Name: Efficient simulations on GPU hardware
Start: 2022-10-24T09:30:00+02:00
End: 2022-10-27T19:00:00+02:00
Location: ETH Zürich

Efficient simulations on GPU hardware

from Monday 24 October 2022 (09:30) to Thursday 27 October 2022 (19:00)

Monday 24 October 2022

13:00 Introduction
Introduction
13:00 - 13:20
Room: HIT E 51
13:20 Keynotes
Keynotes
13:20 - 16:00
Room: HIT E 51
Contributions

13:20 Keynote 1: The Rise of Tightly-Coupled Processor Clusters - Luca Benini (ETHZ)

14:00 Keynote 2: Future Architectures for Computing and AI - Bill McColl (Huawei)

14:40 Keynote 3: Scaling Machine Learning and Scientific Computing: A Data-Centric Perspective - Tal Ben-Nun (ETH Zürich)

15:20 Keynote 4: To the Exascale, and Beyond: Computing Challenges in Lattice QCD - Kate Clark
16:00 Coffee Break
Coffee Break
16:00 - 16:30
Room: In front of HIT E 51
16:30
16:30 - 17:00
Room: HIT E 51
Contributions

16:30 Student Flash-Talk 1: Cachew: Machine Learning Input Data Processing as a Service - Dan-Ovidiu Graur (ETH Zurich)

16:45 Student Flash-Talk 2 - Foteini Strati (ETH Zürich)
17:00 Panel Discussion - Tal Ben-Nun (ETH Zürich) Luca Benini (ETHZ) Fritz Bayer Marina Krstic Marinkovic (ETH Zurich)
Panel Discussion
- Tal Ben-Nun (ETH Zürich)
- Luca Benini (ETHZ)
- Fritz Bayer
- Marina Krstic Marinkovic (ETH Zurich)
17:00 - 17:30
Room: HIT E 51
17:30 Closing words and outlook
Closing words and outlook
17:30 - 17:40
Room: HIT E 51

17:40

17:40 - 19:40
Room: Restaurant Bellavista
Tuesday 25 October 2022
09:30
09:30 - 11:30
Room: HIT E 51
Contributions

09:30 Talk 1: Machine Learning applications to lattice QCD - Andrea Shindler (Michigan State University)

10:45 Talk 2: Applications of normalizing flows as generative models for lattice field theory - Javad Komijani (ETH Zürich)
11:30

11:30 - 12:00
Room: HIT E 51

12:00 Lunch
Lunch
12:00 - 13:00

13:00

13:00 - 13:45
Room: HIT E 51
13:45
13:45 - 17:15
Room: HIT E 51
Contributions

13:45 Talk 3: Machine learning hadron spectral functions in lattice qcd - Heng-Tong Ding

15:00 Talk 4: Hierarchical autoregressive approach to two-dimensional statistical systems - Piotr Korcyl

15:45 Talk 5: Efficient simulations of ML and LQCD - Denis Boyda (Argonne National Laboratory)
Wednesday 26 October 2022
09:30
09:30 - 11:30
Room: HIT E 41.1
Contributions

09:30 Talk 6: Real-time techniques and topological data analysis for non-perturbative phenomena in QFT - Daniel Spitz (Heidelberg University, Institute for Theoretical Physics)

10:45 Talk 7: tmLQCD on GPUs: minimum effort approach to performance-portability - Bartosz Kostrzewa
11:30

11:30 - 12:00
Room: HIT E 41.1

12:00 Lunch
Lunch
12:00 - 13:00
Room: HIT

13:00

13:00 - 13:45
Room: HIT E 41.1
13:45
13:45 - 16:30
Room: HIT E 41.1
Contributions

13:45 Talk 8: ATLAS fast simulation 3 (Atlfast3) is a performance-focused variant of the full ATLAS simulation - Henry Ann Day-Hall

15:00 Talk 9: One Code and Four APIs : performance portable software for lattice field theory - Peter Boyle

15:45 Talk 10: Experience using MILC and QUDA on various GPU systems - Steven Gottlieb
16:30

16:30 - 17:00
Room: HIT E 41.1
Thursday 27 October 2022
09:30
09:30 - 11:30
Room: HIT E 41.1
Contributions

09:30 Talk 11: GPU port of openQCD using CUDA - Felix Ziegler

10:45 Talk 12: HILA lattice simulation framework - write once, run everywhere - Kari Rummukainen
11:30

11:30 - 12:00
Room: HIT E 41.1

12:00 Lunch
Lunch
12:00 - 13:00

13:00

13:00 - 13:30
Room: HIT E 41.1
13:30
13:30 - 16:30
Room: HIT E 41.1
Contributions

13:30 Flash-Talk: Perspectives and challenges in studying the properties of strongly interacting matter in lattice qcd calculations - Jishnu Goswami (Bielefeld University)

13:45 Talk 13: Optimized Deep Learning Inference on High Level Trigger at the LHC: Computing time and Resource assessment - Syed Anwar Ul Hasan (ETH Zurich)

15:00 Talk 14: QUDA: Getting more QCD out of your GPU - Mathias Wagner

15:45 Talk 15: Simulations of Lattice Quantum Chromodynamics on GPUs - Balint Joo
16:30

16:30 - 17:00
Room: HIT E 41.1