Benchmarking device capability and error mitigation

Most quantum computing benchmarks attempt to measure the performance of individual components of the device or are focused on implementing a specific target application. The former category does not account for the potential usage of error mitigation techniques, and the latter is often not scalable to different circuit shapes and sizes. We present a protocol for measuring the capability region of a quantum device that uses error mitigation. This approach takes inspiration from volumetric benchmarks and recent error mitigation experiments to empirically measure the sampling overhead of test circuits. We test this benchmark in noisy simulation for simple application circuits of varying sizes and report the resulting overhead capability region for different noise models with and without error mitigation.