Architecting Quantum Computing Systems in the Presence of Noise
Quantum computers may solve some problems beyond the reach of classical digital computers. However, emerging quantum systems are typically noisy and difficult to control, leaving a significant gap between the exacting requirements of quantum applications and the realities of noisy devices. Bridging this gap is crucial – my work adapts conventional computer systems techniques to meet the critical theoretical and experimental constraints in quantum processors. I divide my talk into three parts: (i) introducing my recent work on systematic noise mitigation for superconducting transmon qubits [MICRO'20], which enhances the robustness of quantum processors through coordination of control instructions; (ii) demonstrating efficient and reliable quantum memory management [ISCA'20], which implements automated tools for allocation, reclamation and reuse of qubits in quantum programs, much like in garbage collection for classical programs; (iii) discussing on-going work on implementing quasi-fault-tolerant rotation gates in quantum error correction, which seeks to provide correctness guarantees for quantum applications by encoding quantum bits in a way that errors can be detected and corrected, analogous to classical error-correcting codes.