REVIEW 3 major objections 4 minor 11 references
Renormalisation Group Equations for 2+1 clover fermions
T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash
Pith's one-line read The paper derives RG equations for 2+1 clover fermions and shows the lattice spacing scales with little dependence on the renormalized quark mass.
desk verdict A clean, honest RG treatment of lattice spacing scaling for 2+1 clover fermions, with a load-bearing assumption about the mass-axis proxy that needs testing at more beta values. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central object is the renormalization-group operator acting at fixed physics, $a\,\partial/\partial a|_{\rm physics}$, expressed in the bare coupling $g_0$ and lattice quark masses $a m_q$. At leading order in $a m_q$ it is controlled by the $\beta$-function $B_0(g_0)=-b_0 g_0^3-\cdots$, the mass anomalous dimension combination $G_0 = 1-\gamma_m^{\rm NS}$, the singlet–non-singlet difference $H_0=\gamma_m^{\rm NS}-\gamma_m^{\rm S}$, and the improvement coefficient $B_1 = b^{\rm lat}_{10}g_0^3+\cdots$ used to define the mass correction $r(g_0)$. Solving the RG equations gives $a = s(g_0)\{1+a\bar m\, r(g_0)\}$ with $s(g_0)$ fixed by the $\beta$-function; the data analysis then rides on the identity that at common $\bar m_{\rm rgi}$ the measured ratio $x(\beta)/x(\beta_{\rm ref})$ equals the lattice-spacing ratio $a^2(\beta)/a^2(\beta_{\rm ref})$. The practical machinery is a two-parameter fit $x=y(A+By)$ in the mass-proxy $y=X_\pi^{{\rm lat}\,2}/X_{t_0}^{{\rm lat}\,2}$, a $[2/2]$ Padé fit to the $\beta$-function, and a one-parameter integral fit for the mass-correction coefficient $D_{\pi/t_0}$.
What would settle it
On existing ensembles, compute the ratio $a^2(\beta,\bar m_{\rm rgi})/a^2(\beta_{\rm ref},\bar m_{\rm rgi})$ directly from separate lattice-spacing determinations at two values of $\bar m_{\rm rgi}$ that differ by about 20% in the proxy $X_\pi^2/X_{t_0}^2$; the RG claim predicts the ratio changes by roughly 2–4%, so a change of 10% or more, or visible curvature in the x-versus-y plots, would contradict the scaling conclusion.
Extended reading notes
Core claim
The central claim is that the lattice spacing in 2+1 flavor clover QCD obeys the renormalization group along lines of constant physics. Concretely, the spacing satisfies $a = s(g_0)\{1 + a\bar m\, r(g_0)\}$, where $s(g_0)$ is the standard $\beta$-function solution and the mass correction $r(g_0)$ is small, of order $-b^{\rm lat}_{10}g_0^2/b_0$ at weak coupling. Using the ratio $X_\pi^{{\rm lat}\,2}/X_{t_0}^{{\rm lat}\,2}$ as a proxy for the renormalized quark mass $\bar m_{\rm rgi}$, the authors show that the ratio of lattice spacings at fixed $\bar m_{\rm rgi}$, $a^2(\beta,\bar m_{\rm rgi})/a^2(\beta_{\rm ref},\bar m_{\rm rgi})$, is nearly flat: a 10% change in the mass proxy moves the coarsest ratio by only about 1–2%. Reconstructing the spacing ratio from $s^2(\beta)/s^2(\beta_{\rm ref})$ plus a fitted linear term $D_{\pi/t_0}$ in the mass proxy gives a smoother set of lattice spacings across $\beta = 5.40, 5.50, 5.65, 5.80, 5.95$ than earlier separate determinations, leading the authors to conclude that scaling holds with little dependence on $\bar m_{\rm rgi}^*$.
Load-bearing premise
The result rests on taking the flavour-singlet combination $X_\pi^{{\rm lat}\,2}/X_{t_0}^{{\rm lat}\,2}$ as a clean stand-in for the renormalised quark mass: if that quantity bends along lines of constant $a\bar m$ at the few-percent level, the derived spacing ratios and the scaling plot would be biased.
Editorial extensions
If this is right
- Lattice spacing ratios across many beta values can be determined more precisely from accurate pion mass and gradient-flow data than from separate fits, since the RG equation fixes the beta-dependence.
- An error in choosing the physical quark mass $\bar m_{\rm rgi}^*$ shifts the whole lattice-spacing set by a nearly constant overall factor, so absolute scale errors decouple from relative spacing errors.
- The RG-guided ratios give a smoother lattice spacing curve across $\beta = 5.40$–$5.95$, which is expected to reduce the noise in $a^2$ continuum extrapolations of hadron masses and matrix elements.
- The numerically small $r(g_0)$ and $b_g(g_0)$ coefficients confirm the standard expectation that the two-loop $\beta$-function describes the coupling dependence of the lattice spacing in this range, with corrections of a few percent at the coarsest lattice.
