Characterization of quantum workloads on SIMD architectures

There is a growing research interest in quantum computing because of its promise to provide significant performance speedups over classical computers at specialized tasks. While there have been many advances in building more capable, robust, and useful quantum algorithms and software, it is not clear how a scalable, high-performance, and area-efficient quantum architecture should be designed for efficient execution of various quantum workloads. This paper attempts to fill this gap by performing a detailed characterization of many real-world quantum algorithms on SIMDstyle quantum architectures. Specifically, we characterize the effect of size and the number of SIMD regions of a quantum SIMD architecture on various metrics such as performance, utilization, qubit (data) movement etc. We hope that the presented insights on the trade-offs and relationships between aforementioned evaluation metrics will be useful towards designing an efficient and scalable quantum architecture.