Optimal multicore quantum computing with few interconnects

TL;DR

This paper investigates how dividing quantum processors into multiple cores with few interconnects can optimize computational complexity, suggesting practical benefits for current quantum device architectures.

Contribution

It introduces a complexity analysis framework for distributed quantum circuits using the majorization criterion, revealing optimal configurations with few interconnects and universal behavior across architectures.

Findings

Optimal complexity achieved with few interconnects

Universal behavior across different architectures

Complexity scales favorably with adding cores of the same size

Abstract

Noisy intermediate-scale quantum processors have produced a quantum computation revolution in recent times. However, to make further advances new strategies to overcome the error rate growth are needed. One possible way out is dividing these devices into many cores. On the other hand, the majorization criterion efficiently classifies quantum circuits in terms of their complexity, which can be directly related to their ability of performing non classically simulatable computations. In this paper, we use this criterion to study the complexity behavior of a paradigmatic universal family of random circuits distributed into several cores with different architectures. We find that the optimal complexity is reached with few interconnects, this giving further hope to actual implementations in nowadays available devices. A universal behavior is found irrespective of the architecture and (approximately) of the core size. We also analyze the complexity properties when scaling processors up by means of adding cores of the same size. We provide a conjecture to explain the results.