High Performance Algorithms for Computational Mechanics
Feng-Nan Hwang, National Central University
Xiao-Chuan Cai, University of Colorado Boulder, USA
In recent years, with the rapid advance of high-performance computers, the numerical simulation of complex biomedical and engineering problems within a reasonable time becomes possible. The key is to replace the algorithm that is sequential in nature by some fully parallel algorithms to enhance the scalability of the kernel solver in the simulator. This mini-symposium is dedicated to the discussion of some recent developments in fully parallel, scalable state-of-the-art solution algorithms, such as domain decomposition method, multilevel or multiscale method, space-time method, and nonlinear preconditioning technique for large sparse linear and nonlinear systems, PDE-constrained optimizations, inverse problems, eigenvalue problems and their applications in computational mechanics
Feng-Nan Hwang, National Central University
Xiao-Chuan Cai, University of Colorado Boulder, USA
In recent years, with the rapid advance of high-performance computers, the numerical simulation of complex biomedical and engineering problems within a reasonable time becomes possible. The key is to replace the algorithm that is sequential in nature by some fully parallel algorithms to enhance the scalability of the kernel solver in the simulator. This mini-symposium is dedicated to the discussion of some recent developments in fully parallel, scalable state-of-the-art solution algorithms, such as domain decomposition method, multilevel or multiscale method, space-time method, and nonlinear preconditioning technique for large sparse linear and nonlinear systems, PDE-constrained optimizations, inverse problems, eigenvalue problems and their applications in computational mechanics