Computational costs of data definition at the quantum - classical interface

TL;DR

This paper analyzes the classical computational costs involved in defining and interpreting data at the quantum-classical interface, highlighting when these costs are significant compared to the quantum computation itself.

Contribution

It characterizes the classical costs of data definition at the quantum-classical interface and identifies cases where these costs dominate the overall computational effort.

Findings

Classical data assignment costs often exceed quantum computation costs.

Data definition costs are comparable to quantum costs except for specific applications.

The paper provides a framework for understanding data-related classical overheads in quantum computing.

Abstract

Model-independent semantic requirements for user specification and interpretation of data before and after quantum computations are characterized. Classical computational costs of assigning classical data values to quantum registers and to run-time parameters passed across a classical-to-quantum application programming interface are derived. It is shown that the classical computational costs of data definition equal or exceed the classical computational cost of solving the problem of interest for all applications of quantum computing except computations defined over the integers and the simulation of linear systems with linear boundary conditions.