Experiment2019

Quantum Supremacy Using a Programmable Superconducting Processor

Auteurs: Frank Arute, Kunal Arya, Ryan Babbush, et al. (Google AI Quantum)

Publié: Nature 574, 505–510 (2019)

En une phrase

Reports the first experimental demonstration that a quantum processor can perform a specific task infeasible for classical supercomputers.

Points clés

  • Google's 53-qubit Sycamore processor sampled random quantum circuits.
  • The task took ~200 seconds on Sycamore versus an estimated thousands of years on classical hardware (a figure later disputed).
  • A milestone for 'quantum supremacy' — a contrived task, not a useful application.

En langage simple

Google ran a task deliberately designed to be easy for its quantum chip and brutally hard for classical computers: sampling the output of random quantum circuits. Sycamore did it in minutes; simulating the same thing classically was estimated to take millennia. Critics argued better classical algorithms shrink that gap, and the task itself is useless — but the experiment showed, for the first time, that quantum machines had crossed into territory classical computers struggle to reach.

Pourquoi c'est important

It was the first concrete evidence that quantum hardware can exceed classical simulation on some task, a psychological and scientific turning point — even though the specific benchmark has no practical use and classical methods later narrowed the gap.

Termes du glossaire associés