We introduce quantum utility, a new approach to evaluating quantum performance that aims to capture the user experience by including overhead costs associated with the quantum computation. A demonstration of quantum utility by a quantum processing unit (QPU) shows that the QPU can outperform classical solvers at some tasks of interest to practitioners, when considering computational overheads. We consider overhead costs that arise in standalone use of the QPU (as opposed to a hybrid computation context). We define three early milestones on the path to broad-scale quantum utility that focus on restricted subsets of overheads: Milestone 0 considers pure anneal time (no overheads) and has been demonstrated in previous work; Milestone 1 includes overhead times to access the QPU (that is, programming and readout); and Milestone 2 incorporates an indirect cost associated with minor embedding. We evaluate the performance of a D-Wave Advantage QPU with respect to Milestones 1 and 2, using a testbed of 13 input classes and seven classical solvers implemented on CPUs and GPUs. For Milestone 1, the QPU outperformed all classical solvers in 99% of our tests. For Milestone 2, the QPU outperformed all classical solvers in 19% of our tests, and the scenarios in which the QPU found success correspond to cases where classical solvers most frequently failed. Analysis of test results on specific inputs reveals fundamentally distinct underlying mechanisms that explain the observed differences in quantum and classical performance profiles. We present evidence-based arguments that these distinctions bode well for future annealing quantum processors to support demonstrations of quantum utility on ever-expanding classes of inputs and for more challenging milestones.

翻译：我们引入量子效用这一评估量子性能的新方法，旨在通过纳入与量子计算相关的开销成本来捕捉用户体验。量子处理单元（QPU）的量子效用演示表明，在考虑计算开销的情况下，QPU能够在某些从业者感兴趣的任务上超越经典求解器。我们考虑在独立使用QPU（而非混合计算场景）时产生的开销成本。我们定义了通往大规模量子效用道路上的三个早期里程碑，这些里程碑聚焦于受限的开销子集：里程碑0考虑纯退火时间（无开销），已在先前工作中得到验证；里程碑1包含访问QPU的开销时间（即编程和读出）；里程碑2则纳入了与微小嵌入相关的间接成本。我们使用包含13个输入类别和七种在CPU及GPU上实现的经典求解器的测试平台，评估了D-Wave Advantage QPU在里程碑1和2下的性能。对于里程碑1，QPU在99%的测试中超越了所有经典求解器。对于里程碑2，QPU在19%的测试中超越了所有经典求解器，且QPU成功的情景对应于经典求解器最常失败的情况。对特定输入测试结果的分析揭示了截然不同的底层机制，这些机制解释了观测到的量子与经典性能特征的差异。我们基于证据的论证表明，这些差异预示着未来退火量子处理器有望在不断扩大输入类别和更具挑战性的里程碑上支持量子效用的验证。