Fabian Czappa

TU Darmstadt, Working area Parallel Programming

Contact

TU Darmstadt
Working area Parallel Programming
Hochschulstr. 10
64289 Darmstadt

parallelism@nhr4ces.de

Biography

Fabian received his Bachelor’s and Master’s degree from the Technical University of Darmstadt, for which he focused on the intersection of massively parallel data processing (both on GPUs and CPUs) and theoretical computer science.

Since 2020, Fabian has been part of the “Parallel Programming” group at the Technical University of Darmstadt and thus is part of the Cross-Sectional Group “Parallelism and Performance”. He is pursuing his Ph.D. in the context of computational neuroscience.

Thematic Advice

High-performance applications intrinsically suffer from communication, be it in the form of actual message exchange or (implicit) synchronization. Making this communication visible is key to understanding the application’s scalability.

Reducing this communication – exchanging the algorithm, moving the computation close to the data, or just restructuring the synchronization points – is a big part of HPC application development that is often overlooked.

Professional Competence

Fabian’s main interests and competencies lie in developing and maintaining high-performance applications utilizing MPI with synchronous and asynchronous communications.

Furthermore, Fabian has a strong background in theoretical computer science, e.g., in complexity theory, (higher-order) computational theory, and branches of model theory.