Pith. sign in

REVIEW 3 major objections 4 minor 66 references

Quantum annealers as programmable thermal machines

T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read A commercial quantum annealer can be described as a programmable thermal machine whose operating mode is selected by the initial thermal ensemble and the reverse-annealing schedule.

desk verdict A useful experimental extension of the D-Wave thermal machine framework, undercut by identifying a fitted effective temperature with the bath temperature in the fluctuation theorem; refrigerator/engine labels are not certified by the one-sided bounds. read the letter →

arxiv 2608.04564 v1 pith:PV2K3DUK submitted 2026-08-05 quant-ph

classification quant-ph
keywords quantumannealingD-WavethermodynamicuncertaintyrelationsreversethermalmachineentropyproductioneffectivetemperatureIsingmodel
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 sets out to show that a commercial quantum annealer is not just an optimizer or sampler but a programmable thermal machine: by choosing the initial inverse temperature $\beta_1$ and the reverse-annealing turning point $s_p$, the same chip can be made to operate as a heater, an accelerator, a refrigerator, or an engine-compatible device. The analysis uses only the measured change in the programmed Ising energy over a closed cycle, together with an effective environment temperature $\beta_2$ fitted to the output bitstrings, to place lower bounds on entropy production, heat, work, and power via thermodynamic uncertainty relations. The interest is that these bounds turn the limited energy readout already available on the hardware into an energy-aware characterization that is independent of solution quality and runtime. If the claim holds, thermodynamic mode becomes a practical diagnostic for comparing schedules and hardware, and a bridge between quantum optimization and energy efficiency.

What carries the argument

The argument is carried by the exchange fluctuation theorem, $p(\Delta E_1,\Delta E_2)/p(-\Delta E_1,-\Delta E_2)=\exp(\beta_1\Delta E_1+\beta_2\Delta E_2)$, combined with thermodynamic uncertainty relations that convert the first two moments of the measured processor energy change $\Delta E_1$ into lower bounds on entropy production, heat, work, and power. The environment inverse temperature $\beta_2$ is obtained by pseudo-likelihood fitting of the output bitstrings to the programmed Ising Hamiltonian, and the sign structure of the cycle-averaged energy exchanges then assigns the thermodynamic regime. The reverse-annealing schedule keeps the initial and final Hamiltonians identical, which makes the two-point energy change $\Delta E_1$ directly measurable and closes the thermodynamic cycle.

What would settle it

Run the same reverse-annealing cycles while independently measuring the heat absorbed by the cryogenic environment (e.g., with calorimetry or power metering at the mixing chamber) and compare with the paper's lower bound on $-\langle Q\rangle$: if the measured heat falls below the bound, the identification of the fitted $\beta_2$ with the fluctuation-theorem temperature fails. A simpler statistical check is to hold out part of the output bitstrings and test whether they are consistent with the Gibbs distribution at the fitted $\beta_2$; a clear failure would mean the sign-based regime labels are not thermodynamically certified.

Watch

Extended reading notes

Core claim

The central discovery claimed is that reverse annealing on D-Wave hardware realizes a closed thermodynamic cycle whose operating regime is programmable. With a fixed problem Hamiltonian, changing only the preparation temperature and the turning point moves the device through heater-, accelerator-, refrigerator-, and engine/accelerator-compatible regimes, and these regimes can be read off from the sign structure of the cycle-averaged energy exchanges without a detailed microscopic bath model. The claim is supported by phase diagrams in the $(\beta_1, s_p)$ plane for one-dimensional chains and two-dimensional Pegasus instances, on two hardware generations, and by the observed scaling of work and power bounds with chain length. The authors further claim that the fitted effective temperature $T_2$ acts as an operational thermometer for the sampled degrees of freedom, and that the thermodynamic mode adds information absent from success probability or runtime alone: it distinguishes driven refinement, net heating, and heat pumping while quantifying their energetic consequences.

Load-bearing premise

All quantitative claims assume that the inverse temperature $\beta_2$ fitted from the output bitstrings is the same temperature that enters the exchange fluctuation theorem; if the output distribution is not close to a Gibbs state, or if the cryogenic environment is not a single thermal reservoir, the bounds and every thermodynamic regime label lose their meaning.

