Paper 7 of this series carried user-written two-dimensional circuits through an operating layer (psf-os) that reads no calibration-quality number, adds its own controls to each job and applies no error mitigation. Here the same policy is carried to a chain that no fixed shape can hold and to three-dimensional lattices that the heavy-hex coupling graph cannot embed. IBM's dual-unitary kicked-Ising chain (arXiv:2411.00765) at its largest width, N = 91, read raw on ibm_marrakesh: the decay rates agree with cos 2h at all three h (largest |k - cos 2h| 0.0037) and every depth up to t = 24 (2 205 CZ) is readable, on a chain placed by a fixed rule round the failed elements; a second run on ibm_fez, with the ruler h = 0 in both jobs, again found all three rates in agreement, while h = 0.15 sat +0.0037 above cos 2h on both devices. In three dimensions, a place rule that judges the place with the highest own mirror replaced a gate that had refused every place at 48 and 72 sites; kicked Ising on 4 x 4 x 3 read at every angle and depth, Heisenberg XXX on 4 x 2 x 2 read with a Neel retention of 0.72 to 0.51. Two new layouts cut the transport CZ of the Heisenberg step by 42%, and were compared with the old one on the same places of two devices; an Ising layout with 46% fewer CZ came closer to the exact answer at the judged 4 x 2 x 2 place but farther on 4 x 3 x 2. On the way we changed our main measure for such comparisons, from the retention of an order parameter to a normalised per-site error, because noise that slows the dynamics can raise the retention. Pair creation in the Schwinger model read in 16 of 18 circuits and beat both trivial guesses in all 18. A light-cone test failed as predicted: every front arrived late, while the shape of the spread was reproduced. Every device result re-scores to the same numbers from its committed raw counts (23 of 23). No quantum advantage is claimed: every answer is computed exactly on a classical machine.
Paper 7 of this series carried user-written two-dimensional circuits through an operating layer (psf-os) that reads no calibration-quality number, adds its own controls to each job and applies no error mitigation. Here the same policy is carried to a chain that no fixed shape can hold and to three-dimensional lattices that the heavy-hex coupling graph cannot embed. IBM's dual-unitary kicked-Ising chain (arXiv:2411.00765) at its largest width, N = 91, read raw on ibm_marrakesh: the decay rates agree with cos 2h at all three h (largest |k - cos 2h| 0.0037) and every depth up to t = 24 (2 205 CZ) is readable, on a chain placed by a fixed rule round the failed elements; a second run on ibm_fez, with the ruler h = 0 in both jobs, again found all three rates in agreement, while h = 0.15 sat +0.0037 above cos 2h on both devices. In three dimensions, a place rule that judges the place with the highest own mirror replaced a gate that had refused every place at 48 and 72 sites; kicked Ising on 4 x 4 x 3 read at every angle and depth, Heisenberg XXX on 4 x 2 x 2 read with a Neel retention of 0.72 to 0.51. Two new layouts cut the transport CZ of the Heisenberg step by 42%, and were compared with the old one on the same places of two devices; an Ising layout with 46% fewer CZ came closer to the exact answer at the judged 4 x 2 x 2 place but farther on 4 x 3 x 2. On the way we changed our main measure for such comparisons, from the retention of an order parameter to a normalised per-site error, because noise that slows the dynamics can raise the retention. Pair creation in the Schwinger model read in 16 of 18 circuits and beat both trivial guesses in all 18. A light-cone test failed as predicted: every front arrived late, while the shape of the spread was reproduced. Every device result re-scores to the same numbers from its committed raw counts (23 of 23). No quantum advantage is claimed: every answer is computed exactly on a classical machine.
