Automated Circuit Depth Reduction of Quantum Subroutines via Compilation

TL;DR

This paper presents a compiler-driven method to automatically optimize quantum circuits by reducing their depth, specifically for GHZ state creation and CNOT/CZ chain decomposition, enhancing parallelism at the cost of increased gate count.

Contribution

It introduces a novel compiler-based approach that achieves constant and logarithmic depth reductions for key quantum subroutines, improving scalability and efficiency.

Findings

Significant reduction in circuit depth for benchmarked algorithms.

Constant-depth GHZ state creation and CZ chain decomposition.

Logarithmic depth recursive CNOT chain decomposition.

Abstract

Optimizing quantum circuits by reducing circuit depth is essential for improving the efficiency and scalability of quantum algorithms, particularly as quantum hardware continues to evolve. This can be achieved by restructuring quantum algorithms to allow more parallelism. A compiler is needed to automatically detect and apply these optimizations. In this work, we focus on the optimization of two fundamental quantum subroutines: GHZ state creation and CNOT/CZ chain decomposition. Traditional implementations of these subroutines suffer from linearly increasing circuit depth, which limits scalability. We propose a compiler-driven approach that automatically detects and optimizes these two fundamental quantum subroutines. Our approach reduces circuit depth through constant-depth GHZ state creation, constant depth CZ chain decomposition, and logarithmic depth recursive CNOT chain decomposition, which enhance parallel execution. Performance analysis of benchmarked algorithms shows significant reductions in depth. However, our solution also results in an increased gate count, which makes our optimization a trade-off. The gate count for the CNOT chains is doubled, where logarithmic depth reduction is achieved. The reduced circuit depth results in more efficient algorithms by reducing execution time.