HPC 2026
High Performance Computing
State of the Art, Emerging Disruptive Innovations and
Future Scenarios
An
International Advanced Workshop
June
22 – 26, 2026, Cetraro, Italy
Final Programme
Sponsors
|
CEREBRAS |
|
|
|
CMCC Euro-Mediterranean
Center on Climate Change |
|
|
|
CSC Finnish Supercomputing Center |
|
|
|
DWAVE Systems |
|
|
|
Juelich Supercomputing Center, Germany |
|
|
|
LENOVO |
|
|
|
NVIDIA |
|
|
|
PLANQC |
|
|
|
PSIQUANTUM |
|
|
|
QUANTUM MACHINES |
|
|
|
RIGETTI Computing |
|
|
|
SAMBANOVA SYSTEMS |
|
|
|
UNIVERSITY OF SALENTO |
|
|
Media Partners
|
HPCwire is a news
portal and weekly newsletter covering the fastest computers in the world and
the people who run them. As the trusted source for HPC news since 1987, HPCwire
has served as the publication of record on the issues, challenges,
opportunities, and conflicts relevant to the global High Performance
Computing space. Its reporting covers the vendors, technologies, users, and
the uses of high performance, AI- and data-intensive computing within
academia, government, science, and industry. Subscribe now
at www.hpcwire.com. |
|
About
insideHPC Founded in 2006, insideHPC
is a global publication recognized for its comprehensive and insightful coverage
across the HPC-AI community, linking vendors, end-users and HPC strategists.
insideHPC has a large and loyal audience drawn from public and private
companies of all sizes, government agencies, research centers, industry
analysts and academic institutions. In short: the buyers and influencers of
HPC, HPC-AI and associated emerging technologies. THE News
Analysis Site for HPC Insiders: Written and edited by seasoned
technology journalists, we’re all about HPC and AI, offering feature stories,
commentary, news coverage, podcasts and video interviews with HPC’s leading
voices. Like the evolving HPC ecosystem, insideHPC’s coverage continually
expands into emerging focus areas to better serve our growing readership and
advertising base, insideHPC in 2023 will deliver an updated format and new
spotlight coverage of of enterprise HPC, HPC-AI, exascale (and post-exascale)
supercomputing, quantum computing, cloud HPC, edge computing, High
Performance Data Analytics and the geopolitical implications of
supercomputing. |
|
Simr, formerly UberCloud, is dedicated to helping manufacturers thrive
by advancing SimOps, a framework of best practices for automating simulation
processes that enhance decision-making, innovation, efficiency, and quality
in product design and engineering. Founded by engineers for engineers, Simr
builds on a versatile, high-performance platform compatible with any compute
infrastructure, setting a new standard in manufacturing design. Our platform
allows engineers to securely design and test product concepts using existing
workflows and tools, ensuring complete control over their simulation
workflows and data. Visit us at simr.com and
follow us on LinkedIn. |
|
ILKAY
ALTINTAS San
Diego Supercomputer Center and Workflows
for Data Science (WorDS) Center of Excellence and WIFIRE
Lab University
of California at San Diego, CA USA PETE
BECKMAN US
DOE Argonne National Laboratory and University
of Chicago and Northwestern
University / Argonne National Lab. Institute for Science and Engineering USA ANDREW BESTWICK RIGETTI Computing USA GIL BLOCH NVIDIA Mellanox Tel Aviv ISRAEL CHARLIE CATLETT Argonne National Laboratory Argonne, Illinois USA TIM CLARKE Sambanova Systems USA FRANCESCO FERRARI Quantum
Computing Solutions – Leonardo S.p.A. ITALY IAN
FOSTER Argonne
National Laboratory Data
Science and Learning Division Argonne,
IL and Dept.
of Computer Science The
University of Chicago Chicago,
IL USA ANDREA GENTILE Pasqal FRANCE WOLFGANG GENTZSCH Executive
Director & Co-founder of SimOps Regensburg GERMANY and Sunnyvale,
CA USA VLADIMIR
GETOV Distributed
and Intelligent Systems Research Group School
of Computer Science and Engineering University
of Westminster London UNITED
KINGDOM VLAD
GHEORGHIU Institute
for Quantum Computing, University of Waterloo and SoftwareQ
Inc, Waterloo Waterloo,
Ontario CANADA NEIL
GILLESPIE Riverlane UNITED
KINGDOM CHRISTOPH
GOHLE Planqc GERMANY ALFREDO
GOLDMAN Sao
Paulo University Institute
of Mathematics and Statistics Sao
Paulo BRAZIL NOBUYASU
ITO RIKEN
Center for Computational Science, Kobe JAPAN HIROAKI
KOBAYASHI Architecture
Laboratory Department
of Computer and Mathematical Sciences Graduate
School of information Sciences Tohoku
University JAPAN LORENZO LEANDRO Quantum Machines Milan ITALY DMITRY
LYAKH NVIDIA USA DYLAN
MAHLER Xanadu CANADA SALVATORE MANDRÀ Google, Mountain View, California USA STEFANO
MARKIDIS KTH
Royal Institute of Technology Computer
Science Department / Computational Science and Technology Division Stockholm SWEDEN ERIC
MUELLER University
of Heidelberg Heidelberg GERMANY KEVIN
OBENLAND Quantum
Information and Integrated Nanosystems Lincoln
Laboratory Massachusetts
Institute of Technology MIT Boston, MA USA TOMAS OPPELSTRUP Cerebras USA IRWAN
OWEN D-Wave
Systems Inc CANADA
- USA VALERIO
PASCUCCI Center
for Extreme Data Management, Analysis and Visualization and Scientific
Computing and Imaging Institute School
of Computing, University of Utah and Laboratory
Fellow, Pacific Northwest National Laboratory USA DAVID RIVAS RIGETTI Computing Berkeley, California USA JOHANNES
SCHEMMEL European
Institute for Neuromorphic Computing and Kirchoff
Institute for Physics Heidelberg
University Heidelberg GERMANY SEETHARAMI
SEELAM IBM
Research USA JEAN
SENELLART Quandela,
Chief Technology and Product Officer FRANCE PETE
SHADBOLT Co-founder PsiQuantum
Corp. Palo Alto, California USA THOMAS STERLING Senior
Research Scientist University
of Texas at Austin Texas
Advanced Computing Center Austin,
TX USA MICHAEL
STREIF Boehringer GERMANY SERGII
STRELCHUK Oxford
University Oxford
UK ANDREA
TABACCHINI Quantum
Brilliance AUSTRALIA
