Accelerating Classical Simulation with Tensor Network Contraction on Sunway Architecture

In a recent study titled SW-TNC: Reaching the Most Complex Random Quantum Circuit via Tensor Network Contraction, published on April 12, 2025, researchers led by Yaojian Chen and Zhaoqi Sun demonstrated significant advancements in classical simulation of quantum circuits using Sunway architecture. Their approach achieved over tenfold acceleration with more than 1024 nodes, showcasing the potential for efficient simulation of increasingly complex quantum circuits.

The study addresses challenges in classical simulation of complex quantum circuits by proposing novel tensor network contraction strategies optimized for Sunway architecture. Key innovations include data reuse schemes to reduce floating-point operations, memory organization techniques to eliminate slicing overhead, and a step fusion strategy enhanced through multi-core cooperation. Fine-grained optimizations such as vectorized permutations and split-K operators further improve performance. These advancements enable a more than 10x acceleration in simulating the Zuchongzhi-60-24 circuit using over 1024 Sunway nodes, demonstrating progress toward efficient classical simulation of increasingly complex quantum circuits.

Classical simulations serve as a critical bridge in quantum computing research. They allow researchers to verify results from quantum experiments without relying solely on quantum systems, which can be error-prone and difficult to debug. This verification process is crucial for ensuring the reliability and accuracy of quantum computations.

Tensor Networks and Mixed Precision

At the heart of these simulations are tensor network algorithms, which efficiently represent high-dimensional data structures. These networks help manage the complexity inherent in quantum states, making simulations feasible even on classical hardware. Recent advancements have introduced multi-tensor contraction techniques, enhancing computational efficiency by reducing memory usage and optimizing operations.

Additionally, mixed precision algorithms have emerged as a game-changer. By combining different numerical precisions—such as single and double precision—they balance accuracy with computational speed. This approach significantly reduces the time required for simulations while maintaining necessary precision levels, making it possible to handle larger and more complex quantum circuits.

Tailoring for Specific Processors

The effectiveness of these methods is further amplified by hardware-specific optimizations. For instance, Alibaba’s SW26010 processor has been optimized using precise performance modeling to overcome program optimization challenges. This tailored approach ensures that simulations run efficiently on specific architectures, maximizing computational resources and minimizing bottlenecks.

Implications for Quantum Computing Research

The advancements in classical simulation techniques are transformative for quantum computing research. These methods provide a robust foundation for further exploration into quantum technologies by enabling accurate verification of quantum experiments. As both software algorithms and hardware continue to evolve, the synergy between classical simulations and quantum computing will play a crucial role in unlocking new possibilities in computational science.

In summary, integrating tensor networks, mixed precision algorithms, and hardware-specific optimisations is driving significant progress in verifying quantum advantage experiments. These developments not only enhance our ability to validate quantum computations but also pave the way for future innovations in this rapidly advancing field.