• 한국정보과학회논문지:시스템및이론
• /
• 제31권8호
• /
• pp.421-431
• /
• 2004

본 논문은 병렬 프로그램을 실행할 때 계산량이 작은 병렬 루프를 병렬로 실행하는 경우에 생기는 프로그램의 성능 저하를 피하기 위하여, 컴파일 시나 실행 시에 성능 예측 모델을 이용하여 병렬 루프의 성능을 예측한 다음 적응형 실행 기법을 이용하여 병렬 프로그램을 실행하는 방법을 소개한다. 성능예측 알고리즘과 적응형 실행 알고리즘은 컴파일러 전처리기에 구현이 되었으며, 이 전처리기는 병렬 루프가 실행되는 방식을 컴파일 시나 실행 시에 결정하는 코드를 원래의 병렬 프로그램에 삽입한다. Fortran77로 씌어진 다섯 개의 대표적인 과학 수치계산 병렬 벤치마크 프로그램을 32개의 프로세서로 구성된 분산 공유 메모리 병렬 컴퓨터(SGI Origin2000)에 실행하여 본 논문에서 제안한 방법의 성능 평가를 하였을 때, 제안한 기법을 적응한 경우가 32, 16, 8, 및 4개의 프로세서에서 원래의 병렬 프로그램 보다 각각 26%, 20%, 16%, 및 10% 빨리 실행되었다. 이중 한 프로그램은 원래 병렬 프로그램 보다 32개 프로세서에서 두 배 이상 빠르게 실행되었다.

• Journal of Computing Science and Engineering
• /
• 제6권1호
• /
• pp.26-38
• /
• 2012

Java has been increasingly used in programming for real-time systems. However, some of Java's features such as automatic memory management and dynamic compilation are harmful to time predictability. If these problems are not solved properly then it can fundamentally limit the usage of Java for real-time systems, especially for hard real-time systems that require very high time predictability. In this paper, we propose to exploit multicore computing in order to reduce the timing unpredictability that is caused by dynamic compilation and adaptive optimization. Our goal is to retain high performance comparable to that of traditional dynamic compilation, while at the same time, obtain better time predictability for Java virtual machine (JVM). We have studied pre-compilation techniques to utilize another core more efficiently, preoptimization on another core (PoAC) scheme to replace the adaptive optimization system (AOS) in Jikes JVM and the counter based optimization (CBO). Our evaluation reveals that the proposed approaches are able to attain high performance while greatly reducing the variation of the execution time for Java applications.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제5권12호
• /
• pp.2294-2314
• /
• 2011

The key issue in providing fast and reliable access on cloud services is the effective management of resources in a cloud system. However, the high variation in cloud service access rates affects the system performance considerably when there are no default routines to handle this type of occurrence. Adaptive techniques are used in resource management to support robust systems and maintain well-balanced loads within the servers. This paper presents an adaptive resource management for cloud systems which supports the integration of intelligent methods to promote quality of service (QoS) in provisioning of cloud services. A technique of dynamically assigning cloud services to a group of cloud servers is proposed for the adaptive resource management. Initially, cloud services are collected based on the excess cloud services load and then these are deployed to the assigned cloud servers. The assignment function uses the proposed proportional ordering which efficiently assigns cloud services based on its resource consumption. The difference in resource consumption rate in all nodes is analyzed periodically which decides the execution of service assignment. Performance evaluation showed that the proposed dynamic service assignment (DSA) performed best in throughput performance compared to other resource allocation algorithms.

• 한국시뮬레이션학회논문지
• /
• 제12권4호
• /
• pp.51-59
• /
• 2003

Real-Time applications on Grid environment have several problems in terms of resource management addressed as follows; (1) dynamic resource allocation to provide QoS objectives, (2) heterogeneous resources that is different scale, or capacity in same unit, and (3) resource availability, and resource needs. This paper describes the techniques of resource manager (RM) handling above problems to support QoS of dynamic real-time applications on Grid. The contributions of this paper to solve problems are as follows: unification of dynamic resource requirements among heterogeneous hosts, control of resources in heterogeneous environments, and dynamic load balancing/sharing. Our heuristic allocation scheme works not only 257% better than random, 142% better than round robin, and 36.4% better than least load in QoS sensitivity, but also 38.6% better than random, 28.5% better than round robin, and 31.6% better than least load in QoS.