and GERMANY WILLIAM
TANG Princeton
University Dept. of Astrophysical Sciences, Princeton
Plasma Physics Laboratory and Center
for Statistics and Machine Learning (CSML) and Princeton
Institute for Computational Science & Engineering (PICSciE) Princeton
University USA |
Workshop
Agenda
Monday,
June 22nd
|
Session |
Time |
Speaker/Activity |
|
9:45 – 10:00 |
Welcome
Address |
|
|
State of the art
and future scenarios |
||
|
|
10:00 – 10:30 |
T. STERLING |
|
|
10:30 – 11:00 |
v. getov Dynamic
Analysis of Emerging Applications, Systems, and Architectures |
|
11:00 – 11:30 |
COFFEE BREAK |
|
|
|
11:30 – 12:00 |
e. mueller Taming
Complexity: Deploying Scientific Ecosystems via Portable HPC Containers |
|
12:00 – 12:30 |
CONCLUDING
REMARKS |
|
|
|
Emerging Computer
Systems and Solutions |
|
|
|
17:30 – 18:00 |
g. bloch Scaling
Out and Across: Networking Innovations for Giga-Scale AI Systems |
|
18:00 – 18:30 |
COFFEE BREAK |
|
|
|
18:30 – 19:00 |
A. GOLDMAN Mapping
the Environmental Impacts of HPC, and Backfilling and Its Blind Spots |
|
19:00 – 19:30 |
CONCLUDING REMARKS |
Tuesday,
June 23rd
|
Session |
Time |
Speaker/Activity |
|
Advanced AI
Processing: Challenges and Perspectives 1 |
||
|
|
10:00 – 10:25 |
i. foster |
|
|
10:25 – 10:50 |
p. beckman |
|
|
10:50 – 11:20 |
COFFEE BREAK |
|
|
11:20 – 11:45 |
T. CLARKE |
|
|
11:45 – 12:10 |
T. OPPLESTRUP |
|
|
12:10 – 12:35 |
C. CATLETT AI
Agents as Scientific Collaborators: Key Safety and Assurance Mechanisms |
|
|
12:35 – 12:50 |
CONCLUDING
REMARKS |
|
|
Advanced AI Processing: Challenges and Perspectives
2 |
|
|
|
17:30 – 18:00 |
V. PASCUCCI |
|
|
18:00 – 18:30 |
COFFEE BREAK |
|
|
18:30 – 19:00 |
J. SCHEMMEL |
|
|
19:00 – 19:30 |
CONCLUDING
REMARKS |
Wednesday,
June 24th
|
Session |
Time |
Speaker/Activity |
|
|
The QUANTUM COMPUTING
Promises 1 |
|
|
|
9:30 – 9:55 |
A. BESTWICK Scaling
a chiplet-based superconducting quantum processor toward fault-tolerance |
|
|
9:55 – 10:20 |
D. LYAKH |
|
|
10:20 – 10:45 |
S. SEELAM A
Reference Architecture for a Quantum-Centric Supercomputer |
|
|
10:45 – 11:10 |
S. mandrà |
|
|
11:10 – 11:40 |
COFFEE BREAK |
|
|
11:40 – 12:05 |
p. shadbolt Scaling
Toward Utility-Scale Fault-Tolerant Quantum Computing with Photonics |
|
|
12:05 – 12:30 |
d. mahler Scaling
Photonic Quantum Computing with Integrated GKP Qubit Sources |
|
|
12:30 – 12:45 |
CONCLUDING REMARKS |
|
|
The QUANTUM
COMPUTING Promises 2 |
|
|
|
17:00 – 17:25 |
i. owen How Quantum Computing is
Transforming Business Today |
|
|
17:25 – 17:50 |
D. RIVAS Rigetti Technology Update |
|
|
17.50 –18:15 |
a. tabacchini |
|
|
18:15 – 18:45 |
COFFEE
BREAK |
|
|
18:45 – 19:10 |
L. LEANDRO |
|
|
19:10 – 19:35 |
n. ito |
|
|
19:35 – 19:50 |
CONCLUDING REMARKS |
Thursday,
June 25th
|
Session |
Time |
Speaker/Activity |
|
|
The Quantum Computing
Prospects and Deployments 1 |
|
||
|
|
10:00 – 10:25 |
S. STRELCHUK Quantum
Pangenomics: From Sequence Assembly to the First Whole-Genome Encoding |
|
|
|
10:25 – 10:50 |
k. obenland Application
Driven Evaluation of Fault-Tolerant Quantum Computing Architectures |
|
|
|
10:50 – 11:15 |
H. KOBAYASHI Ising
Machine-Based Optimal Route Planning for Tsunami Rescue and Relief Resource
Allocation |
|
|
|
11:15 – 11:45 |
COFFEE BREAK |
|
|
|
The
Quantum Computing Prospects and Deployments 2 |
|
|
|
|
11:45 – 12:10 |
S. MARKIDIS Quantum
Accelerated Units: Toward Specialized QPUs for Scientific Computing and AI |
|
|
|
12:10 – 12:35 |
A. GENTILE Towards
Practical Hybrid Quantum Computing: Integrating Hardware, Software, and
Applications |
|
|
|
12:35 – 12:50 |
CONCLUDING REMARKS |
|
|
|
The Quantum
Computing Prospects and Deployments 3 |
|
|
|
|
17:00 – 17:30 |
J. SENEllart |
|
|
|
17:30 – 18:00 |
N. GILLESPIE |
|
|
18:00 – 18:15 |
CONCLUDING REMARKS |
|
|
|
18:15 – 18:45 |
COFFEE BREAK |
|
|
|
18:45 – 19:45 |
Chairperson:
W. Tang Panelists:
E. Mueller, V. Getov, D. Rivas, P. Shadbolt The Intersection of Quantum
Computing and HPC During
the past several decades, supercomputing speeds have gone from Gigaflops to
Teraflops, to Petaflops and Exaflops. As the end of Moore’s law approaches,
the HPC community is increasingly interested in disruptive technologies that
could help continue these dramatic improvements in capability. This
interactive panel will identify key technical hurdles in advancing quantum
computing to the point it becomes useful to the HPC community. Some questions
to be considered:
|
||
|
Is the “belle
époque” of classical High Performance Computer Systems coming at the end? |
|||
Friday,
June 26th
|
Session |
Time |
Speaker/Activity |
|
Key Projects, Novel
Developments and Challenging Applications |
||
|
|
10:00 – 10:30 |
F. FERRARI |
|
|
10:30 – 11:00 |
c. gohle MAQCS at
LRZ: Bringing Neutral-Atom Quantum Computing into Industrial HPC Workflows |
|
11:00 – 11:30 |
COFFEE BREAK |
|
|
|
11:30 – 12:00 |
W. TANG |
|
|
12:00 – 12:30 |
W. GENTZSCH |
|
12:30 – 12:45 |
CONCLUDING REMARKS |
Chairpersons
IAN FOSTER
Argonne National Laboratory
Data
Science and Learning Division
Argonne,
IL
and
Dept.
of Computer Science
The
University of Chicago
Chicago,
IL
USA
PETER
BECKMAN
Northwestern
University
Evanston,
Illinois
USA
VLADIMIR
GETOV
Distributed and Intelligent Systems Research
Group
School of Computer Science and Engineering
University of Westminster, London
UNITED KINGDOM
CHARLIE
CATLETT
Argonne National Laboratory
Argonne, Illinois
USA
WILLIAM
TANG
Princeton
University Dept. of Astrophysical Sciences,
Princeton
Plasma Physics Laboratory
and
Center
for Statistics and Machine Learning (CSML)
and
Princeton
Institute for Computational Science & Engineering (PICSciE)
Princeton University
USA
KEVIN
OBENLAND
Quantum
Information and Integrated Nanosystems
Massachusetts
Institute of Technology
Lincoln
Laboratory
Boston,
MA
USA
DAVID RIVAS
RIGETTI Computing
Berkeley, California
USA
DAVID RIVAS
RIGETTI Computing
Berkeley, California
