Modular Architecture for Classical Simulation of Quantum Circuits

Abstract

Duke ARTIQ Extensions (DAX) provides a framework for writing modular control software for ion-trap quantum systems. DAX allows users to interface with the system at the gate level using an intermediate representation called DAX.program. However, DAX does not have the tools needed to simulate these programs to see quantum state evolutions, as DAX only allows for simulations at the device level. We propose a modular simulation framework, DAX.program-sim (DPS), that can simulate quantum programs at the level of quantum operations. This addition to DAX for classical simulation of quantum systems is vital for testing, benchmarking, and verification of quantum hardware. The DPS pipeline is designed to have input identical to the one that runs on the hardware. Our architecture allows flexible backends for simulation, enabling both gate-level simulations and pulse-level simulations with and without noise. We demonstrate this unified workflow, executing the same DAX.program file in simulation as well as on hardware. As a specific example, we run benchmarking circuits using simulators targeting ion trap quantum computers and compare them to hardware results of the same circuits.

Find the PDF here.

Publication
In 2022 IEEE International Conference on Quantum Computing and Engineering (QCE)