- MASAO Ogino, Daido University
- Amane Takei
- Qinghe Yao
- Sin-Ichiro Sugimoto
To perform the numerical simulations by the finite element method on parallel computers such as modern supercomputers, which consists of multi-core processors, many-core processors, and/or graphical processing units, efficient parallel algorithms are indispensable. The domain decomposition method (DDM) is a well-known algorithm appropriate for the distributed-memory parallel computer, and there has been much research for the numerical method to solve initial boundary value problems of the partial differential equation (PDE). However, the architectures of modern parallel computers are increasingly complex, making it very difficult to optimize the use of computer performance. The purpose of this minisymposium is to discuss wide-ranging topics of parallel computing and large-scale computation. The DDM means not only numerical methods for PDE but a broader sense of parallel computing.