Currently working as a software developer in the ATML Parallel performance at the Jülich Supercomputing Centre
My work focuses on development on the performance analysis infrastructure Score-P. My main interests lie in anything and everything accelerator, compiler and parallel paradigm related. This is reflected in both personal and work projects, trying to bridge a gap between topics being interesting and directly contributing to them, if possible.
Click to open
- Extending support for the OpenMP Tools Interface in Score-P to fully support host-side events (abstract)
- Extending support further to also support the device tracing interface in Score-P (webinar)
- Implementing a basic information dumping tool, used to understand the dispatched event sequence of OpenMP programs with available OpenMP runtimes and handling their pecularities in Score-P
- Adding an additional compiler instrumentation backend to Score-P, based on LLVM IR, with continously keeping up with changes upstream
- Collaboratively bring XRay into Score-P as another compiler instrumentation backend
- Work on re-writing the CUDA instrumentation in Score-P via CUPTI, going from the old CUDA 4.0 instrumentation to CUDA 11.6 and newer
- Work on extending the CUDA adapter of Score-P, introducing features like supporting CUDA graphs, user-defined activities and multi-subscriber support
- Investigating accelerator support of instrumentation frameworks like PTI-SDK
- Extending support for OpenMP v6.0 in Score-P
- Contributing to LLVM OpenMP & Offload to fix issues in the OpenMP Tools Interface and add new features
- 29 years old
- Master of Science in Applied Mathematics and Informatics
- Hobbies: Gaming, Reading, Music (Drum&Bass), Hiking, Bouldering
Find me on:



