Seminar on Efficient Programming of HPC Systems - Frameworks and Algorithms

Wednesday 16 Jul 2025, 09:00 → 18:00 Europe/Berlin

MPCDF Hörsaal

09:00 → 09:15 Welcome

Speaker: Erwin Laure (MPCDF)

09:15 → 09:45 CUDA

Speaker: Max Konstanian

09:45 → 10:15 OpenMP’s Role in Today’s HPC Landscape

Speaker: Markus Fabian Schorer

10:15 → 10:45 Comparative Analysis of GPU-based Approaches to the Traveling Salesman Problem

Speaker: Jonas Hörter

10:45 → 11:00 Break

11:00 → 11:30 Cabana: performance-portable library for particle-based simulations

Speaker: Loic Euzenot

11:30 → 12:00 Achieving Performance Portability Across Heterogeneous HPC Systems Using Alpaka

Speaker: Malte Clement

12:00 → 12:30 Investigating the StarPU Runtime System: A Study in Design, Capabilities, and Drawbacks

Speaker: Mehmet Emre Cakal

12:30 → 13:30 Lunch

13:30 → 14:00 Optimizing Application Performance for HPC: A Comparative Analysis of AoS, SoA, and AoSoA Data Layouts

Speaker: Adrian Simon Würth

14:00 → 14:30 Lossy Data Compression in High Performance Computing

Speaker: Raphael Dichler

14:30 → 15:00 Optimizing High-Performance Applications through Mixed-Precision Computing

Speaker: Mathis Bryan Kengni Fozin

15:00 → 15:15 Break

15:15 → 15:45 SLATE - Software for Linear Algebra Targeting Exascale

Speaker: Damla Tuana Karasozen

15:45 → 16:15 Ginkgo and its Application in Solving Large Sparse Linear Systems

Speaker: Elisei Shushpanov

16:15 → 16:45 Efficient Programming of the FFT on HPC Systems

Speaker: Vicente Serna Adame

16:45 → 17:15 Parallelism Strategies for Efficient LLM Training

Speaker: Hesam Mohammad Rezaei

17:15 → 17:30 Conclusions