Pith. sign in

REVIEW 2 major objections 6 minor 25 references

Distillation and position-space sampling for local multiquark interpolators

T0 review · 2 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A position-space sampling scheme for distillation makes local multiquark interpolators computationally affordable, and a preliminary Tcc study finds a small downward energy shift when local operators are added.

desk verdict A genuinely new random-displacement sparse-grid estimator for distillation; clean numerical evidence on one ensemble, with an untested volume-scaling assumption worth chasing. read the letter →

arxiv 2412.09246 v1 pith:TX4RNB2O submitted 2024-12-12 hep-lat

Andres Stump , Jeremy R. Green This is my paper · ORCID
classification hep-lat
keywords latticeQCDdistillationposition-spacesamplingsparsegridsstochasticestimatortetraquarkTcc(3875)+variationalmethod
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 proposes and tests a way to compute local multiquark operators inside distillation without paying the usual high-rank tensor cost. The idea is to evaluate the smeared quark propagator only on randomly displaced sparse grids, and to rescale the sum so that the result is an unbiased stochastic estimator of the full correlator. On a single $64\times32^3$ ensemble with 32 Laplacian modes, the statistical error from the sampling is negligible compared to the gauge-field Monte Carlo error once the grid spacing $N_{\mathrm{sep}}$ is at most 8, for pions, $D$ mesons, and local tetraquark operators. The method drops the contraction cost from scaling like $N^5$ or $N^7$ to $N(N_s/N_{\mathrm{sep}})^6$, which is a volume scaling of $V^3$ when the physical grid spacing is held fixed. Applying it to $T_{cc}(3875)^+$, adding two local operators to a two-operator bilocal basis lowers the ground-state energy by $a\Delta E = 0.00032(3)$, small compared with the energy error.

What carries the argument

The load-bearing object is the randomly displaced sparse grid $\tilde{\Lambda}_3 = \{a\,\mathbf{n} + \tilde{\mathbf{x}} \mid n_k = 0, N_{\mathrm{sep}}, 2N_{\mathrm{sep}},\dots,N_s-N_{\mathrm{sep}}\}$ with independent random offset $\tilde{\mathbf{x}}\in\Lambda_3$ chosen separately for sink and source, per configuration and time. Evaluated inside the estimator (Eq. 10), this grid replaces the exact position sums by a Monte Carlo average; the random displacement guarantees that no position is systematically excluded, so the momentum projection is exact and the sparse grid changes only the variance, not the mean. It is this object that converts the expensive high-rank mode tensors into cheap contractions whose cost scales with the grid volume $N(N_s/N_{\mathrm{sep}})^6$ rather than with $N^5$ or $N^7$.

What would settle it

Compute the same estimator on a larger spatial volume, e.g. $48^3$ or $64^3$ at the same physical $N_{\mathrm{sep}}$ and with $N$ scaled with volume, and compare the sampled-correlator error to the full-lattice result at $N_{\mathrm{sep}}=1$: if the relative error grows with volume, the claimed $V^3$ advantage over standard contractions fails.

Watch

Extended reading notes

Core claim

In distillation, a local tetraquark correlator normally requires first forming rank-four mode tensors by summing over all positions, making the contraction cost scale as $N^5$ in the number of Laplacian modes, and as $N^7$ for a hexaquark. The paper's claim is that this can be avoided by swapping the summation order: compute $\bar{D}^{-1}_{\mathrm{sm},f}(\mathbf{x}',t';\mathbf{x},t)$ only for $\mathbf{x}'$ and $\mathbf{x}$ on independently chosen, randomly displaced sparse grids, multiply by $|\Lambda_3|^2/(|\tilde\Lambda'_3||\tilde\Lambda_3|)$, and average over the random offsets. Because the offsets are chosen fresh per gauge configuration and source time, the momentum projection is exact and the estimator is unbiased; $N_{\mathrm{sep}}$ controls only the variance. The numerical evidence is that for the pion, $D$, $D^*$, and two local $T_{cc}$ operators, the sampled correlators are statistically equivalent to the full-lattice ones for $N_{\mathrm{sep}}\le 8$, with errors dominated by gauge noise. In the variational analysis, the $T_{cc}$ ground state moves down by $a\Delta E = 0.00032(3)$ when the two local operators are added.

