High-Performance Storage

Infinite Storage

Infinite Possibilities

The research group High-Performance Storage improves the capabilities of storage landscapes applying smart concepts. We speak big data analytics and high-performance computing and apply our knowledge to meet the needs of environmental modeling.

Further information about our mission.

The HPS group is tightly integrated into the GWDG AG Computing group with more than 25 people.

2021-06-01: Julian is now Professor in High-Performance Computing at the Georg-August-Universität Göttingen and the deputy head High-Performance Computing at the GWDG.
2020-09-30: Julian is guest editor of the special issue “High-Performance Computing Algorithms and their Applications 2021” in MDPI's journal Algorithms
2019-09-08: Listen to the Radio Free HPC podcast with the IO-500 team
2019-08-25: BCS interview: Inside the Open Source Community

Email security with Postfix/DKIM/DMARC on Ubuntu 20.04 (2020-07-20 11:56)
Green screen remover (2018-05-01 20:57)
Impact of KPTI on the IO-500 for DKRZ (2018-01-23 10:58)
Fine grained Impact of KPTI on HPC nodes I/O performance (2018-01-23 10:40)
Impact of KPTI on HPC storage performance (2018-01-19 18:06)

Explore our research using the power of machine learning, click on a keyword to see relevant documents.

DAYA: Data-Aware Agentic AI Architecture for Decision-Oriented Analytics on HPC Systems (Michael Bidollahkhani, Orr Shomroni), NHR Conference 2026 - HPC-Driven Decision Science: Optimization, Simulation, and AI for Economics and Management, Heinz Nixdorf MuseumsForum, Paderborn, Germany, 2026-09-15
GWDG AI Services and Requirements for a Scientific AI Service (Michael Bidollahkhani), Bremen AI School, Bremen, Germany, 2026-08-19
Research Exchange at the Cambridge University Press Booth (Michael Bidollahkhani), IJCAI-ECAI 2026, Congress Centrum Bremen, Bremen, Germany, 2026-08-18
KISSKI (Sascha Safenreider), Einstiegsberatung, KISSKI, 2026-06-24 Presentation
KI in der Ausbildung (UMG-Ausbildungsverantwortliche) (Prof. Dr. Julian Kunkel, Sascha Safenreider), Chat AI und RAG, Historische Sternwarte Göttingen, 2026-06-19

Poster: Resilient HPC Operations through Observability-Aware Predictive Maintenance (Michael Bidollahkhani), 2026-09-14 BibTeX URL
Access to AI, Beyond a Liberty (Meisa Kamyab, Michael Bidollahkhani, Julian M. Kunkel), 2026-06 BibTeX
Robust I/O Characterization of Machine Learning Workloads Across Performance Analysis Tools (Zoya Masih, Radita Liem, Julian Kunkel), 2026-05 BibTeX URL DOI
When GPUs Fail Quietly: Observability-Aware Early Warning Beyond Numeric Telemetry (Michael Bidollahkhani, Freja Nordsiek, Julian M. Kunkel), 2026-03-17 BibTeX URL DOI
AI-Powered Smart Cities (Michael Bidollahkhani), 2026 BibTeX URL

Performance Evaluation of LLM Inference Engines, Yining Liu (Bachelor's Thesis), Advisors: Prof. Dr. Julian Kunkel, Dr. Jonathan Decker, 2026-08-21, Thesis BibTeX
Performance Evaluation of vLLM and LMCache for Conversational AI Workloads on HPC, Dennis Treder (Bachelor's Thesis), Advisors: Prof. Dr. Julian Kunkel, Dr. Jonathan Decker, 2026-08-14, Thesis BibTeX
Accelerating Irregular Parallel Algorithms on SpiNNaker2 Hardware, Jakob Aljoscha “Jake” von Dömming (Bachelor's Thesis), Advisors: Prof. Dr. Julian Kunkel, Dr. Jonathan Decker, 2026-08-11, Thesis BibTeX
Exploiting KV Caches to reduce Prefill Time in LLM Inference, Torben Heine (Bachelor's Thesis), Advisors: Prof. Dr. Julian Kunkel, Dr. Jonathan Decker, 2026-08-03, Thesis BibTeX
A Multi-Dimensional Evaluation Framework for LLM Inference Optimization in Production AI-as-a-Service Environments, Vansh Mohan (Master's Thesis), Advisors: Prof. Dr. Julian Kunkel, Dr. Jonathan Decker, 2026-07-31, Thesis BibTeX

News