Editorial extensions

If this is right

  • The same hardware can be moved between heater, accelerator, refrigerator, and engine/accelerator-compatible regimes by changing only the initial inverse temperature $\beta_1$ and the turning point $s_p$.
  • The measured mean processor energy change $\langle\Delta E_1\rangle$ tells whether final configurations improve or worsen the programmed objective on average, so thermodynamic data double as objective-improvement diagnostics.
  • For sampling, the fitted effective temperature $T_2$ gives an operational measure of how strongly probability concentrates on low-energy configurations, and schedule-conditioned shifts in $T_2$ provide a built-in thermometry diagnostic.
  • In the one-dimensional critical window, the TUR-based work and power bounds grow as $L\log L$ with chain length, indicating where dissipation concentrates and how it scales.
  • Refrigerator operation is not automatically good for optimization; it corresponds to computational improvement only at points where $\langle\Delta E_1\rangle < 0$.

Reading between the lines

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

  • Editorial inference: reporting success probability, target-energy probability, sample diversity, and time to solution conditional on the thermodynamic mode could reveal whether mode labels predict practical performance; the paper explicitly leaves this conditional analysis as future work.
  • Editorial inference: folding the chip-level bounds together with programming, readout, control, and refrigeration costs would give an energy-to-solution benchmark for comparing algorithms and hardware.
  • Editorial inference: the sign-structure classification should transfer to other driven open quantum systems with closed cyclic schedules, making the thermodynamic-mode diagnostic a general tool rather than a D-Wave-specific one.
  • Editorial inference: replacing the pseudo-likelihood estimate with a direct Gibbsness test on the output distribution would strengthen the claim that the fitted $\beta_2$ is the thermodynamic temperature, and would show which regions of the phase diagram are robust.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper reports reverse-annealing experiments on D-Wave Advantage and Advantage2 processors, treating each run as a closed thermodynamic cycle. From the measured distribution of the processor energy change ΔE1 and an effective inverse temperature β2 obtained by pseudo-likelihood fitting of the output bitstrings, the authors use an exchange fluctuation theorem and thermodynamic uncertainty relations to derive lower bounds on entropy production, heat, work, and power, and to classify operating regimes as heater, accelerator, refrigerator, or engine. The regime maps are presented as functions of the initial inverse temperature β1 and the reverse-annealing turning point sp, for one-dimensional chains and two-dimensional Pegasus instances, with a companion finite-size scaling analysis. The central claim is that the quantum annealer is a programmable thermal machine whose thermodynamic role can be selected by initialization and schedule and diagnosed from energy statistics alone.

Significance. If the underlying identification of β2 with the environment temperature in the fluctuation theorem were validated, the paper would provide a genuinely useful, device-agnostic methodology for benchmarking quantum annealers on energetic grounds, complementary to solution quality and runtime. The scale of the experiments (multiple D-Wave generations, thousands of qubits, >10^4 cycles per setting), the explicit statement of assumptions in Methods 4.8, the honest use of the label 'engine/accelerator-compatible' in the two-dimensional analysis, and the public data/code repositories are all strengths. However, the central quantitative claims and the four-mode classification rest on an unvalidated identification of a fitted output parameter with the bath temperature of the exchange fluctuation theorem; this makes the significance conditional rather than established.

