The evaluation of quantum computing hardware has traditionally leaned heavily on gate fidelity and error rates—metrics that describe the precision of individual operations. However, as the industry moves toward practical utility, researchers are advocating for a shift toward measuring sustained throughput. A new benchmarking framework has emerged to quantify how many operations a processor can execute consistently over time, providing a more holistic view of hardware performance.
While gate fidelity tells us how 'correct' a single step is, it does not account for the cumulative degradation or the operational bottlenecks that occur during long-running algorithms. The new benchmark focuses on the sustained execution of tasks, offering a clearer picture of the actual computational capacity of a quantum processing unit (QPU). This is critical for developers who need to predict how a system will behave when running multi-step circuits.
By establishing a standard for throughput, the research community can better compare disparate hardware architectures. Whether a system utilizes superconducting qubits, trapped ions, or neutral atoms, a standardized throughput metric allows for an apples-to-apples comparison of their practical processing power. This shift is essential for moving from the era of 'proof of concept' to the era of 'functional utility.'
Ultimately, this development serves as a vital step in the maturation of quantum engineering. By prioritizing sustained throughput, the industry can more effectively identify hardware limitations, optimize compiler efficiencies, and provide users with realistic expectations for the timeline of complex quantum applications.