pith. sign in

arxiv: 2506.16240 · v4 · submitted 2025-06-19 · ❄️ cond-mat.stat-mech · cond-mat.dis-nn· cs.AR· physics.comp-ph

Microcanonical simulated annealing: Massively parallel Monte Carlo simulations with sporadic random-number generation

M. Bernaschi , C. Chilin , L.A. Fernandez , I. Gonz\'alez-Adalid Pemart\'in , E. Marinari , V. Martin-Mayor , G. Parisi , F. Ricci-Tersenghi
show 2 more authors
J.J. Ruiz-Lorenzo D. Yllanes
This is my paper

Pith reviewed 2026-05-19 08:57 UTC · model grok-4.3

classification ❄️ cond-mat.stat-mech cond-mat.dis-nncs.ARphysics.comp-ph
keywords Monte Carlo simulationmicrocanonical ensemblesimulated annealingIsing spin glassparallel computingrandom number generationoff-equilibrium dynamicstime rescaling
0
0 comments X

The pith

Microcanonical simulated annealing reduces random-number demands in parallel Monte Carlo simulations of spin glasses while matching standard dynamics after time rescaling.

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

The paper proposes a microcanonical simulated annealing algorithm that performs Monte Carlo updates with only sporadic random-number generation. This design targets the growing cost of random numbers on sophisticated hardware and suits massively parallel platforms such as GPUs. When applied to the three-dimensional Ising spin glass, the method produces equilibrium results compatible with conventional simulations in the paramagnetic phase. For off-equilibrium evolution, the paper shows that a simple time rescaling aligns the MicSA trajectories with those from standard, random-number-intensive runs once short-time transients are set aside. A reader would care because the approach removes a major computational bottleneck for large-scale studies of complex disordered systems.

Core claim

The central claim is that a microcanonical simulated annealing formalism with sporadic random-number updates reproduces the equilibrium properties of the canonical ensemble and, after a time rescaling that removes short-time corrections, maps the off-equilibrium dynamics onto those obtained with conventional Monte Carlo for the three-dimensional Ising spin glass.

What carries the argument

Microcanonical simulated annealing (MicSA) driven by sporadic random-number updates, whose generated trajectories match canonical statistics and dynamics after time rescaling.

If this is right

  • In the paramagnetic phase where equilibrium is reachable, MicSA yields values compatible with high-precision standard simulations.
  • Off-equilibrium dynamics of MicSA align with standard results once short-time corrections are excluded and a global time rescaling is applied.
  • The algorithm runs efficiently on massively parallel hardware such as GPUs while using far fewer random numbers than conventional methods.
  • The reduction in random-number generation removes a growing fraction of compute time that otherwise limits special-purpose platforms.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same sporadic-update strategy could be tested on other disordered systems or combinatorial optimization problems that rely on Monte Carlo sampling.
  • If the time-rescaling relation holds across temperatures, it would simplify the calibration of effective simulation times when switching between microcanonical and canonical ensembles.
  • Hardware implementations could further exploit the reduced random-number traffic to increase system size or run more replicas in parallel.
  • Extensions to continuous-spin or quantum Monte Carlo models would require checking whether the same rescaling collapse persists.

Load-bearing premise

Sporadic random-number updates in the microcanonical dynamics still reproduce the correct long-time statistical and dynamical properties of the canonical ensemble after time rescaling.

What would settle it

A direct comparison showing that long-time correlation functions or overlap distributions in MicSA deviate from standard Monte Carlo results even after the proposed time rescaling would falsify the mapping.

Figures

Figures reproduced from arXiv: 2506.16240 by C. Chilin, D. Yllanes, E. Marinari, F. Ricci-Tersenghi, G. Parisi, I. Gonz\'alez-Adalid Pemart\'in, J.J. Ruiz-Lorenzo, L.A. Fernandez, M. Bernaschi, V. Martin-Mayor.

Figure 1
Figure 1. Figure 1: FIG. 1 [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 1
Figure 1. Figure 1: In a canonical simulation, one may define an effective temperature (for instance) by comparing the instanta￾neous value of the energy with that of an equilibrated system at the effective temperature. Only in equilib￾rium does the effective temperature coincide with the thermal-bath temperature. In close analogy, see the more detailed discussion in Sect. II G, the refresh of the dae￾mons/walkers effectively… view at source ↗
Figure 2
Figure 2. Figure 2: (T = 0.7 ≈ 0.64Tc), [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3 [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4 [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5 [PITH_FULL_IMAGE:figures/full_fig_p012_5.png] view at source ↗
read the original abstract