Load-bearing premise

The argument depends on the variance of the sparse-grid estimator remaining dominated by gauge-field Monte Carlo noise as the physical volume grows while $N_{\mathrm{sep}}$ is held fixed and $N$ is scaled with volume; the variance tests are done on one $32^3$ spatial lattice at a single lattice spacing.

Editorial extensions

If this is right

  • Local tetraquark and hexaquark interpolators become usable in larger volumes; the contraction cost scaling improves from $V^5$ or $V^7$ to $V^3$ when $N_{\mathrm{sep}}$ is held fixed.
  • Correlators computed this way are unbiased at any $N_{\mathrm{sep}}$, so the method adds no systematic shift; in practice $N_{\mathrm{sep}}=8$ already gives errors that are gauge-noise dominated.
  • The same estimator works for single-meson, heavy-quark meson, and local tetraquark operators, so it is a general replacement for mode doublets and triplets in contractions.
  • Keeping the same random offsets for all sink times preserves correlations across $t'$, which is necessary for reliable plateaus and GEVP analysis.
  • For the $T_{cc}$, the ground-state energy shifts downward by $a\Delta E = 0.00032(3)$ when local operators join the basis, indicating local structure matters, though the shift is small compared with the energy uncertainty.

Reading between the lines

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

  • If the variance remains gauge-dominated on larger volumes, this estimator could be combined with stochastic sources to reduce $N$ itself, a step the paper does not take.
  • The observed shift may reflect an incompletely saturated operator basis: adding more bilocal operators could produce a comparable shift, so the local-operator effect should be re-examined with a larger basis before physical conclusions are drawn.
  • The random-displacement idea may extend beyond position: the same unbiased-sampling logic could be applied to spin, color, or taste contractions where exact summation is costly.
  • The method's practical range likely depends on momentum: the paper sees error rise at $N_{\mathrm{sep}}=16$ for nonzero-momentum mesons, suggesting high-momentum local operators need finer grids.
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

2 major / 6 minor

Summary. The paper presents a position-space sampling method for evaluating local multiquark operators in the distillation framework. The key idea is to replace the full spatial lattice sums in the smeared-propagator contractions by sums over two independently randomly displaced sparse grids, with normalization factors restoring the full-lattice result. The authors argue that the contraction cost then scales as V^3 (with N proportional to V and fixed physical grid spacing) instead of V^5 or V^7. They test the estimator on a single 32^3 CLS ensemble with N=32 Laplace modes: pion, D, and D* effective masses are independent of Nsep up to Nsep=8 (and Nsep=16 at zero momentum), and two local Tcc tetraquark operators show no reduction in plateau-error when Nsep is lowered from 8 to 4. Adding these local operators to a basis of two bilocal DD* operators in a GEVP produces a downward shift aDeltaE=0.00032(3) in the lowest level, which is small compared with the statistical error of the energy.

Significance. If the volume-scaling assumption is correct, the method is a genuinely useful tool: local tetraquark and hexaquark operators become much cheaper in distillation, and the paper provides a clean numerical demonstration on one ensemble, including an exact Nsep=1 reference for mesons. The estimator is simple and is unbiased by construction for independent uniform offsets, and the authors are careful to label the Tcc result as preliminary and to note that more bilocal operators may remove the shift. The principal open question is the volume dependence of the position-sampling variance, which the paper does not address.

