Quantum Instruction Set Design for Performance

• DOI: 10.48550/arxiv.2105.06074
• Publication Date: 2021-01-01
• Authors: Cupjin Huang; Tenghui Wang; Feng Wu; Dawei Ding; Qi Ye; Linghang Kong; Fang Zhang; Xiaotong Ni; Zhijun Song; Yaoyun Shi; Hui‐Hai Zhao; Chunqing Deng; Jianxin Chen
• Keywords: Quantum Fourier transform; Quantum circuit

A quantum instruction set is where hardware and software meet. We develop new characterization compilation techniques for non-Clifford gates to accurately evaluate different designs. specifically apply them our fluxonium processor that supports mainstream $\mathrm{iSWAP}$ by calibrating characterizing its square root $\mathrm{SQiSW}$. measure a gate fidelity of up $99.72\%$ with an average $99.31\%$ realize Haar random two-qubit using $\mathrm{SQiSW}$ $96.38\%$. This error reduction $41\%$ the former $50\%$ latter compared on same processor. shows designing consisting single-qubit such platforms leads performance boost at almost no cost.