Benchmark Tests on ANSYS Parallel Processing Technology
It is extremely important for manufacturing industries to reduce their design process period in order to keep a competitive position in worldwide markets. CAE has been and will have an even more important role in shortening the design process. To effectively use CAE as a part of the design process, the ability of doing analysis in an affordable way, using large-scale real life problems is commonly required. One of the requirements is the development of a finite element method suitable for parallel computational processing using a number of CPUs. A reduction of computational time is achieved by using parallel machines such as multiprocessing servers or network clusters. Consequently the burden of model simplification, a major part of total engineering time, can be reduced enormously. ANSYS Inc. has released parallel processing solvers, DDS (Distributed Domain Solver) and AMG (Algebraic Multigrid Solver) in ANSYS version 5.7 in order to meet the requirements from the market. In this paper, simple tests are first carried out to confirm the features and capabilities of DDS and AMG for different types of analyses. Furthermore, real-life problems such as thermal stress of an engine block (7.3 million DOF) and elastic-plastic-creep behavior of an IC packaging (about 84,000 DOF) are analyzed. The effectiveness of ANSYS parallel processing solvers is also studied.