Edinburgh algorithm lets ordinary computers match quantum boson sampling

[LunarShade65]

A team led by the University of Edinburgh and EPCC, the UK's national supercomputing centre, has developed an algorithm that lets ordinary computers simulate Gaussian boson sampling. UKAuthority's Mark Chillingworth reports that the method needs only around five percent of the computing power previously required

Gaussian boson sampling is a task designed to show off photonic quantum computers. Several experiments have claimed quantum advantage using it, on the basis that classical computers would struggle to reproduce the results. Until now, simulating it needed the world's most powerful supercomputers

In tests, classical machines running the new algorithm matched current quantum devices, and in some cases did better. The related preprint reports beating 100 mode experiments using a single CPU or GPU. The work was led by Dr Raul Garcia-Patron Sanchez of Edinburgh's School of Informatics, with Polytechnique Montreal

The research is published in Science Advances and funded by EPSRC and NSERC. Edinburgh first announced it on 14 October, so this is a slightly late write up, but it is an important result for the debate over quantum advantage

Results like this are a healthy part of science. Every claim of quantum advantage invites classical researchers to try to beat it, and sometimes they succeed. That pushes quantum hardware teams to build bigger and better experiments

Does this weaken past claims of quantum advantage? And is boson sampling still a useful benchmark?

National Supercomputing Centre Makes Quantum Algorithm Breakthroughwww.ukauthority.com

[IronFist21]

Science Advances is a respected journal. Peer review gives the result credibility

Hard to argue with published numbers. Journals add weight to arguments like this