PASC23 Program: Interdisciplinary Dialogue & Awarded Talks

The PASC23 team is delighted to highlight today the conference interdisciplinary dialogue featuring computational physicist Marina Krstic Marinkovic and mechanical engineer Jens Honore Walther, who will discuss research on flow simulations across various scales, touching applications in carbon nanotubes, molecular dynamics, chemical physics and computational fluid dynamics. We are also pleased to announce that we will be hosting talks by the winners of the “PRACE HPC Excellence Award” and “PRACE Ada Lovelace Award for HPC,” as well as by the winner of the “PASC23 Best Paper Award.”

We take the occasion to remind PASC23 program presenters to carefully read the “Guidelines for Presentations” available to registered participants on the event platform prior to the conference start.

Last but not least, we are especially excited to see that the list of PASC23 participants reached 434 registrants, most of whom will be joining the event in-person! A list of registered participants as of today, can be found here.

We look forward to welcoming you all to PASC23 at the end of the month!

PASC23 program spotlight on interdisciplinary dialogue

In line with one of the core principles of the PASC Conference – interdisciplinarity – we are delighted to announce a steady appointment on the PASC Conference program: the interdisciplinary dialogue.

We look forward to hosting Professor Marina Krstic Marinkovic from ETH Zurich in the role of interviewer and Professor Jens Honore Walther from the Technical University of Denmark in the role of interviewee during this year’s dialogue entitled “Flow Simulations Across Scales: From Atoms to Vortices.”

Most phenomena in fluid dynamics are most naturally described by considering the fluid as a collection of particles: Lagrangian elements that carry the properties of the flow and whose motion describes the evolution of the system. The discrete representations range from atomistic models of nano scale systems, to granular and dispersed flow, to continuum models including vortex particle methods and smooth particle hydrodynamics. These methods serve as accurate and efficient tools for fundamental studies in fluid dynamics. For general engineering problems with flows in complex geometries and subject to a multitude of physical processes Eulerian methods prevail. Examples from flow simulations across scales include carbon nanotube membranes bluff body aerodynamics, gravity waves, and combustion.

Additional information can be found here.

PASC23 presentations on awarded research

Plenary sessions on Wednesday morning will focus on outstanding HPC research that has been recently awarded by PRACE, and in the context of the PASC23 Conference Papers Program.

PRACE HPC Excellence Award 2023

Petros Koumoutsakos will open the morning with a presentation on “AI, Computing and Thinking: Algorithmic Alloys for Advancing Scientific Discovery.”

Our generation has experienced more than a billion-fold increase in computer hardware capabilities and a dizzying pace of acquiring and transmitting massive amounts of data. Harnessing this resource has introduced a new form of inquiry: Computing. Computing is transforming our Thinking for solving complex problems and it is fueling the Artificial Intelligence (AI) revolution that is changing our world. I believe that we are at the dawn of a new era in science that would benefit from forming “alloys” of AI, Computing and Thinking.

I will present two examples of this process: Learning the Effective Dynamics of multi-scale systems and a fusion of Multi-Agent Reinforcement Learning and HPC for modeling and control of biological and turbulent flows. I will juxtapose successes and failures and argue that the proper integration of domain knowledge, AI expertise and Computing are essential to advance scientific discovery and technology frontiers.

Petros Koumoutsakos (Harvard University, USA)

PRACE Ada Lovelace Award for HPC 2023

Sarah Neuwirth will follow with a talk entitled “Leveraging HPC Performance Engineering to Support Exascale Scientific Discovery.”

HPC applications are evolving not only to include traditional modeling and simulation bulk-synchronous scale-up workloads but also scale-out workloads including: artificial intelligence, big data analytics, deep learning, and complex workflows. Given the ever-growing complexity of supercomputers and the advent of exascale computing, these trends can create a gap between expected and observed peak performance. Therefore, performance engineering is critical to bridge this gap through reproducible benchmarking, prediction, optimization, and analysis of large-scale HPC workloads. In this talk, I will highlight the challenges and opportunities in leveraging modular HPC performance engineering to support exascale scientific discovery. This will include introducing the key pillars of performance engineering such as user-friendly tool infrastructures, performance modeling, automatic optimization through integration into the application/system lifecycle, and feedback and user engagement.

Sarah Neuwirth (Goethe University Frankfurt, Germany)

PASC23 Best Paper

We are excited to continue on with Thorsten Kurth’s presentation on “FourCastNet: Accelerating Global High-Resolution Weather Forecasting Using Adaptive Fourier Neural Operators.”

Extreme weather amplified by climate change is causing increasingly devastating impacts across the globe. The current use of physics-based numerical weather prediction (NWP) limits accuracy due to high computational cost and strict time-to-solution limits. We report that a data-driven deep learning Earth system emulator, FourCastNet, can predict global weather and generate medium-range forecasts five orders-of-magnitude faster than NWP while approaching state-of-the-art accuracy. FourCastNet is optimized and scales efficiently on three supercomputing systems: Selene, Perlmutter, and JUWELS Booster up to 3,808 NVIDIA A100 GPUs, attaining 140.8 Petaflops in mixed precision (11.9% of peak at that scale). The time-to-solution for training FourCastNet measured on JUWELS Booster on 3,072 GPUs is 67.4 minutes, resulting in an 80,000 times faster time-to-solution relative to state-of-the-art NWP, in inference. FourCastNet produces accurate instantaneous weather predictions for a week in advance and enables enormous ensembles that could be used to improve predictions of rare weather extremes. 

Thorsten Kurth (NVIDIA Inc., Switzerland)

Additional information on these talks can be found here.

Guidelines for presentations at PASC23

The paper and minisymposium sessions will be held in-person at the Congress Center Davos with the possibility of digital participation via Zoom.

When you registered for the conference, you received a confirmation of registration with an invitation to activate your profile (email from info@pasc-conference.org with subject line: “PASC23 Conference registration”). Those credentials will give you access to the platform where PASC23 will be held from June 26 to 28, and where session recordings will be stored post-conference. Zoom connection details will be provided from within that platform.

In-person speakers will also need to share their screens over Zoom when they are presenting.

Please note that remote attendees will see presentation slides and will be able to listen to the talks and to the questions asked in the physical room; however, they will not see in-person presenters. Remote attendees can ask questions live (with audio and video where possible), or use the text chat to ask questions.

Technical information and guidelines for presenters are available to registered participants on the event platform after login with their credentials.