USA
Quantum Brilliance
AUSTRALIA and GERMANY
ALFREDO GOLDMAN
Sao Paulo University
Institute of
Mathematics and Statistics
Sao Paulo
BRAZIL
Panel
Abstracts
|
AI Agents at the Edge — Is your Science
Ready? Pete Beckman Northwestern University, Evanston, Illinois, USA AI is on the move. Large,
sophisticated models are now being pushed to the edge, where smart
infrastructure, intelligent sensors, and advanced scientific instruments are
forming a new kind of AI-enabled computing continuum. The Sage Grande Testbed
is a groundbreaking national resource (www.sagecontinuum.org) designed to support this transformation. The
Sage Grande Testbed offers access to AI-enabled edge computing resources and
integrated sensor systems—including infrared and RGB cameras, microphones,
and atmospheric and air quality instruments—deployed across a range of
environments: natural landscapes, urban areas, and wildfire-prone regions.
Agentic workflows are now becoming possible at the edge. OpenClaw, Hermes, LangGraph and other
frameworks are transforming scientific inquiry. What happens when we Agentic workflows are
pushed to the edge and can control instruments? Come and find out. |
|
Scaling a chiplet-based superconducting
quantum processor toward fault-tolerance Andrew Bestwick Rigetti Computing, USA Rigetti Computing’s
modular quantum processors scale to large qubit counts by interconnecting
high fidelity, small scale superconducting quantum devices via inter-chip
tunable couplers. In this talk, I will discuss Rigetti’s progress in building
and integrating such QPUs by scaling chiplets laterally, allowing for steady
manufacturing yield and good performance up to very large systems. I will
conclude with Rigetti’s roadmap toward fault-tolerance with this
architecture. |
|
Scaling Out and Across: Networking
Innovations for Giga-Scale AI Systems Gil
Bloch NVIDIA Mellanox,
Tel Aviv, Israel As AI factories grow to
hundreds of thousands, and soon millions, of interconnected GPUs, the network
has become the defining architecture of large-scale AI. Achieving
uncompromised performance requires fabrics that not only scale up and out
within a single data center, but scale across multiple sites with predictable
latency, lossless throughput, and end-to-end resilience. In this session, NVIDIA
will discuss the full breadth of networking innovations that enable the
construction of giga-scale AI infrastructure. |
|
AI Agents as Scientific Collaborators: Key
Safety and Assurance Mechanisms Charlie Catlett Argonne National Laboratory, Argonne, Illinois,
USA OpenClaw and other
frameworks for managing AI-based agents have gained rapid momentum, yet basic out-of-the-box systems introduce
significant security and safety vulnerabilities if not carefully managed. We
discuss our safety/security/containment implementation architecture as well
as every significant functionality gaps that we have filled including
scheduling tasks, layered review of actions, a framework for relying on code
for deterministic actions and LLMs for actions requiring judgement or
creativity, and more convenient configuration management tools. |
|
Purpose Built Generative AI Platforms Tim Clarke Sambanova Systems, USA We are entering the era of
‘Inference-First’ AI design. This presentation explores the core technologies
driving this transition, specifically how SambaNova’s Dataflow chips move
beyond traditional constraints to deliver high-reasoning Premium Tokens at
scale. We will analyze the immediate value of maximizing Intelligence per
Joule, demonstrating how this efficiency metric allows enterprises to deploy
SOTA open-weight models sustainably today. Finally, we will look toward the
next horizon: Disaggregated Inference, the architectural shift that will
further decouple compute stages to unlock the next level of global AI
scaling. |
|
Quantum computing at Leonardo: applications
and insights Francesco
Ferrari Quantum Computing
Solutions – Leonardo S.p.A. Quantum computing
represents a rapidly evolving paradigm that holds promise for addressing
classes of problems that remain beyond the reach of conventional digital
computation. In this talk, we will outline Leonardo’s pragmatic effort
towards evaluating the capabilities and limitations of quantum computing for
industrial applications within the aerospace and defense domains. We will
focus mainly on quantum and hybrid quantum-classical approaches applied to
combinatorial optimization tasks. We will showcase recent results on two
representative applications, antenna placement for disaster response and
satellite mission scheduling in space operations, based on quantum approaches
developed in-house. These use cases illustrate the potential pathways for
integrating quantum computing into decision-making workflows. |
|
Accelerating Discovery with Intelligent
Agents Ian Foster Argonne National Laboratory, Data Science and
Learning Division, Argonne, IL and Dept. of Computer Science, The University
of Chicago, USA The convergence of massive
datasets, advanced machine learning, autonomous laboratories, and exascale
computing creates new opportunities for science and engineering. Realizing
this potential, however, requires new approaches to middleware and infrastructure.
This talk explores how embodied AI agents can create adaptive systems that
augment human expertise. These agents synthesize scientific knowledge,
orchestrate physical and computational experiments, and support real-time
problem solving. In doing so, they enable new levels of coordination that
allow researchers to focus on high-level discovery. At the same time,
managing and resourcing such agents introduces new challenges in software and
hardware infrastructure that must be addressed to fully realize AI-assisted
science. |
|
Towards Practical Hybrid Quantum Computing:
Integrating Hardware, Software, and Applications Andrea Gentile Pasqal, France As quantum computing
