Introducing the Quantum Universal Operations Performance System: QUOPS

Introduction

Quantum computing is moving beyond the noisy, intermediate‑scale era toward fault‑tolerant, application‑scale machines. Traditional specs—qubit count, gate fidelity, and gate speed—no longer tell the full story of what a system can actually accomplish. The Quantum Universal Operations Performance System (QUOPS) addresses this gap. Developed with input from Sandia National Laboratories, Quantinuum, and NVIDIA, QUOPS provides a single, architecture‑agnostic benchmark that reports two intuitive metrics:

• Q – the largest circuit size (2 × width × depth) that a system can run successfully within a utility‑motivated region.
• Ω – the effective operations‑per‑second at that circuit size.

Together, these numbers give a two‑dimensional view of capability, translating diverse hardware designs into a common currency: quantum operations.

Why a New Metric Is Needed

Component‑level numbers remain essential for engineering, but they often fail to predict system‑level performance, especially once error correction and fault tolerance are introduced. Logical computation involves many layers—encoding, syndrome measurement, decoding, magic‑state production, routing, and control—each adding overhead that raw qubit counts ignore. As organizations plan for real‑world workloads requiring 10⁹–10¹² operations on thousands of qubits, they need a simple answer to a critical question:

What computation can this machine actually execute successfully?

QUOPS answers by measuring the integrated system rather than inferring performance from isolated components. It runs randomized workloads across a grid of widths and depths, checks whether each workload meets a predefined success threshold, and maps the boundary of the system’s capability region. The resulting Q and Ω values capture both the scale of feasible computation and the speed at which it can be performed.

Practical Benefits for Procurement and Planning

For HPC centers, governments, and commercial users, QUOPS offers a concrete reference point when evaluating quantum hardware:

• Clear comparison – Because QUOPS is architecture‑neutral, buyers can compare trapped‑ion, superconducting, and photonic platforms on equal footing.
• Anti‑gaming – The benchmark includes anti‑gaming provisions that prevent vendors from optimizing solely for a single metric.
• Decision‑making – A procurement target such as “run a trillion error‑free operations” can be translated directly into required Q and Ω values, simplifying resource‑estimate calculations that would otherwise involve complex assumptions about error‑correction codes, magic‑state factories, and decoding latency.

Early Results and Industry Landscape

Initial QUOPS runs have been performed on leading processors from Quantinuum, Google, and IBM. The data illustrate distinct trade‑offs:

• Superconducting systems (e.g., Willow, Boston) – Very fast gate speeds yield high Ω values but limited connectivity restricts the maximum Q.
• Trapped‑ion QCCD systems (e.g., Helios) – All‑to‑all connectivity enables larger Q scores, though slower gates result in lower Ω.

These results confirm a long‑standing intuition: speed can be traded for circuit size, and QUOPS quantifies that trade‑off with a single, comparable metric. The benchmark also accounts for the sampling overhead of general error‑mitigation techniques, providing a realistic view of effective throughput.

Building a Benchmarking Ecosystem

QUOPS is not intended to replace specialized benchmarks that target specific applications, component‑level diagnostics, or hybrid HPC workloads. Instead, it serves as a foundational system‑level yardstick that can be layered with other suites. By establishing a common language for quantum performance, QUOPS helps:

1. Standardize reporting across vendors and research groups.
2. Guide roadmap planning for hardware developers aiming at utility‑scale targets.
3. Inform policy and funding decisions with objective, comparable data.

Closing Thoughts

As quantum computers transition to fault‑tolerant operation, the industry needs metrics that reflect real computational capability rather than isolated hardware specs. QUOPS delivers that by capturing both the size of executable circuits and the speed at which they can be run, all within a single, architecture‑agnostic framework. For buyers, researchers, and vendors alike, QUOPS offers a clear, comparable, and actionable view of quantum performance—turning abstract hardware numbers into concrete, mission‑relevant insight.