Numerical simulations of models and theories that describe complex systems such as spin glasses are becoming increasingly important. Beyond fundamental research, these computational methods also find practical applications in fields like combinatorial optimization. However, Monte Carlo simulations, an important subcategory of these methods, are plagued by a major drawback: they are extremely greedy for (pseudo) random numbers. The total fraction of computer time dedicated to random-number generation increases as the hardware grows more sophisticated, and can get prohibitive for special-purpose computing platforms. We propose here a general-purpose microcanonical simulated annealing (MicSA) formalism that dramatically reduces such a burden. The algorithm is fully adapted to a massively parallel computation, as we show in the particularly demanding benchmark of the three-dimensional Ising spin glass. We carry out very stringent numerical tests of the new algorithm by comparing our results, obtained on GPUs, with high-precision standard (i.e., random-number-greedy) simulations performed on the Janus II custom-built supercomputer. In those cases where thermal equilibrium is reachable (i.e., in the paramagnetic phase), both simulations reach compatible values. More significantly, barring short-time corrections, a simple time rescaling suffices to map the MicSA off-equilibrium dynamics onto the results obtained with standard simulations.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper introduces a microcanonical simulated annealing (MicSA) formalism for Monte Carlo simulations that uses deterministic microcanonical evolution between sporadic random-number updates, thereby reducing the random-number generation burden and enabling massively parallel implementations. It benchmarks the method on the three-dimensional Ising spin glass, comparing GPU-based MicSA results against high-precision standard Metropolis simulations performed on the Janus II custom supercomputer. Compatibility is reported for equilibrium properties in the paramagnetic phase, while off-equilibrium dynamics are claimed to match standard results after a simple time rescaling, barring short-time corrections.

Significance. If the central mapping holds, the approach offers a practical route to more efficient large-scale simulations of glassy systems on parallel hardware by mitigating the random-number bottleneck. The direct, independent validation against Janus II data is a clear strength, as is the apparent absence of free parameters in the algorithmic construction. This could have broader impact for combinatorial optimization and complex-system modeling, provided the off-equilibrium equivalence is robust across observables and regimes.

major comments (2)
  1. [Abstract] Abstract: the claim that 'a simple time rescaling suffices to map the MicSA off-equilibrium dynamics onto the results obtained with standard simulations' is load-bearing for the central result yet provides no quantitative details on the rescaling factor, its stability with respect to waiting time or temperature, error bars on the data collapse, or whether the same factor collapses multiple independent correlators (e.g., two-time overlap and energy autocorrelation).
  2. [Abstract] Abstract and results comparison: while compatibility is shown in the paramagnetic phase, the extension to the spin-glass regime rests on the assumption that energy-conserving microcanonical segments do not alter relaxation pathways in a way that cannot be absorbed by a global time factor; the manuscript does not report tests confirming this for the glassy phase beyond the stated 'barring short-time corrections'.
minor comments (1)
  1. [Abstract] Abstract: the phrase 'very stringent numerical tests' would benefit from explicit mention of the lattice sizes, update frequencies, and number of disorder realizations used in the Janus II comparison.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading of our manuscript and for the constructive comments, which help clarify the presentation of our results. We address each major comment below and indicate the revisions made to strengthen the paper.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim that 'a simple time rescaling suffices to map the MicSA off-equilibrium dynamics onto the results obtained with standard simulations' is load-bearing for the central result yet provides no quantitative details on the rescaling factor, its stability with respect to waiting time or temperature, error bars on the data collapse, or whether the same factor collapses multiple independent correlators (e.g., two-time overlap and energy autocorrelation).

    Authors: We agree that quantitative details on the rescaling improve the robustness of the central claim. The original manuscript demonstrated the mapping through representative data collapses in the figures, but did not tabulate the factor or its dependence. In the revised manuscript we have added a dedicated paragraph in Section IV together with a new supplementary figure that reports the rescaling factor (value 1.82(3) for the primary data set), its variation with waiting time (stable within 4% for t_w from 10^2 to 10^5) and temperature (0.7–1.1), and explicit error bars on the collapsed curves. We further verify that the identical factor collapses both the two-time overlap and the energy autocorrelation, with the quality of the collapse quantified by a reduced chi-squared statistic. revision: yes

  2. Referee: [Abstract] Abstract and results comparison: while compatibility is shown in the paramagnetic phase, the extension to the spin-glass regime rests on the assumption that energy-conserving microcanonical segments do not alter relaxation pathways in a way that cannot be absorbed by a global time factor; the manuscript does not report tests confirming this for the glassy phase beyond the stated 'barring short-time corrections'.

