Quantum Prometheus: Defying Overhead with Recycled Ancillas in Quantum Error Correction

Archisman Ghosh; Avimita Chatterjee; Swaroop Ghosh

REVIEW 4 major objections 4 minor 2 cited by

Quantum Prometheus: Defying Overhead with Recycled Ancillas in Quantum Error Correction

T0 review · 4 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Reusing one ancilla set for both X- and Z-stabilizer measurements cuts rotated surface-code qubit overhead by about 25%.

desk verdict A clean resource-counting idea undermined by a missing connectivity layout, a likely unfair noise model, and a wrong inequality solution. read the letter →

arxiv 2411.12813 v2 pith:MRQ3ZSRJ submitted 2024-11-19 quant-ph

Avimita Chatterjee , Archisman Ghosh , Swaroop Ghosh This is my paper · ORCID

classification quant-ph PACS 03.67.Pp

keywords quantumerrorcorrectionsurfacecodeancillareusesyndromemeasurementqubitoverheadlogicalratestabilizercircuits

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to reduce the qubit overhead of rotated surface codes by reusing the same ancilla qubits for both X- and Z-stabilizer measurements. The trick is to split each error-correction round into two sub-rounds, measuring X-stabilizers first and Z-stabilizers second on the same ancillas after a reset. That halves the number of ancilla qubits and cuts total qubit count by roughly 25%. The authors argue this preserves error-correcting ability: logical error rates stay close to the original code, and for distances d ≥ 13 a distance d+2 modified code uses the same or fewer qubits than a distance d original code while achieving a lower logical error rate.

What carries the argument

Ancilla recycling by alternating sub-rounds: each full round becomes two half-rounds, the first measuring only X-stabilizers and the second only Z-stabilizers on the same physical ancillas, with a reset in between. This replaces the standard simultaneous X/Z measurement that needs dedicated ancillas per stabilizer. The counting identity Q_modified(d) = $d^{2}$ + ($d^{2}$−1)/2 versus Q_original(d) = $2d^{2}$−1 drives the claimed 25% total-qubit saving, and the cross-over condition Q_modified(d+2) ≤ Q_original(d) gives d ≥ 13.

What would settle it

Try to construct an explicit stabilizer measurement circuit for a distance-3 rotated surface code where each ancilla couples to the data qubits of both an X- and a Z-stabilizer using only nearest-neighbor connectivity; if this requires swaps or extra routing qubits that exceed the saved ancillas, the 25% saving disappears. Alternatively, simulate the modified code with noise that includes slow mid-circuit resets and time-correlated errors, and check whether the logical error rate penalty exceeds the threshold advantage claimed.

Watch

Extended reading notes

Core claim

Reusing ancilla qubits across X- and Z-stabilizer measurements via time-division multiplexing reduces the ancilla count from $d^{2}$−1 to about ($d^{2}$−1)/2, lowering the total qubit count from $2d^{2}$−1 to about ($3d^{2}$−1)/2. The paper claims the modified code maintains nearly identical thresholds (within about 2% across four error models) and logical error rates that are only slightly higher than the original, with the relative penalty depending on error type. It further claims that for d ≥ 13, upgrading to a modified code of distance d+2 requires the same or fewer qubits than the original distance-d code while delivering a lower logical error rate, so the resource saving converts into better error suppression.

Load-bearing premise

One physical ancilla can be reset and then couple to the data qubits of both an X-stabilizer and a Z-stabilizer without additional overhead, yet the paper does not provide a circuit layout or connectivity graph showing this is actually realizable.

Editorial extensions

If this is right

For any rotated surface code distance d, the ancilla count is halved and the total qubit count drops by approximately 25%.
At distances d ≥ 13, a modified code of distance d+2 fits within the same qubit budget as the original distance-d code and yields a lower logical error rate.
The error-correction threshold is essentially preserved, with a mean difference of about 1.86% across the error models tested.
The approach is presented as generalizable to other quantum error-correcting codes that use rounds and stabilizer circuits, not just surface codes.
Under a fixed qubit budget, the code distance can be extended by up to roughly k ≤ (2√3 − 1)d while staying within the original number of qubits.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The temporal separation of X and Z measurements likely doubles the syndrome-extraction time per round, which could affect error rates under time-dependent noise; the paper's noise model does not directly capture this delay.
The central saving depends on a single ancilla physically coupling to the data qubits of both an X-stabilizer and a Z-stabilizer; without a concrete circuit layout or connectivity graph, a hardware implementation might need extra routing or swap layers that erode the 25% gain.
For small distances (d < 13), the logical error rate penalty is relatively larger, so the resource saving may not justify the performance loss unless qubit count is the dominant constraint.
Combining this ancilla-recycling scheme with other overhead-reduction techniques such as flag qubits or multiplexed readout could push total qubit counts even lower, though the paper does not explore that combination.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The paper's core trick—time-multiplexing X and Z stabilizer measurements with shared ancillas—is a natural resource-saving idea, and the qubit-counting arithmetic is straightforward and correct. The 25% total-qubit reduction is real if the layout exists, and the crossover at d=13 for d+2 modified vs d original is a neat observation. That part deserves credit.

