HPC 2025
High Performance Computing
State of the Art, Emerging Disruptive Innovations and
Future Scenarios
An
International Advanced Workshop
June
23 – 27, 2025, Cetraro, Italy
Final Programme
Sponsors
|
CEREBRAS |
|
|
|
CMCC Euro-Mediterranean
Center on Climate Change |
|
|
|
CSC Finnish Supercomputing Center |
|
|
|
CSCS |
|
|
|
DWAVE Systems |
|
|
|
EOFS |
|
|
|
Juelich Supercomputing Center, Germany |
|
|
|
LENOVO |
|
|
|
NVIDIA |
|
|
|
PARTEC |
|
|
|
PSIQUANTUM |
|
|
|
QUANTINUUM |
|
|
|
QUANTUM BRILLIANCE |
|
|
|
QUANTUM MACHINES |
|
|
|
RIGETTI Computing |
|
|
|
SAMBANOVA SYSTEMS |
|
|
|
ThinkParQ |
|
|
|
University of Calabria Department of Computer Engineering,
Electronics, and Systems |
|
|
|
UNIVERSITY OF SALENTO |
|
|
|
VastData |
|
|
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. |
|
FRANK
BAETKE EOFS European
Open File System Organization GERMANY MANDY
BIRCH CEO
and founder of TreQ UNITED
KINGDOM GEORGE BOSILCA NVIDIA Tel Aviv ISRAEL VAST DATA GERMANY TIM CLARKE Sambanova Systems Inc. Palo Alto, CA USA MADISON COTTERET University of Groningen Groningen THE NETHERLANDS DANIELE DRAGONI Leonardo S.p.A. High
Performance Computing Lab. Genova ITALY WOLFGANG
GENTZSCH Co-founder
& President of Simr SimOps
Simulation Operations Org. Regensburg GERMANY and Sunnyvale,
CA USA XAVIER GEOFFRET Quandela FRANCE 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 BETTINA HEIM NVIDIA USA ThinkParQ GmbH GERMANY NOBUYASU
ITO RIKEN
Center for Computational Science, Kobe JAPAN MICHAEL
JAMES Cerebras
Systems Sunnyvale,
California USA HIROAKI
KOBAYASHI Architecture
Laboratory Department
of Computer and Mathematical Sciences Graduate
School of information Sciences Tohoku
University JAPAN DHIREESHA
KUDITHIPUDI University
of Texas at San Antonio San Antonio, TX USA LORENZO LEANDRO Quantum
Machines Tel
Aviv ISRAEL PEKKA
MANNINEN CSC
Director of Science and Technology Finnish
IT Center for Science Espoo FINLAND STEFANO
MARKIDIS KTH
Royal Institute of Technology Computer
Science Department / Computational Science and Technology Division Stockholm SWEDEN SATOSHI
MATSUOKA Director
RIKEN Center for Computational Science, Kobe and Tokyo
Institute of Technology, Tokyo JAPAN CHRISTIAN
MAYR Technical
University Dresden Dresden GERMANY LUCAS MENGER University of Frankfurt GERMANY University of Heidelberg Heidelberg GERMANY JOSH
MUTUS Rigetti
Computing Director
Quantum Devices USA/CANADA KEVIN
OBENLAND Quantum
Information and Integrated Nanosystems Lincoln
Laboratory Massachusetts
Institute of Technology MIT Boston,
MA USA IRWAN
OWEN D-Wave
Systems Inc CANADA - USA Nash PALANISWAMY Quantinuum UK - USA TROY
PATTERSON ThinkParQ GERMANY DAVID
RIVAS Rigetti Computing Berkeley, CA USA RAFFAELE SANTAGATI Quantum
Group Boheringen
Ingelheim GERMANY JOHANNES
SCHEMMEL European
Institute for Neuromorphic Computing and Kirchoff
Institute for Physics Heidelberg
University Heidelberg GERMANY THOMAS
SCHULTHESS CSCS
Swiss Center of Supercomputing Lugano SWITZERLAND PETE
SHADBOLT Co-founder PsiQuantum Corp. Palo Alto, California USA SARAH SHELDON IBM Yorktown Heights, NY USA THOMAS
STERLING Senior
Research Scientist University
of Texas at Austin Texas
Advanced Computing Center Austin, TX USA ANNA STOCKKLAUSER Quantum Motion UK SERGII
STRELCHUCK Oxford
University Oxford
UK 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 ZACHARY
VERNON XANADU CANADA ALEKSANDER WENNERSTEEN Pasqal FRANCE |
Workshop
Agenda
Monday,
June 23rd
|
Session |
Time |
Speaker/Activity |
|
9:45 – 10:00 |
Welcome
Address |
|
|
State of the art
and future scenarios |
||
|
|
10:00– 10:30 |
T. STERLING Towards
an Active Memory Architecture for Graph Processing beyond Moore's Law |
|
|
10:30 – 11:00 |
V. GETOV Application-Driven
Development and Evolution of the Computing Continuum |
|
11:00 – 11:30 |
COFFEE BREAK |
|
|
|
11:30 – 12:00 |
E. MUELLER Sustainable
Management of Complex Software Ecosystems for Novel Accelerators |
|
12:00 – 12:30 |
N. PALANISWAMY |
|
|
12:30 – 12:45 |
CONCLUDING
REMARKS |
|
|
|
Emerging Computer
Systems and Solutions |
|
|
|
17:00 – 17:30 |
t.b.a. |
|
|
17:30 – 18:00 |
H. KOBAYASHI Graph-based
Data Analysis of Three-dimensional Electron Diffraction Data |
|
|
18:00 – 18:30 |
W. GENTZSCH SimOps
