Abstract
In this paper, we introduce VeriQBench—an open-source benchmark for quantum circuits. It offers high-level quantum circuit abstractions of various circuit types, including combinational, dynamic, sequential, and variational quantum circuits, covering almost all existing types in the literature. VeriQBench is a versatile benchmark used in verifying quantum software for applications such as circuit verification, simulation, testing, and debugging. All circuits are described in OpenQASM and validated on Qiskit and QCOR simulators.
Introduction
As the quantum computing field enters the Noisy Intermediate-Scale Quantum (NISQ) era, the complexity of quantum software—including compilers, optimizers, and verification tools—is expanding rapidly. To ensure the reliability of these tools, standardized and comprehensive benchmarks are essential. Most existing benchmarks are limited to "combinational" circuits, which consist of fixed sequences of gates. However, real-world quantum computing is becoming increasingly complex, involving mid-circuit measurements, classical feedback loops, and parameterized structures. VeriQBench addresses this gap by providing a diverse, scalable, and standardized library that covers the full spectrum of modern quantum circuit design.
Methods: Circuit Categorization
VeriQBench categorizes quantum circuits into four distinct types, providing a robust framework for testing various software functionalities:
- Combinational Circuits: The fundamental gate-array model, including standard algorithms like Grover’s and the Quantum Fourier Transform (QFT).
- Dynamic Circuits: These involve mid-circuit measurements where subsequent gates depend on classical results, such as in Quantum Teleportation.
- Sequential Circuits: A newer class incorporating clock signals and loops, enabling the modeling of "Repeat-Until-Success" (RUS) protocols and Quantum Walks.
- Variational Circuits (VQCs): Parameterized circuits used in hybrid quantum-classical algorithms like the Variational Quantum Eigensolver (VQE).
The benchmark is built on OpenQASM (supporting both 2.0 and 3.0) and includes Python generation scripts. This allows users to generate circuits ranging from a few qubits to over 50, ensuring the benchmark remains useful as hardware scales.
Datasets and Benchmark Contents
VeriQBench serves as a comprehensive dataset for quantum software researchers. Its library includes:
- Standard Algorithms: Bernstein-Vazirani, QFT, Phase Estimation, Grover’s, and Quantum Adders.
- Reversible Logic: Classical logic gates implemented via quantum gates.
- Randomized Benchmarks: Clifford circuits, Quantum Volume, and Supremacy circuits.
- Complex Control Examples: Teleportation, Semiclassical QFT, and Quantum Walks.
- Variational Templates: Hardware-efficient, Unitary Coupled Cluster (UCC), and Symmetry-Preserving ansatzes.
Key Insights
A major contribution of VeriQBench is the formal inclusion of Sequential Circuits. By treating quantum circuits more like classical synchronous logic with feedback, the authors provide a path for verifying complex, long-running quantum protocols. Additionally, by providing Generation Scripts rather than static files, VeriQBench ensures scalability. Every circuit in the suite was validated using both Qiskit and QCOR simulators, ensuring robustness and portability across programming ecosystems.
Key Results and Applications
The utility of VeriQBench is demonstrated through its application in several high-impact research areas:
- Equivalence Checking: Proving two different circuit representations perform the same mathematical operation.
- Fault Simulation: Testing algorithm survival in the presence of hardware noise.
- Automatic Test Pattern Generation (ATPG): Creating specific inputs to detect hardware faults.
- Qubit Mapping: Evaluating how efficiently a compiler maps logical circuits onto physical chip layouts.
Conclusion
VeriQBench provides a critical standardized resource for the quantum software community. By moving beyond simple gate arrays to include dynamic, sequential, and variational circuits, it offers a comprehensive testing ground for the next generation of quantum tools.
Source: VeriQBench GitHub Repository
Summary updated