Superstaq: Deep Optimization of Quantum Programs (2309.05157v1)

Abstract: We describe Superstaq, a quantum software platform that optimizes the execution of quantum programs by tailoring to underlying hardware primitives. For benchmarks such as the Bernstein-Vazirani algorithm and the Qubit Coupled Cluster chemistry method, we find that deep optimization can improve program execution performance by at least 10x compared to prevailing state-of-the-art compilers. To highlight the versatility of our approach, we present results from several hardware platforms: superconducting qubits (AQT @ LBNL, IBM Quantum, Rigetti), trapped ions (QSCOUT), and neutral atoms (Infleqtion). Across all platforms, we demonstrate new levels of performance and new capabilities that are enabled by deeper integration between quantum programs and the device physics of hardware.

• The paper presents a novel quantum software platform, SuperstaQ, that superoptimizes quantum program compilations at the hardware level, achieving up to 10x performance improvements.
• It employs innovative techniques such as dynamical decoupling to extend qubit lifetimes, quadrupling T1 and doubling T2 coherence in experimental tests.
• The platform demonstrates adaptability across multiple architectures, reducing gate costs and runtime overheads on superconducting, neutral atom, and trapped ion systems.

SuperstaQ: Superoptimization of Quantum Programs

The paper entitled "SuperstaQ: Superoptimization of Quantum Programs" presents a quantum software platform, SuperstaQ, that significantly enhances the execution efficiency of quantum programs. By optimizing components down to the hardware primitive level, SuperstaQ achieves an estimated 10x performance improvement, showcasing experimental success across several quantum hardware platforms, including superconducting, trapped ion, and neutral atom technologies.

Addressing the Gap to Utility-Scale Quantum Computation

The authors elucidate the persistent gap towards achieving utility-scale quantum computation, despite marked progress in quantum hardware, particularly in reducing gate errors across various qubit technologies such as superconducting transmon qubits, neutral atoms, and trapped ions. The enhancement of gate fidelities alone is predicted to require several more years before reaching societally beneficial outcomes. SuperstaQ intervenes through innovative software compilation techniques, acting as a force multiplier to accelerate the realization of practical results in quantum computing. This approach mirrors the advancements in classical computing, where optimized compilation plays a pivotal role in leveraging high-end hardware capabilities.

Theoretical Framework and Implementation

SuperstaQ's superoptimization involves novel strategies in quantum program compilation, focusing on minimizing traditional compilation objectives while incorporating noise-reduction techniques. For instance, Dynamical Decoupling (DD) exemplifies the trade-off between conventional gate optimization and noise suppression, by inserting strategic gate sequences to mitigate qubit decay and dephasing, thereby extending qubit lifetimes significantly. Illustrated experimentally, DD techniques have quadrupled T1 times and doubled T2 coherence times in specific hardware tests.

The paper details SuperstaQ's deployment across diverse platforms, providing notable results in optimized decomposition to native gate sets. Particularly in neutral atom qubit architectures, it reduces two-qubit gate costs by deftly utilizing global rotation gates. Here, the emphasis lies on minimizing the pulse area required for operations, made possible through strategic scheduling and decomposition of single-qubit gates, while ensuring parallel implementation to cut runtimes effectively.

Platforms and Techniques: Empirical Validation

The comparative analysis spans multiple platforms, each evaluating SuperstaQ under different operational and architectural paradigms. Notably:

• On superconducting platforms like AQT, IBM, and Rigetti, SuperstaQ optimizes decompositions to the level of native gatesets, yielding measurable performance gains.
• For neutral atoms, innovative compilation techniques leverage the constraints and capabilities of their unique gatesets, effectively demonstrating significant reductions in gate-time overheads.
• The work with trapped ion systems via QSCOUT reflects advancements in gate decomposition, showcasing SuperstaQ's adaptability across hardware types.

Implications and Future Directions

The findings advocate for a holistic approach to quantum software development, stressing the interplay between compiler design and hardware primitives. The realizations of SuperstaQ herald burgeoning opportunities in quantum optimization, suggesting avenues for similar advancements in emerging quantum technologies. The integration of Dynamical Decoupling with superoptimization underscores future potential in further reducing error rates and enhancing qubit coherence.

Future developments could explore broader applications of superoptimization across diversified qubit technologies and more intricate quantum circuits. Furthermore, refining these techniques in tandem with hardware evolutions could yield even higher efficiency gains, offering tangible prospects in the deployment of scalable, utility-driven quantum computing solutions. The open-source accessibility via cirq-superstaq and qiskit-superstaq enables community engagement and continuous improvement, facilitating shared innovation in the quantum software ecosystem.