Compilers for today's quantum processors choose a layout by reading the day's calibration. We ask how deep a circuit stays readable when no placement reads a quality number. We place a ring of CZ gates on IBM's heavy-hex devices in three ways that see only the coupling graph, the basis and which elements have failed: a formula that puts the ring on a lattice cycle (the rule), the ring folded onto the same qubits (zipper), and Qiskit's level-2 search without error rates (blind Qiskit). The workload is the kicked Ising ring at J = pi/4; the probes are a weight-2 charge conserved by the circuit, a mirror echo with a longitudinal field that breaks integrability, and the global parity. In declared tests on ibm_marrakesh and ibm_kingston (29-30 October 2026), the rule kept more of the charge than both other placements in all 24 comparisons (readable depth 56-90 steps against 15-26), and in the non-integrable echo it won 71, tied 11 and lost none of 144 comparisons, at sizes up to n = 72. Several declared predictions failed, and we report them with the results. On ibm_kingston the rule's n = 40 ring carried a run of 18 sites whose echo was near random or inverted: the same 18 sites in three runs of the byte-identical circuit over three hours, while the calibration showed nothing there and the same qubits were healthy on the n = 20 ring. Gate twirling, which leaves the ideal circuit unchanged, turned the inverted part random, so the cause is on the device, not in the circuit; a light-cone argument bounds how far one fault reaches. These sites set the readable depth at n >= 40 and erased the parity at n = 40. A map read from the calibration could not have avoided them. All declarations, records and raw counts are public in TOQB.
A user writes, in plain Qiskit, a circuit whose answer is known exactly. A small operating layer (psf-os) places it on an IBM Heron r2 device by translation, without reading any calibration-quality number, adds its own controls to the same job, and runs it with no error mitigation. We report how close the raw answers come and how that is shown. Two points of the report are of use beyond our own numbers. First, an absolute criterion |M - M_exact| <= c is biased in both directions: on 5 November a problem that DIFFERS under c = 0.05 kept 0.82-0.86 of its signal, as much as a problem that AGREES (0.73-0.85), and a small-answer problem AGREES while keeping only 0.73 of its signal. We therefore print the declared verdict first and the retention R = M/M_exact beside it on every row. Second, every number can be checked: answers are sealed by hash before each run, raw counts and job identifiers are committed, and one script re-scores all 13 device results of this paper from the raw counts and finds every number the same. The results: 22 practice circuits on two devices (all 22 READ on both; 19 and 18 AGREE), a daily record from 5 to 7 November (7 runs and one refusal), and IBM's dual-unitary kicked-Ising chain (arXiv:2411.00765) at N = 51 and N = 71 sites, where the raw decay rates agree with cos 2h at all three h (largest |k - cos 2h|: 0.0045 at N = 51, 0.0017 at N = 71) and every depth up to t = 24 (1,715 CZ) is readable. Local failure regions, not problem size, set several outcomes. A correction built from the job's own controls, declared in advance, reduced the error on both devices but did not make the answers agree across problems. No quantum advantage is claimed: every answer here is computed exactly on a classical machine. We acknowledge the use of IBM Quantum services for this work.
A circuit placed on a superconducting device stays readable only to some depth. This preprint asks what sets that depth on a given day and device, and how much of it a user-side layer can recover from its own runs, without reading the calibration map.The layer scans the rings it would use with a mirror echo read per site, keeps a dated registry of the paths it found dead, places the next ring around them with the same formula, and, for mirror-type circuits only, applies a local frame flip. Every test was declared before its device data were fetched (SHA-256 and time in the paper's declaration table).Results on IBM's heavy-hex devices (ibm_kingston, ibm_fez, ibm_marrakesh, 30 October to 2 November 2026):- Width: on ibm_kingston, the 40-qubit ring placed around the registry kept 40 of 40 sites usable on three days, against 22 on the standard ring, where the same 18-qubit run was dead on every day (WIDER in 6 of 6 declared comparisons). Scan-then-avoid also widened the ring on ibm_fez and ibm_marrakesh.- Accuracy: on rings already fully usable, forward computations with known answers were slightly more right on the avoiding ring (+0.040 and +0.045 in mean site signal), with all of the gain on the replaced qubits.- A sentinel in the user's own job read the known dead run as DEAD and the avoiding ring as CLEAN, at job time and an hour after the scan. The cost of the same defect depended on the computation: small for a 12-layer forward identity, total for a weight-40 parity.- Limits are stated: three devices over a few days, one placement formula, repair only for mirror circuits; a lower bound for calibration-only placement is given in an appendix and called close to a tautology.
