Efficient parameterised compilation for hybrid quantum programming

TL;DR

This paper introduces a parameterized compilation approach in OpenQL that significantly reduces compilation time for hybrid quantum algorithms, outperforming existing frameworks like PyQuil and Qiskit.

Contribution

The authors implement explicit parameters in OpenQL to prevent full recompilation, enhancing efficiency for iterative hybrid quantum algorithms.

Findings

Compilation time in OpenQL is significantly reduced.

OpenQL_PC is up to twice as fast as PyQuil and Qiskit.

Improved runtime performance for MAXCUT benchmark.

Abstract

Near term quantum devices have the potential to outperform classical computing through the use of hybrid classical-quantum algorithms such as Variational Quantum Eigensolvers. These iterative algorithms use a classical optimiser to update a parameterised quantum circuit. Each iteration, the circuit is executed on a physical quantum processor or quantum computing simulator, and the average measurement result is passed back to the classical optimiser. When many iterations are required, the whole quantum program is also recompiled many times. We have implemented explicit parameters that prevent recompilation of the whole program in quantum programming framework OpenQL, called OpenQL_PC, to improve the compilation and therefore total run-time of hybrid algorithms. We compare the time required for compilation and simulation of the MAXCUT algorithm in OpenQL to the same algorithm in both PyQuil and Qiskit. With the new parameters, compilation time in OpenQL is reduced considerably for the MAXCUT benchmark. When using OpenQL_PC, compilation of hybrid algorithms is up to two times faster than when using PyQuil or Qiskit.