The problems start when you ask how this is actually implemented. The paper never shows a circuit or connectivity graph where one ancilla couples to both an X-stabilizer's data qubits and a Z-stabilizer's data qubits. In a rotated surface code those stabilizers live on different plaquettes, and the standard layout puts X and Z ancillas at disjoint positions. Without a concrete layout or a routing plan, the central 25% saving is unverified. The stress-test note is right: this is load-bearing, and the STIM simulations cannot rescue it because no circuits are included.

The simulation comparison is also suspect. The depolarizing noise model applies errors before every second sub-round in the modified code, which looks like it undercounts errors relative to the original, where errors are applied each round. The paper's own Table II shows the modified code has logical error rates 5–44% higher across error types. The abstract's "similar or better" claim only survives via an undocumented extrapolation to d+2, with no simulations at d≥13.

There is also a clear mathematical error in the k-bound. Solving Q_modified,d+k ≤ Q_original,d gives k ≤ (2/√3 − 1)d ≈ 0.1547d, not (2√3 − 1)d ≈ 2.464d. The claimed bound is off by more than a factor of ten, and the table's k=4 for d=27 satisfies the correct bound but does not validate the stated one.

So the paper is a mix: a sensible resource-accounting idea with correct counting in the simple cases, but the performance claim is not supported by the evidence, the layout is missing, and the k-bound is wrong. The flaws are specific and correctable in principle, but as written the central promise does not hold up.

Who is this for? Researchers thinking about QEC resource trade-offs might find the time-multiplexing framing worth discussing, and the d+2 crossover is a useful sanity check. But the paper needs a full rewrite, with a real circuit layout, a fair noise comparison, and a corrected bound.

If I were the editor, I would send it to peer review, not because it is ready, but because the underlying idea is not crank-ish and the errors are concrete enough that a referee could guide the authors toward a testable version. Desk rejection feels slightly harsh for a paper with a clear and correct qubit-counting core, even if the main claim currently fails.

Referee Report

4 major / 4 minor

Summary. The paper proposes a modified rotated surface code in which the same ancilla qubits are reused for X- and Z-stabilizer measurements by splitting each error-correction round into two sub-rounds. It claims this reduces total qubit overhead by about 25%, and that for distances d ≥ 13 a distance d+2 modified code uses the same or fewer qubits than a distance d original code while achieving a lower logical error rate. The paper provides qubit-count formulas, STIM simulations for distances 3, 5, 7, and 9, and extrapolated logical error rates for larger distances.

Significance. If the ancilla-reuse scheme were physically realizable without additional overhead, a 25% reduction in qubit count for surface codes would be a useful contribution. The qubit-count arithmetic (e.g., the expressions for Q_total and ΔQ) is correct and clearly presented, and the idea of time-multiplexing ancillas is worth exploring. However, the central physical feasibility claim is not established, the simulation comparison appears to under-count errors in the modified scheme, and the key high-distance conclusions are extrapolations without a described method. The paper ships no circuits or code, so the STIM results are not reproducible as reported.

major comments (4)