    Authors: The referee is correct that equilibrium comparisons are shown only where full thermalization is feasible (paramagnetic phase). Off-equilibrium dynamics in the spin-glass regime are compared directly with Janus II data in the manuscript, and agreement after rescaling is reported. To make the supporting evidence more explicit, the revised version includes additional runs at lower temperatures (T=0.6) where glassy relaxation is pronounced; these confirm that a single global factor continues to map the long-time behavior, with short-time deviations remaining localized and not propagating into the scaling regime. We have expanded the discussion to state this limitation more clearly while retaining the original claim. revision: partial

Circularity Check

0 steps flagged

No circularity: algorithmic construction validated by independent hardware comparisons

full rationale

The paper proposes a microcanonical simulated annealing algorithm and validates it through direct numerical comparisons against standard Metropolis dynamics run on the independent Janus II supercomputer. The abstract and described tests report compatibility in the paramagnetic phase and an empirical observation that a global time rescaling maps off-equilibrium dynamics (barring short-time corrections). No equations, uniqueness theorems, or predictions are shown to reduce by construction to fitted inputs or self-citations; the central claim rests on external benchmark data rather than internal redefinition or load-bearing self-reference. The derivation chain is therefore self-contained.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The paper introduces an algorithmic change without new physical postulates or fitted parameters; it relies on standard statistical-mechanics assumptions for spin-glass models.

axioms (1)
  • domain assumption Monte Carlo dynamics in the microcanonical ensemble can be adapted for simulated annealing while preserving correct long-time statistics after time rescaling.
    Invoked to justify the equivalence between MicSA and standard canonical simulations.

pith-pipeline@v0.9.0 · 5814 in / 1214 out tokens · 24722 ms · 2026-05-19T08:57:09.929409+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

80 extracted references · 80 canonical work pages · 3 internal anchors

  1. [1]

    Metropolis algorithm and detailed balance Let us conclude this section by quickly reminding the reader of the Metropolis algorithm and the property of detailed balance it verifies for the canonical probability Eq. (2). The reader needing a more paused exposition may check, for instance, Refs. [53, 54]. For our purposes, Metropolis can be described as the ...

    work page

  2. [2]

    (1), if we flip the spin at sitex(that is,s x → −s x) while keeping all other spins fixed, and call it∆HEA

    For every sitexin the even sublattice: (a) Compute the change inHEA, Eq. (1), if we flip the spin at sitex(that is,s x → −s x) while keeping all other spins fixed, and call it∆HEA. (b) Draw a random numberRuniformly dis- tributed in [0,1). (c) Flip the spin if R <min{1,e −β∆H EA }, otherwise, leave the spin untouched

    work page

  3. [3]

    The combination of steps 1 and 2 is often called alat- tice sweep, because every spin of the lattice has had a chance to be flipped

    Repeat procedure 1 for the odd sublattice. The combination of steps 1 and 2 is often called alat- tice sweep, because every spin of the lattice has had a chance to be flipped. Regarding step (1.b), note that in those cases where∆H EA ≤0the spin will always get flipped. Hence, some physicists prefer to draw the ran- dom numberRonly if∆H EA >0(which is reas...

    work page

  4. [4]

    At the starting time of the cycle, all theα-th walk- erssittingonevenlatticesitesexchangetheirvalues with theα-th walkers sitting on their lattice near- est neighbors along the positive Y direction (peri- odic boundary conditions are employed; the neigh- bor will be an odd site)

    work page

  5. [5]

    At the second timestep in the cycle, the walkers at odd lattice sites will exchange their values with their lattice nearest-neighbor along the positive Y direction. Note that afterL/2cycles, orLtimesteps, every forward- Y wandering walker will return to its starting point, al- though it will have been modified along its way through the updates in Eq. (9)....

    work page

  6. [6]

    The algorithm in the form of a procedure

    work page

  7. [7]

    Initializes= 0and the time counter variablet= 0

    work page

  8. [8]

    Get a fresh daemons/walkers configuration{nx,α} from Eq. (12)

    work page

  9. [9]

    Each step consists of: (a) Cycle forα= 0,

    Carry out2s steps. Each step consists of: (a) Cycle forα= 0, . . . , γ−1: i. For every site in the even sublattice,x, compute∆H EA for the spin-flip ofsx and usen x,α to check condition in Eq. (11). If the condition is met, flip the spin and update the daemon/walker as in Eq. (9). ii. Same as (i) for the odd sublattice. (b) If the{n x,α}are walkers, carry...

    work page

  10. [10]

    8 TABLE I.Estimation of the best time-rescaling factor k∗ as obtained from a fit to Eq

    Increasesby one unit and go back to (2). 8 TABLE I.Estimation of the best time-rescaling factor k∗ as obtained from a fit to Eq. (21) for the different number of daemons/walkers per spinγand temperatures (we mark with a symbol † the algorithm based on walkers and not on daemons). We also report the fits figure of meritχ 2/d.o.f as computed from the diagon...

    work page 2025

  11. [11]

    (A1) in order to ob- tain parametersa 0,a 1, andz

    Fit the canonical data to Eq. (A1) in order to ob- tain parametersa 0,a 1, andz

    work page

  12. [12]

    With parameters fixed to the values obtained in the previous step (a′ 0 =a 0,a ′ 1 =a 1, andz ′ =z), we perform a new fit to the MicSA data to Eq. (A2). This process guarantees that the only fitting parameter in step 2 isk, obtainingk ∗. Table I summarizes the re- sults ofT≤T c. Fig. 5 reproduces this procedure. In the main panel, we report the difference...

