Ancilla reuse in repeated syndrome extraction couples reset quality to logical-cycle latency. We evaluate blind reset -- unitary-only recycling via scaled sequence replay -- on IQM Garnet, Rigetti Ankaa-3, and IonQ under matched seeds, sequence lengths, and shot budgets. Using ancilla cleanliness F_clean=P(|0>), per-cycle latency, and a distance-3 repetition-code logical-error proxy, platform-calibrated simulation identifies candidate regions where blind reset cuts cycle latency by up to 38x under NVQLink-class feedback overhead while maintaining F_clean >= 0.86 for L <= 6. Hardware experiments on IQM Garnet confirm blind-reset cleanliness >= 0.84 at L=8 (1024 shots, seed 42); platform-calibrated simulation for Rigetti Ankaa-3 predicts comparable performance. Architecture-dependent crossover lengths are L* ~ 12 (IQM), ~ 11 (Rigetti), ~ 1 (IonQ), and ~ 78 with GPU-linked external feedback. Two added analyses tighten deployment boundaries: a T1/T2 sensitivity map identifies coherence-ratio regimes, and error-bound validation confirms measured cleanliness remains consistent with the predicted diagnostic envelope. A deployment decision matrix translates these results into backend-specific policy selection.
Ancilla reuse in repeated syndrome extraction couples reset quality to logical-cycle latency. We evaluate blind reset -- unitary-only recycling via scaled sequence replay -- on IQM Garnet, Rigetti Ankaa-3, and IonQ under matched seeds, sequence lengths, and shot budgets. Using ancilla cleanliness F_clean=P(|0>), per-cycle latency, and a distance-3 repetition-code logical-error proxy, platform-calibrated simulation identifies candidate regions where blind reset cuts cycle latency by up to 38x under NVQLink-class feedback overhead while maintaining F_clean >= 0.86 for L <= 6. Hardware experiments on IQM Garnet confirm blind-reset cleanliness >= 0.84 at L=8 (1024 shots, seed 42); platform-calibrated simulation for Rigetti Ankaa-3 predicts comparable performance. Architecture-dependent crossover lengths are L* ~ 12 (IQM), ~ 11 (Rigetti), ~ 1 (IonQ), and ~ 78 with GPU-linked external feedback. Two added analyses tighten deployment boundaries: a T1/T2 sensitivity map identifies coherence-ratio regimes, and error-bound validation confirms measured cleanliness remains consistent with the predicted diagnostic envelope. A deployment decision matrix translates these results into backend-specific policy selection.
In quantum algorithm design, scratchpad auxiliary (ancilla) qubits represent a critical hardwareoverhead. While intermediate measurements can collapse superpositions destructively, Bennett’s principle of reversible uncomputation allows auxiliary registers to be disentangled unitarily and returned to their ground state |0⟩ for reuse. In this paper, we report empirical hardware benchmarks of coherent ancilla uncomputation and full-cycle state restoration across 12-qubit and 20-qubit registers executed on the 156-qubit superconducting processor ibm_kingston . To evaluate uncomputation under non-trivial amplitude configurations, continuous phase rotations were parameterized using entropy bytes extracted from a classical binary stream. The hardware successfully reset 10 entangled ancilla qubits back to the |0000000000⟩ state with an uncomputation fidelity of 90.30% (903/1000 shots). However, executing a complete forward-and-reverse Loschmidt echo protocol across the full 20-qubit register yielded an exact state restoration fidelity of only 21.40%. This degradation quantitatively illustrates how non-unitary NISQ errors— including 2-qubit gate cross-talk, T1/T2 relaxation, and compounding measurement readout error—destroy exact reversibility as circuit depth doubles. We provide full experimental provenance, calibration context, and job identifiers to establish a reproducible baseline for NISQ ancilla management. Keywords: Ancilla recycling, Quantum uncomputation, Loschmidt echo, NISQ benchmarking, Superconducting qubits,Heavy-hex architecture.
Readout errors are one of the dominant sources of noise in current quantum processors, limiting both expectation-value estimation and sampling-based applications. Since they affect only the classical measurement outcomes, they can be addressed using classical coding techniques: immediately before measurement, each data qubit is redundantly encoded with ancilla qubits, and the resulting bit string is decoded either by post-selection or by majority voting. Unlike conventional readout error mitigation, which corrects only aggregate quantities such as expectation values, this approach operates on individual measurement shots and can therefore produce approximately corrected samples. We present a systematic cross-platform and cross-generation experimental evaluation of repetition-code readout error detection and correction. We benchmark the same protocol on IBM Heron r1-r3 superconducting processors and Quantinuum H1 and H2 trapped-ion processors while independently varying the code distance, hardware generation, and encoding layout. We find that both error detection and correction improve readout fidelity on every device and generation tested, even as the unencoded baseline improves substantially across successive hardware releases. At the same time, the value of additional redundancy depends strongly on the underlying hardware. On superconducting processors, the extra gate errors introduced by the encoding rapidly offset its benefits, whereas on trapped-ion processors the much lower gate error rates allow larger code distances to remain advantageous.
