Google is wrong to claim it has reached a turning point in the history of computing by being the first to achieve “quantum supremacy”, according to researchers at IBM.
In a paper published on Monday, five researchers at the US computer maker said Google had overstated a claim that its system, built using the principles of quantum mechanics, could far surpass even the world’s most-powerful supercomputer.
Google’s claim was contained in an unpublished research paper first disclosed by the Financial Times last month. It has been hailed as a big step in the development of quantum computers, which have the potential to solve problems in materials science and other fields that are far beyond traditional computers.
Google reported that its quantum system had taken three minutes and 20 seconds to carry out a calculation that would take the US Department of Energy’s Summit supercomputer 10,000 years to complete. That made it the “first computation that can only be performed on a quantum processor”, according to the paper, adding that this “heralds the advent of a much-anticipated computing paradigm”.
However, the IBM researchers said that, using their own methods, they would be able to program a supercomputer to handle the same task in only two and a half days. They faulted Google for assuming that the supercomputer is limited by the amount of data it can store in system memory, and said the internet company had “failed to account for plentiful disc storage” that could overcome this limitation. It also pointed to other hardware and software advances that, if anything, made its own calculation of two and a half days “a conservative, worst-case estimate”.
IBM, which has been locked in a race with Google to make quantum computing a practical reality, also attacked the company for even trying to claim “quantum supremacy”. It said headlines about the claim would “inevitably mislead the general public” into thinking a new era of computing had been reached, whereas big differences between quantum and classical computers mean they will “work in concert” in future. IBM had already publicly criticised Google’s claim.
Many independent experts hailed the Google achievement as a breakthrough because the company made significant headway on some of the difficult challenges of controlling a quantum system. But some have also warned that it is too soon to declare an end to the race between quantum and classical computers because new ways of programming today’s machines could allow them to catch up.
The Google research paper was briefly posted on a Nasa website last month before being removed, and has yet to be published formally.