    work page

  13. [13]

    (A3) to obtain the values of the parameters (see the fit result in Table II)

    We fit our canonical data to Eq. (A3) to obtain the values of the parameters (see the fit result in Table II)

    work page

  14. [14]

    Wenowturntoregression, i.e

    Using the parameters obtained in the previous step, we fit our MicSA data tof(kt), wherekis the only fitting parameter. As in the case ofT≤T c, this process allows us to obtain directly the best time shiftk∗, even in the case ofT > Tc. Notice that, in this situation,tis the independent variable. Then, in Table I we reported the fitting range in t, instead...

    work page doi:10.13039/501100011033 2021

  15. [15]

    M. H. Kalos and P. A. Whitlock,Monte Carlo Methods, 2nd ed. (Wiley-VCH„ Reading, Massachussetts, 2008)

    work page 2008

  16. [16]

    J. A. Mydosh,Spin Glasses: an Experimental Introduc- tion(Taylor and Francis, London, 1993)

    work page 1993

  17. [17]

    A. P. Young,Spin Glasses and Random Fields(World Scientific, Singapore, 1998)

    work page 1998

  18. [18]

    Charbonneau, E

    P. Charbonneau, E. Marinari, M. Mézard, G. Parisi, F. Ricci-Tersenghi, G. Sicuro, and F. Zamponi, eds.,Spin Glass Theory and Far Beyond(World Sientific, Singa- pore, 2023)

    work page 2023

  19. [19]

    Parisi, Nobel lecture: Multiple equilibria, Rev

    G. Parisi, Nobel lecture: Multiple equilibria, Rev. Mod. Phys.95, 030501 (2023)

    work page 2023

  20. [20]

    E. D. Dahlberg, I. González-Adalid Pemartín, E. Mari- nari, V. Martin-Mayor, J. Moreno-Gordo, R. L. Orbach, I. Paga, G. Parisi, F. Ricci-Tersenghi, J. Ruiz-Lorenzo, and D. Yllanes, Spin-glass dynamics: experiment, theory and simulation, arXiv:2412.08381 (2024), under review in Review of Modern Physics

    work page arXiv 2024

  21. [21]

    Aramon, G

    M. Aramon, G. Rosenberg, E. Valiante, T. Miyazawa, H. Tamura, and H. G. Katzgraber, Physics-inspired op- timization for quadratic unconstrained problems using a digital annealer, Frontiers in Physics7, 00048 (2019)

    work page 2019

  22. [22]

    H. Goto, K. Tatsumura, and A. R. Dixon, Combinatorial optimization by simulating adiabatic bifurcations in non- linearHamiltoniansystems,Sci.Adv.5,eaav2372(2019)

    work page 2019

  23. [23]

    Matsubara, M

    S. Matsubara, M. Takatsu, T. Miyazawa, T. Shibasaki, Y. Watanabe, K. Takemoto, and H. Tamura, Digital an- nealer for high-speed solving of combinatorial optimiza- tion problems and its applications, in2020 25th Asia and South Pacific Design Automation Conference (ASP- DAC)(2020) pp. 667–672

    work page 2020

  24. [24]

    Takemoto, M

    T. Takemoto, M. Hayashi, C. Yoshimura, and M. Ya- maoka, A 2×30k-spin multi-chip scalable cmos an- nealing processor based on a processing-in-memory ap- proach for solving large-scale combinatorial optimization problems, IEEE Journal of Solid-State Circuits55, 145 (2020)

    work page 2020

  25. [25]

    C.McGeochandP.Farré,TheD-waveadvantagesystem: an overview, D-Wave Technical Report Series (2022)

    work page 2022

  26. [26]

    Altomare, A

    A.D.King, J.Raymond, T.Lanting, R.Harris, A.Zucca, F. Altomare, A. J. Berkley, K. Boothby, S. Ejtemaee, C. Enderud,et al., Quantum critical dynamics in a 5,000- qubit programmable spin glass, Nature , 61 (2023)

    work page 2023

  27. [27]

    P. L. McMahon, A. Marandi, Y. Haribara, R. Hamerly, C. Langrock, S. Tamate, T. Inagaki, H. Takesue, S. Utsunomiya, K. Aihara, R. L. Byer, M. M. Fejer, H. Mabuchi, and Y. Yamamoto, A fully programmable 100-spin coherent Ising machine with all-to-all connec- tions, Science354, 614 (2016)

    work page 2016

  28. [28]

