MS402 Parallel computing and domain decomposition methods

undefinedI. Herrera1, P.D. Minev2
1National University of Mexico (UNAM)/MX, 2University of Alberta/CA

 

Large-scale problems with growing computational demands are a common place in many branches of science and engineering. Parallel computing is outstanding among the new computational tools, especially at present when further increases in hardware speed have reached insurmountable barriers. So, during the last decades the emergence of parallel computers prompted on the part of the computational modeling community a continual and systematic effort for harnessing it with the purpose of predicting the behavior of many systems. Very early after such an effort began, it was recognized that domain decomposition methods (DDM) are the most effective techniques for applying parallelism to the solution of partial differential equations, since such approaches drastically simplify the coordination of the many processing units (CPUs or GPGPUs) that carry out the different tasks and also reduces very much the requirements of communication between them. In recent years much of the attention has shifted from overlapping methods to non-overlapping DDMs.

 

The purpose of this mini symposium is to bring together researchers from different fields, to present recent advances of domain decomposition methods as well as vanguard applications of such methods. Software, hardware and theoretical contributions are welcome.

 

>> submit an abstract online

Print Version