Rejection-free Glauber Monte Carlo for the 2D Random Field Ising Model via Hierarchical Probabilistic Counters

Ezio Puppin; Federico Ettori; Giovanni Cerri; Luca Cattaneo; Paolo Biscari

Rejection-free Glauber Monte Carlo for the 2D Random Field Ising Model via Hierarchical Probabilistic Counters

T0 review · 3 major / 2 minor · reviewed 2026-07-13 · grok-4.5

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

Pith's one-line read A rejection-free Glauber Monte Carlo for the 2D random-field Ising model selects spins in O(log N) with hierarchical probabilistic counters, speeding low-temperature dynamics by over 100× while staying faithful to continuous-time evolution.

desk verdict Useful methods claim for 2D RFIM dynamics—rejection-free Glauber via hierarchical counters—but the abstract alone cannot verify the rate-equivalence that makes the speedups fair. read the letter →

arxiv 2603.17032 v1 pith:WC6Z2PJL submitted 2026-03-17 cond-mat.stat-mech cond-mat.dis-nnphysics.comp-ph

Luca Cattaneo , Federico Ettori , Giovanni Cerri , Paolo Biscari , Ezio Puppin This is my paper · ORCID

classification cond-mat.stat-mechcond-mat.dis-nnphysics.comp-ph PACS 05.10.Ln05.50.+q75.10.Nr75.40.Mg

keywords random-fieldIsingmodelGlauberdynamicsrejection-freeMonteCarloBortz-Kalos-Lebowitzhierarchicalcountersevent-drivensimulationdisorderedspinsystemscriticalslowingdown

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 introduces a Monte Carlo algorithm for the two-dimensional random-field Ising model that is both rejection-free and event-driven under Glauber rates. Classical Metropolis sampling becomes painfully slow at low temperature because most proposed flips are rejected; the classic Bortz–Kalos–Lebowitz method avoids rejections for the pure Ising model by grouping spins into energy classes, but random fields destroy those classes. The authors restore rejection-free, continuous-time dynamics by replacing the energy-class bookkeeping with hierarchical probabilistic counters that pick the next spin to flip in O(log N) operations. The resulting scheme therefore remains dynamically faithful even when quenched disorder is present. In the low-temperature, low-disorder regime the method yields wall-clock speed-ups of more than two orders of magnitude relative to Metropolis, and it recovers the expected downward shift of the pseudo-critical temperature as the Gaussian random-field strength is increased. The practical consequence is that both equilibrium thermodynamics and non-equilibrium relaxation of disordered Ising systems become accessible on much larger lattices and longer timescales.

What carries the argument

Hierarchical probabilistic counters: a tree-structured data structure that stores cumulative flip probabilities and permits selection of the next spin (and the waiting time) in O(log N) operations under site-dependent Glauber rates, thereby replacing the energy-class lists of the pure BKL algorithm.

What would settle it

Compare the algorithm’s measured flip-rate histogram and long-time magnetization autocorrelation against an exact enumeration (or a carefully equilibrated Metropolis run) on a small lattice with known Gaussian disorder; any systematic deviation in the waiting-time distribution or in the temperature dependence of the susceptibility peak would falsify dynamical fidelity.

Watch

Extended reading notes

Core claim

An event-driven, rejection-free Monte Carlo that employs hierarchical probabilistic counters can implement Glauber dynamics for the two-dimensional random-field Ising model in O(log N) per accepted flip, delivering more than 100-fold acceleration over Metropolis at low temperature while preserving the correct continuous-time master-equation evolution that classical energy-class BKL cannot maintain once random fields are present.

Load-bearing premise

The hierarchical counters correctly implement the site-dependent Glauber rates without bias and keep the event-driven process equivalent to the underlying continuous-time master equation once random fields destroy the simple energy-class structure.

Editorial extensions

If this is right

Low-temperature and low-disorder RFIM dynamics can be followed for wall-clock times two orders of magnitude longer than with Metropolis.
Pseudo-critical temperatures extracted from Gaussian-disorder sweeps can be mapped with higher statistics, confirming the expected depression with increasing field strength.
Both equilibrium sampling and non-equilibrium aging or coarsening protocols become practical for larger two-dimensional lattices.
The same counter hierarchy can be reused for other disordered Ising models whose transition rates lack a simple energy-class partition.