major comments (3)
  1. [Methods 4.5–4.6 (Eqs. 6–12)] The β2 appearing in the exchange fluctuation theorem, Eq. (7), is the inverse temperature of the environment in the initial factorized state (6). In the applications, Eqs. (8)–(10), the same symbol β2 is replaced by the maximizer of the pseudo-likelihood (11)–(12) fitted to the final output configurations. This is not a harmless notational choice: the paper itself states in Section 2 and the Discussion that T2 is an effective temperature of the sampled Ising degrees of freedom that depends on freeze-out, calibration, schedule, and residual non-equilibrium dynamics, whereas a bath temperature is a property of the cryogenic environment. Unless the output distribution is exactly Gibbs at the fitted β2 and the environment acts as a single reservoir at that same temperature, the bounds (8)–(10) and every regime label derived from them are not consequences of Eq. (7). The concessions in Appendix A ('these diagnostics do not prove that the full high dimensional distribution is exactly Gibbsian') and the Discussion ('a fitted temperature alone does not establish that the complete output distribution is thermal') mark exactly the unvalidated premise. This issue is load-bearing for all quantitative claims in the paper and needs either an independent validation of β2 (for example, comparison with the vendor's freeze-out effective temperature or with direct cryogenic thermometry) or a systematic reframing of the results as conditional effective-temperature phenomenology rather than thermodynamic bounds.
  2. [Methods 4.6, Eq. (9), Table 2, and Figs. 4–6] Equation (9) is a lower bound on −⟨Q⟩, equivalently an upper bound on ⟨Q⟩. It can certify that Q<0 (the environment gains energy), but it cannot certify Q>0 (the environment loses energy). The refrigerator and engine assignments in Table 2 require, for some of the temperature orderings, a definite sign of ⟨ΔE2⟩ or ⟨W⟩ that is not accessible from Eq. (9) or Eq. (10): Eq. (10) is a lower bound on ⟨W⟩ and cannot certify W<0, which the engine class requires. The body text is careful in the two-dimensional sections to use the combined label 'engine/accelerator-compatible,' but the abstract and Introduction state that refrigerator- and engine-compatible regimes are realized and mapped on the same device (Abstract; Section 1: 'we implement protocols that realize all four behaviours on the same device'). Given the one-sided nature of the available bounds, the data can certify heater and accelerator operation directly, but refrigerator and engine operation are at best compatible with the bounds, not certified. The abstract and the summary sentences in the Discussion should be revised to state precisely which labels are certified by measurements and which are only allowed by the bounds.
  3. [Section 2, 'Scaling with system size', Eq. (1) and Fig. 2] The claim that the per-spin work and power bounds grow logarithmically with system size, and the total bounds as L log L, is based on Fig. 2, which shows four chain lengths without error bars, fit lines, or goodness-of-fit statistics. The text says the data are 'consistent with a logarithmic finite size enhancement,' which is a much weaker statement than the subsequent 'the empirical LlogL law is the expected finite size signature.' If this scaling is presented as a quantitative result, the authors should provide a fit with uncertainties or explicitly label it as a qualitative trend. This is not the central claim of the paper, but as written the section overstates the evidential support.
minor comments (4)
  1. [Fig. 3 caption] The caption reads 'The experiment was performed the D-Wave Advantage6.4 system'; 'performed' should be 'performed on'.
  2. [Methods 4.6, Eq. (13)] The expression for Ep(sp,σ) is hard to read; the factor 'hP 109 2 BGHz(sp)' appears to mean hP × 10^9 × (1/2) × BGHz(sp) but the typesetting is ambiguous. Please rewrite with explicit powers of ten.
  3. [References] References [8] and [21] are the same paper, and references [22] and [26] are the same arXiv preprint; please merge or distinguish them.
  4. [Methods 4.6] The phrase 'device-certified, bath model-free lower bounds' is misleading: the bounds are model-free only with respect to the microscopic bath dynamics, but they still require the single-reservoir assumption and the identification of β2 with the bath temperature. Please rephrase to avoid this overstatement.

Circularity Check

1 steps flagged · score 4.0 of 10

The inverse temperature beta2 entering the exchange fluctuation theorem and TUR bounds is fitted to the final output bitstrings, so the 'environment temperature' is an output-fit parameter; the thermodynamic bounds and regime labels are partially circular, though the measured Delta E1 moments enter independently.

  1. fitted input called prediction [Methods 4.5-4.6, Eqs. (6)-(12); Discussion thermometry paragraph; Appendix A]
    "Consider the compound (system + environment) initialized in the factorized thermal state ρ0 = e−β1Hp/Zp(β1) ⊗ e−β2HE/ZE(β2). ... p(∆E1,∆E2)/p(−∆E1,−∆E2) = exp(β1∆E1 +β2∆E2). ... The inverse temperature β2 of the environment is obtained by fitting the sampled spin configurations to the Boltzmann distribution encoded by the programmed couplings and fields [13, 63, 64]."

    In Eq. (7), β2 is the initial environment inverse temperature of state (6), but the only operational β2 is the pseudo-likelihood maximizer over final output bitstrings, Eq. (12): β2 = arg max Λ(β). This same β2 enters the TUR bounds (9)-(10) and, via Table 2, the heater/accelerator/refrigerator/engine label assigned to that same output. So the 'environment temperature' is not an independent bath property; it is the temperature making the output look most Gibbs-like, and the output's own moments are then certified against bounds containing that fitted temperature. The measured ⟨∆E1⟩ moments enter independently, so the bounds are not tautological, but every quantitative bound and regime label is conditional on the output-fit β2 equaling the true initial bath temperature.

full rationale

The paper's derivation chain is mostly self-contained: the exchange fluctuation theorem (Eq. 7) and the TUR bounds (Eqs. 8-10) are standard external results, the measured Delta E1 statistics come directly from hardware readout, and the sign conventions in Table 2 define the thermodynamic modes operationally. The one load-bearing circular link is the temperature beta2: Methods 4.6 obtains beta2 by pseudo-likelihood fitting to the final output bitstrings, yet Eq. (7) requires beta2 to be the initial environment inverse temperature. Reusing the output-fitted beta2 in the bounds and in the mode classification means the same output distribution supplies both the temperature and the fluctuation statistics used to certify its own thermodynamic regime. This is a fitting loop rather than a full equivalence, because the first two moments of Delta E1 enter independently and the fluctuation-relation mathematics is not derived from the fit. The paper's explicit concessions in Appendix A and the Discussion do not remove the loop; they identify the assumption on which every quantitative claim rests. No load-bearing self-citation chain or imported uniqueness theorem was found; the self-citations (refs. 7, 14) support context but are not the logical core. Accordingly, the circularity score is moderate: 4.

