Ultrametric Quantum Computation for Direct Integer Factorization: A Resource Analysis with a Largely Negative Verdict

In the paper's words

Factoring large integers underpins RSA, and the best-known polynomial-time quantum method, Shor's algorithm, extracts the period of modular exponentiation through a quantum phase-estimation detour that dominates circuit depth. A conjecture in the ultrametric research program holds that a quantum computer whose native state space is p-adic rather than Archimedean could access that period directly, bypassing phase estimation. This paper formalizes the conjecture as two concrete register models — a tensor-product adelic register and a Bruhat–Tits tree register — and costs every component exactly.

The abstract, verbatim, and its opening sentence is the claim. Every line below is quoted from the text.

Numbers it reports

1. The Tianyan cloud platform operates superconducting processors with 105 qubits at single-qubit, two-qubit, and readout fidelities of $99.90\%$, $99.56\%$, and $98.7\%$ respectively, and demonstrates random circuit sampling on a 74-qubit region as a demonstration of quantum-advantage capability [2]; Devices of this class remain in the noisy intermediate-scale regime, and the depth of phase estimation is one of the dominant obstacles.

   from 1. Introduction

2. Hardware context bounds what any new register model must eventually map onto. [2] describes the Tianyan cloud platform operating superconducting processors with 105 qubits at the fidelities quoted above, and demonstrates random circuit sampling on a 74-qubit region as a demonstration of quantum-advantage capability.

   from 2. Background and Related Work

3. For the tree register at a $b$-bit modulus with $p=2$ and $d=2b$, the register has $3 \cdot 2^{2b}-2$ vertices and the transform costs $3(3 \cdot 2^{2b}-2)$ operations, against $b(2b+1)$ quantum Fourier transform gates in Shor's control register: the ratio is $4336.9$ at $b=8$ and approximately $10^{1227}$ at $b=2048$.

   from Abstract

4. The ultrametric-quantum-computation program proposes Bruhat–Tits trees as a physical substrate, Hecke operators as gates, and a threshold theorem for fault tolerance on the tree [10].

   from 1. Introduction

5. An exact resource comparison showing that direct $p$-adic period-finding on the tree register for a $b$-bit modulus costs $3(3 \cdot 2^{2b}-2)$ transform operations against $b(2b+1)$ QFT gates in Shor's control register, a ratio of $4336.9$ at $b=8$ and approximately $10^{1227}$ at $b=2048$ (Sections 4 and 5).

   from 1. Introduction

6. A full reproduction of the toy adelic register arithmetic — dimension $810{,}000$, structured cost $2{,}888$, dense baseline $6.561 \times 10^{11}$, separation $2.27 \times 10^{8}$ — with the dense baseline re-derived as a dense matrix-vector product of size $\dim^2$, and with the structural-versus-physical caveat made explicit (Sections 4 and 5).

   from 1. Introduction

How it is done

1. Factoring large integers underpins RSA, and the best-known polynomial-time quantum method, Shor's algorithm, extracts the period of modular exponentiation through a quantum phase-estimation detour that dominates circuit depth.

   from Abstract

2. The canonical polynomial-time quantum method, Shor's algorithm, reduces factoring to period-finding — finding the multiplicative order $r$ of a unit $a$ modulo $N$ — and executes it by preparing a superposition over exponents, computing $a^z \bmod N$ into a second register, and applying the quantum Fourier transform (QFT) over roughly $N^2$ points to estimate an eigenphase from which $r$ is recovered by continued fractions.

   from 1. Introduction

3. The mathematical foundation for taking ultrametric state spaces seriously is well developed. [1] proves ultrametric analogues of three classical metric-geometry results: the Arens–Eells isometric embedding theorem, the Hausdorff extension theorem, and the Niemytzki–Tychonoff characterization of compactness, working over ultra-normed modules over arbitrary commutative rings and replacing the Dugundji extension theorem with a zero-dimensional Michael continuous selection theorem.

   from 2. Background and Related Work

4. Second, a structured transform on a tensor factor of dimension $q = p^{k_p}$ implemented by a radix-$p$ Cooley–Tukey algorithm costs $q \log_p q = k_p\, q$ operations.

   from 3.4 Costing conventions

What would falsify it

1. Falsification criteria and a minimal de-risking pathway, combining a classical exact-cost calculation with a three-qubit $p$-adic phase-gate prototype (Sections 6 and 7).

   from 1. Introduction

2. We now show this vector cannot determine the residue class modulo $r$, by explicit counterexample.

   from 4.8 Structural obstruction: valuations do not determine residues modulo the period

3. Falsification of the main result.

   from 6. Discussion

Where the claim stops

1. What the result does not rule out.

   from 6. Discussion

2. The costing assumes the tree transform must be applied at every vertex and costs $p+1$ operations per vertex.

   from 6. Discussion

3. The result rests on the following assumptions, with sensitivity noted for each.

   from 6. Discussion

Where to read

1. Abstract222 words · 1 min
2. 1. Introduction657 words · 3 min
3. 2. Background and Related Work1,075 words · 5 min
4. 3. Methods534 words · 2 min
5. 4. Analysis1,424 words · 6 min
6. 5. Results253 words · 1 min
7. 6. Discussion632 words · 3 min
8. 7. Practical Implications and Engineering Pathways672 words · 3 min
9. 8. Conclusion254 words · 1 min
10. Provenance and AI Involvement225 words · 1 min