[Paper Review] Redefining the Quantum Supremacy Baseline With a New Generation Sunway Supercomputer

Xin Liu, Chu Guo+12 more|arXiv (Cornell University)|Nov 1, 2021

Quantum Computing Algorithms and Architecture4 citations

TL;DR

This paper demonstrates that the most challenging quantum supremacy task—sampling from a 53-qubit random quantum circuit—can be fully simulated on the new-generation Sunway supercomputer in just 440 seconds using a customized tensor network contraction (TNC) algorithm. The simulation collapses the original quantum supremacy claim by showing classical systems can achieve comparable performance, while also enabling precise verification of XEB fidelities for circuits up to 14 cycles.

ABSTRACT

A major milestone in the era of noisy intermediate scale quantum computers is extit{quantum supremacy} [Nature extbf{574}, 505 (2019)] claimed on the Sycamore quantum processor of $53$ qubits, which can perform a random circuit sampling task within $200$ seconds while the same task is estimated to require a runtime of $10,000$ years on Summit. This record has been renewed with two recent experiments on the Zuchongzhi $2.0$ ($56$ qubits) and Zuchongzhi $2.1$ ($60$ qubits) quantum processors. On the other front of quantum supremacy comparison, there has also been continuous improvements on both the classical simulation algorithm as well as the underlying hardware. And a fair justification of the computational advantages for those quantum supremacy experiments would require to practically simulate the same problems on current top supercomputers, which is still in lack. Here we report the full-scale simulations of these problems on new generation Sunway supercomputer, based on a customized tensor network contraction algorithm. Our benchmark shows that the most challenging sampling task performed on Sycamore can be accomplished within $1$ week, thus collapsing the quantum supremacy claim of Sycamore. Additionally, we show that the XEB fidelities of the extit{quantum supremacy circuits} with up to $14$ cycles can be verified in minutes, which also provides strong consistency check for quantum supremacy experiments. Our results redefine quantum supremacy baseline using the new generation Sunway supercomputer.

Motivation & Objective

• To re-evaluate the quantum supremacy claim by performing full-scale classical simulations of the most challenging quantum supremacy tasks on state-of-the-art supercomputers.
• To close the gap in benchmarking by simulating the same problems that were claimed to be intractable for classical systems, using both advanced classical algorithms and top-tier hardware.
• To provide a rigorous consistency check for quantum supremacy experiments by computing exact XEB fidelities for circuits with up to 14 cycles.
• To assess the current limits of classical simulation capacity in light of rapid progress in both quantum processors and classical computing infrastructure.

Proposed method

• Implemented a customized tensor network contraction (TNC) algorithm optimized for the Sunway supercomputer’s architecture to simulate random quantum circuits efficiently.
• Leveraged mixed-precision arithmetic (single and half-precision) to improve performance while maintaining numerical stability.
• Used a highly parallelized, memory-bandwidth-aware TNC strategy to minimize communication overhead and maximize computational throughput on the Sunway system.
• Applied the TNC algorithm to simulate full random circuit sampling tasks for Sycamore-20 and Zuchongzhi-2.0 circuits, achieving high-accuracy results.
• Computed exact amplitudes for up to a million bitstrings to calculate precise XEB fidelities, comparing them with estimated values from quantum experiments.
• Validated the results against the theoretical Porter-Thomas distribution using the XEB fidelity formula: $ P_l(x|\mathcal{F}_{\text{XEB}}) = (\mathcal{F}_{\text{XEB}}x + (1 - \mathcal{F}_{\text{XEB}}))e^{-x} $, where $ x = Np $.

Experimental results

Research questions

• RQ1 Can the most challenging quantum supremacy task—random circuit sampling on a 53-qubit Sycamore processor—be classically simulated within a practical timeframe on a modern exascale supercomputer?
• RQ2 To what extent does the performance of classical simulators, especially using tensor network contraction, close the gap with near-term quantum processors?
• RQ3 How accurate are the estimated XEB fidelities reported in quantum supremacy experiments compared to exact classical computations?
• RQ4 What is the role of classical simulators in verifying the consistency and fidelity of quantum supremacy experiments?
• RQ5 Can classical systems with high memory bandwidth and optimized TNC algorithms outperform the original claims of quantum advantage in sampling tasks?

Key findings

• The full simulation of the Sycamore-20 random circuit sampling task was completed in 440 seconds using single-precision arithmetic on the new-generation Sunway supercomputer.
• With mixed-precision arithmetic, the same simulation was completed in 276 seconds, demonstrating significant performance gains through numerical optimization.
• The most challenging quantum supremacy task—generating one million bitstrings with 0.2% fidelity—can be achieved within one week on the Sunway system, directly challenging the original claim of quantum supremacy.
• Exact XEB fidelities were computed for Sycamore-12, Sycamore-14, and Zuchongzhi 2.0-12, yielding values of (1.34 ± 0.1)%, (0.73 ± 0.1)%, and (0.27 ± 0.1)% respectively, which are consistent with but slightly lower than the estimated values.
• The distribution of bitstring probabilities matched the theoretical Porter-Thomas distribution under the computed XEB fidelities, confirming the consistency of quantum supremacy experiments.
• The authors project that with improved memory bandwidth and better tensor contraction ordering, the simulation time for Sycamore-20 could be reduced to approximately 1.5 hours, representing a 100-fold improvement over current results.