Introduces HPC Software Stack with HPC Software Certification and
Training |
|
18:30 – 19:00 |
COFFEE BREAK |
|
|
|
19:00 – 19:30 |
F. BAETKE Needs of
the HPC Community vs. Computer Science Curricula – a Widening Gap |
|
19:30 – 19:45 |
CONCLUDING REMARKS |
Tuesday,
June 24th
|
Session |
Time |
Speaker/Activity |
|
Advanced AI Processing:
Challenges and Perspectives |
||
|
|
9:30 – 9:55 |
W. TANG AI-Powered
Machine Learning for Accelerating Scientific Grand Challenges |
|
|
9:55 - 10:20 |
M. JAMES Scaling
Physics Simulations on AI Hardware: Insights from Molecular Dynamics and
Planetary Modeling |
|
|
10:20 – 10:45 |
G. BOSILCA Unlocking the Full
Potential of AI With Next-Gen Networking |
|
|
10:45 – 11:15 |
COFFEE BREAK |
|
|
11:15 – 11:40 |
P. MANNINEN |
|
|
11:40 – 12:05 |
S. BREUNER The
VAST Data Platform: Modern HPC and AI Storage as it should be |
|
|
12:05 -12:30 |
T. CLARKE Agentic
AI at Scale: How SambaNova’s Architecture Powers Next-Gen LLM Workflows in
National Labs |
|
|
12:30 – 12:55 |
T. PATTERSON and F. HEROLD Challenges
for Software-defined storage in the modern HPC and AI World |
|
|
12:55 – 13:10 |
CONCLUDING
REMARKS |
|
|
Neuromorphic
Computing |
|
|
|
17:00 – 17:30 |
D. KUDITHIPUDI THOR: The Neuromorphic
Commons |
|
|
17:30 – 18:00 |
C. MAYR |
|
|
18:00 – 18:30 |
COFFEE BREAK |
|
|
18:30 – 19:00 |
J. SCHEMMEL |
|
|
19:00 – 19:30 |
M. COTTERET Vector-Symbolic
Architectures for Scalable Neuromorphic Systems Design |
|
|
19:30 – 19:45 |
CONCLUDING
REMARKS |
Wednesday,
June 25th
|
Session |
Time |
Speaker/Activity |
|
|
The QUANTUM COMPUTING
Promises 1 |
|
|
|
10:00 – 10:30 |
S. SHELDON t.b.a. |
|
|
10:30 – 11:00 |
J. MUTUS Superconducting
qubits at the utility scale: the potential and limitations of modularity |
|
|
11:00 – 11:30 |
COFFEE BREAK |
|
|
11:30 – 12:00 |
P. SHADBOLT Progress
towards large-scale fault-tolerant quantum computing with photons |
|
|
12:00 – 12:30 |
Z. VERNON Photonic
fault-tolerant quantum computing: Scaling, networking, and modularity |
|
|
12:30 – 12:45 |
CONCLUDING REMARKS |
|
Session VI |
|
The QUANTUM
COMPUTING Promises 2 |
|
|
17:30 – 18:00 |
K. OBENLAND Logical
Resource Analysis of Utility-Scale Quantum Applications using pyLIQTR |
|
|
18:00 – 18:30 |
V. GHEORGHIU Embedding
classical logic into quantum computation, or, mid-circuit operations on
steroids |
|
|
18:30 – 19:00 |
COFFEE
BREAK |
|
|
19:00 – 19:30 |
I. OWEN |
|
|
19:30 – 19:45 |
CONCLUDING REMARKS |
Thursday,
June 26th
|
Session |
Time |
Speaker/Activity |
|
|
Quantum Computing Challenging
Applications |
|
||
|
|
9:30 – 10:00 |
R. SANTAGATI Accelerating
Quantum Chemistry Simulations on Quantum Computers |
|
|
|
10:00 – 10:30 |
D. DRAGONI |
|
|
|
10:30 – 11:00 |
S. STRELCHUCK |
|
|
|
11:00 – 11:30 |
COFFEE BREAK |
|
|
|
The
Quantum Computing Prospects and Deployments |
|
|
|
|
11:30 – 12:00 |
B. HEIM Defining the quantum
accelerated supercomputer |
|
|
|
12:00 – 12:30 |
A. WENNERSTEEN |
|
|
|
12:30 – 13:00 |
A. TABACCHINI |
|
|
|
13:00 – 13:15 |
CONCLUDING REMARKS |
|
|
|
The Quantum
Computing Prospects and Deployments 2 |
|
|
|
|
17:00 – 17:25 |
L. LEANDRO The
Research Driving Control System Innovation Towards Utility Scale Quantum
Computing |
|
|
|
17:25 – 17:50 |
M. birch Preparing
HPC for Quantum Acceleration: Why Open Architecture Matters |
|
|
|
17:50 –18:15 |
S. MARKIDIS Rethinking
Quantum Parallelism: From Superposition to Speed-Up |
|
|
18:15 – 18:45 |
COFFEE BREAK |
|
|
|
18:45 – 19:45 |
Chairperson:
t.b.a. Panelists: Vladimir
Getov, University of Westminster, UK William
Tang, Princeton University, USA Pete
Shadbolt, PsiQuantum Corp., USA (t.b.c.) David
Rivas, Rigetti Computing, USA 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 27th
|
Session |
Time |
Speaker/Activity |
|
Key Projects, Novel
Developments and Challenging Applications |
||
|
|
10:00 – 10:30 |
S. MATSUOKA t.b.a. |
|
|
10:30 – 11:00 |
N. ITO |
|
11:00 – 11:30 |
COFFEE BREAK |
|
|
|
11:30 – 12:00 |
L. MENGER |
|
|
12:00 – 12:30 |
X. GEOFFRET Bridging
HPC and Quantum Computing: Quandela’s full-stack approach |
|
12:30 – 12:45 |
CONCLUDING REMARKS |
Chairpersons
WOLFGANG
GENTZSCH
Co-founder
& President of Simr
SimOps
Simulation Operations Org., Regensburg
GERMANY
and
Sunnyvale,
CA
USA
t.b.a.
