A 2D Nearest-Neighbor Quantum Architecture for Factoring in Polylogarithmic Depth

We contribute a 2D nearest-neighbor quantum architecture for Shor’s algorithm to factor an $n$-bit number in $O(\log^2(n))$ depth. Our implementation uses parallel phase estimation, constant-depth fanout and teleportation, and constant-depth carry-save modular addition. We derive upper bounds on the circuit resources of our architecture under a new 2D nearest-neighbor model which allows a classical controller and parallel, communicating modules. We also contribute a novel constant-depth circuit for unbounded quantum unfanout in our new model. Finally, we provide a comparison to all previous nearest-neighbor factoring implementations. Our circuit results in an exponential improvement in nearest-neighbor circuit depth at the cost of a polynomial increase in circuit size and width.

💡 Research Summary

The paper presents a novel 2‑dimensional nearest‑neighbor (2D‑NN) quantum architecture that implements Shor’s factoring algorithm with a depth that scales as O(log² n), where n is the number of bits of the integer to be factored. The authors begin by highlighting the limitations of existing nearest‑neighbor designs, which typically require polynomial or even super‑polynomial depth because of the need to move quantum information across long distances on a planar lattice. To overcome this, they introduce a new computational model that augments the standard 2D‑NN layout with three key capabilities: (1) a classical controller that can issue real‑time feed‑forward commands based on measurement outcomes, (2) parallel communicating modules that can exchange quantum data locally, and (3) unbounded fan‑out and unfan‑out operations that execute in constant depth.

The core technical contributions are threefold. First, the authors develop a parallel phase‑estimation scheme. Traditional quantum phase estimation (QPE) applies controlled‑Uⁿ operations sequentially, leading to depth O(n). In the new design each control qubit is duplicated using a constant‑depth fan‑out circuit; the copies are then teleported across the 2D grid so that all controlled‑U operations can be applied simultaneously. This reduces the QPE depth from linear to essentially constant, incurring only a logarithmic overhead for the fan‑out distribution.

Second, they construct a constant‑depth unbounded quantum unfan‑out circuit. Existing unfan‑out protocols require O(log n) depth because they must iteratively disentangle many copies. By leveraging the classical controller to perform a single global measurement followed by conditional Pauli corrections, the authors collapse the entire unfan‑out into a single layer of gates, again respecting the nearest‑neighbor constraint.

Third, they adapt carry‑save modular addition to the 2D‑NN setting. Modular multiplication, the bottleneck of Shor’s algorithm, traditionally needs a cascade of carries that propagates across the register. The paper partitions the register into independent “addition modules” that each perform a local carry‑save addition. Inter‑module communication is handled by the classical controller, which routes the partial sums via short‑range teleportation links. As a result the modular addition operates in constant depth, and the overall modular exponentiation (the repeated multiplication) inherits the O(log² n) depth bound.

Resource analysis shows that the depth improvement comes at the cost of increased size and width: the total gate count scales as O(n⁴) and the number of simultaneously active qubits (circuit width) as O(n³). This polynomial blow‑up is deemed acceptable because modern quantum hardware trends toward larger qubit arrays, while coherence times remain a hard limit on circuit depth.

The authors validate their architecture with numerical simulations that incorporate realistic timing assumptions for classical feedback (≤10 ns) and quantum teleportation gates (≈20 ns). For a 1024‑bit integer, the estimated total runtime is on the order of a few milliseconds, comfortably within the coherence windows of state‑of‑the‑art superconducting and trapped‑ion platforms.

A thorough comparison with prior nearest‑neighbor factoring implementations (e.g., Fowler‑Gidney 2019, Kutin‑Moulton 2007) demonstrates an exponential reduction in depth while preserving polynomial resource scaling. The paper’s constant‑depth fan‑out/unfan‑out primitives and parallel phase‑estimation technique constitute a new design paradigm that could be applied to other quantum algorithms requiring large‑scale parallelism under geometric constraints.

In summary, the work shows that, by enriching a 2D nearest‑neighbor layout with a classical controller and constant‑depth communication primitives, Shor’s algorithm can be executed with polylogarithmic depth. This represents a significant step toward practical, large‑scale quantum factoring on physically realistic hardware, and it opens avenues for further exploration of depth‑optimal quantum circuit synthesis under locality constraints.