Dr. Marcus Mohr
Scientist
After obtaining a diploma in mathematics from TU Munich in 1997, I worked at the University of Augsburg (Group for Applied Analysis and Numerical Mathematics) and the Friedrich-Alexander-University of Erlangen-Nuremberg (Doctoral Student in the Group for System Simulation at the Dept. of Informatics, PostDoc at the Dept. for Sensor Technology). In 2006 I joined Geophysics at LMU Munich.
CV
1997 Diploma in Mathematics from the Technical University of Munich
1997 - 1998 Graduate Research Assistant, Group for Applied Analysis and Numerical Mathematics, University of Augsburg
1998 - 2003 Graduate Research Assistant, Group for System Simulation, Friedrich-Alexander-University of Erlangen-Nuremberg
2003 - 2005 PostDoc within the DFG Junior Research Group: Inverse Problems in Piezoelectricity, Department for Sensor Technology, Friedrich-Alexander-University of Erlangen-Nuremberg
since 2006 Senior Scientist, Department of Earth and Environmental Sciences, Ludwig-Maximilians-Universität München
- Stellvertretender Leiter der Rechnerbetriebsgruppe Geophysik
- Fachstudiengangskoordinator Geophysik (International Masters Programme in Geophysics & Nebenfach Geophysik)
- Mitglied der Auswahlkommission für den Masterstudiengang Geophysics
- Vertreter des wissenschaftlichen Personals in der Kommission zur Vergabe der Studienzuschüsse (Geowissenschaften)
- Ansprechpartner der Fakultät 20 für die Einführung von campusLMU
Research
My research interests are numerical algorithms and their efficient implementation for high-performance computing (HPC). The main focus here are methods for the discretisation of partial differential equations and the parallel solution of the resulting systems of equations especially with matrix-free techniques and multigrid. I have also done research in the context of inverse problems.
TerraNeo meshing approach and result in the DFG Math Calendar 2014 | © B. Gmeiner, U. Rüde, C. Waluga, and B. Wohlmuth (used with permission)
TerraNeo: A Next-Generation Mantle Convection Code
While composed of solid rocks the mantle of our planet moves very slowly on geologic time-scales through processes such as diffusion and dislocation creep. This convective movement is the main driving force behind large scale geologic phenomena such as mountain building, motion of tectonic plates and dynamic topography that shape the surface of our planet. While the mantle covers a volume of nearly 1012 km3, features of interest such as sinking slabs and rising plumes are one the order of a few kilometres only. Simulating convection of the mantle, thus, requires meshes with extremely fine resolution and is a challenging endeavour even on the largest supercomputers. TerraNeo is a next-generation C++ code for this purpose. The software is based on the Hybrid Tetrahedral Grids (HyTeG) framework for extreme-scale high performance finite element simulations that we develop together with colleagues from TU Munich, FAU Erlangen and the Leibniz Supercomputing Center (LRZ).
Running Project
- Multi-physics simulations for Geodynamics on heterogeneous Exascale Systems (CoMPS) (2022-2025, BMBF/BMFTR)
- TerraNeo Project webpage
- HyTeG framework
Böhm, Fabian and Bauer, Daniel and Kohl, Nils and Alappat, Christie L. and Thönnes, Dominik and Mohr, Marcus and Köstler, Harald and Rüde, Ulrich (2025), Code Generation and Performance Engineering for Matrix-Free Finite Element Methods on Hybrid Tetrahedral Grids, 47(1), B131-B159, doi: 10.1137/24m1653756
Nils Kohl and Marcus Mohr and Sebastian Eibl and Ulrich Rüde (2022), A Massively Parallel Eulerian-Lagrangian Method for Advection-Dominated Transport in Viscous Fluids , 44(3), C260-C285, doi: 10.1137/21M1402510
Bauer, S., et al. (2020), TerraNeo-Mantle Convection Beyond a Trillion Degrees of Freedom Lecture Notes in Computational Science and Engineering, vol. 136, pp. 569-610, Springer, Editors: Bungartz, H.-J. and Reiz, S. and Uekermann, B. and Neumann, P. and Nagel, W., doi: 10.1007/978-3-030-47956-5_19