VLADIMIR
GETOV
Distributed and Intelligent Systems Research
Group
School of Computer Science and Engineering
University of Westminster, London
UNITED KINGDOM
WOLFGANG
GENTZSCH
Co-founder
& President of Simr
SimOps
Simulation Operations Org., Regensburg
GERMANY
and
Sunnyvale,
CA
USA
RAFFAELE SANTAGATI
Quantum Group
Boheringen Ingelheim
GERMANY
JOSH
MUTUS
Rigetti
Computing
Berkeley,
CA
USA
t.b.a.
VLAD GHEORGHIU
University of Waterloo
Institute for Quantum Computing
Waterloo, Ontario
CANADA
HIROAKI KOBAYASHI
Architecture Laboratory
Department of Computer and Mathematical
Sciences
Graduate School of information Sciences
Tohoku University
JAPAN
VLADIMIR
GETOV
Distributed and Intelligent Systems Research
Group
School of Computer Science and Engineering
University of Westminster
London
UNITED KINGDOM
Panel
Abstracts
|
Needs of the HPC Community vs. Computer
Science Curricula – a Widening Gap Frank Baetke EOFS, European Open File System Organization,
Germany The talk will summarize
observations made at an ISC panel and three EOFS workshops in 2022, 2024 and
2025 with representatives of the academic community and major European HPC
Centers. Core components of any HPC
installation such as operating systems, storage, I/O and file systems are no
longer considered interesting and important topics in computer science or
information technology curricula. Related lectures at several universities
have been abandoned in favor of AI, web services and other areas that are
considered more relevant and/or interesting. Communication between
users of large HPC centers and IT staff responsible for system scheduling and
storage/file-system management is becoming an increasing problem as many
application developers and/or users are unaware or uninterested in the
operational aspects of large application programs and the associated
challenges such as multiple storage hierarchies, demand scheduling, etc. This disconnect often
causes issues related to load balancing, resource efficiency and system
responsiveness and leads to frustration on both sides. |
|
Preparing HPC for Quantum Acceleration: Why
Open Architecture Matters Mandy Birch CEO & Founder, TreQ, United Kingdom As quantum computing
advances, its role in high-performance computing (HPC) is becoming clear—not
as a standalone solution, but as a specialized accelerator integrated with
classical systems. This talk explores how
open-architecture design principles can future-proof HPC environments for
quantum integration. We’ll examine how modularity, hardware diversity, and
upgradeability enable adaptation to rapidly evolving quantum technologies
across qubit modalities, control systems, and software layers. Through practical examples
and architectural strategies, we’ll discuss: - Aligning infrastructure
decisions with long-term interoperability - Supporting heterogeneous
quantum-classical workflows - Enabling experimentation
with configurable system components As a concrete example,
we’ll share insights from a UK-based testbed project that integrates multiple
quantum processors, control stacks, and software environments—yielding eight
distinct configurations in a single system. This work illustrates the practical
value of modular design and the development of open specifications that span
the quantum stack. This session is intended for system architects, research leaders, and
HPC infrastructure strategists aiming to stay adaptive while managing
long-term risk. |
|
The VAST Data Platform: HPC and AI Storage as
it should be Sven Breuner Field CTO International @ VAST Data Storage in HPC lacks innovation
and has gotten way too complicated and inconvenient over the last years - for
the system administrators that have to manage a whole zoo of different
systems and for the researchers that should be able focus on their research
instead of data management, access patterns and protocols. Time for a change
before it gets even worse over the next years! |
|
Agentic AI at Scale: How SambaNova’s
Architecture Powers Next-Gen LLM Workflows in National Labs Tim Clarke Account Executive for Public Sector SambaNova
Systems, USA As Large Language Models
(LLMs) evolve from standalone tools to interconnected Agentic AI systems,
traditional GPU-based infrastructure struggles with the demands of
multi-model orchestration, memory constraints, and rapid task-switching. This
presentation explores how SambaNova’s Reconfigurable Dataflow Unit
(RDU)—designed for trillion-parameter workloads—enables National Labs to
deploy state-of-the-art AI with unmatched efficiency. We’ll highlight: World-record inference
performance and unified training/inference on a single system, eliminating
GPU bottlenecks. TBs of memory per node,
allowing hundreds of models (or trillion-parameter LLMs) to reside in-memory
simultaneously-critical for Agentic AI’s complex workflows. Microsecond model
switching (100x faster than GPUs), enabling dynamic multi-model pipelines for
scientific research, data analysis, and operational automation. Real-world use cases from
National Labs leveraging these capabilities to accelerate tasks like
hypothesis generation, knowledge synthesis, and scalable simulation. |
|
Vector-Symbolic Architectures for Scalable
Neuromorphic Systems Design Madison Cotteret University of Groningen, Netherlands As neuromorphic hardware
grows in scale and complexity, their development is limited by the ease of
configuring large neural systems to realise high-level behaviours. In
conventional digital hardware, such complexity is managed through a hierarchy
of abstraction layers, which permit development at one layer (e.g. machine
code) without requiring expert knowledge of the lower levels (e.g. gate
implementations). Neuromorphic computing sorely lacks a similar
robust-yet-flexible hierarchy of abstractions, making it prohibitively
difficult to build up to complex high-level function. I present vector-symbolic
architectures (VSAs) as a candidate abstraction layer for neuromorphic
computing. VSAs form distributed representations of symbolic data structures
using high-dimensional random vectors, and are inherently fault tolerant.
