Emuplat: Framework-Agnostic Quantum Hardware Emulation with Validated Four-Stage Transpiler-to-Pulse Pipeline

Quantum computing promises revolutionary advances, but bridging the gap between software and the complex hardware remains a significant challenge. Jun Ye and Jun Yong Khoo, from the Institute of High Performance Computing (A*STAR), alongside their colleagues, now present EmuPlat, a new platform designed to overcome this interoperability issue. EmuPlat uniquely functions as a framework-agnostic emulator, enabling seamless integration across diverse quantum computing ecosystems and implementing a validated pipeline that transforms abstract quantum circuits into precise hardware instructions. This achievement delivers a crucial infrastructure component, demonstrated through benchmarks achieving exceptionally high fidelity in key quantum operations and paving the way for accelerated development of hybrid quantum-classical algorithms and improved hardware-software co-design.

Scalable Quantum Emulation With Realistic Noise

This research details Emu, a new quantum computing emulator designed for flexibility and scalability. Quantum computers are rapidly developing, but access to physical hardware remains limited and expensive, making emulators crucial for algorithm development, testing, and education. Emu addresses these limitations by accurately simulating the noise present in real quantum devices, significantly impacting algorithm performance. Emu’s modular architecture allows researchers to easily swap out components, such as different noise models or simulation engines, without major code changes, enabling the addition of new features and optimizations.

The emulator handles a larger number of qubits than many existing systems and incorporates a comprehensive set of noise models, including T1 and T2 decoherence, simulating energy relaxation and dephasing in superconducting qubits. It also models crosstalk between qubits, imperfections in quantum gate operations, and inaccuracies in control pulses. Emu operates at the pulse level, simulating the actual control signals applied to the qubits, allowing for more accurate modeling of noise and control imperfections. It works with popular quantum programming frameworks like Qiskit and Cirq. The core of Emu is a state vector simulator, applying noise after each gate operation, mimicking real quantum hardware.

Validation involved comparing Emu’s results to those obtained from real quantum hardware and other emulators. This work delivers a modular and scalable quantum computing emulator with a comprehensive set of noise models and pulse-level simulation capabilities. Emu’s integration with existing quantum programming frameworks makes it a valuable tool for researchers, developers, and educators, enabling more accurate algorithm testing, performance analysis, and the development of noise mitigation techniques. Emu provides a powerful and flexible platform for simulating quantum computers, bridging the gap between theoretical algorithms and the realities of noisy quantum hardware.

Unified Quantum Hardware Emulation

EmuPlat, a new quantum hardware emulation platform, addresses a critical gap in interoperability between high-level programming frameworks and hardware-specific control systems. This work introduces a unified infrastructure enabling seamless integration across diverse quantum computing ecosystems, including CUDA-Q and Qibolab. The platform implements a complete transpiler-compiler pipeline systematically transforming abstract quantum circuits through four validated stages, optimizing for hardware constraints at each step. Comprehensive benchmarks on superconducting transmon architectures demonstrate EmuPlat’s capabilities.

Specifically, the team achieved 99. 958% fidelity in Bell state preparation using hardware-calibrated noise models, demonstrating precise control and accurate simulation of quantum states. Furthermore, successful implementation of a 4-qubit Quantum Fourier Transform demonstrates the platform’s ability to handle scalable circuit execution, a crucial step towards practical quantum computation. The compiler generates optimized pulse sequences, reducing pulse count by 30-50% through the implementation of virtual Z operations, which accumulate phase offsets without requiring physical pulses. Validation experiments, utilizing parameters from Anyon Technologies’ superconducting quantum processors, confirm the platform’s accuracy.

These experiments, featuring transmon qubits with coherence times of T1 ≈24μs and T2 ≈33μs, demonstrate the platform’s ability to accurately simulate quantum dynamics under realistic noise conditions. The team’s comprehensive validation framework, TransformationValidator, verifies correctness and performance at each stage of the transpiler-compiler-simulation pipeline, ensuring mathematical equivalence and quantifying the impact of hardware-specific optimizations. Fidelity calculations confirm the accuracy of simulated entangled state preparation. The platform’s production-ready implementation establishes EmuPlat as essential infrastructure for accelerating hybrid quantum-classical algorithm development and hardware-software co-design.