http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
소프트웨어 Dynamically Connected Transport의 설계와 구현
박지웅(Jiwoong Park),염헌영(HeonYoung Yeom) 한국정보과학회 2018 정보과학회 컴퓨팅의 실제 논문지 Vol.24 No.12
기존 RDMA 모델에서 사용하는 RC QP와 UD QP는 확장성, 사용성 측면에서 각각 단점을 가지고 있다. 이러한 문제를 해결하기 위하여 제시된 Dynamically Connected Transport는 하드웨어의 지원이 필요하다는 문제점을 가지고 있다. 본 연구에서는 Dynamically Connected Transport를 소프트웨어적으로 구현함으로써 하드웨어에 의존적인 기존의 문제점을 해결하면서도 추가적인 오버헤드를 거의 발생시키지 않을 수 있음을 보였다. 또한, DC Transport를 활용할 수 있는 예시를 제시하고, 기존 RC QP기반 응용에 적용하였을 때 얻을 수 있는 성능적 이점이 있음을 실험을 통해 증명하였다. Most of the RDMA applications use either RC QP(Reliable Connected Queue Pair) or UD QP(Unreliable Datagram Queue Pair). RC QP has advantages with regards to usability and disadvantages with respect to scalability; while UD QP is scalable but hard to use. DCT (Dynamically Connected Transport) has been introduced to address the limitations when using one of the existing QPs. However, DCT is only available with hardware support. In this paper, we have designed and implemented a totally software-based DCT prototype to eliminate hardware dependency. Our experimental results show that our prototype adds only negligible overhead while providing the advantages of both RC QP and UD QP. Furthermore, we present and evaluate write and evaluate a DC QP-based application to prove its effectiveness compared to the RC QP-based application.
하드웨어 캐시 파티셔닝과 소프트웨어 캐시 파티셔닝의 성능 비교
박지웅(JiWoong Park),염헌영(HeonYoung Yeom),엄현상(Hyeonsang Eom) Korean Institute of Information Scientists and Eng 2015 정보과학회논문지 Vol.42 No.2
The era of multi-core processors has begun since the limit of the clock speed has been reached. These days, multi-core technology is used not only in desktops, servers, and table PCs, but also in smartphones. In this architecture, there is always interference between processes, because of the sharing of system resources. To address this problem, cache partitioning is used, which can be roughly divided into two types: software and hardware cache partitioning. When it comes to dynamic cache partitioning, hardware cache partitioning is superior to software cache partitioning, because it needs no page copy. In this paper, we compare the effectiveness of hardware and software cache partitioning on the AMD Opteron 6282 SE, which is the only commodity processor providing hardware cache partitioning, to see whether this technique can be effectively deployed in dynamic environments.
풍력블레이드 비정상 공력하중 해석을 위한 자유후류기법 개발 및 실험적 연구
신형기(Shin, Hyungki),박지웅(Park, Jiwoong),김호건(Kim, Hogeon),이수갑(Lee, Soogab) 한국신재생에너지학회 2005 한국신재생에너지학회 학술대회논문집 Vol.2005 No.06
A critical issue in the field of the rotor aerodynamics is the treatment of the wake. The wake is of primary importance in determining overall aerodynamic behavior, especially, a wind turbine blade includes the unsteady air loads problem. In this study, the wake generated by blades are depicted by a free wake model to analyse unsteady loading on blade and a new free wake model named Finite Vortex Element(FVE hereafter) is devised in order to include a wake-tower interact ion. In this new free wake model, blade-wake-tower interaction is described by cutting a vortex filament when the filament collides with a tower. This FVE model is compared with a conventional free wake model and verified by a comparison with NREL and SNU wind tunnel model. A comparison with NREL and SNU data shows validity and effectiveness of devised FVE free wake model and an efficient.
후류와 타워의 영향을 고려한 수평축 풍력발전기 블레이드의 비정상 하중 예측을 위한 새로운 자유후류기법의 연구
신형기(Shin, Hyungki),박지웅(Park, Jiwoong),이수갑(Lee, Soongab),김주언(Kim, Jueon) 한국신재생에너지학회 2005 신재생에너지 Vol.1 No.1
A critical issue in the field of the rotor aerodynamics is the treatment of the wake. The wake is of primary importance in determining overall aerodynamic behavior, especially, a wind turbine blade includes the unsteady airloads problem. In this study, the wake generated by blades are depicted by a free wake model to analyse unsteady loading on blade and a new free wake model named Finite Vortex Element(FVE hereafter) is devised in order to include a wake-tower interaction. In this new free wake model, blade-wake-tower interaction is described by cutting a vortex filament when the filament collides with a tower. This FVE model is compared with a conventional free wake model and verified by a comparison with NRELand SNU wind tunnel model. A comparison with NREL and SNU data shows validity and effectiveness of devised FVE free wake model and an efficient.