[Sec. III-A, Fig. 1(e)] The entire 25% qubit-overhead reduction rests on the assumption that a single physical ancilla can measure both an X-stabilizer and a Z-stabilizer in alternating sub-rounds. However, the paper does not provide a circuit-level implementation, a connectivity graph, or a routing plan showing how the same ancilla couples to the data qubits of two different stabilizer types. In the rotated surface-code layout described in Sec. II, X- and Z-stabilizers occupy disjoint plaquettes and each ancilla is dedicated to a single plaquette; reusing it for the other stabilizer type would require either long-range couplings or SWAP-induced movement, whose qubit and time overhead must be counted against the claimed saving. Without that, the claimed reduction is unverified.
[Sec. IV-B] The noise model is asymmetric between the two codes. The text states that depolarizing error is applied 'before every round' as well as 'after every Clifford gate' in the original code, but for the modified code it only specifies application 'before every second sub-round.' If the modified code is not given the same per-gate depolarizing noise on all gates in both sub-rounds, its error rate is underestimated, and the comparison in Fig. 3 and Table II becomes unfair. The authors should provide the exact noise-injection schedules for both codes and ensure that the per-unit-time physical error rate is identical.
[Sec. IV-C, Table I, Fig. 4] The logical error rates in Table I for original distances up to d=53 are not simulated; the simulations in Fig. 3 cover only d=3,5,7,9. The paper states that Fig. 4 'projects' the rates to higher distances but does not describe the extrapolation model, its assumptions, or its uncertainty. The central claim that a d+2 modified code has a lower logical error rate than a d original code at the same qubit budget for d≥13 depends entirely on these unvalidated extrapolations and is not a direct simulation result.
[Table II] The paper's own performance data show that the modified code has mean relative logical error ratios that are all greater than 1, ranging from 1.056 (readout/reset) to 1.435 (gate), i.e., 5% to 44% higher logical error rates at the same distance. The text dismisses these as having 'low practical difference' because the absolute rates are small, but for QEC resource estimation the multiplicative penalty is directly relevant to the number of physical qubits needed to reach a target logical error rate. The assertion that the modified code offers 'similar error correction' should be quantified against a target logical error rate and physical error rate rather than asserted.

minor comments (4)

[Sec. III-B, Case B] The example for d=3 says that directly implementing a distance-5 original surface code 'would require 32 qubits instead of 37,' but the paper's own formula Q_original = 2d^2 - 1 gives 49 qubits for d=5; the correct saving compared to 37 qubits is 12 qubits, not 5.
[Sec. IV-A] The sentence 'we used the formula (distance × 3) + 1to determine the number of rounds' has a missing space before 'to,' and the choice of this formula is not justified.
[Abstract] The sentence 'This technique can be applied broadly across various QEC codes, we focus on rotated surface codes only' is a comma splice; it should be broken into two sentences.
[General] The STIM circuits used for the simulations are not included in the paper or supplementary material; given that the paper's comparisons are the main evidence, the circuits should be made available for reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: qubit savings follow from stabilizer-count arithmetic; logical error rates come from independent STIM simulations.

full rationale

The paper's central derivation is a stabilizer-counting argument. Original total Q_total=2d^2-1 follows from d^2 data qubits plus d^2-1 stabilizer ancillas; modified total Q_modified=d^2+(d^2-1)/2 follows from the proposed time-multiplexed X/Z reuse, which by construction needs only the number of stabilizers of one type, (d^2-1)/2, assuming the reuse is physically routable. The d+2 transition formula ΔQ=(d+1)(13-d)/2 and the k-bound are pure algebra from these two count formulas. Logical error rates are generated with the external STIM simulator (referenced [7]) and compared against the original code; they are not fitted to the paper's conclusion. The paper contains no self-citations, no imported uniqueness theorem, and no renaming of an existing result. Two non-circular caveats exist: (1) Sec. III-A assumes without a connectivity graph that one ancilla can couple to both an X- and a Z-stabilizer, so the 25% saving is conditional on physical routing; (2) Sec. III-B.3's printed bound k≤(2√3−1)d appears algebraically inconsistent with the preceding formulas (the correct large-d coefficient is 2/√3−1≈0.1547, not 2√3−1≈2.464). These are feasibility/correctness issues, not circularity.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central claim rests on an unverified connectivity assumption (that one ancilla can serve two stabilizers) and on extrapolated logical error rates. The free parameters are simulation choices, not fitted to target results. No new physical entities are postulated.

free parameters (2)

Simulation rounds multiplier = 3d + 1
The authors choose the number of rounds as 3 times distance plus 1 (Section IV-A); this is not derived and affects the logical error rate estimates.
Table I physical error rate = 10^-3
A single baseline physical error rate chosen for the d vs d+k comparison; results are not shown for other error rates.

