• 제목/요약/키워드: Workload Uniformity

검색결과 3건 처리시간 0.019초

작업영역의 동적 할당을 통한 고화질 애니메이션의 병렬 렌더링 (Parallel Rendering of High Quality Animation based on a Dynamic Workload Allocation Scheme)

  • 이윤석
    • 한국컴퓨터정보학회논문지
    • /
    • 제13권1호
    • /
    • pp.109-116
    • /
    • 2008
  • 고화질 입체 영상의 효과적인 재생을 위해 PC클러스터를 활용한 여러 형태의 병렬화 기법이 제안되었지만, 영상을 구성하는 객체의 분포가 균일하지 않은 경우 충분한 성능을 발휘하지 못하였다. 본 연구에서는 POV-Ray 렌더러를 채택한 PC클러스터 기반의 병렬 렌더링 시스템을 구축하고, 병렬화 성능을 높이기 위한 효과적인 부하 균형 기법을 개발하였다. 특히 애니메이션을 구성하는 연속 프레임 작업에서 프레임간의 연관성(coherence)이 높다는 사실에 근거하여, 임의 프레임의 각 분할 영역에 소요된 계산량을 바탕으로 다음 프레임의 부하 분포를 예측하고 이에 맞게 각 프로세서의 작업 영역을 재조정하는 기법을 제안하였다. 제안 기법의 성능을 평가하기 위해, 충분하지는 않지만 2개의 실제 애니메이션 데이터에 대한 적용 결과, 정적 분할에 비해 약 40% 가량의 성능 향상을 보였다. 또한 다양한 부하 분포에 대한 각 기법의 성능을 추정하기 위해 수행한 모의실험에서, 정적 분할 기법에 대해 부하균형, 확장성 측면에서 우월한 것으로 예측되었다.

  • PDF

실 규모 터널 주행실험을 통한 터널 경계부 휘도 기준의 적정성 검증 (Verification of the Appropriateness of the Standard for Tunnel Luminance in the Threshold Zone Through a Full-scale Tunnel Driving Test)

  • 박원일;조원범;정준화
    • 한국도로학회논문집
    • /
    • 제17권6호
    • /
    • pp.105-115
    • /
    • 2015
  • PURPOSES : This study aimed to evaluate the appropriateness of safety with the standard for threshold zone luminance as specified in the Recommendation for Lighting of Traffic Tunnel, which has been widely adopted worldwide. METHODS : A driving test of the subject in a full-scale road tunnel was conducted. The adaptation luminance and threshold zone luminance, which should be known for the driver to perceive an object within stopping sight distance, were obtained. These values were compared with the adaptation luminance and threshold zone luminance obtained by the existing reduced model test and tunnel lighting standard that has served as a guideline for the current threshold zone luminance standard. RESULTS : According to this study, threshold zone luminance should be increased to at least 1.8 times the value proposed in the existing studies and to twice the domestic tunnel lighting standard (KS C 3703: 2014). CONCLUSIONS : The threshold zone luminance proposed in this study differs largely from that obtained from indoor tests and from the current tunnel lighting standard used worldwide; this difference may be attributed to the fact that the indoor tests did not incorporate driving workload, non-uniformity of luminance distribution in terms of sight, and factors that reduce the visibility of the driver, such as the light reflected into the driver's eyes. Hence, it is necessary to further review the factors that reduce the visibility of drivers approaching tunnels in order to determine the rational tunnel threshold zone luminance.

유전자 알고리즘을 이용한 프로젝트 포트폴리오 투입인력 최적화 모델에 관한 연구 (A Study on the Optimization Model for the Project Portfolio Manpower Assignment Using Genetic Algorithm)

  • 김동욱;이원영
    • 한국IT서비스학회지
    • /
    • 제17권4호
    • /
    • pp.101-117
    • /
    • 2018
  • Companies are responding appropriately to the rapidly changing business environment and striving to lead those changes. As part of that, we are meeting our strategic goals through IT projects, which increase the number of simultaneous projects and the importance of project portfolio management for successful project execution. It also strives for efficient deployment of human resources that have the greatest impact on project portfolio management. In the early stages of project portfolio management, it is very important to establish a reasonable manpower plan and allocate performance personnel. This problem is a problem that can not be solved by linear programming because it is calculated through the standard deviation of the input ratio of professional manpower considering the uniformity of load allocated to the input development manpower and the importance of each project. In this study, genetic algorithm, one of the heuristic methods, was applied to solve this problem. As the objective function, we used the proper input ratio of projects, the input rate of specialist manpower for important projects, and the equal load of workload by manpower. Constraints were not able to input duplicate manpower, Was used as a condition. We also developed a program for efficient application of genetic algorithms and confirmed the execution results. In addition, the parameters of the genetic algorithm were variously changed and repeated test results were selected through the independent sample t test to select optimal parameters, and the improvement effect of about 31.2% was confirmed.