“Euro-Q-Exa will take us into unknown territories”
Technologie:Quantum Computing
Forschungsbereich:Future Computing
Within the User Enablement and Applications team the excitement is high: The specialists support the researchers who work with the quantum systems at the LRZ– soon also with Euro-Q-Exa. With its 54 qubits, the European quantum computer is the most powerful system in the LRZ portfolio. “With the current memory capacity of SuperMUC-NG we can simulate the quantum state of systems with up to 42 qubits”, explains Luigi Iapichino, physicist (PhD) and head of the team. “As a result, it’s becoming increasingly difficult to reproduce the results from Euro-Q-Exa using classical high-performance computing.” At the interface between high performance computing (HPC) and quantum computing, new questions arise and new methods must be developed – an exciting field for users, support teams, and research. Luigi Iapichino explains what researchers can expect from Euro-Q-Exa and how they can apply for access in this interview.
The LRZ is already operating a quantum computer based on superconducting circuits from IQM, Q-Exa. Now Euro-Q-Exa starts into operation, what makes this system so special?
Dr. Luigi Iapichino: Quantum computing is still a young technology and under constant development. At LRZ we integrate quantum computers into supercomputers so that we can actually better compute with them. With 54 qubits, Euro-Q-Exa offers more than twice as many qubits as Q-Exa and thus significantly higher computing resources, because every additional qubit in principle doubles the computational capacity. In addition, the European system is based on much improved and further developed hardware, with respect to earlier generations of products. While Euro-Q-Exa does not yet achieve quantum advantage, with this system we are ideally crossing the line of computations which can be simulated with exact techniques on classical supercomputers: based on our experience, with the current memory capacity of SuperMUC-NG we can simulate the quantum state of systems with up to 42 qubits. This means that it is becoming increasingly difficult to reproduce the results from Euro-Q-Exa using classical high-performance computing. Euro-Q-Exa will take us into unknown territories.”
Besides computing power, are there other differences between the systems?
Iapichino: A crucial feature of the systems is their software. Euro-Q-Exa runs the latest version of the Munich Quantum Software Stack, MQSS for short. It includes adapters to widely used quantum software such as PennyLane, Qiskit, CUDA-Q, and allows easy access to a wide range of quantum computing resources within a unified environment. Within the Munich Quantum Valley, we also collaborate with partners that will provide further programs and tools, including those closer to scientific applications. And there is another difference between the two quantum computers: Q-Exa is integrated into SuperMUC-NG Phase 1, while Euro-Q-Exa is planned to be connected to Phase 2 of SuperMUC-NG, and the planned second system with more than 100 qubits is expected to be linked with the upcoming Blue Lion in 2027.
What can we compute with a quantum computer offering 54 qubits?
Iapichino: The number of qubits alone does not convey the full story about the potential of a quantum computer, it has to be considered together with the fidelity of the system, its usability through the available software, the integration with the HPC ecosystem. Systems of that size and quality allow interesting proof-of-concept simulations, especially when HPC and quantum resources work in combination within a hybrid workflow.
Who can use Euro-Q-Exa, and what is required?
Iapichino: Euro-Q-Exa will be available to researchers from across Europe, through the Munich Quantum Portal and as accelerator for HPC workflows making use of quantum resources. The Munich Quantum Portal also provides access to the LRZ’s other quantum resources – Q-Exa and an ion-trap system from AQT. The access through the HPC system enables hybrid workloads that combine quantum technologies with classical methods. An application process for the European quantum computers of the EuroHPC Joint Undertaking is still being developed. We plan to open the regular user operation of Euro-Q-Exa in June, after a test period, and by then we will describe the access process in the LRZ documentation.
Where is quantum computing better than high performance computing? And are the results already reliable?
Iapichino: There are a few areas where quantum algorithms promise a scaling which looks more favourable than the classical counterpart. A typical example is the simulation of quantum systems in computational chemistry. Another promising area is optimisation. In both fields there are interesting results which testify how development of algorithms is of key importance for exploiting even limited hardware resources. One should add for clarity that we are talking about results which are still well within reach of classical computational resources. However, the competition is open.
Which research disciplines will benefit from these advantages?
Iapichino: Although the current list of algorithms which are promising for advantage is not very long, their applications are long-reaching. The simulation of quantum systems, besides chemistry, can impact fields like material science, medicine, solid state physics, to name a few. Optimisation has a very broad spectrum of applications from the supply chain management to the energy sector and finance. Fields like Quantum Machine Learning are getting traction. An additional research line has to do with systems of partial differential equations, which are ubiquitous in computational fluid dynamics, climate research and astrophysics. There is active research and the hope is that a combination of hardware improvements alongside algorithm development will bring us step by step closer to a breakthrough.
Do-you see interest in working with Euro-Q-Exa, and what can researchers expect from the LRZ?
Iapichino: Yes, of course — we already have initial interest from research groups across Europe. With them, we will begin a Friendly User Phase in March. During a kick-off with IQM Quantum Computers, participants will receive an initial overview of the technical capabilities of the hardware and software. Afterwards, a solution architect from our team will support each research group, implement additional software if needed, and assist with any issues that arise. In this first usage phase, we will also evaluate what researchers need for their activities using quantum computers. During ISC26 in June we aim to present the first experiences and results.
What about training opportunities?
Iapichino: We already offer various training sessions on quantum computing and will now observe what researchers are interested in, what knowledge they need, and which topics we should add. We also plan to implement new offerings — such as software or tools — on the quantum systems and present their strengths in workshops with industry partners. Interview: vs | LRZ