When combined with attractor network theory, they give a reliable method to
embed symbolic computational structures (e.g. state machines, generalised
line attractors) into neuromorphic hardware, independent of the underlying
neural representations. This is a significant step towards a hierarchical
theory of neural processing suitable for programming large-scale neuromorphic
systems that are capable of performing extended cognitive tasks. |
|
Quantum Path at Leonardo: Enabling Innovation Daniele
Dragoni Leonardo S.p.A.,
High Performance Computing Lab, Genova, Italy Quantum Computing
(QC) represents a transformative paradigm with the
potential to address computational challenges that remain out of reach for
classical systems. Although a clear quantum advantage in real-world
applications is yet to be demonstrated, the momentum is growing, and several
industries are investing early to explore its disruptive potential and secure
future competitiveness. In my presentation, I will
outline Leonardo's strategic approach to thoroughly evaluate the capabilities
and limitations of QC within the aerospace, security, and defense domains. I
will delve into our stance on QC both from an industrial end-user perspective, showcasing ongoing initiatives and practical applications pursued
through integrated HPC and QC methodologies in alignment with national
strategic objectives, and from the viewpoint of a technology provider developing
capabilities to support clients interested in targeted quantum
experimentation. |
|
SimOps Introduces HPC Software Stack with HPC
Software Certification and Training Wolfgang Gentzsch Co-founder & President of Simr, SimOps
Simulation Operations Org., Regensburg, Germany and Sunnyvale, CA, USA SimOps (Simulation
Operations Automation) recently introduced the SimOps Software Stack, a suite
of HPC tools designed to simplify, optimize, and automate the use, operation,
and management of HPC infrastructures—both in the cloud and on-premises—for
HPC system administrators and simulation engineers. The SimOps Software Stack
is a curated collection of HPC tools, services, and platforms that enable the
implementation of SimOps principles to improve the engineering simulation and
HPC operations lifecycle. SimOps defines tools that support these principles
as ‘SimOps-compliant.’ Much like the DevOps
Software Stack - used to optimize software development workflows - the SimOps
Software Stack includes components across multiple layers such as
provisioning and HPC middleware, platform and access layers, HPC
infrastructure and workload management, workflow automation, data management,
analytics, visualization and observability, CI/CD and DevOps for SimOps, as
well as security and compliance tools. Together, these components accelerate
and automate engineering simulations on optimized HPC infrastructures.<'text-align:center'>Quantum Path at Leonardo: Enabling Innovation Daniele
Dragoni Leonardo S.p.A.,
High Performance Computing Lab, Genova, Italy Quantum Computing
(QC) represents a transformative paradigm with the
potential to address computational challenges that remain out of reach for
classical systems. Although a clear quantum advantage in real-world
applications is yet to be demonstrated, the momentum is growing, and several
industries are investing early to explore its disruptive potential and secure
future competitiveness. In my presentation, I will
outline Leonardo's strategic approach to thoroughly evaluate the capabilities
and limitations of QC within the aerospace, security, and defense domains. I
will delve into our stance on QC both from an industrial end-user perspective, showcasing ongoing initiatives and practical applications pursued
through integrated HPC and QC methodologies in alignment with national
strategic objectives, and from the viewpoint of a technology provider developing
capabilities to support clients interested in targeted quantum
experimentation. |
|
SimOps Introduces HPC Software Stack with HPC
Software Certification and Training Wolfgang Gentzsch Co-founder & President of Simr, SimOps
Simulation Operations Org., Regensburg, Germany and Sunnyvale, CA, USA SimOps (Simulation
Operations Automation) recently introduced the SimOps Software Stack, a suite
of HPC tools designed to simplify, optimize, and automate the use, operation,
and management of HPC infrastructures—both in the cloud and on-premises—for
HPC system administrators and simulation engineers. The SimOps Software Stack
is a curated collection of HPC tools, services, and platforms that enable the
implementation of SimOps principles to improve the engineering simulation and
HPC operations lifecycle. SimOps defines tools that support these principles
as ‘SimOps-compliant.’ Much like the DevOps
Software Stack - used to optimize software development workflows - the SimOps
Software Stack includes components across multiple layers such as
provisioning and HPC middleware, platform and access layers, HPC
infrastructure and workload management, workflow automation, data management,
analytics, visualization and observability, CI/CD and DevOps for SimOps, as
well as security and compliance tools. Together, these components accelerate
and automate engineering simulations on optimized HPC infrastructures. What is SimOps? SimOps
explores the potential of streamlining and automating on-premises and
cloud-based simulation infrastructures, which are vital for enhancing and
accelerating scientific inquiries and engineering designs. SimOps is a new
community initiative and non-profit organization (you could say in short: the
“DevOps of HPC”) bringing simulation experts and IT operations experts closer
together by developing best practices, guidelines, and educational training
courses, for setting up, operating, maintaining, supporting and efficiently
using HPC/AI infrastructures for complex scientific and engineering
applications. SimOps aims at automating and accelerating simulation
processes, significantly increasing scientific and engineering productivity
and organizational contributions (to innovation and competitiveness). SimOps
will examine and collaborate on ways to reduce the complexities traditionally
associated with high-performance computing environments. |
|
Bridging HPC and Quantum Computing:
Quandela’s full-stack approach Xavier Geoffret Quandela, France As quantum computing (QC)
advances at a fast pace, its integration with high-performance computing
(HPC) is becoming a critical topic for both researchers and industry leaders.
