efficiency of system which is concerned about collection and transmission of data from sensor network. In this paper, RF sensing node which smooth communication inter system in the RTOS that used ATmega platform is designed and implemented, and more productively control system is proposed for the control of device from making practice the detection of stability and realizing the light weighted platform. Besides, Typical TinyOS and MicroC/OS are compared and analyzed for the verification. Therefore, The result of this paper is expected to contribute in the development of RF network products which suit about characteristic of ATmega chip in the process of development the RF sensor network products which could display variable information.