matures, the central challenge is no longer only improving individual
devices, but making quantum systems usable within larger hybrid computing
workflows. This requires progress
across the full stack, from hardware capability to software integration and
application execution. In this talk, we present a
broad overview of recent progress at Pasqal from that full-stack perspective.
We will discuss elements of the software and systems layer, including the
middleware, runtime, and workflow approaches needed to connect neutral-atom
quantum processors to practical hybrid execution environments. We exemplify this approach
by recent advances at the hardware and application levels, including using
256 qubits for quantum simulation of a real material. |
|
SimOps - An Engineering Simulation Framework
For Multi-Physics, Multi-Cloud, Machine Learning Workflows Wolfgang Gentzsch Executive Director & Co-founder of SimOps,
Regensburg, Germany and Sunnyvale, CA, USA Traditional on-premises
HPC architectures often fail to deliver the breakthrough improvements
required for modern, compute- and data-intensive engineering simulations. To
address this challenge, the authors and Dassault Systèmes developed a fully
automated, self-service multi-cloud simulation platform designed specifically
for high-performance cloud computing (HPC Cloud) and engineering simulations
and applied it to patient-specific cardiac repair with a machine learning
workflow for the Living Heart Project. This novel environment, which we
implemented on the Google Cloud Platform and Microsoft Azure, enhances
engineering productivity, bolsters IT security, and minimizes administrative
overhead while maintaining granular control over corporate assets. One of the
results of these simulations was a set of lessons learned and recommendations
which we included in the SimOps Simulation Operations list of best practices. |
|
Dynamic Analysis of Emerging Applications,
Systems, and Architectures Vladimir Getov Distributed and Intelligent Systems Research
Group, University of Westminster, London, U.K. Since the IRDS (International
Roadmap for Devices and Systems) inception in 2017, the Applications
Benchmarking (AB) and Systems and Architectures (SA) international teams have
contributed the top-down guidance to the IRDS communities. AB shared insight
into what we need to compute and tracked the major application domains while
SA provided insight on where that computing occurs across a second dimension
– system categories – from IoT Edge beacons to exascale supercomputers. The
rapid adoption of generative AI has provided an unprecedented stochastic
shock to the entire computing ecosystem, which demands changes to our
methodology and has led to the merger of the two teams into Applications,
Systems, and Architectures (ASA). At present,
application-specific attributes and requirements represent a third dimension
in the ASA considerations and analysis space. The stochastic shock of the
wide scale adoption of generative AI since 2023 marked the beginning of a new
era. The voraciousness of LLM training and inference for all resources:
water, energy, talent, information, die area, and wafer starts have
engendered unprecedented change in rate and diversity of the change
necessitated in our methodology. Key to this new process is the realization
that the changes are being implemented dynamically at scale much faster than
before. Given the heterogeneity of approaches, we now find ourselves having
to cast wider nets for information on trends across both the applications and
systems spaces and the holistic co-design is the required motivation to
address those new challenges. |
|
The Classical Bottleneck in Fault-Tolerant
Quantum Computing: Latency, Throughput, and the Decoder Problem Neil Gillespie Riverlane, United Kingdom Fault-tolerant quantum computing requires not just good qubits, but a
real-time classical processing system capable of keeping pace with the
quantum hardware. Any quantum error correction (QEC) system must
simultaneously satisfy three fundamental constraints: high throughput, low
latency, and high accuracy. Meeting all three at scale is a hard classical
computing problem - one that sits squarely at the intersection of HPC and
quantum. In this talk I will describe how Riverlane's Deltaflow system addresses
these constraints, focusing on the design of our Local Clustering Decoder and
its deployment on real quantum hardware. I will present experimental results
demonstrating real-time decoding in practice, and discuss what these results
tell us about the engineering challenges ahead on the path to fault-tolerant
quantum computing. |
|
MAQCS at LRZ: Bringing Neutral-Atom Quantum
Computing into Industrial HPC Workflows Cristoph Gohle PlanQC, Germany This talk presents MAQCS,
the project to integrate a neutral-atom quantum computer into the Leibniz
Supercomputing Centre (LRZ). The focus is on what this means in practice for
an HPC environment: turning a quantum processor into an accessible computing resource
rather than treating it as a stand-alone machine. A key aspect is the role
of neutral atoms in this context. Their scalability, programmability, and
operational flexibility make them a promising platform for integration into
HPC settings. Using MAQCS as a case study, the talk will discuss how such a
system can be connected to classical infrastructure and made accessible to
users. More broadly, MAQCS illustrates a realistic path for quantum computing
in HPC through system integration and operational deployment. |
|
Mapping the Environmental Impacts of HPC, and
Backfilling and Its Blind Spots Alfredo Goldman São Paulo University, Institute of Mathematics and
Statistics, Brazil High-Performance Computing
underpins much of modern science. Its environmental footprint extends well