Two registered experiments with the QITE equilibrium law on the IBM Heron processor ibm_marrakesh, one day apart in nothing but the place where the law runs: outside the processor on a microcontroller board, and inside one shot in the processor's own feed-forward. Each experiment has its own registration, hashed before its first registered job, its own frozen scripts, complete raw data, verdict and write-up; the two archives in this record are independent of each other. QITE Hold (zenodo_qite_hold_20260923.zip, SHA-256 97853e20d86c146c16d6011191e9c038dddc743dfd5392476993187654f6e1d3; write-up custodia_paper_v1.pdf; working name Custodia). A microcontroller board (ESP32-C3) holds one pair of numbers per qubit for 144 qubits and turns each pair into a rotation angle. Every tick, one circuit of 144 single-qubit rotations is measured 4096 times; the fraction of ones per qubit goes to the board, the board updates its numbers by its fixed law and returns the angles of the next circuit; it never reads the processor's calibration. A registered programme of eight blocks ran over 280 ticks and 290 jobs: square waves, a hold, a staircase, a checkerboard of 18 zones up and 18 down, two state shocks written into the board, two disturbances written into the circuits, and a comparison with an open-loop reference of frozen angles. Of the seven binding lines, six were run and held; the seventh (drift against the open-loop reference) could not be run because the device did not drift by the registered minimum. The square waves came back with a gain of 0.231 of the commanded swing (the law: 0.25), the checkerboard put 98.5 % of (tick, zone) pairs on the commanded side, the disturbances were rejected to 0.557 and 0.539 of their open-loop size (the law: 4/7). A first registration (v1.0) ended at tick 130 through a design flaw of the host and is included in full. Registration SHA-256 1de82ab12a53c1da349196e8ef07851e6b2f176efc6bd1ca6ba14cacd1f7d54b. QITE In-Shot (zenodo_qite_inshot_20260923.zip, SHA-256 a087f88a95f17c6f8023d5ff40aff35901b5e3e87db531dcd76728511f7ae944; write-up qite_inshot_paper_v1.pdf). On each of 54 qubits, twenty rounds of a measurement followed by a rotation conditioned on that measurement, with angles that set the target level and the memory g; the records of a qubit form a two-state Markov chain whose second eigenvalue is 1 - g. The lines test that eigenvalue, not the level: the autocorrelations of the records for g = 0.3, 0.6 and 1.0 were +0.607, +0.366 and +0.046, the two memories in the ratio 1.657 (the law: 1.75); the relaxation from |0> came back at 0.690 and 0.446 per round (the law: 0.70 and 0.40, lowered by T1 during a round); a step of the target proceeded at 0.658 per round and, with g = 1, completed in one round; a kick of the state was followed by a return to within 0.012 of the unkicked hold; over the same twenty rounds an open-loop state lost 0.451 while the closed loop held within 0.006; 98 % of the qubits with a clean read-out showed their own rate within the registered band. All seven binding lines held; 17 jobs. The size is what the service's compiler accepted on the day (144, 72 and 64 qubits failed with an internal compiler error; every attempt is recorded with its job id). Registration SHA-256 51933de4196bfdd7998e6a5305f7576bb7d35ac137e8b6ce0109965372b4fe5c. Nothing here is a claim of quantum advantage: the observables are classical bits and the circuits contain no two-qubit gate. The claim of both experiments is one of control: a fixed law, registered in advance, holding a public quantum processor at commanded levels with the numbers the law says, once from a board over minutes and once inside the coherence time of the qubits. The board's firmware and the serial link layer are not part of the record; both are identified by their SHA-256 in the registration of QITE Hold. Licence: CC BY 4.0.