Reading between the lines
- If the scaling claim holds at finer couplings, the same RG-guided interpolation could set lattice spacings at intermediate beta values from accurate pion and t0 data alone, reducing the number of dedicated scale-setting runs.
- The same machinery, with appropriate anomalous dimensions, could be applied to hadron matrix elements rather than just the lattice spacing, giving RG-guided joint continuum extrapolations.
- A direct test of the constancy assumption would be to check whether $X_s^{{\rm lat}\,2}$ shows curvature along $a\bar m = \text{const}$ lines once more SU(3)-symmetric data at finer lattices are included; the paper's own plots show little evidence for it now.
- Repeating the analysis with an independent third value of the physical mass proxy (beyond the two averages used here) would test whether the reported insensitivity to $\bar m_{\rm rgi}^*$ persists at the percent level.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This Lattice 2024 proceedings contribution derives renormalisation-group equations for the lattice spacing of 2+1 clover fermions, including the leading O(am_q) mass-dependence terms, and applies them to five lattice spacings (beta = 5.40, 5.50, 5.65, 5.80, 5.95) using pion mass and gradient-flow observables. The central construction is the RG solution s(g0) for the beta function and r(g0) for the mass dependence of the lattice spacing, Eqs. (12)-(16). On the data side, the authors use y = X_pi^lat^2 / X_t0^lat^2 as a proxy for the RGI quark mass, fit x = X_pi^lat^2 = y(A+By) at each beta, extract s^2(beta)/s^2(beta_ref) and the mass-slope coefficient D_{pi/t0}, fit the beta function with a [2/2] Pade, and reconstruct a^2(beta,mbar*)/a^2(beta_ref,mbar*) at two values of mbar*. They find only a small mass dependence, mainly at beta = 5.40, and a smoother lattice-spacing curve than their previous separate determinations.
Significance. The paper proposes a conceptually attractive way to use RG equations to constrain lattice-spacing ratios across beta, which could lead to smoother continuum extrapolations. The derivation in Section 2 is clean and standard: the equations for u(g0), v(g0), s(g0) and r(g0) are internally consistent, and the reduction to the usual beta-function form is explicit. The authors are also honest that the results are preliminary and that the mass proxy is a calibration assumption. If the cross-beta calibration of y with mbar_rgi is confirmed at the few-percent level, the claimed scaling result is a useful methodological step. The paper does not overclaim an independent prediction, and the data are standard for the collaboration, though no reproducibility package is provided for this proceedings article.
major comments (3)
- [Section 3, Eq. (21), Fig. 2] The identification of y = X_pi^lat^2 / X_t0^lat^2 with a beta-independent function of the renormalised quark mass is the load-bearing assumption of the analysis, and the evidence offered for it is not sufficient for the few-percent accuracy claimed. The stationarity of X_s^lat under flavour-symmetry breaking (Fig. 1, left) is demonstrated at a single beta = 5.50 and only shows that X_s^lat is constant along lines of constant a mbar at that beta; it does not control O(a^2) discretisation errors in the ratio X_pi^lat^2 / X_t0^lat^2 that can differ from one beta to another. Consequently, equal y at different beta values need not correspond to equal mbar_rgi, and the interpolated values of x, the ratios s^2(beta)/s^2(beta_ref), the coefficient D_{pi/t0}, and the final scaling plot in Fig. 4 would all inherit the resulting bias. I would like to see a quantitative estimate of this effect, for example by comparing y against a directly determined mbar_rgi at two or more beta values, or by repeating the analysis after dropping one beta and checking the stability of the extracted ratios.
- [Section 5, Eq. (24), Fig. 4] The final a^2 ratios are reconstructed from the same fits that define the parameters entering Eq. (24): A and B per beta from the x = y(A+By) fits, b2^eff from the beta-function fit, and the proportionality constant in Eq. (28). The agreement between the RG-guided curve and the magenta points from QCDSF15 is therefore a consistency check of the fit model rather than an independent determination, and the manuscript should say so explicitly. To make the claim of a smoother lattice-spacing set quantitative, the authors should report the uncertainty on the reconstructed ratios that propagates from all fit parameters, and/or perform a leave-one-beta-out exercise.
- [Section 5, Eq. (27)] The [2/2] Pade form for the beta function is introduced as the simplest choice, but the sensitivity of s^2(beta)/s^2(beta_ref) and D_{pi/t0} to this model choice is not quantified. Since b2^eff is an effective parameter and the two-loop result is visibly different in Fig. 3, the analysis should include an estimate of the model uncertainty, for instance by using alternative Pade orders or by allowing b2 to vary within a reasonable range. Without such an estimate, the fit to B0(g0) cannot be distinguished from a simple empirical interpolation.
minor comments (4)
- [Section 5, Eq. (28)] In the last expression of Eq. (28), the integration variable is written as 1/(2 b0 g0^2) without stating the substitution; please define it explicitly so that the incomplete-Gamma form is transparent.