beyond operational energy use. It also includes embodied carbon, water,
materials, and ecological impacts. This talk presents a systematic mapping
study of research on the environmental impacts of HPC. The study spans
sixteen years of literature and a final corpus of 62 reports. It classifies
each report along four facets: which impact is studied, when in the lifecycle
it occurs, where in the system stack it appears, and how it is managed. The
map shows a field concentrated in one corridor. Most work addresses carbon
emissions, the operational stage, the facility level, and measurement. Water,
materials, and biodiversity remain thin. These gaps may reflect the limits of
available signals more than the limits of HPC’s actual footprint. Water-aware
scheduling is one possible example, and it is one direction we wanted to
explore. The second half of the
talk builds on this. It introduces greenfilling, a set of backfilling-based
heuristics that respond to carbon and water-intensity signals. These
heuristics were explored through simulations. Together, the two parts attempt
to move from a structural account of what the field measures toward its blind
spots. |
|
HPC Simulation of Quantum Computer Nobuyasu Ito RIKEN Center for Computational Science, Kobe,
JAPAN Development of quantum
computers has been remarkable, but both quantum hardware and practical
applications are still in an early stage. Current QPUs remain at the NISQ
level, are expensive to operate, and exhibit relatively high error rates.
Classical simulation of quantum circuits therefore continues to play an
important role in developing quantum applications and validating QPU
executions. This talk introduces RIKEN
braket, a quantum circuit–level simulator developed within the JHPC-Quantum
project and designed for massively parallel HPC systems. Recent advances in
quantum computing simulation technologies will also be reviewed. |
|
Ising Machine-Based Optimal Route Planning
for Tsunami Rescue and Relief Resource Allocation Hiroaki Kobayashi Tohoku University, Japan In large-scale tsunami disasters
caused by major earthquakes, extensive coastal inundation forces many
residents to evacuate to emergency evacuation towers and shelters. Under such
conditions, local governments and emergency response agencies are required to
efficiently visit evacuation facilities using multiple rescue boats in order
to deliver relief supplies and rescue stranded victims under severe time and
resource constraints. This problem can be
generalized as a combinatorial optimization problem related to the Traveling
Salesman Problem (TSP) and the Vehicle Routing Problem (VRP). However,
practical optimization for tsunami disaster response must additionally
consider disaster-specific constraints such as inundated areas, shelter
capacities, the availability of multiple rescue boats, and priority shelters
that must be serviced earlier even at the expense of longer travel distances. In this talk, we present a
formulation of multi-boat routing and resource allocation problems for
tsunami disaster response and report computational results obtained using an
Ising machine. |
|
The Open Acceleration Stack: Large-scale
calibrations, quantum error correction, and the quest for quantum-classical
integration Lorenzo
Leandro Quantum Machines,
Milan, Italy We present Quantum
Machines’ Open Acceleration Stack, a modular framework integrating Quantum
Machines’ hybrid controller with a flexible choice of classical accelerators,
including GPUs, CPUs, FPGAs, and superchips, at microsecond latencies. The
stack is designed around a key principle: the quantum program itself
orchestrates classical compute resources, enabling intelligent and dynamic
distribution of workloads across the full quantum-classical system. We
demonstrate the Open Acceleration Stack through real use cases spanning
closed-loop calibration on real qubits and error-corrected quantum
algorithms, showing how the framework adapts to the demands of each workload
in practice. This modular, open approach provides a concrete and scalable
path to the integrated quantum-classical systems that fault-tolerant quantum
computing demands. |
|
Designing and Operating QPUs with AI
Supercomputing Dmitry Lyakh NVIDIA, USA AI and GPU supercomputing
are accelerating a broad range of quantum computing workflows, from QPU design,
optimization, calibration and tuning, to real-time decoding, quantum error
correction, and algorithm design and verification. We will highlight recent
progress in how AI models are transforming quantum computing workloads, and
tight integration of QPUs and GPUs building accelerated quantum
supercomputers. First, we will show how CUDA math libraries can accelerate
quantum EDA and QPU modeling at the analog level. Then, we will demonstrate
how the CUDA-Q platform and cuQuantum SDK can streamline QPU programming and
HPC quantum circuit simulation. We will also highlight how the NVIDIA Ising
family of open models are bringing the latest in AI to QPU calibration and
quantum error correction. Finally, we will outline how the NVQLink
interconnection architecture allows the construction of hybrid
quantum-classical systems. |
|
Scaling Photonic Quantum Computing with
Integrated GKP Qubit Sources Dylan Mahler Xanadu, Canada As the pursuit of
utility-scale quantum computing accelerates, photonics offers a compelling
pathway driven by mass manufacturability and seamless networkability. This
talk introduces Xanadu’s approach to continuous-variable photonic quantum
computing, which leverages squeezed light and modular architectures to build
a scalable, fault-tolerant platform. A central pillar of this
framework is the use of Gottesman-Kitaev-Preskill (GKP) states. We will present recent experimental
breakthroughs in synthesizing these crucial GKP states using an
ultra-low-loss integrated photonic chip on a custom silicon nitride platform.