There are in fact significant challenges for making hybrid HPC-QC workflows practical
and scalable, and empower the broad community of users to QC. This talk will explore the
key challenges in bridging HPC and QC, including hardware and software
integration, workload partitioning, and the need for new programming
paradigms. We will discuss how quantum processors can complement HPC for
applications in optimization, machine learning, and simulation, as well as
the technical and economic considerations for HPC centers looking to
incorporate QC into their infrastructure. |
|
Application-Driven Development and Evolution
of the Computing Continuum Vladimir Getov Distributed and Intelligent Systems Research Group
University of Westminster, London, U.K. Over the last decade, a new
concept – the computing continuum – has been gaining attention amongst the
professional community. It encompasses the growing variety of interconnected
computational, network, and storage resources across multiple layers of a
high-speed distributed infrastructure. The most important components of the
computing continuum include specialized cyber-physical systems, personal
augmentation equipment, cloud and high-performance computing data centers, as
well as Internet-of-Things edge devices. At present, we still have
centralized and limited visibility over the system performance, quality of
service, and quality of data. Meanwhile, the rapidly evolving computing
fabric is already composed of all traditional and emerging computational
resources. A seamless integration of the computing continuum infrastructure
leverages the best of each component. The representative application domains,
such as artificial intelligence, physical system simulation, cryptography,
machine learning, and multimedia, can be characterized by their service level
objectives and requirements, which specify the development and evolution of
the computing continuum components. Application domains with similar
objectives and requirements can be merged to reduce the overall number of
application domains under consideration. Some other domains, although
distinctive and representative for specific applications, are negligible due
to lower customer interest – e.g. they may not be recognized as “market
drivers” – and can be left out of consideration. We constantly need to
understand better and improve the relationship between service-level
objectives/requirements and the underlying architectures. Since both the
application domains and the computing continuum components are rapidly
developing and evolving entities, the most appropriate development approach
is the application-architecture co-design which is considered and described
in this presentation. |
|
Embedding classical logic into quantum
computation, or, mid-circuit operations on steroids Vlad Gheorghiu Institute for Quantum Computing, University of
Waterloo and SoftwareQ Inc, Waterloo, Ontario, Canada We present our innovative
solution on integrating arbitrary classical logic into quantum circuits at
compile time. This feature reduces the complexity of quantum circuit design,
particularly for fundamental tasks such as syndrome extraction, mid-circuit measurements,
and variational quantum algorithms. It also lays a foundation for the
seamless hybridization of classical and quantum computing. Live
demonstrations will be provided using our open-source quantum computing
framework, Quantum++, https://github.com/softwareqinc/qpp. |
|
Challenges
for Software-defined storage in the modern HPC and AI World Frank Herold and Troy Patterson ThinkParQ, Germany This session will cover
some of the challenges faced by the end users however by the providers who create the solution infrastructure such
as ThinkParQ with its Parallel file system BeeGFS. This session will highlight
traditional HPC and AI Environments that have been utilizing BeeGFS for their
workflows, and will also showcase the latest data management features of the
newly launched BeeGFS 8. |
|
JHPC-quantum project: HPC and QPU hybrid
challenge of RIKEN Nobuyasu Ito RIKEN Center for Computational Science, Kobe,
Japan The development of quantum
information technology in recent years has been remarkable, and its
applications are expanding the scope of computation beyond current
computational and environmental limits.So far, QPUs have been operated as
standalone processors, which has hindered the smooth and efficient
development of QPU applications. Now is the time to introduce quantum
computers into the IT environment, and last year, RIKEN, with financial
support from NEDO JAPAN, began an effort to fuse QPUs with HPC. This project
covers everything from hardware preparation to the development of industrial
applications. In this presentation, we will provide an overview and current
status of the JHPC-quantum project. |
|
Scaling Physics Simulations on AI Hardware:
Insights from Molecular Dynamics and Planetary Modeling Michael James Cerebras Systems, Sunnyvale, California, USA The evolution of AI
hardware platforms is unlocking new opportunities for scaling physics
simulations, offering capabilities previously absent in CPU and GPU-based
platforms. Notably, these platforms provide the bandwidth necessary for
high-utilization PDE solutions and network capabilities that support strong
scaling. With the advancements in
modern AI hardware, we can now extend the reach of traditional
high-performance computing (HPC) methods. In this talk, we will explore how
AI-driven architectures can revolutionize physics simulations, enabling us to
approach problems that have been beyond the reach of exascale platforms. We
will delve into the Cerebras wafer-scale platform, showing its capabilities
with examples in molecular dynamics and planetary modeling. Bio: Michael is the
Founder and Chief Architect of Advanced Technologies at Cerebras, the company
renowned for creating the world’s largest and most powerful computer
processor. He leads the initiative to redefine the algorithmic foundations
for the next generation of AI technologies. Before Cerebras, Michael was a
Fellow at AMD, where he developed adaptive and self-healing circuits based on
cellular automata, enhancing the reliability of distributed fault-tolerant
machines. Throughout his career, Michael has focused on the intersection of
natural phenomena, mathematics, and engineered systems. His degree from UC
Berkeley is in Molecular Neurobiology, Computer Science, and Mathematics. |
|
Graph-based Data Analysis of
Three-dimensional Electron Diffraction Data Hiroaki Kobayashi Tohoku University, Japan In this talk, I will present
a graph-based data analysis of three-dimensional electron diffraction data.
Three-dimensional electron diffraction is an emerging technique that allows
researchers to obtain detailed molecular structures from their small crystals
using transmission electron microscopy. However, due to the relatively poor
signal-to-noise ratio of many diffraction images, a large amount of data is
generated, including data that misrepresents the molecule. Therefore, the
development of automatic molecular structure identification from a large
amount of data containing both correct and incorrect structure data is
crucial. We will mention an automatic
graph generation that represents the molecular structure, and an
identification method using multiple features that extracts features of
correct molecular structures. |
|
The Research
Driving Control System Innovation Towards Utility Scale Quantum Computing Lorenzo Leandro Quantum Solutions Physicist at Quantum Machines Scaling quantum processors
introduces new requirements on control, such as ensuring high-fidelity qubit
operations by optimizing the analog front-end, automating calibration
workflows, and integrating hybrid control for quantum error correction. To
make significant progress, we need clear understanding of both present
technology and the demands of future large-scale quantum computers. Deep
research is needed, both in academia and industry, to unveil the important
bottlenecks and their possible solution. In this talk, we will explore key
technical challenges and focus on how the research done in QM facilitates
informed definition of the control system requirements paving the way towards
useful quantum computing. |
|
Towards AI supercomputing with LUMI-AI Pekka Manninen CSC Finland As a part of the AI
Innovation Package of the European Union, the EuroHPC Joint Undertaking is
currently planning a set of AI supercomputers in Europe, deployed within the
announced 13 AI Factories. In this talk, we will discuss the largest of these
upcoming systems, LUMI-AI, to be located in Kajaani, Finland. LUMI-AI will be
one of the most powerful and advanced quantum-accelerated supercomputing
systems in the world at the time of its completion. In this talk, I will
present the 6-country consortium behind it, some history, the technical
vision of the LUMI-AI infrastructure, its current status, as well as plans
and ambitions. |
|
Rethinking Quantum Parallelism: From
Superposition to Speed-Up Stefano Markidis KTH Royal Institute of Technology, Computer
Science Department, Stockholm, Sweden Quantum computing's power
lies in its ability to explore multiple computational paths simultaneously
through quantum parallelism, a concept often misunderstood or oversimplified.