Reading between the lines

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

The same O(log N) selection structure should extend immediately to three-dimensional RFIM or to models with random bonds, provided the rate tree can be updated after each flip.
Because waiting times are drawn from the exact continuous-time distribution, the method supplies a natural clock for measuring physical aging exponents without arbitrary Monte-Carlo-step rescaling.
A parallel or GPU realization of the hierarchical counters would further enlarge accessible system sizes for finite-size scaling of the RFIM critical line.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Punchline: this is a computational methods paper that tries to give you rejection-free, continuous-time Glauber dynamics for the 2D random-field Ising model by replacing classical BKL energy classes with hierarchical probabilistic counters that pick a spin in O(log N). If the bookkeeping is exact, that is a real tool for low-T and non-equilibrium RFIM work; classical BKL does not carry over cleanly once every local field is unique.

What looks new is the combination, not the ingredients. BKL, Glauber rates, and tree-based event selection are known. The abstract’s point is that the hierarchy keeps an exact, updatable sum of heterogeneous flip rates so the sampler stays dynamically faithful under Gaussian disorder, where energy-class BKL fails. They report the expected drop of the pseudo-critical temperature with disorder and >100× speedups versus Metropolis at low T. For a methods paper those are the right qualitative checks, and the circularity burden is low: they are not fitting a quantity and then re-predicting it.

The soft spot is load-bearing and currently invisible. Once random fields destroy energy classes, you need recursive update rules and a short argument that the sampled rates and the exponential waiting time match the master equation. The abstract asserts that and stops. Without the full text we cannot see the rate bookkeeping, complexity analysis, error bars, or code. That is not a manufactured flaw; it is simply what an abstract-only read leaves open. Nothing in the abstract is internally contradictory, and the stress-test concern is exactly the right one to demand from the paper, not a reason to dismiss the idea.

Who this is for: people who actually run disordered Ising dynamics at low T or out of equilibrium. It is not a physics breakthrough; it is subfield infrastructure. I would send it to a serious referee rather than desk-reject. If the proofs and reproducible artifacts are in the manuscript, it is a useful tool paper. If they are hand-waved, the referee will catch it. Worth a look when the full text is available; not something I would cite from the abstract alone.

Referee Report

3 major / 2 minor

Summary. The manuscript proposes a rejection-free, event-driven Monte Carlo algorithm for the two-dimensional Random Field Ising Model (RFIM). It combines the Bortz–Kalos–Lebowitz (BKL) event-driven framework with Glauber transition probabilities and introduces hierarchical probabilistic counters that select spins in O(log N) operations. The abstract claims that this construction remains dynamically faithful even when a random field destroys the energy-class structure exploited by classical BKL, that Gaussian-RFIM runs reproduce the expected reduction of the pseudo-critical temperature with increasing disorder, and that speedups exceed two orders of magnitude relative to Metropolis in the low-temperature regime.

Significance. If the hierarchical-counter construction is rigorously equivalent to continuous-time Glauber dynamics for heterogeneous rates, the method would supply a practically useful tool for both equilibrium and non-equilibrium studies of disordered Ising systems in regimes where Metropolis suffers severe critical slowing down. An O(log N) rejection-free sampler that does not rely on a small number of energy classes would also be of broader algorithmic interest beyond the RFIM. The claimed qualitative consistency with the known disorder-induced drop of the pseudo-critical temperature is a necessary but not sufficient validation target.

major comments (3)

[Abstract] The abstract asserts that hierarchical probabilistic counters implement unbiased Glauber rates and preserve exact continuous-time RFIM dynamics once random fields render every flip rate unique. Classical BKL equivalence rests on a finite set of identical energy classes; the abstract supplies neither the recursive rate-update rules, a proof that the sampled probabilities remain proportional to the local Glauber rates, nor a demonstration that the exponential waiting-time clock remains exact. This equivalence is load-bearing for both the “dynamically faithful” claim and the fairness of the reported speedups; without it the method could be a fast approximate sampler.
[Abstract] Speedups “exceeding two orders of magnitude” versus Metropolis are stated without system sizes, temperature and disorder ranges, wall-clock versus Monte-Carlo-step accounting, error bars, or a precise definition of the Metropolis baseline (single-spin vs. sweep, continuous-time vs. discrete-time). These details are required to judge whether the comparison is fair and whether the hierarchical overhead is correctly amortized.
[Abstract] Reproduction of the expected reduction of the pseudo-critical temperature with Gaussian disorder is only a qualitative consistency check. The abstract does not report quantitative comparison to established literature values, finite-size scaling analysis, or any dynamical observable (autocorrelation times, domain-growth exponents) that would test continuous-time fidelity rather than mere equilibrium sampling.

minor comments (2)