By utilizing high-efficiency photon-number-resolving detectors and advanced
chip packaging, this work validates a critical milestone: moving complex
state preparation to scalable nanophotonics. This advancement paves the way
for the mass-manufacturable arrays of GKP sources required to power future
fault-tolerant quantum computers. |
|
A Heuristic for Matrix Product State
Simulation of Out-of-Equilibrium Dynamics of Transverse-Field Ising Models Salvatore
Mandrá Google Quantum
AI, Mountain View, CA, USA Simulating the
out-of-equilibrium dynamics of quantum systems is classically challenging due to rapid
entanglement growth. This presentation introduces a heuristic to accelerate
the convergence of Matrix Product State (MPS) simulations by rescaling
expectation values at low bond dimensions using a factor derived solely from
the MPS wave function fidelity. Applying this technique, we classically
simulate the dynamics of the Transverse-Field Ising Model on a 7 x 8 grid with
periodic boundary conditions using a single A100 GPU, achieving
excellent agreement with results from digital quantum processors. This approach
provides a practical path for simulating many-body systems at intermediate
energy densities where standard methods struggle. |
|
Quantum Accelerated Units: Toward Specialized
QPUs for Scientific Computing and AI Stefano Markidis KTH Royal Institute of Technology, Computer
Science Department / Computational Science and Technology Division,
Stockholm, Sweden Quantum computing may
become most useful not as a new accelerator layer for selected computational
kernels. In this talk, I will present a vision for specialized QPUs designed
around small, repeated micro-kernels for stencil computations and machine
learning. The main idea is to treat the QPU like a future quantum analog of a
GPU tensor core: not general-purpose, but highly effective for specific
operations such as sampling, convex sums, categorical choices, expectations,
and local stencil updates. I will discuss how this paradigm applies to
explicit PDE solvers through QPU micro-kernels for stencil computations, and
how the same concepts extend naturally to probabilistic and sampling-based
machine learning workloads. Reference: Markidis, S., Pennati, L.,
Netzer, G., Pasquale, M., & Peng, I. (2026, January). QPU Micro-Kernels
for Stencil Computation. In Proceedings of the Supercomputing Asia and
International Conference on High Performance Computing in Asia Pacific Region
(pp. 68-80). |
|
Taming Complexity: Deploying Scientific
Ecosystems via Portable HPC Containers Eric Mueller University of Heidelberg, Heidelberg, Germany Deploying software on HPC
systems is notoriously difficult. When dealing with massive software
ecosystems across multiple sites, the process can turn into a logistical
nightmare of incompatibilities and performance bottlenecks. The EBRAINS Software
Distribution (ESD) actively mitigates this by providing a highly curated,
versioned stack that can be automatically built and tested against different
compilers and base software installations early in the release cycle. In this talk, we detail
our efforts to build and deploy the ESD into Apptainer-based container images
to tame this complexity. Using these container images on multiple HPC sites,
we demonstrate that close-to-bare-metal performance can be maintained. We will also outline our
immediate next steps—expanding our benchmarks to recently installed systems
(ARM64/GPU), automating container builds within our CI pipelines, and
introducing user-friendly tools to easily validate runtime performance in
user-managed environments. Finally, we will look
ahead to the next challenges: adapting our deployment strategy to handle the
increasing influx of software packages (especially AI-based tools), our
pathway to integrate with EESSI, and the implementation of trusted deployment
workflows. These secure, verifiable container pipelines are becoming
critically important as the EBRAINS Research Infrastructure expands into
clinical applications handling sensitive patient data. |
|
Application Driven Evaluation of
Fault-Tolerant Quantum Computing Architectures Kevin Obenland Quantum Information and Integrated Nanosystems,
Massachusetts Institute of Technology, Lincoln Laboratory, Boston, MA, USA The last year has shown
many new proposals for scalable utility-scaled quantum computing
architectures. These new fault-tolerant architectures are driven by the
requirements of Quantum Error Correction (QEC) codes and the applications
that are intended for these platforms. Including details of QEC and
application implementation in the evaluation of architectural design
trade-offs will increase the fidelity of the assessments of these
architectures and enable co-design of the hardware with QEC and algorithm design.
