Quantum computational advantage via 60-qubit 24-cycle random circuit sampling.

Area of Science:

• Quantum Computing
• Superconducting Quantum Systems
• Quantum Information Science

Background:

• Continuous improvements in classical algorithms and hardware necessitate upgrades in quantum hardware for sustained quantum computational advantage.
• Superconducting quantum computing systems are a leading platform for achieving quantum advantage.

Purpose of the Study:

• To demonstrate an upgraded superconducting quantum computing system, Zuchongzhi 2.1.
• To showcase enhanced capabilities in large-scale random quantum circuit sampling.
• To significantly extend the quantum computational advantage over classical simulation.

Main Methods:

• Developed Zuchongzhi 2.1, a 66-qubit superconducting quantum processor with a tunable coupler architecture.
• Improved readout fidelity to an average of 97.74%.
• Performed large-scale random quantum circuit sampling experiments with up to 60 qubits and 24 cycles.

Main Results:

• Achieved a fidelity of F_XEB=(3.66±0.345)×10^-4 for the sampling task.
• The Zuchongzhi 2.1 sampling task is approximately 6 orders of magnitude more difficult than Sycamore and 3 orders of magnitude more difficult than Zuchongzhi 2.0 in classical simulation.
• Classically simulating the experiment would take approximately 4.8×10⁴ years, while Zuchongzhi 2.1 completed it in about 4.2 hours.

Conclusions:

• Zuchongzhi 2.1 significantly enhances quantum computational advantage.
• The system's performance demonstrates a substantial leap in quantum computing capabilities.
• The demonstrated advancements pave the way for future, more complex quantum computations.