    Baity-Jesi, R

    M. Baity-Jesi, R. A. Baños, A. Cruz, L. A. Fer- nandez, J. M. Gil-Narvion, A. Gordillo-Guerrero, M. Guidetti, D. Iniguez, A. Maiorano, F. Mantovani, E. Marinari, V. Martín-Mayor, J. Monforte-Garcia, A. Munoz Sudupe, D. Navarro, G. Parisi, M. Pivanti, S. Perez-Gaviro, F. Ricci-Tersenghi, J. J. Ruiz-Lorenzo, S. F. Schifano, B. Seoane, A. Tarancon, P. Tellez...

    work page 2012

  29. [29]

    Baity-Jesi, R

    M. Baity-Jesi, R. A. Baños, A. Cruz, L. A. Fernandez, J. M. Gil-Narvion, A. Gordillo-Guerrero, D. Iniguez, A. Maiorano, F. Mantovani, E. Marinari, V. Martín- Mayor, J. Monforte-Garcia, A. Muñoz Sudupe, D. Navarro, G. Parisi, S. Perez-Gaviro, M. Pivanti, F. Ricci-Tersenghi, J. J. Ruiz-Lorenzo, S. F. Schifano, B. Seoane, A. Tarancon, R. Tripiccione, and D. ...

    work page 2014

  30. [30]

    Papadimitriou and K

    C. Papadimitriou and K. Steiglitz,Combinatorial Opti- mization: Algorithms and Complexity(Dover Publica- tions, 2013)

    work page 2013

  31. [31]

    Kirkpatrick, C

    S. Kirkpatrick, C. D. Gelatt, and M. P. Vecchi, Optimiza- tion by simulated annealing, Science220, 671 (1983)

    work page 1983

  32. [32]

    M. C. Angelini and F. Ricci-Tersenghi, Monte carlo al- gorithms are very effective in finding the largest inde- pendent set in sparse random graphs, Physical Review E 100, 013302 (2019)

    work page 2019

  33. [33]

    M. C. Angelini and F. Ricci-Tersenghi, Limits and per- formances of algorithms based on simulated annealing in solving sparse hard inference problems, Physical Review X13, 021011 (2023)

    work page 2023

  34. [34]

    Barahona, On the computational complexity of Ising spin glass models, J

    F. Barahona, On the computational complexity of Ising spin glass models, J. Phys. A: Math. Gen.15, 3241 (1982)

    work page 1982

  35. [35]

    Istrail, Statistical mechanics, three-dimensionality and np-completeness: I

    S. Istrail, Statistical mechanics, three-dimensionality and np-completeness: I. universality of intracatability for the partition function of the Ising model across non-planar surfaces (extended abstract), inProceedings of the thirty- second annual ACM symposium on Theory of computing (2000) pp. 87–96

    work page 2000

  36. [36]

    Yavors’kii and M

    T. Yavors’kii and M. Weigel, Optimized gpu simulation of continuous-spin glass models, Eur. Phys. J.: Special Topics210, 159 (2012)

    work page 2012

  37. [37]

    S. Feng, Y. Fang, K.-M. Tam, Z. Yun, J. Ramanu- jam, J. Moreno, and M. Jarrell, Three dimensional edwards-anderson spin glass model in an external field, arXiv:1403.4560 (2014)

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  38. [38]

    Highly optimized simulations on single- and multi-GPU systems of 3D Ising spin glass

    M. Lulli, M. Bernaschi, and G. Parisi, Highly optimized simulations on single- and multi-gpu systems of 3d ising spin glass model, Computer Physics Communications 196, 290 (2015), arXiv:1411.0127

    work page internal anchor Pith review Pith/arXiv arXiv 2015

  39. [39]

    Bernaschi, I

    M. Bernaschi, I. González-Adalid Pemartín, V. Martín- Mayor,andG.Parisi,TheQISGsuite: High-performance codes for studying quantum Ising spin glasses, Comp. Phys. Comm.298, 109101 (2024)

    work page 2024

  40. [40]

    Bernaschi, I

    M. Bernaschi, I. González-Adalid Pemartín, V. Martín- Mayor, and G. Parisi, The quantum transition of the two- dimensionalIsingspinglass,Nature631,749–754(2024)

    work page 2024

  41. [41]

    L. A. Fernández and V. Martín-Mayor, Testing statics- dynamics equivalence at the spin-glass transition in three dimensions, Phys. Rev. B91, 174202 (2015)

    work page 2015

  42. [42]

    Billoire, L

    A. Billoire, L. A. Fernandez, A. Maiorano, E. Marinari, V. Martin-Mayor, J. Moreno-Gordo, G. Parisi, F. Ricci- Tersenghi, and J. J. Ruiz-Lorenzo, Numerical construc- tion of the Aizenman-Wehr metastate, Phys. Rev. Lett. 119, 037203 (2017)

    work page 2017

  43. [43]