In this talk, we revisit the fundamental nature of quantum parallelism, drawing
analogies with classical parallel computing models such as data and task
parallelism. We introduce the concept of quantum dataflow diagrams as a tool
for visualizing and quantifying parallelism in quantum circuits. By analyzing
quantum algorithms, such as the Quantum Fourier Transform and Amplitude
Amplification, we examine how quantum interference, both constructive and
destructive, impacts algorithm efficiency. Furthermore, we challenge the
direct applicability of classical parallelism laws (Amdahl's and Gustafson’s)
in quantum computing, highlighting the unique role of classical-quantum I/O
and the non-trivial relationship between parallelism and speed-up. This talk
aims to deepen our understanding of quantum parallelism and its implications
for algorithm design and performance evaluation. Reference: Markidis,
Stefano. "What is quantum parallelism, anyhow?" ISC High
Performance 2024 Research Paper Proceedings (39th International Conference),
2024. |
|
Neuromorphic Computing at Cloud Level Christian Georg Mayr TU
Dresden/SpiNNcloud Systems, Germany AI is having an increasingly large impact on our
daily lives. However, current AI hardware and algorithms are still only
partially inspired by the major blueprint for AI, i.e. the human brain. In
particular, even the best AI hardware is still far away from the 20W power
consumption, the low latency and the unprecedented large scale,
high-throughput processing offered by the human brain. In this talk, I will describe our bio-inspired AI
hardware, in particular our award-winning SpiNNaker2 system, which achieves a
unique fusion of GPU, CPU, neuromorphic and probabilistic components. It
takes inspiration from biology not just at the single-neuron level like
current neuromorphic chips, but throughout all architectural levels. On the algorithm front, I will give examples on how
to use general neurobiological computing principles (hierarchy, asynchronity,
dynamic sparsity and distance-dependent topologies/hierarchical computing) to
reframe conventional AI algorithms, usually achieving an order of magnitude
improvement in energy-delay product, for both inference and training. |
|
Realizing Hybrid Quantum-Classical
Applications in OmpSs-2 Lucas Menger MSQC @ Goethe Universität Frankfurt, Germany High-Performance Computing
increasingly explores heterogeneous architectures to accelerate demanding
workloads. We present an extension of the OmpSs-2 programming model that
enables offloading computations to quantum computers in an HPC context. By
modifying the Clang compiler and the Nanos6 runtime, we integrate quantum
devices into the OmpSs-2 ecosystem, allowing developers to write hybrid
quantum-classical applications in a unified way. A custom-built simulator
models quantum nodes in a networked environment, receiving and executing
offloaded jobs. We illustrate the approach with four representative use
cases: random number generation, a mean-field ansatz parameter scan, a
variational quantum-classical algorithm, and a hybrid neural network for
handwritten digit recognition. |
|
Sustainable Management of Complex Software
Ecosystems for Novel Accelerators Eric Mueller University of Heidelberg, Germany Accelerators are
ubiquitous in today's HPC environments. However, the integration
of non-numerical accelerators - such as those leveraging physical computing
principles - presents new and unique challenges. These accelerators often
require novel programming paradigms that differ significantly from
traditional numerical computing, both in the way they interface with
conventional systems and in the way computation is expressed. The software ecosystems
that support these accelerators can be particularly complex, with deep
dependency trees and high coupling between modules, posing significant
challenges to the development and deployment processes. In addition, the lack of
standardized approaches to managing these environments adds to the difficulty
especially in the HPC context. This talk will focus on
the software ecosystem for novel accelerators and present sustainable
strategies for managing, building, deploying, and containerizing these
complex systems. We will use a real-world
case study to illustrate best practices for addressing the challenges of
modern accelerator-driven HPC environments. |
|
Superconducting qubits at the utility scale:
the potential and limitations of modularity Josh Mutus Rigetti Computing, Director Quantum Devices,
USA/Canada The development of fault-tolerant
quantum computers (FTQCs) is receiving increasing attention within the quantum
computing community. Like conventional digital computers, FTQCs, which utilize error correction
and millions of physical qubits, have the potential to address some of
humanity’s grand challenges. However, accurate estimates of the tangible
scale of future FTQCs, based on transparent assumptions, are uncommon. How
many physical qubits are necessary to solve a practical problem intractable
for classical hardware? What costs arise from distributing quantum
computation across multiple machines? We present an architectural model of a
potential FTQC based on superconducting qubits, divided into discrete modules
and interconnected via coherent links. We employ a resource estimation
framework and software tool to assess the physical resources required to
execute specific quantum algorithms compiled into their graph-state form and
arranged onto a modular superconducting hardware architecture. |
|
Logical Resource Analysis of Utility-Scale
Quantum Applications using pyLIQTR Kevin Obenland MIT Lincoln Laboratory, USA As part of the DARPA
Quantum Benchmarking program, MIT Lincoln Laboratory is actively evaluating
proposed applications in physical science for their utility and amenability
to fault-tolerant quantum computing platforms. Our team is developing a tool
called pyLIQTR, which provides implementations of important quantum
algorithms and encodings used in the workflows of applications in physical
science. With the implementations provided by our tool, one can measure the
quantum logical resources required for applications at utility scale in a
number of different ways. pyLIQTR can provide a breakdown and count of gates
used in an application, and it can produce a detailed time-schedule of the
execution of the logical quantum circuit. Our logical circuits can also be
used as the input to resource analysis that targets physical platforms. In
this talk, we will describe the logical circuit implementations available in
pyLIQTR and demonstrate the tool’s logical resource estimation capabilities
by showing analysis of particular applications developed in the QB program. |
|
Quantum Realised: The Energy-Efficient
Frontier For HPC Irwan Owen VP of Business Development, D-Wave Systems Inc.,
CANADA - USA At this presentation Irwan
Owen, VP of Business Development, will provide updates to D-Wave’s progress
in its technology roadmap and commercial use-cases. He will also discuss why today’s HPC
centres require energy-efficient compute for hard problems with quantum
technology. Topics will include AI, research, and the quantum technology