Quantum computation of trapped spectra for nuclear scattering requires resource-efficient circuits and reliable energy inputs. We present a cross-platform benchmark of ancilla-recycled RODEO filtering for a four-dimensional effective Hamiltonian describing a trapped neutron--proton model. Static circuits on IonQ Forte-1 and dynamic circuits on IBM Aachen implement the same ideal ten-cycle all-zero accepted filter using twelve and three active qubits, respectively. Mid-circuit measurement and reset reduce the active qubit count by 75\% while retaining all controlled evolutions. The interleaved IBM peak-center offset from the exact level is −0.017±0.507keV, compared with −0.568±0.694keV for IonQ (finite-shot bootstrap uncertainties). At ω=3.7MeV, the modified effective range expansion maps both centers to finite-confinement scattering inputs consistent with the exact reference within propagated uncertainties. Three IBM batches exhibit center variations beyond finite-shot uncertainties (p=0.015), despite comparatively stable postselection amplitudes. Calibration-derived noisy simulations do not consistently reproduce the measured centers. Complete sampled support enables sample-based quantum diagonalization to recover the exact effective-space level, whereas larger-space simulator tests expose substantial errors from omitted configurations. The calculation validates one finite-confinement input, not a free-space phase shift. This benchmark provides a reusable circuit and validation protocol separating spectral precision, hardware repeatability, and sampled-subspace completeness for multilevel nuclear-scattering calculations.
One of the most promising paths towards large scale fault tolerant quantum computation is the use of quantum error correcting codes. The majority of popular codes are stabilizer codes, in which error information is extracted via the repeated execution of a set of commuting stabilizer measurement circuits and then decoded. Just like every other quantum circuit, these circuits must be compiled to hardware in a way to minimize the total physical error introduced into the system, for example either due to high latency execution or excessive gates to meet connectivity limitations of the target hardware. However, unlike arbitrary quantum circuits, all syndrome extraction circuits have several common properties, for example they have a bipartite connectivity graph, consist only of commuting subcircuits, and have a set of ancilla which are indistinguishable between measurement rounds, among others. For the most part, compilation methods have aimed at being generic, able to map any input circuit into executables on the hardware, and therefore cannot appropriately exploit these properties and result in executables which have higher physical error. In the case of modular trapped ion systems, specifically QCCDs, this corresponds to the insertion of excessive shuttling operations necessary to realize arbitrary qubit interactions. We propose a compilation scheme explicitly tailored for the structural regularity of quantum error correction code physical circuits, with the primary objective of minimizing circuit latency from shuttling operations. Our compiler's success is predicated on several key observations: 1. only ancilla or data (but not both) should be shuttled, 2. stabilizers can be executed in any order meaning we can dynamically modify circuit execution on a per-cycle basis 3. ancilla are indistinguishable meaning any can be selected to begin a stabilizer measurement and since only ancilla are shuttled, we retain a fixed-point mapping between cycles, and 4. QCCD hardware limits the number of parallel operations equal to the number traps in the system bounding the rate of stabilizer measurement meaning fewer ancilla are necessary and can be easily reused. Our resulting compiler, leads to QEC circuits which are on average3.38×faster to execute (by up to5.24×) and therefore incur lower physical error. For codes with good decoders, e.g. the color code and surface code, these improvements lead to up to two orders of magnitude of improvement in logical error rates with realistic physical error rates.
Fault-tolerant quantum memories depend on the syndrome extraction circuit as much as on the underlying code. Ancilla-free or dynamic circuits are an effective way to improve this circuit layer. For the 6.6.6 honeycomb Floquet code, making the circuit dynamic raises the threshold and lowers the qubit overhead, but at the cost of halving the spatial code distance. A dynamic construction for the 4.8.8 lattice layout was conjectured to preserve full distance. I confirm this and give a dynamic measurement circuit for the CSS 4.8.8 Floquet code. To benchmark it, I construct and compare four circuit-level implementations on a torus, including two dynamic variants (with and without mid-circuit resets), the standard ancilla-based circuit, and a pipelined ancilla-based circuit. Under circuit-level depolarising noise, the reset dynamic circuit reaches a per-round threshold of 0.463%(0.490%) with MWPM (BP+matching), while the no-reset variant reaches the highest threshold of all four circuits at 0.512%(0.574%). The standard ancilla-based circuit only achieves 0.228%(0.240%), but the pipelined schedule reaches 0.478%(0.489%). The reset dynamic circuit also has a faster-growing timelike distance, with 2≤dt/nqec≤3 asymptotically against a tight 3/2 for the other three, and running it for fewer rounds gives the smallest spacetime volume in the fast-reset regime, while the no-reset variant is smallest in the slow-reset regime. The 4.8.8 dynamic circuits therefore see the expected threshold gain and overhead reduction without the spatial-distance cost, demonstrating the advantage of dynamic syndrome extraction in Floquet codes.
Quantum error correction (QEC) decoders produce corrections but provide noper-shot signal of correction reliability. We introduce syndrome-consistency gating: adecoder-independent shot acceptance filter that uses redundant syndrome copies to identifyunreliable decoder outputs. The gate is veto-only (it may block shots but never overrides thedecoder) and fails closed (blocks all shots on gate failure). We validate this approach on twoindependent hardware vendors. The primary dataset comprises 768,000 shots on IBM Torino(Heron r2, 156 qubits); a cross-vendor dataset of 64,000 shots on Rigetti Ankaa-3 (84qubits) provides supporting evidence; and an exploratory dataset of 64,000 shots on IBMFez preceded both. Total: ~896,000 hardware shots. All experiments use repetition codes.On IBM Torino with a distance-5 code, strict gating with two redundant copies (f=2) achieves100% logical error rate (LER) reduction (0.188% to 0.00%) at 36.3% allow rate, with all 30observed logical errors concentrated exclusively in blocked shots. Under practicalparameters, LER reduction reaches 85.6% (0.081% to 0.012%) at 53.3% allow rate with13.7× error concentration. On Rigetti Ankaa-3, gated active broadcast achieves zero logicalerrors in 16,076 accepted shots versus 9 errors in 24,000 passive shots (p = 0.014). Wecharacterise the allow-rate scaling as exponential decay (R² ≥ 0.98) and identify f=2 as thepractical operating point.