Static load balancing for CFD distributed simulations [electronic resource].
- Washington, D.C. : United States. Dept. of Energy, 2001.
Oak Ridge, Tenn. : Distributed by the Office of Scientific and Technical Information, U.S. Dept. of Energy.
- Physical Description:
- PDF-FILE: 22 ; SIZE: 0.3 MBYTES pages
- Additional Creators:
- Lawrence Livermore National Laboratory, United States. Department of Energy, and United States. Department of Energy. Office of Scientific and Technical Information
- Restrictions on Access:
- Free-to-read Unrestricted online access
- The cost/performance ratio of networks of workstations has been constantly improving. This trend is expected to continue in the near future. The aggregate peak rate of such systems often matches or exceeds the peak rate offered by the fastest parallel computers. This has motivated research towards using a network of computers, interconnected via a fast network (cluster system) or a simple Local Area Network (LAN) (distributed system), for high performance concurrent computations. Some of the important research issues arise such as (1) Optimal problem partitioning and virtual interconnection topology mapping; (2) Optimal execution scheduling and load balancing. CFD codes have been efficiently implemented on homogeneous parallel systems in the past. In particular, the helicopter aerodynamics CFD code TURNS has been implemented with MPI on the IBM SP with parallel relaxation and Krylov iterative methods used in place of more traditional recursive algorithms to enhance performance. In this implementation the space domain is divided into equal subdomain which are mapped to the processors. We consider the implementation of TURNS on a LAN of heterogeneous workstations. In order to deal with the problem of load balancing due to the different processor speeds we propose a suboptimal algorithm of dividing the space domain into unequal subdomains and assign them to the different computers. The algorithm can apply to other CFD applications. We used our algorithm to schedule TURNS on a network of workstations and obtained significantly better results.
- Report Numbers:
- E 1.99:ucrl-jc-142207
- Other Subject(s):
- Published through SciTech Connect.
15th International Conference on Supercomputing, Sorrento (IT), 06/18/2001--06/21/2001.
Wissink, A; Chronopoulos, A T; Grosu, D; Benche, M.
- Funding Information:
View MARC record | catkey: 14349995