At Lincoln Laboratory we have developed the pyLIQTR framework, an
application-driven set of tools supporting application implementation and
architectural analysis. In this talk, I will describe pyLIQTR and then show
an analysis across a wide range of application domains and emerging
fault-tolerant architectures. Applications including: lattice spin dynamics
and ground-state problems, chemical catalysis, NMR, and electronic structure
systems. For architectures, we consider simple lattice-surgery based surface
code architectures to architectures based on qLDPC codes. |
|
The strong scaling game: hardware, AI, and
simulation Tomas Oppelstrup Cerebras, USA Generative AI inference,
iterative solvers, and time-stepped simulation all reduce to propagating
state through long dependency chains. Even though the computational demands
of AI now drive hardware design, novel hardware also benefits applications in
scientific computing. In this presentation, we
will explore how the Cerebras Wafer Scale Engine (WSE) can enable unprecedented
speed in AI inference and many HPC applications. We will show examples in
non-linear solvers and particle simulation, where we have achieved orders of
magnitude performance improvements over current state-of-the-art. |
|
The National Science Data Fabric and its
evolution into the ATLAS omnicloud for AI development of medical devices Valerio Pascucci Center for Extreme Data Management, Analysis and
Visualization; Scientific Computing and Imaging Institute School of
Computing, University of Utah; Laboratory Fellow, Pacific Northwest National
Laboratory, USA
NSDF has successfully supported a diverse range of scientific
endeavors, including experiments at the Cornell Synchrotron Light Source,
Dark Matter studies utilizing SLAC data, weather and climate research at
NASA, support for autonomous laboratories at DOE national labs, and the
development of AI tools for advanced precision surgery. The talk will also discuss how the NSDF tools and experience allowed
the introduction of “ATLAS—Advanced Training and Learning at Scale” a
national omnicloud for complete AI development in medicine. ATLAS, which
involves over 30 institutions, features a comprehensive cyberinfrastructure
offering the following services: ●
Data Marketplace: A complete
representation of the US population's data. ●
Computing and Storage:
Virtually unlimited growth potential in computing and storage, leveraging
hyperscalers like Google and leadership computing environments such as TACC
and ALCF. ●
AI Tools and Services: A
full suite of tools for data searching, data deduplication and harmonization,
annotation services for AI training, and clinical and regulatory consulting. ●
Federated Learning Framework: The
ATLAS system enables the use of images from multiple hospitals and secure
enclaves without requiring any data to be moved outside the institution of
origin. The talk will close with a few considerations on how to reconnect
ATLAS to the original NSDF design and make most of its services
domain-agnostic so they can be applied to other applications with societal
impact, such as manufacturing and materials science. BIOGRAPHY Valerio Pascucci is the founding Director of
the Center for Extreme Data Management Analysis and Visualization (CEDMAV), a
Faculty of the Scientific Computing and Imaging Institute (Inaugural John R.
Parks Endowed Chair), a Professor of the School of Computing of the
University of Utah, and DOE Laboratory Fellow of PNNL. Valerio has received
the 2022 IEEE VGCT Visualization Technical Achievement Award and the
2022-2023 Distinguished Research Award (DRA) from the University of Utah and
has been inducted into the IEEE VGTC Visualization Academy in 2022. Valerio is also the President of ViSOAR LLC,
a University of Utah spin-off, and the founder of Data Intensive Science, a
501(c) nonprofit providing outreach and training to promote the use of
advanced technologies for science and engineering. Valerio's research
interests include Big Data management and analytics, progressive
multi-resolution techniques in scientific visualization, discrete topology,
and compression. Valerio is the coauthor of more than three hundred refereed
journal and conference papers and was an Associate Editor of the IEEE
Transactions on Visualization and Computer Graphics.
|
|
BrainScaleS – Networks of Analog Neuromorphic
Processors Johannes Schemmel European Institute for Neuromorphic Computing and
Kirchoff Institute for Physics, Heidelberg University, Heidelberg, Germany Brain-inspired event-based
neuromorphic computing is a promising technology for energy-efficient
bio-inspired AI. It also enables continuous learning based on local learning
algorithms. For maximum energy
efficiency, a brain-like in-memory realization is desirable. The Heidelberg
BrainScaleS platform is an example of a neuromorphic architecture that
combines true in-memory computing with hardware support for continuous local
learning. In addition, its analog implementation of the core neuronal
operations enables complex neuron models without an energy or area penalty. For real-world
applications as well as neuroscience, certain minimum network sizes are
required. To realize the necessary upscaling, BrainScaleS pioneered
wafer-scale integration. For future generations of
BrainScaleS, this will not be feasible due to the high mask costs of modern
semiconductor processes. This talk presents alternative solutions and how
they can be utilized by the current BrainScaleS platform in EBRAINS. EBRAINS is a European
research infrastructure for neuroscience. Its BrainScaleS neuromorphic
service democratizes access to analog neuromorphic computing. |
|
A Reference Architecture for a
Quantum-Centric Supercomputer Seetharami Seelam IBM Research, USA Quantum computers have
demonstrated utility in simulating quantum systems beyond brute-force
classical approaches. As the community builds on these demonstrations to
explore using quantum computing for applied research, algorithms and
workflows have emerged that require leveraging both quantum computers and
classical high-performance computing (HPC) systems to scale applications,
especially in chemistry and materials, beyond what either system can simulate
alone. Today, these disparate
systems operate in isolation, forcing users to manually orchestrate
workloads, coordinate job scheduling, and transfer data between systems -- a
cumbersome process that hinders productivity and severely limits rapid
algorithmic exploration. These challenges motivate the need for flexible and
high-performance Quantum-Centric Supercomputing (QCSC) systems that integrate
Quantum Processing Units (QPUs), Graphics Processing Units (GPUs), and
Central Processing Units (CPUs) to accelerate the discovery of hybrid
quantum-classical algorithms across applications. These systems will be
co-designed across quantum and classical HPC infrastructure, middleware, and
application layers to accelerate the adoption of quantum computing for
solving critical computational problems. We envision QCSC evolution through
three distinct phases: (1) quantum systems as specialized compute offload
engines within existing HPC complexes; (2) heterogeneous quantum and
