Salishan 2025 Program
Monday, April 21
4:30pm : Registration opens
Welcome/Keynote Address
6:00pm : HPC: Learn from the Past, Build the Future | Daniel Reed, University of Utah
5:00pm : Random Access Submissions Open
7:00pm : Reception and Informal Discussions
Welcome/Keynote Address
6:00pm : HPC: Learn from the Past, Build the Future | Daniel Reed, University of Utah
5:00pm : Random Access Submissions Open
7:00pm : Reception and Informal Discussions
Tuesday, April 22
7:30am : Breakfast
Session 1 - Discussion Panel: Reflections on Exascale
8:30am : ECP Post-Mortem: What We Got Right, What We Got Wrong | Doug Kothe, Sandia National Laboratories
8:42am : Exascale vs the Exascale Report: Why so long? And Are We Really There Yet? | Peter Kogge, University of Notre Dame
8:54am : AI and (Traditional) HPC Divergence - Existential Crisis for HPC, or just a Myth? | Satoshi Matsuoka, RIKEN Center for Computational Science
9:06am : Software Ecosystems for Exascale and Beyond | Lois Curfman McInnes, Argonne National Laboratory
9:18am : The Lasting Impact of ECP at LANL | Tim Randles, Los Alamos National Laboratory
9:30am : 20/20 hindsight: What did DOE get right (and wrong) about exascale? | Rob Neely, Lawrence Livermore National Laboratory
9:42am: An update on PEZ focusing on the experience of "E" Exascale | Bob Wisniewski, Hewlett Packard Enterprise
10:00am : Break
10:30am : Q&A Panel Session
12:00pm : Lunch
Session 2 - Applications using Exascale Machines
1:30pm : Hardware limits & opportunities for the future of biomolecular simulation | Erik Lindahl, National Academic Infrastructure for
Supercomputing, Sweden
2:00pm : Exascale Simulation of Turbulence | Kenneth Jansen, University of Colorado, Boulder
2:30pm : Asynchronous-Many-Task Systems: Challenges and Opportunities - Scaling an AMR Astrophysics Code on Exascale machines using Kokkos and HPX
| Patrick Diehl, Los Alamos National Laboratory
3:00pm : Break
3:30pm : Multiphysics at the Exascale: Portable Astrophysical Turbulence, Transport, and Kinetics with Flash-X | Bronson Messer,
Oak Ridge National Laboratory
4:00pm : Q&A Panel Session
5:00pm : Random Access Submissions Close
6:00pm : Working Dinner
Introduction to the Physics of Baseball Pitches | Grey Wilburn, Philadelphia Phillies of Major League Baseball
Session 1 - Discussion Panel: Reflections on Exascale
8:30am : ECP Post-Mortem: What We Got Right, What We Got Wrong | Doug Kothe, Sandia National Laboratories
8:42am : Exascale vs the Exascale Report: Why so long? And Are We Really There Yet? | Peter Kogge, University of Notre Dame
8:54am : AI and (Traditional) HPC Divergence - Existential Crisis for HPC, or just a Myth? | Satoshi Matsuoka, RIKEN Center for Computational Science
9:06am : Software Ecosystems for Exascale and Beyond | Lois Curfman McInnes, Argonne National Laboratory
9:18am : The Lasting Impact of ECP at LANL | Tim Randles, Los Alamos National Laboratory
9:30am : 20/20 hindsight: What did DOE get right (and wrong) about exascale? | Rob Neely, Lawrence Livermore National Laboratory
9:42am: An update on PEZ focusing on the experience of "E" Exascale | Bob Wisniewski, Hewlett Packard Enterprise
10:00am : Break
10:30am : Q&A Panel Session
12:00pm : Lunch
Session 2 - Applications using Exascale Machines
1:30pm : Hardware limits & opportunities for the future of biomolecular simulation | Erik Lindahl, National Academic Infrastructure for
Supercomputing, Sweden
2:00pm : Exascale Simulation of Turbulence | Kenneth Jansen, University of Colorado, Boulder
2:30pm : Asynchronous-Many-Task Systems: Challenges and Opportunities - Scaling an AMR Astrophysics Code on Exascale machines using Kokkos and HPX
| Patrick Diehl, Los Alamos National Laboratory
3:00pm : Break
3:30pm : Multiphysics at the Exascale: Portable Astrophysical Turbulence, Transport, and Kinetics with Flash-X | Bronson Messer,
Oak Ridge National Laboratory
4:00pm : Q&A Panel Session
5:00pm : Random Access Submissions Close
6:00pm : Working Dinner
