In 2019, Google claimed to have achieved the coveted quantum supremacy level of computing with its Sycamore system. But the problem Google used to make that claim has just been solved by its accelerator of choice, the GPU. As reported by chemistry, A Chinese researcher recently revealed the same computational problems that made Google claim the title, despite having a “minor” 512 GPU with some smart changes to the original algorithm. was successfully resolved. However, the whole notion of quantum supremacy refers to the moment when quantum computers solve problems that classical computers cannot.
At the time, Google said it would take 10,000 years for its then-fastest supercomputer, the IBM-supplied Summit, to solve the same calculations that the company’s Sycamore quantum computer did in 200 seconds. His 512 GPU on the Chinese team took him 15 hours to do the same thing.
This is a reminder that both time and quantum computing are relative. This is understandable given the relative state of the technology in its early stages.
Google’s claim of quantum supremacy is based on the discovery of interference patterns in qubit values. Quantum computing is a fickle master, so all current approaches to quantum computing are prone to decoherence. Decoherence refers to the way in which the environment and qubit design and manipulation introduce errors into computations.
From these operating errors, by running the same algorithm on Sycamore for 200 seconds (and millions of iterations), Google extrapolated the results and showed patterns of deviation from the exact and correct values that the processor should output. I was. These deviations occurred because the error made certain outputs more likely than others. This pattern was finally visualized by a reliably reproducible spiky graph.
This graphical representation of the relationship between error and output claims that Google has given quantum supremacy. And this same graph was achieved by Chinese scientists. To achieve that, they expressed the problem via his 3D mathematical array (matrix). This allowed the 512 GPU’s special tensor cores to solve the problem by simply multiplying the values in the array.
“I think they are right that if they had access to a big enough supercomputer, they could have simulated the task… in seconds.” Scott Aaronson, a computer scientist at the University of Texas, said: chemistryThe Chinese team puts this estimate at a computation time of 12 seconds.
To be fair, the Google scientist left a warning in the paper. Sergio Boixo, principal scientist at Google Quantum AI, said in his email: chemistry that “Classical Algorithms Will Be Improved”And their improvements, perhaps a little too quickly, have dulled the edge of Google’s claims, ultimately proving IBM’s objections correct.
However, Google engineers emphasized one point: technology is forever evolving, and quantum computing is currently undergoing a quantum leap in fewer occurrences in traditional systems. If Google’s Sycamore was able to provide spiky outlines with higher fidelity than that (0.2%), better error correction would allow today’s quantum computers to do it better.
The low fidelity achieved by Sycamore is exactly the bit that gave the Chinese scientists some leeway, only improving the computational fidelity to 0.37%. This is enough to beat Sycamore, but far from what is theoretically possible. Due to that fact and the developing nature of quantum computing, Sergio added:
While that’s very likely correct, it seems Google has to remove the quantum supremacy trophy from its walls. Other hands are sure to rise to claim it.