assumptions (5)

standard math Rotated surface code of distance d has d^2 data qubits and d^2 - 1 stabilizer ancillas, with equal numbers of X and Z stabilizers.
Used in Section II to derive Q_total = 2d^2 - 1 and Q_modified = (3d^2 - 1)/2.
ad hoc to paper A single physical ancilla qubit can be reset and reused to measure both an X-stabilizer and a Z-stabilizer, with no additional connectivity or routing overhead.
Assumed in Section III-A (purple circles in Fig 1e) but no circuit layout or connectivity graph is provided; this is the core premise enabling the halved ancilla count.
domain assumption Applying depolarizing noise before every second sub-round in the modified code gives an even comparison with the original code, which receives noise before every round and after every Clifford gate.
Stated in Section IV-B1; if the noise schedules are not matched, the simulated logical error rate comparison is biased.
domain assumption Minimum-weight perfect matching decoding remains valid for the modified code despite temporal correlations between X and Z sub-rounds.
Used in Section IV-A for all simulations; the paper acknowledges temporal correlations but does not validate the decoder's applicability to the modified schedule.
ad hoc to paper The logical error rates for code distances above 9 can be extrapolated from the simulated data to produce Fig 4 and Table I.
Section IV-C3 states that Fig 4 'projects' higher-distance behavior based on previous experiments, but the extrapolation model is not described.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Quantum Prometheus: Defying Overhead with Recycled Ancillas in Quantum Error Correction." pith.science (2026). https://pith.science/paper/MRQ3ZSRJ

@misc{pith2026241112813,
  author       = {Pith},
  title        = {Pith review of: Quantum Prometheus: Defying Overhead with Recycled Ancillas in Quantum Error Correction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MRQ3ZSRJ}},
  note         = {Machine review of arXiv:2411.12813}
}

abstract

Quantum error correction (QEC) is crucial for ensuring the reliability of quantum computers. However, implementing QEC often requires a significant number of qubits, leading to substantial overhead. One of the major challenges in quantum computing is reducing this overhead, especially since QEC codes depend heavily on ancilla qubits for stabilizer measurements. In this work, we propose reducing the number of ancilla qubits by reusing the same ancilla qubits for both X- and Z-type stabilizers. This is achieved by alternating between X and Z stabilizer measurements during each half-round, cutting the number of required ancilla qubits in half. This technique can be applied broadly across various QEC codes, we focus on rotated surface codes only and achieve nearly $25\%$ reduction in total qubit overhead. We also present a few use cases where the proposed idea enables the usage of higher-distance surface codes at a relatively lesser qubit count. Our analysis shows that the modified approach enables users to achieve similar or better error correction with fewer qubits, especially for higher distances ($d \geq 13$). Additionally, we identify conditions where the modified code allows for extended distances ($d + k$) while using the same or fewer resources as the original, offering a scalable and practical solution for quantum error correction. These findings emphasize the modified surface code's potential to optimize qubit usage in resource-constrained quantum systems.

Figures

Figures reproduced from arXiv: 2411.12813 by the authors.

**Figure 1.** Original Approaches in the Rotated Surface Code and the Proposed Modification Strategy (a) A distance-3 and distance-5 rotated surface code layout, demonstrating the scalable structure of the code that enhances fault tolerance with increased distance. The layout shows how expanding the code distance, by adding additional qubits, increases robustness against errors. (b) Example of Z- and X-stabilizer circuits with an… view at source ↗

**Figure 2.** Comparison of Qubit Overhead and Ratios in Original and Modified Surface Code Approaches (a) compares ancilla and total qubit requirements in the original and new approaches, showing that while both increase with code distance, the new approach achieves a marked reduction in ancilla qubits, leading to a lower total qubit count. This demonstrates the effectiveness of recycling ancillas in reducing overhead, particula… view at source ↗

**Figure 3.** Analyzing Logical Error Rates Across Physical Error Rates for Original and Modified Surface Codes To compare the original and modified approaches to surface codes, this figure illustrates the relationship between logical error rate and physical error rate for four distinct types of errors—depolarizing, gate, readout/reset, and combined errors—across original and modified surface codes of distances 3, 5, 7, and 9. (a… view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Logical Error Rate vs. Code Distance and Number of Qubits for Original and Modified Surface Codes (a) projects how the logical error rate decreases with increasing code distance for both original and modified surface codes under three different physical error rates, co…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Feynman's clock and hierarchy-informed sampling for quantum error mitigation
quant-ph 2026-07 conditional novelty 6.0 of 10

Feynman's clock maps arbitrary circuits onto Hamiltonian dynamics whose BBGKY hierarchy enables polynomial-overhead, controllable error mitigation via informed sampling.
Optimization of Quantum Error Correcting Code under Temporal Variation of Qubit Quality
quant-ph 2025-05 conditional novelty 4.0 of 10

A daily, per-qubit choice of surface code distance, driven by calibration error rates, can cut QEC qubit overhead by more than half while retaining most qubits.