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

The central analysis rests on a two-temperature model in which the prepared classical Gibbs distribution (beta1) and the output-fitted effective temperature (beta2) are inserted into an exchange fluctuation theorem. Beta2 is a fitted parameter, not a measured bath temperature, and the entire thermodynamic classification is conditional on the pseudo-likelihood fit and on the single-reservoir assumption.

free parameters (2)
  • beta2 (effective inverse temperature) = inferred 5-120 mK equivalent across scans
    Pseudo-likelihood fit to output spin configurations (Eq 12); enters every heat, work, power, and entropy bound, and the regime classification.
  • beta1 (initial inverse temperature) = 0.1 to 5.2 (scanned)
    Prepared input Gibbs temperature; chosen by hand as a control knob; sets the energy bias of the initial state and enters all bounds.
assumptions (5)
  • domain assumption The compound system plus environment is initialized in a factorized thermal state and obeys the exchange fluctuation theorem (Eq 7).
    Used to derive the TUR bounds and the second-law constraint. If the initial state or the bath model is wrong, the bounds do not apply.
  • domain assumption The environment can be described by a single inverse temperature beta2 that enters the exchange fluctuation theorem.
    A real D-Wave chip has a complex cryogenic environment; beta2 is an effective fit parameter, not a directly measured temperature. This is the central modeling assumption.
  • domain assumption The pseudo-likelihood maximum (Eq 12) gives the true beta2 of the environment.
    Requires the output distribution to be sufficiently close to a Gibbs distribution of the programmed Ising model for the fitted parameter to have thermodynamic meaning.
  • standard math The thermodynamic uncertainty relations (Eqs 8-10) are valid for the exchange fluctuation framework.
    The authors cite refs 59-62 for the TUR inequalities; the bounds are conditional on the fluctuation theorem and the two-temperature setup.
  • domain assumption The classical Gibbs sampler used to prepare initial states converges to the intended distribution.
    Appendix A checks low-order observables (energy and magnetization) but does not establish that the full high-dimensional distribution is exactly Gibbsian.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Quantum annealers as programmable thermal machines." pith.science (2026). https://pith.science/paper/PV2K3DUK

@misc{pith2026260804564,
  author       = {Pith},
  title        = {Pith review of: Quantum annealers as programmable thermal machines},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PV2K3DUK}},
  note         = {Machine review of arXiv:2608.04564}
}
read the original abstract

Programmable quantum annealers are used for optimization, probabilistic sampling, and simulation, but their performance is commonly reported without the energy exchanged during computation. Here we characterize the D-Wave quantum annealer as a closed thermodynamic cycle. From initial and final Ising energies and an effective temperature fitted to the output distribution, we obtain lower bounds on entropy production, environment energy exchange, work, and power. By varying the prepared distribution and the reverse annealing turning point, we map heater-, accelerator-, refrigerator-, and engine-compatible regimes in one dimensional chains and higher connectivity instances, and apply the same analysis to Advantage and Advantage2 hardware. For an encoded optimization problem, the measured processor energy change states whether final candidates improve or worsen the programmed objective on average. For sampling, the fitted temperature provides an operational measure of how strongly probability is concentrated among low energy configurations. The thermodynamic mode therefore adds information absent from solution quality or runtime alone: it distinguishes driven refinement, net heating, and heat pumping while quantifying their energetic consequences. This framework connects quantum optimization, probabilistic computing, statistical physics simulation, hardware diagnostics, and energy-aware assessment without assuming that a thermodynamic label alone determines computational performance.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

66 extracted references · 43 canonical work pages

  1. [1]

    & Gardas, B

    Hanussek, P., Paw lowski, J., Mzaouali, Z. & Gardas, B. Quantum-inspired dynamical models on quantum and classical annealers.Phys. Rev. Appl.25, 044055 (2026). URL https://link.aps.org/doi/ 10.1103/rrf3-jm5m

