Quantum Computers Gain Efficiency with 10% Fewer Operation Steps

Researchers Bruno Avritzer and Nathan Sankary have developed a new method for reducing the complexity of quantum computations across multiple processors. The technique allocates qubits within a colour code family, achieving a 10% reduction in processor-nonlocal gates and offering potential for greater advantages as qubit numbers increase. It minimises operations susceptible to noise between quantum processors, addressing a key challenge in distributed quantum computing, and explores efficient strategies for universal gate sets including magic state distillation, code switching, and logical swaps. These advancements represent a strong step towards scalable and effective error correction in modular quantum architectures

Distributed logical qubits enable sharp reductions in inter-processor communication complexity

A ten percent reduction in processor-nonlocal gates, the connections between processing units in quantum computers, is now possible, a feat previously unattainable due to the challenges of minimising inter-processor communication. Scaling quantum computations beyond a limited size was previously hampered by the escalating complexity of these connections and the associated error rates. Quantum computers, unlike their classical counterparts, are profoundly susceptible to errors arising from environmental noise and imperfections in quantum control. As the number of qubits increases, so too does the probability of encountering these errors, making error correction paramount for reliable computation. Strategically distributing logical qubits, the fundamental units of quantum information encoded using multiple physical qubits for error protection, across multiple processors, rather than confining them to a single unit, offers benefits that increase alongside the number of qubits. This approach shifts the focus from managing error rates within a single, large processor to mitigating the errors introduced between processors during communication.

More efficient methods for universal gate sets, including magic state distillation and code switching, are now enabled by this approach, opening avenues for more scalable and robust quantum computations. Distributing logical qubits across multiple processors reduces the number of processor-nonlocal gates required; a ten percent reduction was achieved using a colour code family, and this benefit scales with increasing qubit numbers. The improvement stems from the relationship between the number of nonlocal gates and code distance, where transversal gate execution grows quadratically while gates resulting from stabilizer splits grow linearly, a trend validated through constraint programming simulations. The ‘code distance’ refers to the error-correcting capability of the quantum code; a higher code distance allows for the detection and correction of more errors. Transversal gates operate directly on the encoded logical qubits without needing to decode them, offering a significant advantage in terms of speed and reduced error propagation. Stabilizer splits, however, require decoding and recoding, introducing additional complexity and potential for errors. The use of constraint programming simulations allowed the researchers to systematically explore different qubit allocation strategies and verify the observed scaling behaviour. Furthermore, the team explored universal gate sets, finding that techniques like magic state distillation and code switching become more efficient in this distributed setting, with code switching offering particularly low overhead by using transversal CNOT gates on a 3D code. Magic state distillation is a process used to create high-fidelity ancillary states required for certain quantum gates, while code switching involves transitioning between different quantum error-correcting codes to optimise performance.

Strategic qubit placement lowers error-inducing communication in quantum processors

Modular quantum computers promise scalability through interconnection, but realising this potential demands minimising communication between processing units. Strategic allocation of qubits across a system makes a ten percent reduction in these essential, yet error-prone, connections, processor-nonlocal gates, possible. This improvement, however, relies on performing syndrome extraction after every logical gate, a technique dictated by current hardware limitations. Syndrome extraction is the process of measuring the error syndrome, information about the errors that have occurred, without disturbing the quantum state itself. This is crucial for error correction, but it adds overhead to the computation. The necessity of syndrome extraction after each gate highlights the current limitations of quantum hardware, where error rates are still relatively high and frequent error correction is essential.

Even modest gains in reducing the demand on inter-processor communication are vital for building larger, more stable quantum computers, paving the way for substantial improvements as technology advances. The current generation of quantum processors are limited in the number of qubits they can reliably control. Modular architectures offer a pathway to overcome this limitation by connecting multiple smaller processors together. However, the communication between these processors introduces new sources of error. Reducing the number of processor-nonlocal gates directly addresses this challenge. This work establishes foundational techniques for qubit allocation, applicable beyond the specific colour code family tested, and offers a clear path towards minimising errors in future modular systems. The colour code is a specific type of quantum error-correcting code known for its relatively high threshold for error tolerance. However, the principles behind the qubit allocation techniques developed in this study can be generalised to other colour codes and potentially even to different types of quantum codes. Careful allocation of qubits, the fundamental units of quantum information, across multiple processors is required to minimise errors in distributed quantum computing. These techniques reduce processor-nonlocal gates, connections between processing units that introduce opportunities for error, and achieving a ten percent reduction demonstrates a pathway to more scalable systems. Current hardware necessitates frequent error checks via syndrome extraction, meaning this improvement isn’t simply a numerical gain, but addresses a core challenge in building larger quantum computers by lessening the burden on individual processors. The reduction in processor-nonlocal gates translates to a reduction in the number of times data needs to be transferred between processors, thereby reducing the overall error rate and improving the reliability of the computation. This is particularly important as quantum computers move towards tackling more complex and computationally intensive problems.

The researchers demonstrated a ten percent reduction in processor-nonlocal gates through optimised qubit allocation within a modular quantum computing architecture. This matters because minimising these gates, connections between processors, directly reduces errors in quantum computations, a key challenge in scaling up these systems. The techniques developed are applicable to colour codes and potentially other quantum codes, offering a versatile approach to error mitigation. This work establishes foundational methods for distributing quantum information and improving the reliability of multi-qubit operations.