COMA Research Group - Group Retreat 2026
Chair of Computational Mathematics
We are part of the TUM School of Computation, Information and Technology at TUM Campus Heilbronn. We design numerical algorithms and high-performance software that let scientific applications make full use of modern supercomputers, from sparse linear algebra, preconditioning and mixed-precision methods to open-source libraries such as Ginkgo that run portably on CPUs and GPUs.
Latest News & Events
- We welcome our new PhD students Jonas Bachmann, Lukas Petermann and Friedrich Tschirpke to the group!
Research at a glance
Sparse kernels
Fast, load-balanced sparse matrix kernels for modern CPUs and GPUs.
Solvers & preconditioners
Krylov methods with parallel ILU, sparse approximate inverse, multigrid and domain decomposition.
Mixed precision
Store data compactly and compute accurately: less data movement at the same accuracy.
Batched solvers
Thousands of small, independent linear systems solved in a single GPU kernel.
Asynchronous methods
Iterative methods that avoid global synchronization on multi-GPU systems.
Get involved
Theses
Write your thesis with us
Bachelor's and Master's theses in high-performance computing, sparse linear algebra and scientific software.
Teaching
Courses this semester
Lectures, seminars and practical courses on numerical methods, parallel and GPU computing.
Open source
Use our software
Ginkgo, NeoN and more: open-source libraries for portable high-performance numerics.
Collaborations & Partners
