이더넷과 인피니밴드 네트워크 기반의 분산 메모리 시스템에서 병렬성능 분석

다국어 초록 (Multilingual Abstract)

In this study, a parallel performance of CUPID-code has been investigated for both Ethernet and Infiniband network system to examine the effect of cache memory and network-speed. Bi-conjugate gradient solver of CUPID-code has been parallelised by using domain decomposition method and message passing interface (MPI). It is shown that the parallel performance of Ethernet-network system is worse than that of Infiniband-network system due to the slow network-speed and a small cache memory. It is also found that the parallel performance of each system deteriorates for a small problem due to the communication overhead, but the performance of Infiniband-network system is better than Ethernet-network system due to a much faster network-speed. For a large problem, the parallel performance depends less on network system.

참고문헌 (Reference)

1 강성우, "이산화된 Navier-Stokes 방정식의 영역분할법을 위한 병렬 예조건화" 대한기계학회 27 (27): 753-765, 2003

2 윤인수, "기가비트 이더넷상에서의 M-VIA 구현" 한국정보과학회 29 (29): 648-654, 2002

3 정인형, "고성능 PC 클러스터 시스템을 위한 VIA 기반 RDMA 메커니즘 구현" 한국정보과학회 31 (31): 635-642, 2004

4 "http://www.open-mpi.org"

5 "http://mvapich.cse.ohio-state.edu"

6 "http://glaros.dtc.umn.edu/kghome/views/metis"

7 전병진, "Symmetric Multi-Processing 시스템에서 다양한 병렬 기법 모델을 적용한 병렬 CUPID 코드의 성능분석" 대한기계학회 38 (38): 71-79, 2014

8 Kirk, D.B, "Programming Massively Parallel Processors" Elsevier Inc 19-27, 2010

9 Shipman, G.M., "Infiniband Scalability in Open MPI" 1-8, 2006

10 Liu, J, "High Performance RDMA-based MPI Implementation over Infiniband" 32 : 167-198, 2004

11 KAERI, "CUPID code 1.7 manual, Vol.1"

12 Lee, B.S, "Building and Benchmarking of Linux Cluster Systems for Computational Fluid Dynamics" 201-206, 1999

13 Mininni, P.D., "A hybrid MPI-OpenMP Scheme for Scalable Parallel Pseudospectral Computations for Fluid Turbulence" 37 : 316-326, 2011