NFDIxCS – National Research Data Infrastructure for and with Computer Science

NFDIxCS promotes the implementation of the FAIR Data Principles for computer science (CS) research data and software artifacts. To this end, it establishes an organizational, technical, and interoperable infrastructure that connects services, tools, and stakeholders from and for computer science, enabling sustainable reuse, sharing, and discovery of research data and software across institutions and countries.

Within NFDIxCS, ZIH (TU Dresden)contributes its expertise in HPC performance analysis and data management. We develop automated, standardized, and FAIR-compliant workflows for performance metadata by extending established open-source tools such as Score-P, CUBE, OTF2, and center-level systems like PIKA. These extensions capture metadata on hardware, software, execution environments, and energy consumption, complemented by anonymization mechanisms to ensure responsible data sharing.

A core outcome of this work is the standardization of performance metadata schemas and reduced performance data summaries, enabling scalable storage and cross-site comparison of performance experiments. The resulting data are made accessible through a performance data portal integrated with NFDI-wide authentication and authorization infrastructures. Through data campaigns, documentation, and training activities, TU Dresden supports a shared ecosystem for performance data exchange, strengthening reproducibility and transparency in computational science.

Partners

https://nfdixcs.org/partners

Duration

01/2023–12/2027

Funding

DFG

ZIH Contact

Dr. Matthias Lieber