QMill achieves record-breaking quantum circuit compression with LUMI supercomputer
The quantum computing company QMill has set a world record in quantum circuit compression and launched its AI-powered QMill Circuit Compression service in record time. The computations were carried out using the LUMI supercomputer.
Founded less than two years ago, QMill, a quantum computing company based in Espoo, Finland, is developing quantum algorithms and software for near-future quantum computers. The practical benefits of quantum computing are rapidly approaching as hardware and software development progresses. QMill’s record-breaking quantum circuit compression method helps utilize quantum devices more efficiently. When quantum computing tools are ready and investments start flowing into the industry, it’s crucial to be prepared.
“Our mission [ET5.1]at QMill is to bring the moment of practical quantum advantage closer in time so that different industries can start benefiting from the power of quantum computing as soon as possible. By compressing circuits users can achieve higher fidelity, faster runtimes, and more reliable execution which are critical improvements from the perspective of today’s hardware,” says Hannu Kauppinen, QMill’s CEO and one of its founders.
From idea to quantum product in six months
The AI-powered service was trained using the computing resources of the LUMI supercomputer. The easy-to-use QMill Circuit Compression service has received excellent feedback from pilot users. It offers researchers and developers a record-breaking AI-based compression method that typically reduces the number of gates and circuit depth by 20–50 percent. Gates are the basic instructions that control the operation of a quantum computer. The gates are arranged sequentially into a quantum circuit, whose depth indicates how many sequential operations are needed to execute a given algorithm. Quantum computers are still sensitive to noise and errors, so the fewer an algorithm requires, the more reliable and practically useful the outcome will be.
“Our team has recently achieved record-breaking levels in circuit compression with the help of an AI-powered method across gate sets of various vendors. This is one of those cases where you suddenly realize you have fantastic research results, with usefulness that is immediately recognizable. I think that the speed at which we have turned our research results into a product is itself another record. The LUMI supercomputer was an essential tool in this development,” says Mikko Möttönen, the Chief Scientist and Co-Founder of QMill.
The LUMI supercomputer played a central role in the work. Developing quantum algorithms also requires supercomputing.
“We use LUMI to test whether our quantum circuits work correctly and efficiently, and how they scale as we increase the number of qubits. We also use LUMI to compare algorithms – how well a problem can be solved using classical algorithms on a supercomputer versus a quantum algorithm,” explains Ville Kotovirta, QMill’s CTO. “This gives us a strong foundation for broader algorithm and software development.”
Collaboration with QMill has been fruitful also for CSC.
“Working with QMill has helped us develop our services with a customer-centric approach. In this case, the result is both an improved service platform and a truly interesting end product. Quantum circuit compression is a prime example of how classical high-performance computing can significantly enhance quantum computing. This kind of synergy will grow rapidly as the power of quantum computers connected to our computing environment continues to increase. When the raw computing power of supercomputers is combined with the distinctive strengths of quantum computing, hybrid computing opens entirely new doors for research and application development,” says Mikael Johansson, Development Manager of Quantum Technologies at CSC.
Photo:QMill
Read more
Dan Still
Dan Still works with building industrial partnerships and networks to boost industrial HPC use.