Toward Practical Quantum Computing Systems with Electronic Design Automation

Abstract

Quantum computing holds transformative potential across scientific and industrial domains, yet its practical deployment remains constrained by noise, limited scalability, and fragmented design methodologies across abstraction layers. In superconducting quantum platforms, critical gaps persist between quantum devices and system architectures, as well as between hardware systems and application-level algorithms. These disconnects hinder holistic optimization and impede the realization of robust, scalable quantum computing. This dissertation establishes Electronic Design Automation (EDA) as a unifying paradigm to systematically bridge these gaps and enable practical quantum computing systems.To bridge the divide between the quantum devices and the quantum system architectures, this dissertation develops quantum-native physical design automation methodologies that embed device-level physics into system-level optimization. QPlacer is proposed as a frequency-aware analytical placement framework that integrates frequency-domain isolation with spatial layout optimization. By analogizing quantum components as charged particles and introducing frequency-repulsive forces, QPlacer jointly optimizes spatial isolation and substrate compactness, mitigating inter-qubit, resonator, and substrate crosstalk within a unified EDA framework.Building upon global placement, qGDP is introduced as a quantum-aware legalization and detailed placement engine that enforces quantum-specific spatial constraints while preserving system-level layout quality. Unlike classical legalizers, qGDP integrates resonator integrity, crossing reduction, and airbridge minimization into post-placement refinement. Integrated as a cohesive suite, QPlacer and qGDP establish a pioneering quantum-native physical design automation flow, effectively bridging device-level physics and scalable quantum system realization.Beyond hardware-level challenges, this dissertation bridges the gap between quantum systems and quantum applications, particularly for variational quantum algorithms (VQAs) operating on noisy intermediate-scale quantum (NISQ) devices. Dynamic noise drift destabilizes iterative optimization and disrupts application-level convergence. To address this, DISQ is proposed as a dynamic iteration-skipping framework that detects and filters noise-drift-corrupted iterations through multi-reference validation and Pauli-term subsetting. By constructing a more stable optimization landscape at runtime, DISQ aligns system-level noise characteristics with application-level training robustness.Collectively, this work establishes EDA as a cross-layer co-design methodology for quantum computing, extending its principles beyond classical integrated circuits. Bridging quantum devices to systems and systems to applications lays the foundation for robust, scalable, and practical superconducting quantum computing platforms.