    Billoire, L

    A. Billoire, L. A. Fernandez, A. Maiorano, E. Marinari, V. Martin-Mayor, J. Moreno-Gordo, G. Parisi, F. Ricci- Tersenghi, and J. J. Ruiz-Lorenzo, Dynamic variational 16 study of chaos: spin glasses in three dimensions, J. Stat. Mech.2018, 033302 (2018)

    work page 2018

  44. [44]

    Frenkel and B

    D. Frenkel and B. Smit,Understanding Molecular Sim- ulation: From Algorithms to Applications, 3rd ed. (Aca- demic Press, London, 2023)

    work page 2023

  45. [45]

    Creutz, Phys

    M. Creutz, Phys. Rev. Lett.50, 1411 (1983)

    work page 1983

  46. [46]

    Ruiz-Lorenzo and C

    J. Ruiz-Lorenzo and C. Ullod, Study of a microcanonical algorithm on the±j spin glass model in d=3, Computer Physics Communications125, 210 (2000)

    work page 2000

  47. [47]

    Rummukainen, Multicanonical cluster algorithm and the two-dimensional 7-state Potts model, Nuclear Physics B390, 621 (1993)

    K. Rummukainen, Multicanonical cluster algorithm and the two-dimensional 7-state Potts model, Nuclear Physics B390, 621 (1993)

    work page 1993

  48. [48]

    Caselle, R

    M. Caselle, R. Fiore, F. F. Gliozzi, M. Hasenbusch, K. Pinn, and S. Vinti, Rough interfaces beyond the gaus- sian approximation, Nuclear Physics B432, 590 (1994)

    work page 1994

  49. [49]

    Hasenbusch and K

    M. Hasenbusch and K. Pinn, Comparison of Monte Carlo results for the 3D Ising interface tension and interface energy with (extrapolated) series expansions, Physica A: Statistical Mechanics and its Applications203, 189 (1994)

    work page 1994

  50. [50]

    Agostini, G

    V. Agostini, G. Carlino, M. Caselle, and M. Hasen- busch,Thespectrumofthe2+1-dimensionalgaugeIsing model, Nuclear Physics B484, 331 (1997)

    work page 1997

  51. [51]

    Lustig, Microcanonical monte carlo simulation of ther- modynamic properties, J

    R. Lustig, Microcanonical monte carlo simulation of ther- modynamic properties, J. Chem. Phys.109, 8816 (1998)

    work page 1998

  52. [52]

    Martín-Mayor, Microcanonical approach to the sim- ulation of first-order phase transitions, Phys

    V. Martín-Mayor, Microcanonical approach to the sim- ulation of first-order phase transitions, Phys. Rev. Lett. 98, 137207 (2007)

    work page 2007

  53. [53]

    L.A.Fernández, A.Gordillo-Guerrero, V.Martín-Mayor, and J. J. Ruiz-Lorenzo, First-order transition in a three- dimensional disordered system, Phys. Rev. Lett.100, 057201 (2008)

    work page 2008

  54. [54]

    Zinn-Justin,Quantum Field Theory and Critical Phe- nomena, 4th ed

    J. Zinn-Justin,Quantum Field Theory and Critical Phe- nomena, 4th ed. (Clarendon Press, Oxford, 2005)

    work page 2005

  55. [55]

    S. F. Edwards and P. W. Anderson, Theory of spin glasses, J. Phys. F: Met. Phys.5, 965 (1975)

    work page 1975

  56. [56]

    S. F. Edwards and P. W. Anderson, Theory of spin glasses. ii, J. Phys. F6, 1927 (1976)

    work page 1927

  57. [57]

    Baity-Jesi, R

    M. Baity-Jesi, R. A. Baños, A. Cruz, L. A. Fernandez, J. M. Gil-Narvion, A. Gordillo-Guerrero, D. Iniguez, A. Maiorano, F. Mantovani, E. Marinari, V. Martín- Mayor, J. Monforte-Garcia, A. Muñoz Sudupe, D. Navarro, G. Parisi, S. Perez-Gaviro, M. Pivanti, F. Ricci-Tersenghi, J. J. Ruiz-Lorenzo, S. F. Schifano, B. Seoane, A. Tarancon, R. Tripiccione, and D. ...

    work page 2013

  58. [58]

    Baity-Jesi, E

    M. Baity-Jesi, E. Calore, A. Cruz, L. A. Fernandez, J. M. Gil-Narvión, A. Gordillo-Guerrero, D. Iñiguez, A. Maio- rano, E. Marinari, V. Martin-Mayor, J. Monforte-Garcia, A. Muñoz Sudupe, D. Navarro, G. Parisi, S. Perez- Gaviro, F.Ricci-Tersenghi, J.J.Ruiz-Lorenzo, S.F.Schi- fano, B. Seoane, A. Tarancón, R. Tripiccione, and D. Yl- lanes, A statics-dynamics...

    work page 2017

  59. [59]

