Researchers at the University of Edinburgh and EPCC have developed an algorithm that lets ordinary computers run quantum simulations that previously required some of the world’s most powerful supercomputers.
The software advance could make these simulations significantly more accessible to researchers working to improve quantum computing technology.
The algorithm reduces the computing power needed to run the simulations to around 5% of the previous requirement, according to the research team.
It also enables standard computers to perform Gaussian boson sampling, a simulation used to demonstrate the capabilities of photonic quantum computers.
The findings could lower the cost of investigating how quantum devices behave in practice, helping scientists identify their limitations and inform the development of larger, more reliable machines.
Dr Raul Garcia-Patron Sanchez, of the University of Edinburgh’s School of Informatics, explained: “Classical simulation of large-scale quantum computing circuits has been achieved so far via a combination of clever algorithms and the intensive use of supercomputing facilities.
“Our work has contributed significant improvements in memory use, running time and energy costs in the simulation of photonic quantum computing devices, making mainstream what was before only accessible to a handful of facilities in the world.”
What are quantum computers?
Quantum computers are devices that use quantum principles to perform calculations, potentially tackling problems beyond the capabilities of even the most advanced conventional supercomputers.
However, existing quantum computers remain relatively small and error-prone, limiting their performance.
Improving their capabilities requires researchers to understand how these machines operate under real-world conditions, including the challenges that could affect their reliability as they grow in size.
Quantum simulations provide a way to investigate this behaviour before designing more advanced devices.
The problem has been the computing resources required to run them, which have traditionally been available at only a small number of facilities worldwide.
The new algorithm could help overcome that barrier by letting more researchers run complex simulations without relying on specialist supercomputing infrastructure.
Simulating photonic quantum computers
The research was led by scientists at the University of Edinburgh and EPCC, the UK’s first National Supercomputing Centre, which is part of the University.
Their work focuses on Gaussian boson sampling, a computational task associated with photonic quantum computers.
These devices use light to carry out quantum operations and represent one of several approaches being explored by researchers in academia and industry.
The simulation is designed to demonstrate the superior capabilities of light-powered quantum devices. Running it on standard computers could give scientists a more accessible way to investigate these systems’ performance.
In tests, conventional computers using the new algorithm matched, and in some cases outperformed, certain existing quantum devices in running the simulations.
This result highlights the potential of improved software to expand what researchers can do with the computing resources already available to them.
Scaling up quantum simulations
The team is now working to scale up the algorithm, aiming to run substantially larger simulations without increasing the required computing power.
If successful, this could let researchers model photonic quantum computers in unprecedented detail, offering further insight into how these devices might perform.
The work could therefore support efforts to identify the limitations of current quantum technology and guide the design of improved systems.
By reducing the resources needed for quantum simulations, the algorithm offers researchers a more practical route to investigating complex quantum devices.
Its further development could help make detailed modelling more widely accessible as scientists work towards larger and more reliable quantum computers.