Professor Leigh Lapworth is a Rolls-Royce Fellow in Computational Science; and an Honorary Professorial Fellow at the University of Warwick, he discusses how important high performance computing is
Speed gives you choices
High performance computing is critical to our business and its impact extends across the life of our products. It’s crucial to how we design and test jet engines and enables us to provide market-leading data-driven services to our customers for the life of the power products they buy. Overall, it’s made us a better, more efficient engineering business, with the ability to use computational power to provide a competitive advantage.
Constant computing
On average we launch a high-fidelity computational fluid dynamics (CFD) simulation once every 30 seconds and a stress analysis once every second. You only have to look at the Trent XWB to see all the areas where high performance computing has helped us to design the world’s most efficient engine. We have some of the world’s most evolved tools for designing subsystems, like a compressor. Yet, there are still things we don’t and can’t yet model. We have an ambition to model an entire gas turbine in high fidelity; and beyond that, new systems such as the electromagnetic interaction between a gas turbine driving an alternator that can power low carbon hybrid electric flight. These are calculations that require thousands of times more computing resources than we currently use and would take many months to complete on even the most powerful of today’s high performance computers.
A solution on the horizon
This is where quantum computing comes in. Ever since Nobel Laureate Richard Feynman proposed the idea in 1981, we have known quantum computers have the potential to revolutionise computing in much the same way as the integrated circuit did in the 1960s and 70s. Whole engine calculations that take months could be completed in days or even hours. This means faster results, which means a faster design; huge reductions in testing and development costs; and even better products that are faster to market, and therefore faster to generate revenue. As my old boss, Ian Martindale, used to say, “speed gives you choices”.
Hype or reality?
Today’s quantum computers are still research devices and achieving the scale needed for whole engine simulations is many years away. It’s sometimes hard to separate the quantum hype and reality. But fundamentally computers are for nought if they don’t have software to run on them. And good quality engineering software also takes years to develop. Not forgetting that quantum computing demands an entirely new way of thinking about programming. That’s where our Innovate UK-funded programme comes in. With our partners we are pushing the limits of what is currently possible.
Colder than deep space
This is why the announcement from our partners, Universal Quantum, is such a breakthrough. To scale quantum computes up to the level of whole engine simulations means they must be built from modules. I have spent much of my 30-year career worrying how to pass data efficiently between modules of a classical computer. Quantum takes this challenge to a whole new level. Many quantum computers operate at temperatures colder than deep space due to the fragility of the quantum state. To be able to pass quantum data from one module to another with only 0.000007% data loss is truly astounding and gives real hope that quantum computers will realise the promise they offer.
Comments