Quantum Computing: “We Need to Develop Realistic Applications Now.”
Jülich, 16 March 2021 – Quantum computers could one day outperform today’s supercomputers to an extent that can only be imagined at present. But there is still a lot of pioneering work to be done. Currently, important decisions are being taken regarding the development of this future technology. We discussed the latest developments with Frank Wilhelm-Mauch, who is coordinating construction at Forschungszentrum Jülich of the first freely programmable European quantum computer as part of the European flagship project OpenSuperQ.
Professor Frank Wilhelm-Mauch, everyone is chasing the quantum computer. What is the current state of development?
With the demonstration of quantum advantage by Google in 2019 and by the Chinese Academy of Sciences’ with 50 qubits, or 100 light particles, in 2020, quantum computers have been shown to outperform the biggest classical supercomputers in artificial tests. We need to develop realistic applications now, and not wait to start with 1 million qubits.
When we talk about quantum computers today, we’re referring to large-scale instruments for research and development. They must become successively bigger, with developers using them to optimize their applications and thus make them smaller and more efficient to ensure that hardware performance and software requirements eventually become aligned. Where exactly this alignment occurs between 100 and 1 million qubits is a matter of scientific debate.
How can performance be improved?
For the foreseeable future, we first need to reduce the probability of error, which is a much greater problem in quantum computers than in classical ones. The 50-plus qubits currently being achieved by leading platforms are of hardly any use. Whenever you run the aforementioned realistic applications or benchmarks on these computers, they use no more than 12 of the 50-plus qubits because errors start to become too prevalent beyond that point.
There are a range of (fundamental) research approaches to reduce errors. It only really makes sense to increase the number of qubits after pursuing these approaches. Since the 1990s, it has been possible to pack many Josephson junctions – the basic components of a superconducting quantum chip – on a chip, but the real trick is in ensuring they can all be utilized by the computer.
US companies like IBM and Google are already planning quantum computers with thousands of qubits. Germany’s quantum roadmap is currently only targeting 500 qubits. Can Germany really achieve a competitive advantage pursuing this goal?
I can only stress here that all the qubits must be usable – in the sense of the low error rate that I mentioned earlier. Above all, however, applications and hardware need to be tightly interlinked in order to exploit the various possibilities of customized hardware, particularly in the area between gate-based, universal quantum computers and adiabatic quantum computers, which are often referred to as quantum annealers.
The latter have a much simpler structure and can only be used for specific problems, such as the optimization of traffic flows. Their development is therefore at a much more advanced stage, meaning that these machines are likely to be used much sooner for realistic applications than universal quantum computers. In Germany, with our existing expertise in basic research, we have the know-how to develop and utilize the best of both worlds.
The German Federal Government is looking to invest an additional € 2 billion in the development of quantum technologies over the next four years as part of an economic stimulus package. Last month, an initial € 120 million of funding was invested in the development of quantum processors in Germany. What impact do you think this will have?
We have waited for this hardware call for a long time. It allows us to combine our strengths at a national level and to take quantum computing in Germany from basic to applied research. We can also work together with newly appointed colleagues and companies like Infineon, IQM, and HQS.
Jülich is involved in three projects, with decisions still pending on other projects – so it’s been an outstanding call. We’re making progress with several hardware platforms and are demonstrating how in Germany, our expertise is essential in different areas.
The quantum computing roadmap created by a group of experts on behalf of the German Federal Government proposes the establishment of “quantum hubs” as a key part of its strategy. Why do we now need to concentrate on just a few locations?
I wouldn’t describe it as concentrating on certain locations, but rather as concentrating on effective teams. The hubs can and should bring together the best minds in Germany. The quantum hubs have specialized areas, but are not all found in the same location. We need this concentration to cope with the great challenge of this topic and to proceed with a coordinated approach. A quantum computer is based on many components and technologies (many of which come from classical physics) and this approach ensures that we can develop them in a coordinated and compatible manner, maintaining the correct degree of redundancy – as a safety reserve, but not too much. In the face of the complexity of the topic, this is also what we need at an international level to ensure the German science community realizes its potential.
What role could Forschungszentrum Jülich play in the future development of quantum computers?
At Jülich, we focus on solid-state qubits in close collaboration with research institutions and companies predominantly based in Germany as part of a coherently organized programme. We work on both superconducting and semiconductor qubits, and have ensured that the important ecosystem is closely coupled with application, theory, and supporting technology. Together with a number of strategically important partners, we can become the German hub for these platforms.
Prof. Dr. Frank Wilhelm-Mauch
Peter Grünberg Institute – Quantum Computing Analytics (PGI-12)
Tel: +49 2461 61-6106
Tel: +49 2461 61-4771