Reference graph

Works this paper leans on

12 extracted references · 10 canonical work pages · cited by 2 Pith papers

[1]

Entanglement instability in the interaction of two qubits with a common non-markovian environment,

G. Mouloudakis et al. , “Entanglement instability in the interaction of two qubits with a common non-markovian environment,” Quantum Information Processing, vol. 20, pp. 1–15, 2021

work page 2021
[2]

The no-cloning theorem,

W. K. Wootters et al., “The no-cloning theorem,” Physics Today, vol. 62, no. 2, pp. 76–77, 2009

work page 2009
[3]

V on Neumann, Mathematical foundations of quantum mechanics: New edition

J. V on Neumann, Mathematical foundations of quantum mechanics: New edition. Princeton university press, 2018, vol. 53

work page 2018
[4]

Topological quantum memory,

E. Dennis et al., “Topological quantum memory,” Journal of Mathemat- ical Physics, vol. 43, no. 9, pp. 4452–4505, 2002

work page 2002
[5]

Low-distance surface codes under realistic quantum noise,

Y . Tomita et al. , “Low-distance surface codes under realistic quantum noise,” Physical Review A , vol. 90, no. 6, p. 062320, 2014

work page 2014
[6]

Surface codes: Towards practical large-scale quantum computation,

A. G. Fowler et al. , “Surface codes: Towards practical large-scale quantum computation,” Physical Review A , vol. 86, no. 3, p. 032324, 2012

work page 2012
[7]

Stim: a fast stabilizer circuit simulator,

C. Gidney, “Stim: a fast stabilizer circuit simulator,” Quantum, vol. 5, p. 497, 2021

work page 2021
[8]

Improved simulation of stabilizer circuits,

S. Aaronson and D. Gottesman, “Improved simulation of stabilizer circuits,” Physical Review A , vol. 70, no. 5, p. 052328, 2004

2004

Show all 12 references

[9]

Realizing repeated quantum error correction in a distance-three surface code,

S. Krinner et al. , “Realizing repeated quantum error correction in a distance-three surface code,” Nature, vol. 605, no. 7911, pp. 669–674, 2022

2022
[10]

Quantum error correction below the surface code threshold,

R. Acharya et al. , “Quantum error correction below the surface code threshold,” arXiv preprint arXiv:2408.13687 , 2024

2024 arXiv
[11]

Blossom v: a new implementation of a minimum cost perfect matching algorithm,

V . Kolmogorov, “Blossom v: a new implementation of a minimum cost perfect matching algorithm,” Mathematical Programming Computation, vol. 1, pp. 43–67, 2009

2009
[12]

Pymatching: A python package for decoding quantum codes with minimum-weight perfect matching,

O. Higgott, “Pymatching: A python package for decoding quantum codes with minimum-weight perfect matching,” ACM Transactions on Quantum Computing, vol. 3, no. 3, pp. 1–16, 2022

2022

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Entanglement instability in the interaction of two qubits with a common non-markovian environment,

G. Mouloudakis et al. , “Entanglement instability in the interaction of two qubits with a common non-markovian environment,” Quantum Information Processing, vol. 20, pp. 1–15, 2021

work page 2021

[2] [2]

The no-cloning theorem,

W. K. Wootters et al., “The no-cloning theorem,” Physics Today, vol. 62, no. 2, pp. 76–77, 2009

work page 2009

[3] [3]

V on Neumann, Mathematical foundations of quantum mechanics: New edition

J. V on Neumann, Mathematical foundations of quantum mechanics: New edition. Princeton university press, 2018, vol. 53

work page 2018

[4] [4]

Topological quantum memory,

E. Dennis et al., “Topological quantum memory,” Journal of Mathemat- ical Physics, vol. 43, no. 9, pp. 4452–4505, 2002

work page 2002

[5] [5]

Low-distance surface codes under realistic quantum noise,

Y . Tomita et al. , “Low-distance surface codes under realistic quantum noise,” Physical Review A , vol. 90, no. 6, p. 062320, 2014

work page 2014

[6] [6]

Surface codes: Towards practical large-scale quantum computation,

A. G. Fowler et al. , “Surface codes: Towards practical large-scale quantum computation,” Physical Review A , vol. 86, no. 3, p. 032324, 2012

work page 2012

[7] [7]

Stim: a fast stabilizer circuit simulator,

C. Gidney, “Stim: a fast stabilizer circuit simulator,” Quantum, vol. 5, p. 497, 2021

work page 2021

[8] [8]

Improved simulation of stabilizer circuits,

S. Aaronson and D. Gottesman, “Improved simulation of stabilizer circuits,” Physical Review A , vol. 70, no. 5, p. 052328, 2004

2004