platform supporting production applications for industry use. Irwan Owen is Vice
President of Business Development at D-Wave, responsible for building
strategic relationships with D-Wave’s largest customers and partners He is a
technology industry veteran with 30 years of international experience in
computing, web and mobile markets and has been instrumental in the
commercialization of a number of ubiquitous technologies including the UNIX
operating system and the Java platform. Prior to D-Wave, Irwan held sales
leadership roles at Red Bend Software (now part of Samsung), Palm Inc., and
Symbian Ltd. He also spent five years at Sun Microsystems, where he was a
founding member of JavaSoft Europe. Irwan holds a BSc. (Hons) in Computation
from the University of Manchester Institute of Science and Technology, and
has held roles in engineering, pre-sales support, product marketing, sales
and business development. |
|
Reinventing HPC
with Quantum – A year in perspective Nash Palaniswamy Quantinuum, UK – USA Quantum is here — and like
AI in the past, it is reinventing HPC. I share my learnings and perspectives
on some key questions that I have encountered over the past year – such as What does good like? how does
quantum work with AI and HPC? What is the buying criteria? What are the
Benchmarks? What can I do with these machines? And many more. |
|
Accelerating Quantum Chemistry Simulations On
Quantum Computers Raffaele Santagati Quantum Group, Boheringen Ingelheim, Germany Quantum chemistry
simulations represent one of the most promising applications for
fault-tolerant quantum computers. While recent algorithmic advancements, such
as qubitization, and improved Hamiltonian representations, like tensor
hyper-contraction, have significantly reduced resource requirements,
achieving practical runtimes for industrially relevant systems remains
challenging. To address this, we
combine these advancements with a novel active volume (AV) compilation
technique. This technique optimizes resource utilization by eliminating the
overhead associated with idling logical qubits, though it necessitates a
specialized AV architecture. When paired with modifications to the tensor
hyper-contraction method, AV compilation achieves substantial runtime
reductions by two orders of magnitude. We apply this approach to
a challenging cytochrome P450 system, a key enzyme in drug metabolism. This
demonstration highlights the potential of our combined strategy to bring
quantum computing closer to practical applications in pharmaceutical research
and other industries. |
|
Scaling Analog Neuromorphic Hardware Johannes Schemmel European Institute for Neuromorphic Computing and Kirchoff
Institute for Physics Heidelberg University, Germany 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. For real-world
applications as well as neuroscience, some minimum network sizes are
required. To realize the necessary upscaling, BrainScaleS has invented Wafer
Scale integration. For the future generations
of BrainScaleS, this will not be feasible due to high mask-costs of modern
semiconductor processes. This talk presents an alternate solution based on
Chiplet technology. It introduces concepts
that not only allow to scale BrainScaleS-based networks but will also provide
a general platform for upscaling all kinds of neuromorphic technologies. |
|
Progress towards large-scale fault-tolerant
quantum computing with photons Pete
Shadbolt Co-founder PsiQuantum Corp., Palo
Alto, California, USA In this talk we will
describe progress towards large-scale, fault-tolerant quantum computing with
photons. This talk will span materials innovations for high-performance
photonics, improvements in photonic component performance with an emphasis on
improved optical loss, prototype systems of entangled photonic qubits, qubit
networking, and novel high-power cryogenic cooling solutions designed for
future datacenter-scale quantum computers. We will show new prototype systems
designed to progressively overcome the key challenges to scaling up photonic
quantum computers. We will also give an overview of the architecture of
fusion-based photonic quantum computers, describe near-term systems
milestones, and give a view on the long-term roadmap to useful, fault-tolerant
machines. |
|
Towards an Active
Memory Architecture for Graph Processing beyond Moore’s Law Thomas Sterling Senior Research Scientist, University of Texas at
Austin, Texas Advanced Computing Center, Austin, TX, USA Three significant
challenges constrain future development of semiconductor-based computing
architecture including 1) end of Denard and Moore's scaling, 2) current
limited parallel execution models, and 3) lack of fine-grain graph
processing. The Active Memory Architecture (AMA) addresses these issues
through its innovative memory-centric non von Neumann parallel computer
architecture. AMA is under development at TACC as a smart memory component
eliminating conventional processor cores, dramatically advancing the parallel
execution model, exploiting semantics for dynamic graph processing, and
incorporating dynamic scheduling and resource management runtime. This presentation
will describe the new principles and innovative mechanisms being pursued at
the Texas Advanced Computing Center at near nanoscale. |
|
Sergii Strelchuk Department of Applied Mathematics and Theoretical
Physics and Centre for Quantum Information and Foundations University of
Cambridge and University of Warwick, Computer Science Department, Warwick
Quantum Centre, UK Genomics is a
transformational technology for biology, driving a massive improvement in our
understanding of human biology and disease. Pangenomics is an important next
step on this journey, as understanding variation across many genomes is key
to unravelling how genetic traits can affect health outcomes. Building and
analysing a pangenome is computationally intensive. Many essential tasks in
genomic analysis are extremely difficult for classical computers due to
problems inherently hard to solve efficiently with classical (empirical)
algorithms. Quantum computing offers novel possibilities with algorithmic
techniques capable of achieving speedups over existing classical exact
algorithms in large-scale genomic analyses. Funded by the Wellcome
Leap Q4Bio program (https://wellcomeleap.org/q4bio/), we pursue two main
research thrusts.1. Algorithm Development: We design novel quantum algorithms
for multiple sequence alignment subproblems and investigate heuristic methods
(QAOA) for de novo assembly. 2. Data Encoding and State Preparation: We aim
to develop efficient quantum circuits to encode genomic data and reduce the
computational overhead with a variety of techniques, including tensor network
representations. It facilitates data encoding into quantum states for a
variety of machine-learning applications. |
|
The Diamond Integrated Quantum Chip
Revolution Andrea Tabacchini VP Quantum Brilliance Solutions, Australia and
Germany Nitrogen-vacancy (NV)
centers in diamond have long stood out as a compelling platform for quantum
technologies due to their exceptional properties - such as long coherence
times at room temperature, high-fidelity operations, and high-speed gates.
