Nowadays, it is very common for scientists and engineers to have access to a supercomputer with thousands or many more processing units. However, hardware only tells part of the story. These supercomputers require algorithms that maintain the same efficiency when the workload grows (scalability), and when the problems become more difficult (robustness). The objective of this internship is to study state of the art parallel solvers for stochastic partial differential equations on supercomputers, apply them to the simulation of porous media flow and compare them.
Topics covered by the internship include: Stochastic PDE, Numerical Analysis, Linear solvers, Domain Decomposition, High Performance Computing and Parallel programming (Python and/or C++).
The internship will be supervised by Nicole Spillane (CNRS, CMAP, Ecole Polytechnique) and Olivier Le Maitre (CNRS, CMAP, Ecole Polytechnique and Inria Saclay). It may be followed by a PhD project funded by ANR.