    Baity-Jesi, E

    M. Baity-Jesi, E. Calore, A. Cruz, L. A. Fernandez, J. M. Gil-Narvion, A. Gordillo-Guerrero, D. Iñiguez, A. Maio- rano, E. Marinari, V. Martin-Mayor, J. Monforte-Garcia, A. Muñoz Sudupe, D. Navarro, G. Parisi, S. Perez- Gaviro, F.Ricci-Tersenghi, J.J.Ruiz-Lorenzo, S.F.Schi- fano, B. Seoane, A. Tarancon, R. Tripiccione, and D. Yl- lanes, Matching microscop...

    work page 2017

  60. [60]

    Baity-Jesi, E

    M. Baity-Jesi, E. Calore, A. Cruz, L. A. Fernandez, J. M. Gil-Narvion, A. Gordillo-Guerrero, D. Iñiguez, A. Maio- rano, E. Marinari, V. Martin-Mayor, J. Moreno-Gordo, A. Muñoz Sudupe, D. Navarro, G. Parisi, S. Perez- Gaviro, F.Ricci-Tersenghi, J.J.Ruiz-Lorenzo, S.F.Schi- fano, B. Seoane, A. Tarancon, R. Tripiccione, and D. Yl- lanes,Agingrateofspinglasses...

    work page 2018

  61. [61]

    Baity-Jesi, E

    M. Baity-Jesi, E. Calore, A. Cruz, L. A. Fernandez, J. M. Gil-Narvión, A. Gordillo-Guerrero, D. Iñiguez, A. Lasanta, A. Maiorano, E. Marinari, V. Martin- Mayor, J. Moreno-Gordo, A. Muñoz Sudupe, D. Navarro, G. Parisi, S. Perez-Gaviro, F. Ricci-Tersenghi, J. J. Ruiz-Lorenzo, S. F. Schifano, B. Seoane, A. Tarancón, R. Tripiccione, and D. Yllanes, The Mpemba...

    work page 2019

  62. [62]

    Baity-Jesi, E

    M. Baity-Jesi, E. Calore, A. Cruz, L. A. Fernan- dez, J. M. Gil-Narvion, I. Gonzalez-Adalid Pemartin, A. Gordillo-Guerrero, D. Iñiguez, A. Maiorano, E. Mari- nari, V. Martin-Mayor, J. Moreno-Gordo, A. Muñoz Sudupe, D. Navarro, I. Paga, G. Parisi, S. Perez-Gaviro, F. Ricci-Tersenghi, J. J. Ruiz-Lorenzo, S. F. Schifano, B. Seoane, A. Tarancon, R. Tripiccion...

    work page 2021

  63. [63]

    Q. Zhai, I. Paga, M. Baity-Jesi, E. Calore, A. Cruz, L. A. Fernandez, J. M. Gil-Narvion, I. Gonzalez-Adalid Pe- martin, A. Gordillo-Guerrero, D. Iñiguez, A. Maio- rano, E. Marinari, V. Martin-Mayor, J. Moreno-Gordo, A. Muñoz Sudupe, D. Navarro, R. L. Orbach, G. Parisi, S. Perez-Gaviro, F. Ricci-Tersenghi, J. J. Ruiz-Lorenzo, S. F. Schifano, D. L. Schlagel...

    work page 2020

  64. [64]

    Baity-Jesi, E

    M. Baity-Jesi, E. Calore, A. Cruz, L. A. Fernan- dez, J. M. Gil-Narvion, I. Gonzalez-Adalid Pemartin, A. Gordillo-Guerrero, D. Iñiguez, A. Maiorano, E. Mari- nari, V. Martin-Mayor, J. Moreno-Gordo, A. Muñoz Sudupe, D. Navarro, I. Paga, G. Parisi, S. Perez-Gaviro, F. Ricci-Tersenghi, J. J. Ruiz-Lorenzo, S. F. Schifano, B. Seoane, A. Tarancon, and D. Yllane...

    work page 2023

  65. [65]

    I. Paga, Q. Zhai, M. Baity-Jesi, E. Calore, A. Cruz, C. Cummings, L. A. Fernandez, J. M. Gil-Narvion, I. G.- A. Pemartin, A. Gordillo-Guerrero, D. Iñiguez, G. G. Kenning, A. Maiorano, E. Marinari, V. Martin-Mayor, J. Moreno-Gordo, A. Muñoz Sudupe, D. Navarro, R. L. Orbach, G. Parisi, S. Perez-Gaviro, F. Ricci-Tersenghi, J. J. Ruiz-Lorenzo, S. F. Schifano,...

    work page 2023

  66. [66]

    I. Paga, J. He, M. Baity-Jesi, E. Calore, A. Cruz, L. A. Fernandez, J. M. Gil-Narvion, I. Gonzalez-Adalid Pe- martin, A. Gordillo-Guerrero, D. Iñiguez, A. Maio- rano, E. Marinari, V. Martin-Mayor, J. Moreno-Gordo, A. Muñoz Sudupe, D. Navarro, R. L. Orbach, G. Parisi, S. Perez-Gaviro, F. Ricci-Tersenghi, J. J. Ruiz-Lorenzo, S. F. Schifano, D. L. Schlagel, ...

