| quantum volume | |
|---|---|
| Name | Quantum volume |
| Caption | Schematic depiction of quantum circuit benchmarking |
| Field | Quantum computing |
| Introduced | 2018 |
| Developers | IBM |
| Units | dimensionless |
quantum volume
Quantum volume is a composite metric used to quantify the effective computational capability of a noisy intermediate-scale quantum computer. It combines factors such as qubit count, gate fidelity, connectivity, and circuit depth to estimate how large and complex a quantum circuit a device can reliably run. The metric matters in Quantum Physics and applied quantum information science because it provides a single-number guide for comparing diverse quantum hardware architectures and tracking progress toward fault-tolerant quantum error correction.
Quantum volume was introduced to capture the realistic performance of quantum processors beyond raw qubit numbers. Rather than counting physical qubits alone, it measures the largest size of a square circuit (equal numbers of qubits and circuit depth) that can be implemented with acceptable success probabilities on a given device. The metric relates to theoretical concepts in quantum information theory such as circuit complexity and the threshold theorem for fault-tolerant quantum computation. As a practical tool, quantum volume influences investment, procurement, and research priorities across industry actors like IBM, Google, Rigetti Computing, and IonQ, and research institutions such as MIT, University of Maryland, Oxford University, and Caltech.
Quantum volume is computed through randomized benchmarking of model circuits that stress key hardware features. Core components include: - Qubit count: the number of logical qubits available for the tested circuit; physical qubit counts alone are insufficient without quality metrics. - Gate fidelity: single- and two-qubit gate error rates measured by protocols like randomized benchmarking and cross-entropy benchmarking used by groups at IBM Research, Google AI Quantum, and the Joint Center for Quantum Information and Computer Science. - Connectivity: topology of qubit interactions (e.g., linear chain, heavy-hexagon, all-to-all) as in superconducting qubit arrays and trapped ion chains; limited connectivity increases circuit compilation overhead. - Circuit depth: the allowable sequence length before errors overwhelm the output; depth interacts with coherence times (T1, T2) and readout error rates characterized in labs such as National Institute of Standards and Technology and NIST collaborations. The methodology uses heavy circuits and statistical hypothesis testing to infer whether circuits of a given width and depth produce output distributions distinguishable from noise. Implementations depend on software toolkits like Qiskit and Cirq and rely on classical simulation capacity provided by high-performance computing centers (e.g., Oak Ridge National Laboratory).
IBM formalized a quantum volume protocol describing circuit families, sampling strategies, and success thresholds; the approach was presented in peer-reviewed venues and white papers from IBM Research. Alternatives and complementary benchmarks include randomized benchmarking (RB), interleaved RB, cross-entropy benchmarking used in Google's quantum supremacy experiments, and task-specific metrics such as quantum advantage demonstrations. Standards efforts involve international entities like the Institute of Electrical and Electronics Engineers (IEEE) and coordination among consortia including the Quantum Economic Development Consortium and national initiatives such as the US National Quantum Initiative. Independent evaluations by academic groups (for example, researchers at University of Sydney and University of Waterloo) have proposed modifications to account for compiler optimizations, error mitigation strategies, and different gate sets.
Quantum volume is one of several indicators; it complements but does not replace metrics like raw qubit count, coherence times, gate fidelities, and application-specific performance (e.g., variational quantum eigensolver accuracy). Critics note limitations: it aggregates diverse characteristics into a single scalar that can obscure hardware trade-offs; it may be sensitive to compilation techniques and calibration procedures; and achieving higher quantum volume does not guarantee practical advantage for specific algorithms such as Shor's algorithm or quantum simulation workloads. Researchers propose multi-dimensional benchmarking suites and standardized challenges to provide a fuller picture. The metric also assumes access to repeated circuit runs and classical verification capabilities, potentially favoring well-resourced labs and companies.
Emphasizing quantum volume shapes research agendas, funding decisions, and market competition among firms like IBM, Google, Intel, Honeywell, and startups such as PsiQuantum and D-Wave Systems. While driving rapid engineering improvements, reliance on a single metric risks privileging technologies that score well under its assumptions, potentially entrenching incumbent players and limiting diversity in hardware approaches. Equity and access concerns arise when benchmarking resources, cloud access, and classical verification infrastructure are concentrated in wealthy institutions or countries, reinforcing global inequalities in scientific capacity. Advocacy from academic consortia and public science funding agencies emphasizes open datasets, transparent protocols, and community-driven benchmarks to democratize participation in quantum research.
Quantum volume informs procurement by governments, collaborations between industry and academia, and national science and technology policy under programs like the US National Quantum Initiative and the European Quantum Flagship. Researchers use the metric to report device progress in publications and conference presentations at venues such as the Conference on Quantum Information Processing and APS March Meeting. Policymakers and funders reference quantum volume in roadmaps and calls for proposals to prioritize investments in error correction, fabrication infrastructure, and workforce training. As the field matures, policy debates focus on balancing performance-driven competition with open standards, data sharing, and equitable distribution of access to quantum computing resources for education, research, and socially beneficial applications.
Category:Quantum computing Category:Benchmarks