[Abstract] The abstract should briefly distinguish hierarchical probabilistic counters from standard binary trees or Fenwick trees already used in other rejection-free Monte Carlo schemes, so that the claimed novelty is clear.
[Abstract] Clarify whether the continuous-time clock is advanced by the exact sum of all rates or by an approximation, and whether the method is intended for equilibrium sampling only or also for genuine non-equilibrium trajectories.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: methods paper with external validation target; abstract-only review finds no self-definitional or fitted-as-prediction steps.

full rationale

This is an abstract-only methods paper proposing a rejection-free event-driven Monte Carlo algorithm for the 2D RFIM that combines BKL-style continuous-time dynamics with hierarchical probabilistic counters for O(log N) spin selection under heterogeneous Glauber rates. The claimed results are algorithmic efficiency (speedups >100x vs Metropolis at low T) and dynamical fidelity, validated against an external, previously known qualitative feature of the RFIM: reduction of the pseudo-critical temperature with increasing Gaussian disorder. No equation, fit, or self-citation chain is present in the available text that would make a 'prediction' equivalent to its inputs by construction. The reader's noted risk (that hierarchical counters correctly implement unbiased Glauber rates once random fields destroy energy-class structure) is a correctness/assumption concern, not circularity: the abstract asserts the construction works and is superior to classical BKL for RFIM, but does not redefine a fitted quantity as a first-principles result, nor import a uniqueness theorem from the authors' prior work to force the claim. Self-reference risk is limited to ordinary methods development. Score 0 is the honest finding for a self-contained algorithmic contribution whose validation target is external.

Assumptions & free parameters 0 free parameters · 3 assumptions · 1 invented entities

Abstract-only review: free parameters and invented physical entities are not introduced. The claim rests on standard Monte Carlo and master-equation assumptions plus the unproved (in the abstract) correctness of the hierarchical-counter rate bookkeeping under random fields. No fitted physical constants appear; any algorithmic tuning constants are not disclosed here.

assumptions (3)

domain assumption Glauber single-spin flip rates define the correct continuous-time Markov dynamics of the RFIM.
Standard choice for kinetic Ising models; abstract adopts it as the target dynamics the algorithm must match.
ad hoc to paper Event-driven rejection-free sampling with hierarchical probabilistic counters is statistically equivalent to the underlying Glauber master equation when random fields are present.
This is the load-bearing algorithmic premise; classical BKL relies on energy classes that RF breaks. Abstract asserts equivalence but does not prove it in the available text.
standard math Standard Monte Carlo ergodicity and detailed-balance requirements for equilibrium sampling on finite 2D lattices.
Background theory assumed for any claim that the method samples equilibrium RFIM correctly.

invented entities (1)

Hierarchical probabilistic counters (algorithmic data structure)
purpose: Select the next spin to flip under Glauber rates in O(log N) without rejection, enabling event-driven RFIM dynamics.
Presented as the enabling technique. It is a computational construct, not a new physical particle or force; independent physical evidence is not applicable. Correctness must be shown by rate accounting and tests, which the abstract only sketches.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Rejection-free Glauber Monte Carlo for the 2D Random Field Ising Model via Hierarchical Probabilistic Counters." pith.science (2026). https://pith.science/paper/WC6Z2PJL

@misc{pith2026260317032,
  author       = {Pith},
  title        = {Pith review of: Rejection-free Glauber Monte Carlo for the 2D Random Field Ising Model via Hierarchical Probabilistic Counters},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WC6Z2PJL}},
  note         = {Machine review of arXiv:2603.17032}
}

read the original abstract

We present an efficient Monte Carlo algorithm for the simulation of the two-dimensional Random Field Ising Model (RFIM). The method combines the event-driven, rejection-free character of the Bortz Kalos-Lebowitz (BKL) algorithm with Glauber transition probabilities, introducing hierarchical probabilistic counters to perform spin selection in O(log N) operations. This enables efficient sampling of the system's dynamics, especially in the low-temperature and low-disorder regime, where traditional Metropolis updates suffer from critical slowing down. Furthermore, this approach allows a proper dynamical simulation of the Ising system's behavior even in the presence of a Random Field (RF), unlike the BKL method. RFIM simulations with Gaussian field distributions reproduce the expected reduction of the pseudo-critical temperature with increasing disorder. Benchmarking shows speedups exceeding two orders of magnitude compared to the Metropolis algorithm in the low-temperature regime. The proposed method provides an efficient and dynamically faithful tool for studying both equilibrium and non-equilibrium phenomena in disordered spin systems.

Discussion (0). Continue with ORCID to comment.

Pith tools

Reviewed July 13, 2026 · model on record in the stance chip above.