  2. [2]

    & Gardas, B

    Paw lowski, J., Tarasiuk, P., Tuziemski, J., Pawela, L. & Gardas, B. Toward quantum scaling advantage in approximate optimization.Phys. Rev. Appl.26, 014024 (2026). URL https://link.aps.org/doi/10. 1103/qcws-8kgm

  3. [3]

    & Gardas, B

    Tuziemski, J., Paw lowski, J., Tarasiuk, P., Pawela, L. & Gardas, B. Limits of quantum run-time advantage.Phys. Rev. Appl.25, 044084 (2026). URL https://link.aps.org/doi/10.1103/gpsf-pn1x

  4. [4]

    URL https://doi.org/10

    Bando, Y.et al.Probing the universality of topological defect formation in a quantum annealer: Kibble– Zurek mechanism and beyond.Physical Review Research2, 033369 (2020). URL https://doi.org/10. 1103/PhysRevResearch.2.033369

  5. [5]

    Sathe, P.et al.Classical criticality via quantum annealing.Nature Communications17, 856 (2026)

  6. [6]

    D.et al.Beyond-classical computation in quantum simulation.Science388, 199–204 (2025)

    King, A. D.et al.Beyond-classical computation in quantum simulation.Science388, 199–204 (2025). URL https://doi.org/10.1126/science.ado6285. 17

  7. [7]

    & Deffner, S

    Gardas, B. & Deffner, S. Quantum fluctuation theorem for error diagnostics in quantum annealers. Scientific reports8, 17191 (2018). URL https://doi.org/10.1038/s41598-018-35264-z

  8. [8]

    Y., Albash, T

    Nelson, J., Vuffray, M., Lokhov, A. Y., Albash, T. & Coffrin, C. High-quality thermal gibbs sampling with quantum annealing hardware.Phys. Rev. Applied17, 044046 (2022). URL https://link.aps.org/ doi/10.1103/PhysRevApplied.17.044046

Show all 66 references
  1. [9]

    Amin, M. H. Searching for quantum speedup in quasistatic quantum annealers.Phys. Rev. A92, 052323 (2015). URL https://link.aps.org/doi/10.1103/PhysRevA.92.052323

  2. [10]

    & Rieffel, E

    Marshall, J., Venturelli, D., Hen, I. & Rieffel, E. G. Power of pausing: Advancing understanding of thermalization in experimental quantum annealers.Phys. Rev. Applied11, 044083 (2019). URL https://link.aps.org/doi/10.1103/PhysRevApplied.11.044083

  3. [11]

    & Ohzeki, M

    Kadowaki, T. & Ohzeki, M. Experimental and theoretical study of thermodynamic effects in a quantum annealer.J. Phys. Soc. Jpn.88, 061008 (2019). URL https://doi.org/10.7566/JPSJ.88.061008

  4. [12]

    & Campisi, M

    Buffoni, L. & Campisi, M. Thermodynamics of a quantum annealer.Quantum Science and Technology 5, 035013 (2020). URL https://iopscience.iop.org/article/10.1088/2058-9565/ab9755

  5. [13]

    & Buffoni, L

    Campisi, M. & Buffoni, L. Improved bound on entropy production in a quantum annealer.Phys. Rev. E104, L022102 (2021). URL https://link.aps.org/doi/10.1103/PhysRevE.104.L022102

  6. [14]

    & Gardas, B

    ´Smierzchalski, T., Mzaouali, Z., Deffner, S. & Gardas, B. Efficiency optimization in quantum computing: balancing thermodynamics and computational performance.Scientific Reports14(2024). URL https: //www.nature.com/articles/s41598-024-55314-z

  7. [15]

    URL https://doi.org/10.1088/1367-2630/ae6e98

    Doucet, E.et al.Thermodynamic significance of qubo encoding on quantum annealers.New Journal of Physics28, 054512 (2026). URL https://doi.org/10.1088/1367-2630/ae6e98

  8. [16]

    & Campbell, S.Quantum Thermodynamics2053-2571 (Morgan & Claypool Publishers, 2019)

    Deffner, S. & Campbell, S.Quantum Thermodynamics2053-2571 (Morgan & Claypool Publishers, 2019). URL https://doi.org/10.1088/2053-2571/ab21c6

  9. [17]

    Blok, M. S. & Landi, G. T. Quantum thermodynamics for quantum computing.Nature Physics21, 187–188 (2025). URL https://doi.org/10.1038/s41567-024-02764-x

  10. [18]