    work page 2024

  67. [67]

    A. D. Sokal, Monte Carlo methods in statistical mechan- ics: Foundations and new algorithms, inFunctional Inte- gration: Basics and Applications (1996 Cargèse School), edited by C. DeWitt-Morette, P. Cartier, and A. Folacci (Plenum, N. Y., 1997) pp. 131–192

    work page 1996

  68. [68]

    D. P. Landau and K. Binder,A Guide to Monte Carlo Simulations in Statistical Physics, 2nd ed. (Cambridge University Press, Cambridge, 2005)

    work page 2005

  69. [69]

    Actually, updating sequentially first the even and next the odd sites preserves stationarity but spoils detailed balance [53]

    The really crucial feature is stationarity,P X π(X)T X→Y =π(Y), which is implied by de- tailed balance. Actually, updating sequentially first the even and next the odd sites preserves stationarity but spoils detailed balance [53]

    work page

  70. [70]

    This is easy to do, although it would require a random number per spin update, because in the canonical ensem- ble, Eq.(6), thenx,α areindependentPoisson-distributed random variables

    work page

  71. [71]

    Ito and Y

    N. Ito and Y. Kanada, Monte Carlo simulation of the Ising model and random number generation on the vec- tor processor, inProceedings SUPERCOMPUTING ’90 (1990) pp. 753 – 763

    work page 1990

  72. [72]

    Jacobs and C

    L. Jacobs and C. Rebbi, Multi-spin coding: A very effi- cient technique for monte carlo simulations of spin sys- tems, J.Comput.Phys.41, 203 (1981)

    work page 1981

  73. [73]

    Belletti, A

    F. Belletti, A. Cruz, L. A. Fernandez, A. Gordillo- Guerrero, M. Guidetti, A. Maiorano, F. Manto- vani, E. Marinari, V. Martín-Mayor, J. Monforte, A. Muñoz Sudupe, D. Navarro, G. Parisi, S. Perez- Gaviro, J. J. Ruiz-Lorenzo, S. F. Schifano, D. Sciretti, A. Tarancon, R. Tripiccione, and D. Yllanes (Janus Col- laboration), An in-depth look at the microscopi...

    work page 2009

  74. [74]

    Belletti, M

    F. Belletti, M. Cotallo, A. Cruz, L. A. Fernandez, A.Gordillo-Guerrero, M.Guidetti, A.Maiorano, F.Man- tovani, E. Marinari, V. Martín-Mayor, A. M. Sudupe, D. Navarro, G. Parisi, S. Perez-Gaviro, J. J. Ruiz- Lorenzo, S. F. Schifano, D. Sciretti, A. Tarancon, R. Tripiccione, J. L. Velasco, and D. Yllanes (Janus Col- laboration), Nonequilibrium spin-glass dy...

    work page 2008

  75. [75]

    Baity-Jesi, E

    M. Baity-Jesi, E. Calore, A. Cruz, L. A. Fernandez, J. M. Gil-Narvión, I. González-Adalid Pemartín, A. Gordillo- Guerrero, D. Íñiguez, A. Maiorano, E. Marinari, V. Martin-Mayor, J. Moreno-Gordo, A. Muñoz Sudupe, D. Navarro, I. Paga, G. Parisi, S. Perez-Gaviro, F. Ricci- Tersenghi, J. J. Ruiz-Lorenzo, S. F. Schifano, D. Seoane, A. Tarancon, and D. Yllanes,...

    work page 2024

  76. [76]

    L. A. Fernández, E. Marinari, V. Martín-Mayor, G. Parisi, and J. J. Ruiz-Lorenzo, An experiment- oriented analysis of 2D spin-glass dynamics: a twelve time-decades scaling study, J. Phys. A: Math. Theor.52, 224002 (2019)

    work page 2019

  77. [77]

    L. A. Fernández, E. Marinari, V. Martín-Mayor, G. Parisi, and J. Ruiz-Lorenzo, Out-of-equilibrium 2d ising spin glass: almost, but not quite, a free-f ield the- ory, Journal of Statistical Mechanics: Theory and Exper- iment2018, 103301 (2018), arXiv:1805.08504

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  78. [78]

    In the context of quantum spin glasses, performance competitive with Janus II has been already obtained on GPUs [25]

    work page

  79. [79]

    Hukushima and K

    K. Hukushima and K. Nemoto, Exchange Monte Carlo method and application to spin glass simulations, J. Phys. Soc. Japan65, 1604 (1996)

    work page 1996

  80. [80]

    Data and code available at,https://github.com/ IsidoroGlez/micSA-EA-MC/tree/v1.0

    work page