An Improved Volumetric Metric for Quantum Computers via more Representative Quantum Circuit Shapes
In this work, we propose a generalization of the current most widely used quantum computing hardware metric known as the quantum volume. The quantum volume specifies a family of random test circuits defined such that the logical circuit depth is equal to the total number of qubits used in the computation. However, such square circuit shapes do not directly relate to many specific applications for which one may wish to use a quantum computer. Based on surveying available resource estimates for known quantum algorithms, we generalize the quantum volume to a handful of representative circuit shapes, which we call Quantum Volumetric Classes, based on the scaling behavior of the logical circuit depth (time) with the problem size (qubit number).