Scientific computing
To conduct the high-fidelity simulations necessary to study the complex problems of interest, we develop stable, accurate and efficient numerical methods for complex flows, including discontinuities (shocks, material interfaces, contacts), turbulence and viscoelasticity. Our focus is on high-order accurate finite difference, finite volume, and discontinuous finite element methods. We are pursuing novel paradigms for high-performance computing (including variable-precision computing) and exploring optimal design of experiments.Past and current computing projects
- Numerical analysis and models
- High-order methods, including finite difference and finite volume methods (WENO, hybrid central/shock-capturing schemes), as well as discontinuous finite element methods (Discontinuous Galerkin)
- Adaptive methods (h/p)
- Shock and interface capturing
- Phase fields
- Compressible multiphase models
- High-performance computing
- Multi-GPU parallelism
- Algorithms for post-exascale computing
- Optimal experimental design