    Quantum computers, quantum computing, and quantum thermodynamics.Frontiers in Quan- tum Science and TechnologyV olume 3 - 2024(2024)

    Cleri, F. Quantum computers, quantum computing, and quantum thermodynamics.Frontiers in Quan- tum Science and TechnologyV olume 3 - 2024(2024). URL https://www.frontiersin.org/journals/ quantum-science-and-technology/articles/10.3389/frqst.2024.1422257

  11. [19]

    & Perdomo-Ortiz, A

    Benedetti, M., Realpe-G´ omez, J., Biswas, R. & Perdomo-Ortiz, A. Estimation of effective tempera- tures in quantum annealers for sampling applications: A case study with possible applications in deep learning.Physical Review A94, 022308 (2016)

  12. [20]

    & Andriyash, E

    Raymond, J., Yarkoni, S. & Andriyash, E. Global warming: Temperature estimation in annealers. Frontiers in ICT3, 23 (2016)

  13. [21]

    Y., Albash, T

    Nelson, J., Vuffray, M., Lokhov, A. Y., Albash, T. & Coffrin, C. High-quality thermal Gibbs sampling with quantum annealing hardware.Physical Review Applied17, 044046 (2022)

  14. [23]

    Quantum technologies need a quantum energy initiative.PRX Quantum3, 020101 (2022)

    Auff` eves, A. Quantum technologies need a quantum energy initiative.PRX Quantum3, 020101 (2022). URL https://link.aps.org/doi/10.1103/PRXQuantum.3.020101

  15. [24]

    arXiv:2504.20145

    Campbell, S.et al.Roadmap on quantum thermodynamics (2025). arXiv:2504.20145

  16. [25]

    & Brunner, N

    Lipka-Bartosik, P., Perarnau-Llobet, M. & Brunner, N. Thermodynamic computing via autonomous quantum thermal machines.Science advances10, eadm8792 (2024)

  17. [26]

    & Nisoli, C

    Grattan, G., Sathe, P. & Nisoli, C. Classical thermometry of quantum annealers (2025). URL https: //arxiv.org/abs/2512.03162. arXiv:2512.03162. 18

  18. [27]

    PRX Quantum4, 040319 (2023)

    Fellous-Asiani, M.et al.Optimizing resource efficiencies for scalable full-stack quantum computers. PRX Quantum4, 040319 (2023). URL https://link.aps.org/doi/10.1103/PRXQuantum.4.040319

  19. [28]

    Linpeng, X.et al.Quantum energetics of a noncommuting measurement.Phys. Rev. Res.6, 033045 (2024). URL https://link.aps.org/doi/10.1103/PhysRevResearch.6.033045

  20. [29]

    arXiv:2501.17069

    Dassonneville, R.et al.Amplifying microwave pulses with a single qubit engine fueled by quantum measurements (2025). arXiv:2501.17069

  21. [30]

    Linpeng, X.et al.Energetic cost of measurements using quantum, coherent, and thermal light.Phys. Rev. Lett.128, 220506 (2022). URL https://link.aps.org/doi/10.1103/PhysRevLett.128.220506

  22. [31]

    Stevens, J.et al.Energetics of a single qubit gate.Phys. Rev. Lett.129, 110601 (2022). URL https://link.aps.org/doi/10.1103/PhysRevLett.129.110601

  23. [32]

    Quantum7, 1099 (2023)

    Maffei, M.et al.Energy-efficient entanglement generation and readout in a spin-photon interface. Quantum7, 1099 (2023). URL https://hal.science/hal-03687615/

  24. [33]

    & Lidar, D

    Chen, H. & Lidar, D. A. Why and when pausing is beneficial in quantum annealing.Phys. Rev. Appl. 14, 014100 (2020). URL https://link.aps.org/doi/10.1103/PhysRevApplied.14.014100

  25. [34]

    Gonzalez Izquierdo, Z.et al.Advantage of pausing: Parameter setting for quantum annealers.Phys. Rev. Appl.18, 054056 (2022). URL https://link.aps.org/doi/10.1103/PhysRevApplied.18.054056

  26. [35]

    & Jin, F

    Mehta, V., De Raedt, H., Michielsen, K. & Jin, F. Unraveling reverse annealing: A study of d- wave quantum annealers.Phys. Rev. A112, 012414 (2025). URL https://link.aps.org/doi/10.1103/ vfxh-gjgy

  27. [36]

