Leigh Lapworth from Rolls-Royce presented his early results on hybrid classical-quantum CFD algorithms at the recent Quantum.Tech conference in Boston, USA. Rolls-Royce is one of the partners in Innovate UK’s Quantum Error Correction (QEC) Project, a consortium creating end-to-end quantum computing solution with early commercial application in CFD. Leigh highlighted QEC and discussed the Technical Challenges being addressed by the project as shown below.
Leigh also presented results from his recent arXiv pre-print and discussed the importance of the classical processing time taken in preparing vectors and matrices for running on the (emulated) quantum device. The results have shown that it is possible, at least in emulation mode, for the HHL (Harrow, Hassidim, Lloyd) quantum linear equation solver to give almost identical results to the classical solver.
Commenting on his findings, Leigh said “Early unoptimized HHL results show that hybrid pre-processing alone can exceed the cost of a classical solution” also emphasising that there is a lot of scope for improvement, which is why Rolls-Royce is making the test matrices open for others to request and work on. “The nice thing is that these are classical computing problems that we can work on now - we don’t have to wait for a fault-tolerant quantum computer”.
“Early quantum advantage in CFD will be just as dependent on classical supercomputers as it will be on quantum computers”.
Commenting on the Quantum Tech event, Leigh said “I think people were a little surprised that we are going straight for a fault tolerant approach. But that’s the advantage of the QEC consortium, the full stack approach allows us to tailor our algorithms for fault tolerant devices and vice-versa with the hardware. As the early results show, there Is so much classical computing to be developed just to be ready for the first fault tolerant machines”.
And the next challenge? “We are working with the STFC Hartree Centre to produce full circuit implementations and run the larger test matrices on their ATOS QLM. We are also working with Riverlane on alternatives to HHL such as Quantum Singular Value Transformation. There’s so much to do, we’ve really only made the first tentative step towards a fault tolerant hybrid CFD solver.”
The QEC Project is tackling a major challenge in making quantum computers commercially viable, correcting the errors that quantum bits – qubits – are prone to. UK start-up Universal Quantum is leading the consortium and their scalable hardware will be used by Riverlane, STFC Hartree Centre and Rolls-Royce to map this specific CFD use case to Universal Quantum’s computer.
By bringing together quantum hardware and software experts with world-class engineers and UK businesses, the QEC Project is creating a new quantum ecosystem. As Leigh says, there’s still a long way to go, but this consortium is pushing forward towards unlocking the full commercial potential of quantum computers.
ความคิดเห็น