Thermalization and Criticality on an Analog-Digital Quantum Simulator, Part 2: benchmarking

ORAL

Abstract



Achieving quantum simulations of complex physical phenomena that are intractable for classical computers is a major milestone in quantum computing. We demonstrate such a regime by studying the analog time-evolution of a 2D XY spin model implemented on a 69-qubit quantum processor. To accurately benchmark this evolution, we introduce a novel and scalable calibration technique specifically designed for analog quantum simulation. Our experiments, validated by cross-entropy benchmarking at systems up to 35 qubits, exhibit error rates under 0.1% per qubit per cycle, rapid entanglement saturation and convergence to a robust Porter-Thomas bitstring distribution.

The combination of fast dynamics and low error rates results in a very pure and entangled state quantified by the maximum reached mixed-state entanglement entropy. This combination ensures high computational complexity even using state-of-the-art classical techniques.

*Google LLC

Publication: https://arxiv.org/abs/2405.17385

Presenters

  • Nikita Astrakhantsev

    • Google Quantum AI

Authors

  • Nikita Astrakhantsev

    • Google Quantum AI
  • Trond Ikdahl Andersen

    • Google LLC
  • Amir H Karamlou

    • Google Quantum AI
  • Aaron M S Szasz

    • Google LLC
  • Julia Berndtsson

    • Google Quantum AI
  • Johannes Motruk

    • University of Geneva
  • Jonathan A Gross

    • Google LLC
  • Tom Westerhout

    • Radboud University
  • Alexander Schuckert

    • University of Maryland College Park
  • Xiao Mi

    • Google LLC
  • Dmitry Abanin

    • Google LLC
  • Guifre Vidal Bonafont

    • Google LLC
  • Pedram Roushan

    • Google LLC
  • Andreas M Laeuchli

    • Univ of Innsbruck