We report a set of pre-registered experiments on the three 156-qubit IBM Heron r2 processors reachable on a single open-plan account (ibm_kingston, ibm_marrakesh, ibm_fez), carried out on 14–15 October 2026. The object of study is not computational speedup but channel reliability: whether fixed symbolic messages can be delivered intact through a present-day noisy processor when the processor's per-wire readout response is measured fresh in the same sitting, and whether two such processors can exchange messages with per-position verification. Messages are 17-symbol sequences drawn from a 13-symbol test alphabet whose preparations collapse to nine distinguishable readout classes. A receiver combining per-wire selection from a fresh characterization, grammar-constrained decoding, and a corpus-frequency tie-break delivered 36/36 messages on ibm_kingston in two independent campaigns, 36/36 on ibm_marrakesh, and 35/36 on ibm_fez, against stale-plan and uncharacterized controls ranging from 2/36 to 30/36 depending on backend and control type. The grammar-constrained receiver without the corpus prior scored 30/36, 30/36, and 29/36 on the three backends; its losses are a single systematic mechanism — a deterministic tie on one battery message — identified from saved counts and closed by the prior. In a separate experiment, ibm_kingston and ibm_marrakesh exchanged self-introduction streams authored by each backend from its own fresh characterization, through a six-stage symmetric relay, with 204/204 position-flights received exactly and zero substitutions. In a counts-only blind test, an independent decoder with no shared codebook recovered the discreteness and grouping structure of two authored streams (pooled pairwise grouping agreement 0.971 against a pre-registered 0.90 strong-pass bar) while overestimating the alphabet size (reported 8, true value 6). A census of repeated characterizations shows that the ranking of a device's clearest wires turns over within hours (same-day top-17 wire overlap of 1–2 of 17 on ibm_marrakesh), while per-wire response quality moves by only ~0.01, so characterizations are within-sitting instruments. No speedup, no machine networking, and no machine understanding is claimed. The full receiver implementation is withheld as a trade secret; verification is offered through the published job IDs, retained raw counts, and tiered third-party reproduction.
Quantum processors have crossed the one-hundred-qubit mark, but noise continues to limit circuit performance, while full quantum error correction remains too costly for routine use. Error suppression and mitigation therefore play an important role in extracting value from current hardware, yet independent comparisons of commercial solutions on identical workloads and devices remain scarce. We benchmark IBM Qiskit Runtime, Q-CTRL Performance Management, and Qedma QESEM on IBM Pittsburgh, a 156-qubit IBM Quantum Heron r3 processor. For Sampler workloads, we run Bernstein-Vazirani, quantum phase estimation, GHZ-state preparation, and randomized mirror circuits with up to 100 measured qubits, comparing raw execution, IBM measurement twirling, and Q-CTRL. For Estimator workloads, we measure chain-averaged magnetization and correlation observables of an eight-layer transverse-field Ising circuit at 25, 50, and 75 qubits against an exact matrix-product-state reference, comparing IBM raw execution, IBM TREX plus twirling, Q-CTRL, and QESEM. Q-CTRL produced the best results on the three structured Sampler workloads while keeping reported QPU times within the same order as the IBM configurations. Across six Ising observable and system-size cases, aggregate mean absolute error was 0.0883 for IBM raw execution, 0.0807 for IBM TREX plus twirling, 0.0285 for Q-CTRL, and 0.0188 for QESEM. Relative to raw execution, Q-CTRL and QESEM reduced aggregate error by factors of 3.10 and 4.70, respectively, while QESEM used 7.5 to 11.1 times the reported QPU time of Q-CTRL. These results show that managed error suppression and mitigation can substantially improve current hardware performance, but with distinct accuracy and execution-time tradeoffs.