Dr. Nils Kohl
Scientist
I earned my BSc in Mathematics (2014) and MSc in Computer Science (2017) at FAU Erlangen-Nürnberg, where I also completed my PhD (Dr.-Ing.) in Computer Science. Since 2024, I have been a Lecturer (AkadR. a. Z.) in the Department of Earth and Environmental Sciences at LMU Munich.
Research Interests
My research focuses on developing massively parallel algorithms for the numerical approximation of partial differential equations. In particular, I work on matrix-free finite element methods, multigrid solvers, and high-performance computing techniques, with an emphasis on reduced- and mixed-precision computations for next-generation architectures.
Block-structured grid approximating Earth’s mantle - highlighting a refined tetrahedral coarse grid element. See https://terraneo.fau.de for more information. | © Nils Kohl
TerraNeo - Towards Mantle Circulation Models at Kilometer Resolution
My research focuses on massively parallel numerical methods for the approximation of partial differential equations at the extreme scale. Motivated by applications such as mantle convection, where kilometer-scale resolution leads to linear systems with trillions (1012) of unknowns, I develop matrix-free finite element methods and geometric multigrid solvers optimized for execution on hundreds of thousands of processes. A central part of this work is the development of the opensource software projects HyTeG and TerraNeo, which combine unstructured coarse meshes with hierarchical refinement to enable efficient, massively parallel finite element implementations. My contributions span algorithmic optimization, code generation, and performance modeling, providing a sustainable foundation for next-generation extreme-scale simulations in geophysics and beyond.
Kohl, N., & Rüde, U. (2022). Textbook efficiency: massively parallel matrix-free multigrid for the Stokes system. SIAM Journal on Scientific Computing, 44(2), C124-C155. https://doi.org/ 10.1137/20M1376005
Kohl, N., Mohr, M., Eibl, S., & Rüde, U. (2022). A Massively Parallel Eulerian-Lagrangian Method for Advection-Dominated Transport in Viscous Fluids. SIAM Journal on Scientific Computing, 44(3), C260-C285. https://doi.org/10.1137/21M1402510