    & Jin, F

    Mehta, V., De Raedt, H., Michielsen, K. & Jin, F. Understanding the physics of d-wave annealers: From schr¨ odinger to lindblad to markovian dynamics.Phys. Rev. A112, 032616 (2025). URL https: //link.aps.org/doi/10.1103/3yk8-qn64

  28. [37]

    & Hasegawa, Y

    Ishida, N. & Hasegawa, Y. Quantum-computer-based verification of quantum thermodynamic uncer- tainty relation.Phys. Rev. E112, 034124 (2025). URL https://link.aps.org/doi/10.1103/9xpz-62cw

  29. [38]

    & Yamasaki, H

    Meier, F. & Yamasaki, H. Energy-consumption advantage of quantum computation.PRX Energy4, 023008 (2025). URL https://link.aps.org/doi/10.1103/PRXEnergy.4.023008

  30. [39]

    Goold, J., Huber, M., Riera, A., Rio, L. d. & Skrzypczyk, P. The role of quantum information in thermodynamics—a topical review.Journal of Physics A: Mathematical and Theoretical49, 143001 (2016). URL http://dx.doi.org/10.1088/1751-8113/49/14/143001

  31. [40]

    Landi, G. T. & Paternostro, M. Irreversible entropy production: From classical to quantum.Rev. Mod. Phys.93, 035008 (2021). URL https://link.aps.org/doi/10.1103/RevModPhys.93.035008

  32. [41]

    & Levy, A

    Kosloff, R. & Levy, A. Quantum heat engines and refrigerators: Continuous devices.Annual Review of Physical Chemistry65, 365–393 (2014). URL https://www.annualreviews.org/content/journals/10. 1146/annurev-physchem-040513-103724

  33. [42]

    The quantum open system as a model of the heat engine.Journal of Physics A12(1979)

    Alicki, R. The quantum open system as a model of the heat engine.Journal of Physics A12(1979). URL https://api.semanticscholar.org/CorpusID:120159113

  34. [43]

    & Whitney, R

    Benenti, G., Casati, G., Saito, K. & Whitney, R. S. Fundamental aspects of steady-state conversion of heat to work at the nanoscale.Physics Reports694, 1–124 (2017). URL https://www.sciencedirect. com/science/article/pii/S0370157317301540. Fundamental aspects of steady-state c...

  35. [44]

    M., Abah, O

    Myers, N. M., Abah, O. & Deffner, S. Quantum thermodynamic devices: From theoretical proposals to experimental reality.A VS Quantum Science4, 027101 (2022). URL https://doi.org/10.1116/5.0083192

  36. [45]

    M., Bhadra, C

    Cangemi, L. M., Bhadra, C. & Levy, A. Quantum engines and refrigerators.Physics Reports1087, 1–71 (2024). URL http://dx.doi.org/10.1016/j.physrep.2024.07.001. 19

  37. [46]

    The one-dimensional ising model with a transverse field.ANNALS of Physics57, 79–90 (1970)

    Pfeuty, P. The one-dimensional ising model with a transverse field.ANNALS of Physics57, 79–90 (1970)

  38. [47]

    & Vicari, E

    Campostrini, M., Pelissetto, A. & Vicari, E. Finite-size scaling at quantum transitions.Phys. Rev. B 89, 094516 (2014). URL https://link.aps.org/doi/10.1103/PhysRevB.89.094516

  39. [48]

    Sachdev, S.Quantum Phase Transitions2 edn (Cambridge University Press, 2011)

  40. [49]

    & Oshikawa, M

    Liu, Y., Shimizu, H., Ueda, A. & Oshikawa, M. Finite-size corrections to the energy spectra of gapless one-dimensional systems in the presence of boundaries.SciPost Physics17(2024). URL http://dx. doi.org/10.21468/SciPostPhys.17.4.099

  41. [50]

    Pang, Y., Coffrin, C., Lokhov, A. Y. & Vuffray, M. The potential of quantum annealing for rapid solution structure identification (2020). URL https://arxiv.org/abs/1912.01759. arXiv:1912.01759

  42. [51]

    & Kendon, V

    Chancellor, N. & Kendon, V. Experimental test of search range in quantum annealing.Physical Review A104, 012604 (2021)

  43. [52]

    & Roy, A

    Boothby, K., Bunyk, P., Raymond, J. & Roy, A. Next-generation topology of d-wave quantum proces- sors. Tech. Rep. 14-1026A-C, D-Wave Systems Inc. (2019). URL https://www.dwavesys.com/media/ jwwj5z3z/14-1026a-c next-generation-topology-of-dw-quantum-processors.pdf

  44. [53]