Despite these advantages, the field has historically regarded NV-based
systems as limited by formidable engineering and scalability challenges. At Quantum Brilliance, we
are redefining those assumptions. With the founders’ 20+ years scientific
experience in the field, sustained R&D, and strategic collaborations, we
have made significant progress in overcoming the key hurdles to practical NV-diamond-based
quantum technologies. Central to this progress is our Integrated Quantum Chip
(IQC), a compact, scalable architecture envisioned to support applications
ranging from quantum sensing to communication to computing. This talk presents a
high-level look at our proprietary five-step process for engineering
high-performance quantum diamond materials, as well as recent experimental
breakthroughs that validate our approach. I will also outline our technology
roadmap for the IQC, highlighting key challenges and recent progresses. |
|
AI-Powered Machine Learning for Accelerating
Scientific Grand Challenges William
Tang Princeton University
Dept. of Astrophysical Sciences, Princeton Plasma Physics Laboratory; Center
for Statistics and Machine Learning (CSML) and Princeton Institute for
Computational Science & Engineering (PICSciE), Princeton University, USA This invited presentation
represents an updated version of the Sidney Fernbach Memorial Award keynote
talk at the international Supercomputing Conference (SC’24) in Atlanta,
GA. It deals with “Artificial
Intelligence-Powered Machine Learning for Accelerating Scientific Grand
Challenges” with highlights including
the deployment of recurrent and convolutional neural networks in Princeton's
Deep Learning Code -- "FRNN" – that enabled the first adaptable
predictive deep learning tool for carrying out efficient "transfer
learning" between experimental facilities while delivering validated
predictions of disruptive events across prominent tokamak devices. Moreover, the AI/DL capability can provide
not only the “disruption score,” as an indicator of the probability of an
imminent dangerous disruption but also a “sensitivity score” in real-time to
indicate the underlying reasons for the predicted disruption. A real-time prediction and control
capability has been significantly advanced with a novel surrogate model/HPC
(high performance computing) simulator ("SGTC") -- a
first-principles-based prediction and control surrogate necessary for
projections to future experimental devices (such as targeted Fusion Power
Plants (FPP's) and indeed ITER) for which no "ground truth"
observational data exist at present.
The near future will feature findings from the deployment of real-time
Surrogates – fast HPC simulators supported by newly validated 1st
principles-based results enabled by using the exciting exascale class high-performance
computing systems that include • FRONTIER (ORNL) and
Aurora (ANL) Exaflop computers in the US; • ALPS with 5000 NVIDIA
Grace-Hopper "Superchips" at the Switzerland’s Supercomputing
Center (CSCS); and the upcoming larger • JUPITER Exascale System
at the German Supercomputing Center (JSC). References: [1] Julian
Kates-Harbeck, Alexey Svyatkovskiy, and William Tang, "Predicting
Disruptive Instabilities in Controlled Fusion Plasmas Through Deep
Learning," NATURE 568, 526 (2019) [2] William Tang, et
al., Special Issue on Machine Learning Methods in Plasma Physics,
Contributions to Plasma Physics (CPP), Volume 63, Issue 5-6, (2023). [3] Ge Dong, et al., 2021, Deep
Learning-based Surrogate Model for First-principles Global Simulations of
Fusion Plasmas, Nuclear Fusion 61 126061 (2021). |
|
Photonic fault-tolerant quantum computing:
Scaling, networking, and modularity Zachary Vernon Chief Technology Officer—Hardware, Xanadu, Canada I will discuss Aurora,
Xanadu's latest photonic quantum computer, showcasing the scalability of our
architecture through modularity and networking. I will also discuss some more
recent hardware developments as we advance towards fault-tolerance. |
|
Towards
Quantum-Classical Supercomputers with Neutral Atoms Aleksander Wennersteen Pasqal, France Quantum computing promises
to accelerate select high-performance computing (HPC) workloads. Realizing this potential
will require deep integration of quantum and classical resources, demanding
close collaboration between the quantum and HPC communities to develop truly
hybrid supercomputing systems. In this talk, we focus on
quantum processing units (QPUs) based on neutral atoms and discuss how they
are being integrated into HPC centers. From the physical infrastructure to
co-processing workflows and scheduling. We present the
capabilities of Pasqal’s devices and platform and then outline our near-term
hardware developments and explain how these advances shape and support the
broader goal of building a hybrid quantum-classical computing platform. |