classical HPC systems coupled through advanced middleware, enabling seamless
execution of hybrid quantum-classical algorithms; and (3) fully co-designed
heterogeneous quantum-HPC systems for hybrid computational workflows. In this
talk, we will present a reference architecture and roadmap for these QCSC
systems. |
|
Photonic Quantum Computing for HPC and AI:
From Today’s Hybrid Architectures to Fault-Tolerant Systems Jean Senellart Quandela, Chief Technology and Product Officer,
France The convergence of HPC,
AI, and quantum computing opens the way to new hybrid architectures where
quantum processors act as complementary accelerators within large-scale
classical workflows rather than as isolated cloud resources. This talk will
discuss what such hybridization can mean in practice, from quantum-assisted
AI and optimization to future fault-tolerant workflows supported by HPC for
simulation, compilation, control, and orchestration. Photonic quantum
computing will be highlighted as a promising modality for this evolution,
thanks to its native connectivity, compatibility with data-center deployment,
fast sampling, persistent optical configurations, and potential for tight
integration with GPU- and HPC-based pipelines. |
|
Scaling Toward Utility-Scale Fault-Tolerant
Quantum Computing with Photonics Pete
Shadbolt Co-founder PsiQuantum
Corp., Palo Alto, California, USA This talk will highlight
recent progress toward utility-scale, fault-tolerant quantum computing using
photonics. We will cover advances across the full technology stack, including
materials innovations for high-performance photonics, improvements in photonic
component performance with a focus on reducing optical loss, prototype
systems for entangled photonic qubits, qubit networking, and high-power
cryogenic cooling solutions designed for future datacenter-scale quantum
computers. We will present new
prototype systems that progressively address the core engineering challenges
associated with scaling photonic quantum computing, alongside an overview of
fusion-based quantum computing architectures and the principles underpinning
large-scale fault tolerance. The talk will also outline key near-term system
milestones and provide a roadmap toward commercially useful, utility-scale
quantum computers. |
|
Active Memory Architecture: A Step Beyond
Moore’s Law Thomas Sterling, Ph.D. Texas Advanced Computing Center, Austin, TX, USA The last four years has
shown an extreme reduction in advancement of the fastest world-class systems
at least as demonstrated by the Top-500 List. Only one system per year has
been added in the range of Exascale and all such machines have clustered in
the range of one to two Exaflops Rmax. The growth rate is by far the lowest
ever witnessed in more than thirty years that data has been collected.
Feature size of logic integration is at single digit nano-scale and unlikely
to progress further. Clock rates have been stalled at between 1 and 3
Gigahertz (approximately) for years due to thermal demands. Despite these
barriers to future growth upon which the exponential improvement over the
first six decades has been based, alternative dimensions of the design space
can be extended, ultimately by three orders of magnitude to the performance
regime of Zettaflops. These potentially dramatic improvements are in the
domain of parallel computer architecture, specifically in the form of non von
Neumann memory-centric structures. The Active Memory Architecture (AMA) is in
the 2nd year of development at the Texas Advanced Computing Center. AMA
comprises a plethora of very small, unified execution units (memory, logic,
and communication) exposing fine-grain parallelism with asynchronous
communication. It is dynamic in resource allocation and incorporates
low-level graph constructs for AI. This presentation describes principles and
design details of the AMA. |
|
Scientific Challenges to Apply Quantum
Computing in Drug Design Michael Streif Boehringer, Germany Quantum computers hold the
promise of outperforming classical systems and are widely regarded as the
future of quantum chemistry and drug design. However, realizing this
potential in real-world pharmaceutical applications requires addressing a far
broader set of challenges than those typically explored - well beyond
ground-state energy calculations for single molecules. In this talk, we
examine the scientific challenges that must be overcome to make quantum
computing a practical tool for drug design. We will share our latest work on
fault-tolerant algorithms and discuss strategies for tackling these
challenges. |
|
Quantum Pangenomics: From SequenceAssembly to
the First Whole-Genome Encoding Sergii Strelchuk Oxford University, United Kingdom Genomics has transformed
our understanding of biology and disease, and pangenomics—the study of
variation across many genomes simultaneously—is the natural next step toward
decoding how genetic diversity shapes human health. Building and analysing
pangenomes, however, gives rise to computational problems that are
notoriously hard for classical algorithms, making them a natural target for
quantum computing. Working within the
Wellcome Leap Quantum for Bio programme, our consortium—led from Oxford in
collaboration with the Wellcome Sanger Institute and Cambridge—has spent the
past phase running real pangenomic workloads on quantum computers. I will
discuss several applications: first, we recast pangenome-guided sequence
assembly as a quantum optimisation problem; second, we developed novel
data-encoding schemes that led to new quantum-inspired algorithms for global
sequence alignment, sending a clear signal to the life-sciences community
about the growing reach of current and emerging quantum hardware. |
|
Distributed Quantum Accelerators: Why
Parallel, Not Monolithic, Is the Preferred Path for Commercial Quantum
Computing Andrea Tabacchini Quantum Brilliance, Australia and Germany Quantum computing will
reach widespread commercial relevance only when deployed as a classically
orchestrated, horizontally scaled accelerator within heterogeneous HPC
environments. This talk argues that distributed, parallel quantum computing -
not ever-larger monolithic processors - best matches how real workloads are
structured and monetized, a conclusion independent of qubit technology and
unchanged by fault tolerant machines. We support it with the first public
results from a collaboration with Oak Ridge National Laboratory: Fragment
Molecular Orbital (FMO) calculations across a series of helium molecules, run
in parallel on a three-QPU cluster operating onsite at ORNL. These
experiments directly compare parallel against serialized execution and show a
clear advantage for the parallel approach. |
|
William Tang Princeton University Dept. of Astrophysical
Sciences, Princeton Plasma Physics Laboratory and Center for Statistics and
Machine Learning (CSML) and Princeton Institute for Computational Science
& Engineering (PICSciE), Princeton University, USA It will focus on new
elements emphasizing the importance of real-time demonstrations of clean
energy fusion using AI/ML and Exascale Computing. Advances with Foundation
Models featuring experimental and simulation data will also be discussed. |