    Boothby, K., King, A. D. & Raymond, J. Zephyr topology of d-wave quantum processors. Tech. Rep. 14-1056A-A, D-Wave Systems Inc. (2021). URL https://www.dwavesys.com/media/2uznec4s/ 14-1056a-a zephyr topology of d-wave quantum processors.pdf

  45. [54]

    Performance gains in the d-wave advantage2 system at the 4,400-qubit scale

    D-Wave Quantum Inc. Performance gains in the d-wave advantage2 system at the 4,400-qubit scale. Tech. Rep. 14-1083A-A, D-Wave Quantum Inc. (2025). URL https://www.dwavequantum.com/media/ wakjcpsf/adv2 4400q whitepaper-1.pdf

  46. [55]

    Nonequilibrium equality for free energy differences.Phys

    Jarzynski, C. Nonequilibrium equality for free energy differences.Phys. Rev. Lett.78, 2690–2693 (1997). URL https://link.aps.org/doi/10.1103/PhysRevLett.78.2690

  47. [56]

    Crooks, G. E. Entropy production fluctuation theorem and the nonequilibrium work relation for free energy differences.Phys. Rev. E60, 2721–2726 (1999). URL https://link.aps.org/doi/10.1103/ PhysRevE.60.2721

  48. [57]

    & Talkner, P

    Campisi, M., H¨ anggi, P. & Talkner, P. Colloquium: Quantum fluctuation relations: Foundations and applications.Rev. Mod. Phys.83, 771–791 (2011). URL https://link.aps.org/doi/10.1103/ RevModPhys.83.771

  49. [58]

    & Mukamel, S

    Esposito, M., Harbola, U. & Mukamel, S. Nonequilibrium fluctuations, fluctuation theorems, and counting statistics in quantum systems.Rev. Mod. Phys.81, 1665–1702 (2009). URL https://link.aps. org/doi/10.1103/RevModPhys.81.1665

  50. [59]

    Barato, A. C. & Seifert, U. Thermodynamic uncertainty relation for biomolecular processes.Phys. Rev. Lett.114, 158101 (2015). URL https://link.aps.org/doi/10.1103/PhysRevLett.114.158101

  51. [60]

    R., Horowitz, J

    Gingrich, T. R., Horowitz, J. M., Perunov, N. & England, J. L. Dissipation bounds all steady-state current fluctuations.Phys. Rev. Lett.116, 120601 (2016). URL https://link.aps.org/doi/10.1103/ PhysRevLett.116.120601

  52. [61]

    Horowitz, J. M. & Gingrich, T. R. Thermodynamic uncertainty relations constrain non-equilibrium fluctuations.Nature Physics16, 15–20 (2020)

  53. [62]

    & Van Vu, T

    Hasegawa, Y. & Van Vu, T. Uncertainty relations in stochastic processes: An information inequality approach.Phys. Rev. E99, 062126 (2019). URL https://link.aps.org/doi/10.1103/PhysRevE.99. 062126

  54. [63]

    & Perdomo-Ortiz, A

    Benedetti, M., Realpe-G´ omez, J., Biswas, R. & Perdomo-Ortiz, A. Estimation of effective tempera- tures in quantum annealers for sampling applications: A case study with possible applications in deep learning.Phys. Rev. A94, 022308 (2016). URL https://link.aps.org/doi/10.1103...

  55. [64]

    & Ekeberg, M

    Aurell, E. & Ekeberg, M. Inverse ising inference using all the data.Phys. Rev. Lett.108, 090201 (2012). URL https://link.aps.org/doi/10.1103/PhysRevLett.108.090201

  56. [65]

    Annealing implementation and controls

    D-Wave Quantum Inc. Annealing implementation and controls. D-Wave Quantum Computing Products Documentation (2026). URL https://docs.dwavequantum.com/en/latest/quantum research/annealing. html. Accessed 2 August 2026

  57. [66]

    Per-qpu solver properties and schedules

    D-Wave Quantum Inc. Per-qpu solver properties and schedules. D-Wave Quantum Computing Products Documentation (2026). URL https://docs.dwavequantum.com/en/latest/quantum research/ solver properties specific.html. Accessed 2 August 2026

  58. [67]

    D-Wave Ocean SDK Documenta- tion (2026)

    D-Wave Quantum Inc.freezeout effective temperature. D-Wave Ocean SDK Documenta- tion (2026). URL https://docs.dwavequantum.com/en/latest/ocean/api ref system/generated/dwave. system.temperatures.freezeout effective temperature.html. Accessed 2 August 2026. Acknowledgments The ...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.