中圖分類號：TN201;TP38 文獻標志碼：A DOI: 10.16157/j.issn.0258-7998.257119
中文引用格式： 王鶯時，邵鵬宇，蔣金虎. 基于SIMD并行的量子切分模擬加速優(yōu)化[J]. 電子技術(shù)應(yīng)用，2025，51(12)：1-7.
英文引用格式： Wang Yingshi，Shao Pengyu，Jiang Jinhu. Quantum cutting simulation acceleration optimization based on SIMD parallelism[J]. Application of Electronic Technique，2025，51(12)：1-7.

Quantum cutting simulation acceleration optimization based on SIMD parallelism

Abstract： Quantum computing is a cutting-edge topic that has attracted much attention in the current computing field. Quantum cutting is a highly promising computing framework proposed to break through the current hardware limitations and increase the number of available quantum bits. In the research and verification of it using classical computers, quantum simulation of subcircuits is the most important bottleneck link in computing power. In order to improve the efficiency of quantum simulation, this study found that the naive linear algebra method used in the key matrix calculation link of Qiskit Aer, a quantum simulation framework commonly used in quantum projects, has room for optimization using Single Instruction Multiple Data (SIMD). Based on this, this paper uses AVX2 instructions to optimize the code of the matrix calculation part, and tests the control variables through the containerization method. The SIMD optimization method successfully gave a significant improvement in the quantum simulation efficiency of Qiskit Aer. It has been verified that this improvement is stable, reliable, and reproducible, and will not introduce unknown effects on irrelevant links outside the quantum simulation. The research results of this paper improve the efficiency of quantum cutting simulation, provide a faster tool for the study of quantum cutting, and provide experience and paradigms that can be used for further optimization of the Qiskit framework.

Key words : quantum computing；quantum circuit cutting；quantum simulation；SIMD；Qiskit