Introduction to the Physics of Baseball Pitches | Grey Wilburn, Philadelphia Phillies of Major League Baseball
Wednesday, April 23
7:30am : Breakfast
8:00am : Random Access voting opens
Session 3 - AI Enabling Scientific Workflow
8:30am : Not for Scale but for (Biological) Discovery: Small AI | Amarda Shehu, George Mason University
9:00am : AI-Accelerated Materials Discovery at Meta | Kyle Michel, Meta
9:30am : The Role of AI in Scientific Workflow Management | Ewa Deelman, University of Southern California
10:00am : Break
10:30am : Harnessing Exascale: Transforming Science and Innovation with El Capitan | Luc Peterson, Lawrence Livermore National Laboratory
11:00am : Q&A Panel Session
11:00am: Random Access voting closes
12:00pm : Lunch/Dinner on your own
5:00pm : Random Access (Talks order will be determined on the fly)
Characterize Data & identify outliers, nearly instantly | Donpaul Stephens
Flux Fiction: Simulating HPC Scheduling Without Running a Single Job | Michela Taufer
Make MPI Great Again | Patrick Bridges
Moving Multi-tenancy into Production | Jack Lange
Pushing Molecular Dynamics Strong Scaling with GPU-Initiated Communication and Multi-node NVLINK | Szilárd Páll
ARMing GPUs for Impactful Science with the GH200 Superchip (and true Exascale) | Torsten Hoefler
So, what do we do now? Possible Strategies for Systems in the Near-Term | Katherine Riley
Sparking Scientific Breakthroughs: Agent-Driven Idea Generation in AgentSpace | Bill Magro
The Artificial Scientist -- in-transit Machine Learning of Plasma Simulations | Sunita Chandrasekaran
The Business of Open Source: The Super Highway of Science | Bill Hoffman
Unexpected challenges of training AI at unprecedented scale | Glenn K. Lockwood
Use cheap RISC-V vector CPUs attached to your laptop! | Ron Minnich
We Do Not Discriminate—Unless It’s a Quantum State or Rethinking Cryogenic Control for Scalable Quantum Systems | Anastasiia Butko
“Why do electric cars have frunks?”: An introduction to Neuromorphic computing. | Srideep Musuvathy
8:00pm : Student Poster Session/Reception and Informal Discussions
Enabling Performant Inter-Node Communication for Kokkos Views | Nicole Avans, Tennessee Technological University
Harnessing Large Language Models for Exascale Code Generation | Nichole Etienne, Emory University
Performance Optimization of an Exascale Implicit Kinetic Plasma Simulation on El Capitan | Ian Lumsden, University of Tennessee, Knoxville
Discrete Diffusion Monte Carlo in Imp | Vincent Novellino, North Carolina State University
Role Models: Speaking the Same Language with LLMs | Toloupe Olatunbosun, Rochester Institute of Technology
Radiation Hydrodynamics at Scale with FleCSI-HARD | Alexander Strack, University of Stuttgart
Reduced Floating-Point Precision Implicit Monte Carlo for Thermal Radiation Transport | Simon Butson, Oregon State University
8:00am : Random Access voting opens
Session 3 - AI Enabling Scientific Workflow
8:30am : Not for Scale but for (Biological) Discovery: Small AI | Amarda Shehu, George Mason University
9:00am : AI-Accelerated Materials Discovery at Meta | Kyle Michel, Meta
9:30am : The Role of AI in Scientific Workflow Management | Ewa Deelman, University of Southern California
10:00am : Break
10:30am : Harnessing Exascale: Transforming Science and Innovation with El Capitan | Luc Peterson, Lawrence Livermore National Laboratory
11:00am : Q&A Panel Session
11:00am: Random Access voting closes
12:00pm : Lunch/Dinner on your own
5:00pm : Random Access (Talks order will be determined on the fly)
Characterize Data & identify outliers, nearly instantly | Donpaul Stephens
Flux Fiction: Simulating HPC Scheduling Without Running a Single Job | Michela Taufer
Make MPI Great Again | Patrick Bridges
Moving Multi-tenancy into Production | Jack Lange
Pushing Molecular Dynamics Strong Scaling with GPU-Initiated Communication and Multi-node NVLINK | Szilárd Páll
ARMing GPUs for Impactful Science with the GH200 Superchip (and true Exascale) | Torsten Hoefler
So, what do we do now? Possible Strategies for Systems in the Near-Term | Katherine Riley
