Cycle Benchmarking for Scalable Verification of Quantum Circuit Performance

Chao Lu, Navnil Choudhury, Kanad Basu

Quantum circuit transpilation facilitates the decomposition of complex operations into simpler hardware-compatible gates. This results in significant changes to the input circuit, and consequently, it becomes imperative to verify the correctness of the transpilation process. Quantum circuit equivalence checking is used to verify the operational equivalence of large quantum circuits to ensure the correctness of computation. Although small quantum circuits can be simulated easily on a classical computer without transpilation, this approach is unscalable for larger, more complex quantum circuits. Furthermore, a scalable quantum equivalence checking algorithm exhibiting subpar performance when applied to larger quantum circuits lacks reliability. Therefore, an equivalence-checking algorithm with scalability and acceptable performance is necessary to verify large quantum circuits. In this paper, we propose QuaSi, a simulation-based verification methodology designed to generate distinct test cases to achieve a scalable and reliable quantum circuit equivalence-checking. Our proposed approach surpasses existing state-of-the-art simulation-based methodologies by boosting the success rate of verification, in addition to reducing the verification latency. When evaluating using large benchmarks with up to 130 qubits, QuaSi furnished significantly improved performance compared to existing approaches in all types of scenarios. Furthermore, we proposed an algorithm to locate the erroneous gates to facilitate debugging the transpilation process.

Megan Lilly, Travis S. Humble

Jordan Hines, Timothy J Proctor

Current benchmarks for midcircuit measurements (MCMs) are limited in scalability or the types of error they can quantify, necessitating new techniques for quantifying MCM performance. Here, we introduce a theory for learning stochastic Pauli noise in MCMs and use it to create MCM cycle benchmarking, a scalable method for benchmarking MCMs. MCM cycle benchmarking extracts detailed information about the rates of errors in randomly compiled layers of MCMs and Clifford gates, and we demonstrate how its results can be used to quantify correlated errors during MCMs on current quantum hardware. Our method can be integrated with existing Pauli noise learning techniques to scalably characterize the errors in wide classes of circuits containing MCMs.

Timothy J Proctor, Kenneth Rudinger, Mohan Sarovar +3 more

Yu‐Fang Chen, Kai-Min Chung, Ondřej Lengál +3 more

As quantum computing hardware advances, the demand for scalable, precise, and fully automated verification techniques for quantum circuits grows. This paper introduces a novel automata-based framework tailored for the verification of quantum circuits. In our approach, the problem is framed as a triple { P } C { S } and the question is whether, given a set P of quantum states on the input of a circuit C , the set of quantum states on the output is equal to (or included in) a set S . Our framework leverages tree automata to compactly represent sets of quantum states and we develop transformers to implement the semantics of quantum gates over this representation. We implemented the proposed approach in a prototype tool and evaluated its performance against various benchmarks from the literature. The evaluation shows that our approach is quite scalable, for example, we managed to verify a large circuit with 40 qubits and 141,527 gates, or catch bugs injected into a circuit with 320 qubits and 1,758 gates, where all tools we compared with failed. Additionally, our work bridges quantum program verification and automata, opening new possibilities to exploit the richness of automata theory in quantum computing.

Mohan Sarovar