Name: Seminar on Efficient Programming of HPC Systems - Frameworks and Algorithms
Start: 2026-07-08T09:00:00+02:00
End: 2026-07-08T18:00:00+02:00
Location: MPCDF

Seminar on Efficient Programming of HPC Systems - Frameworks and Algorithms

Wednesday, 8 July 2026 - 09:00

Monday, 6 July 2026
Tuesday, 7 July 2026
Wednesday, 8 July 2026
09:00 Introduction - Erwin Laure (MPCDF)
Introduction
- Erwin Laure (MPCDF)
09:00 - 09:05
Room: Hörsaal (Ground Floor)
09:05 HPX - A modern C++ task parallelization framework - Guillermo Marcos Lara
HPX - A modern C++ task parallelization framework
- Guillermo Marcos Lara
09:05 - 09:30
Room: Hörsaal (Ground Floor)
09:30 Portable GPU Programming with OpenMP Target Offloading - Denys Myshak
Portable GPU Programming with OpenMP Target Offloading
- Denys Myshak
09:30 - 09:55
Room: Hörsaal (Ground Floor)
09:55 Julia for High-Performance Computing - Daniel Singh
Julia for High-Performance Computing
- Daniel Singh
09:55 - 10:20
Room: Hörsaal (Ground Floor)
10:20 Break
Break
10:20 - 10:35
Room: Hörsaal (Ground Floor)
10:35 Seminar Paper on HPC Storage and Lustre - Daymon Schodits
Seminar Paper on HPC Storage and Lustre
- Daymon Schodits
10:35 - 11:00
Room: Hörsaal (Ground Floor)
11:00 IO-Aware Attention Across GPU Generations: How FlashAttention Tracks the Moving Bottleneck - Tom Osterfeld
IO-Aware Attention Across GPU Generations: How FlashAttention Tracks the Moving Bottleneck
- Tom Osterfeld
11:00 - 11:25
Room: Hörsaal (Ground Floor)
11:25 Balancing Performance and Accuracy: Mixed-Precision Algorithms for Linear Algebra in HPC - Felix Weißleder
Balancing Performance and Accuracy: Mixed-Precision Algorithms for Linear Algebra in HPC
- Felix Weißleder
11:25 - 11:50
Room: Hörsaal (Ground Floor)
11:50 Lunch break
Lunch break
11:50 - 12:45
Room: Hörsaal (Ground Floor)
12:45 Alpaka - Bora Uygar Özyurt
Alpaka
- Bora Uygar Özyurt
12:45 - 13:10
Room: Hörsaal (Ground Floor)
13:10 An MLIR-Based Approach to HPC Portability - Yannick Schürmann
An MLIR-Based Approach to HPC Portability
- Yannick Schürmann
13:10 - 13:35
Room: Hörsaal (Ground Floor)
13:35 JAX/XLA for Accelerator Performance Portability: Compiler Mechanisms and Quantitative Evidence - Pau Marín Roig
JAX/XLA for Accelerator Performance Portability: Compiler Mechanisms and Quantitative Evidence
- Pau Marín Roig
13:35 - 14:00
Room: Hörsaal (Ground Floor)
14:00 Break
Break
14:00 - 14:15
Room: Hörsaal (Ground Floor)
14:15 Mitigating Load Imbalance in Molecular Dynamics through Adaptive Task Parallelism - Audrey Kyrene Chen Kartamihardjo
Mitigating Load Imbalance in Molecular Dynamics through Adaptive Task Parallelism
- Audrey Kyrene Chen Kartamihardjo
14:15 - 14:40
Room: Hörsaal (Ground Floor)
14:40 ADIOS in High-Performance Computing Architecture and Performance Improvements for Scientific Simulations - Thomas Krachten
ADIOS in High-Performance Computing Architecture and Performance Improvements for Scientific Simulations
- Thomas Krachten
14:40 - 15:05
Room: Hörsaal (Ground Floor)
15:05 Data Layouts on Heterogeneous Systems - Lukas Englhauser
Data Layouts on Heterogeneous Systems
- Lukas Englhauser
15:05 - 15:30
Room: Hörsaal (Ground Floor)
15:30 Break
Break
15:30 - 15:45
Room: Hörsaal (Ground Floor)
15:45 Non-IEEE and Reduced-Precision Number Formats for AI and HPC - Paul Fleischmann
Non-IEEE and Reduced-Precision Number Formats for AI and HPC
- Paul Fleischmann
15:45 - 16:10
Room: Hörsaal (Ground Floor)
16:10 Model parallelization strategies for inference and training - Nils Marvin Quiring
Model parallelization strategies for inference and training
- Nils Marvin Quiring
16:10 - 16:35
Room: Hörsaal (Ground Floor)
16:35 Scaling Massive Models Efficiently: Fully Sharded Data Parallelism in PyTorch - Sai Krishna Sriyash Kommalapati
Scaling Massive Models Efficiently: Fully Sharded Data Parallelism in PyTorch
- Sai Krishna Sriyash Kommalapati
16:35 - 17:00
Room: Hörsaal (Ground Floor)
17:00 Conclusions
Conclusions
17:00 - 17:20
Room: Hörsaal (Ground Floor)

Choose timezone