Sparking Scientific Breakthroughs: Agent-Driven Idea Generation in AgentSpace | Bill Magro
The Artificial Scientist -- in-transit Machine Learning of Plasma Simulations | Sunita Chandrasekaran
The Business of Open Source: The Super Highway of Science | Bill Hoffman
Unexpected challenges of training AI at unprecedented scale | Glenn K. Lockwood
Use cheap RISC-V vector CPUs attached to your laptop! | Ron Minnich
We Do Not Discriminate—Unless It’s a Quantum State or Rethinking Cryogenic Control for Scalable Quantum Systems | Anastasiia Butko
“Why do electric cars have frunks?”: An introduction to Neuromorphic computing. | Srideep Musuvathy
8:00pm : Student Poster Session/Reception and Informal Discussions
Enabling Performant Inter-Node Communication for Kokkos Views | Nicole Avans, Tennessee Technological University
Harnessing Large Language Models for Exascale Code Generation | Nichole Etienne, Emory University
Performance Optimization of an Exascale Implicit Kinetic Plasma Simulation on El Capitan | Ian Lumsden, University of Tennessee, Knoxville
Discrete Diffusion Monte Carlo in Imp | Vincent Novellino, North Carolina State University
Role Models: Speaking the Same Language with LLMs | Toloupe Olatunbosun, Rochester Institute of Technology
Radiation Hydrodynamics at Scale with FleCSI-HARD | Alexander Strack, University of Stuttgart
Reduced Floating-Point Precision Implicit Monte Carlo for Thermal Radiation Transport | Simon Butson, Oregon State University
Thursday, April 24
7:30am : Breakfast
Session 4 - Algorithms
8:30am : Revolutionizing Supercomputers: Post-Exascale Insights from AI Architecture | Michael James, Cerebras Systems
9:00am : Scaling Laws in HPC and AI: Yesterday, Today and Tomorrow | Rio Yokota, Institute of Science Tokyo
9:30am : Automatic Differentiation as an Enabling Technology for Simulation, Analysis, and Scientific Machine Learning | Eric Phipps,
Sandia National Laboratories
10:00am : Break
10:30am : Dynamic Contextual Sparsity: The Next Paradigm for Scaling GenAI | Anshumali Shrivastava, Rice University
11:00am : Q&A Panel Session
12:00pm : Lunch
Session 5 - Systems of the Future
1:30pm : Exascale AI Research Resources Federation and Security: Yesterday, Today and Tomorrow | Sadaf Alam, Bristol Centre for
Supercomputing (BriCS)
2:00pm : Will the Cloud Crush HPC? Lessons from Ultra Ethernet | Keith Underwood, Hewlett Packard Enterprise
2:30pm : Complexity perplexity: How should HPC centers evolve to support the scientific workflows of the future? | Debbie Bard, National Energy Research
Scientific Computing Center (NERSC)
3:00pm : Break
3:30pm : Grasping the Opportunities in Front of Us: HPC Systems Built for Purpose | Dan Ernst, NVIDIA
4:00pm : Q&A Panel Session
5:00pm : Reception and informal Discussions
Session 4 - Algorithms
8:30am : Revolutionizing Supercomputers: Post-Exascale Insights from AI Architecture | Michael James, Cerebras Systems
9:00am : Scaling Laws in HPC and AI: Yesterday, Today and Tomorrow | Rio Yokota, Institute of Science Tokyo
9:30am : Automatic Differentiation as an Enabling Technology for Simulation, Analysis, and Scientific Machine Learning | Eric Phipps,
Sandia National Laboratories
10:00am : Break
10:30am : Dynamic Contextual Sparsity: The Next Paradigm for Scaling GenAI | Anshumali Shrivastava, Rice University
11:00am : Q&A Panel Session
12:00pm : Lunch
Session 5 - Systems of the Future
1:30pm : Exascale AI Research Resources Federation and Security: Yesterday, Today and Tomorrow | Sadaf Alam, Bristol Centre for
Supercomputing (BriCS)
2:00pm : Will the Cloud Crush HPC? Lessons from Ultra Ethernet | Keith Underwood, Hewlett Packard Enterprise
2:30pm : Complexity perplexity: How should HPC centers evolve to support the scientific workflows of the future? | Debbie Bard, National Energy Research
Scientific Computing Center (NERSC)
3:00pm : Break
3:30pm : Grasping the Opportunities in Front of Us: HPC Systems Built for Purpose | Dan Ernst, NVIDIA
4:00pm : Q&A Panel Session
5:00pm : Reception and informal Discussions