- [Section 4, Eq. (22)] The quantity X_pi^2(mbar_rgi)/X_t0^2(mbar_rgi) is used in Eq. (24) and in Fig. 4 before being defined; a sentence introducing this ratio as the continuum counterpart of the proxy y would improve readability.
- [Section 3, paragraph after Fig. 1] The statement that X_s^lat^2 is constant along any a mbar = const line is made without an error budget; please quote the size of the observed scatter at beta = 5.50 and, if available, at other beta values.
- [References] Reference [8] is given as an arXiv preprint; if a published version exists, it should be cited instead or in addition.
Circularity Check
No significant circularity; the RG-guided scaling analysis is an explicit fit/reconstruction, not a prediction forced by construction.
full rationale
The paper's derivation chain is self-contained. The RG equations in Section 2 are derived from standard definitions (Eqs. (1)-(19)), with the beta function and mass anomalous dimension inputs stated explicitly. The mass proxy y = X_pi^lat^2 / X_t0^lat^2 is introduced via Eq. (21) as a leading-order chiral relation, not as a fitted parameter disguised as a prediction. The two-parameter fits x = y(A+By) in Fig. 2 determine the lattice-spacing ratios, and Eq. (24) is an algebraic rearrangement of Eq. (12); the final Fig. 4 is explicitly a reconstruction ('we wish to re-construct a^2... via eq. (24)'), not an independent confirmation. Self-citations to prior QCDSF work ([5], [6], [8]) provide data and empirical evidence for the stationarity/constancy used in the mass-proxy argument, but that evidence is also shown in Fig. 1 and is not an unverified uniqueness theorem or ansatz imported solely through citation. The central scaling claim is a stated fit outcome ('we find scaling, with little dependence on mbar_rgi*'), which is a data-analysis result rather than a circular derivation. No step reduces by definition to its own input, and no fitted parameter is renamed as a prediction.
Assumptions & free parameters
free parameters (3)
- b2^eff (effective third beta-function coefficient) =
not stated
- D_pi/t0 proportionality constant =
not stated
- A and B per beta in x = y(A + B y) =
not stated
assumptions (4)
- domain assumption Leading-order truncation at O(a m_q)
- domain assumption X_s^lat^2 is constant along a ¯m = const lines
- ad hoc to paper [2/2] Padé form for the beta-function
- standard math PCAC relation for the mass proxy
Cite this review
Pith. "Pith review of Renormalisation Group Equations for 2+1 clover fermions." pith.science (2026). https://pith.science/paper/KTRPO56K
@misc{pith2026250116920,
author = {Pith},
title = {Pith review of: Renormalisation Group Equations for 2+1 clover fermions},
year = {2026},
howpublished = {\url{https://pith.science/paper/KTRPO56K}},
note = {Machine review of arXiv:2501.16920}
}
read the original abstract
Many lattice QCD simulations now have many lattice spacings available, and it is of interest to investigate how they scale. In this talk we first derive renormalisation group equations appropriate for 2+1 clover fermions. This is then used together with pion mass and gradient flow results at five lattice spacings to study scaling.
Figures
Figures from the paper (1 more)
Reference graph
Works this paper leans on
-
[1]
QCDSF Collaboration, in preparation
-
[2]
M. Lüscheret al., Nucl. Phys. B478 (1996) 365, [arXiv:hep-lat/9605038 [hep-lat] ]
arXiv 1996
-
[3]
S. Sintet al., Nucl. Phys. B465 (1996) 71, [arXiv:hep-lat/9508012 [hep-lat] ]
arXiv 1996
-
[4]
Dalla Bridaet al., JHEP2024(2024) 188, [arXiv:2401.00216 [hep-lat] ]
M. Dalla Bridaet al., JHEP2024(2024) 188, [arXiv:2401.00216 [hep-lat] ]
arXiv 2024
-
[5]
W. Bietenholzet al., Phys. Rev. D84(2011) 054509, [arXiv:1102.5300 [hep-lat] ]
arXiv 2011
-
[6]
V. G. Bornyakovet al., Phys. Lett. B767 (2017) 366, [arXiv:1612.04798 [hep-lat]]
arXiv 2017
-
[7]
N. Cundyet al., Phys. Rev. D79 (2009) 094507, [arXiv:0901.3302 [hep-lat] ]
arXiv 2009
-
[8]
V. G. Bornyakovet al., arXiv:1508.05916 [hep-lat]
Show all 11 references
-
[9]
Aokiet al., [Flavour Lattice Averaging Group (FLAG)], arXiv:2411.04268 [hep-lat]
Y. Aokiet al., [Flavour Lattice Averaging Group (FLAG)], arXiv:2411.04268 [hep-lat]
-
[10]
T. R. Haaret al., EPJ Web Conf.175 (2018) 14011, [arXiv:1711.03836 [hep-lat] ]
2018 arXiv
-
[11]
140(2005)832,[ arXiv:hep-lat/0409003]
R.G.Edwards et al.,Nucl.Phys.BProc.Suppl. 140(2005)832,[ arXiv:hep-lat/0409003]. 9
2005 arXiv
Reviewed August 10, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.