major comments (2)
  1. [Sec. 3, cost scaling] The claim that keeping a fixed physical point separation a Nsep yields a V^3 scaling of the computational cost is conditional on the position-sampling variance remaining negligible relative to the gauge-field Monte Carlo noise as the volume grows. The variance tests in Secs. 4.2 and 4.3 are performed on a single 32^3 ensemble at one lattice spacing, so the volume dependence of this variance is not established. If the position-sampling variance grows with volume, Nsep would need to be reduced and the V^3 scaling would degrade toward V^5 or V^7. Please either provide a theoretical variance bound, add a test on at least one additional volume or lattice spacing, or explicitly restrict the larger-volume claim to the ensemble studied.
  2. [Sec. 4.3, tetraquark Nsep comparison] For the two local tetraquark operators, the conclusion that the statistical error is dominated by the Monte Carlo error is based only on comparing Nsep=4 with Nsep=8, because the Nsep=1 full-lattice reference is not available for these contractions. Since both estimates are computed on the same gauge configurations and the fitted plateau values are not the actual finite-volume energies, this comparison has limited power to detect a moderate position-sampling variance contribution. I recommend adding a quantitative variance estimate, for example from several independent offset samples on fixed gauge configurations, or a Nsep=2 comparison for at least one tetraquark interpolator, or explicitly stating that the tetraquark variance has not yet been established.
minor comments (6)
  1. [Sec. 3, Eq. (11)] The definition of the randomly displaced sparse grid should state that the shifted coordinates are taken modulo the spatial lattice extent; as written, an arbitrary offset \tilde{x}\in\Lambda_3 can push some grid points outside the lattice.
  2. [Sec. 3, Eq. (10)] The unbiasedness statement is deferred with 'one can show'; since this is the theoretical foundation of the estimator, a one-sentence proof (average over independent uniform offsets for fixed gauge field) should be included.
  3. [Sec. 4.2, Fig. 1] The statement that there is 'no increase in error' up to Nsep=8 or 16 is based on a visual comparison of error bars; please provide the numerical values of the errors or a ratio test to support this conclusion.
  4. [Sec. 4.3, Fig. 2] Please state the plateau fit ranges and the number of effective-mass points used in the fits, since the comparison between Nsep=4 and Nsep=8 depends on this choice.
  5. [Sec. 4.4, Fig. 3] The word 'significant' for the shift a\Delta E=0.00032(3) should be qualified as statistical significance within the chosen operator basis; the authors' own caveat that additional bilocal operators could produce a similar shift means that the result is not yet evidence for a physical local-tetraquark component.
  6. [References] In reference [3], the collaboration name contains a typo ('W ASA-at-COSY Collaborationcollaboration').

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the position-space sampling estimator is validated against exact Nsep=1 results and the physical shift is a measured GEVP difference.

full rationale

The paper's central claim is the unbiasedness of a position-space sampling estimator using randomly displaced sparse grids. This is established by a self-contained expectation-value argument: the estimator in Eq. (10) is paired with the random subspace definition in Eq. (11), and the paper states that splitting the expectation value as <...>_{G,Ltilde} = <<...>_G>_{Ltilde} shows the estimator is unbiased. No fitted parameter enters this derivation, and the method is checked against the exact full-lattice Nsep=1 result for pion and D-meson two-point functions, so the method's validity does not reduce to its own inputs. The choice of Nsep=8 is an efficiency decision made from error comparisons, not a fitted parameter that forces the tetraquark result; the paper explicitly notes that reducing Nsep below 8 gives no further error reduction. The tetraquark shift aDeltaE = 0.00032(3) is a measured difference between GEVP energies with and without the local operators, and the paper identifies it as small compared to the energy error, so it is not presented as a fitted prediction. The only self-citation, Ref. [6] by one of the authors, is used for background on baryon distillation and is not load-bearing for the sampling method. The concern that sparse-grid variance might grow with volume is a robustness assumption about scaling, not a circular derivation, and is therefore outside the circularity assessment.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on algorithmic parameters (Nsep, N, t0) and on established lattice-technique assumptions. No new physical entities are introduced. The most fragile item is the unbiasedness lemma, which is asserted rather than derived, and the untested volume-scaling of the sampling variance.

free parameters (3)
  • Nsep (sparse grid point separation) = 8 (production); 4, 8, 16 tested
    Chosen to make sampling variance negligible; selected by comparing errors as a function of Nsep in Figures 1 and 2, rather than derived from first principles.
  • N (number of Laplacian eigenvectors) = 32
    Fixed simulation parameter; the cost of distillation contractions scales with N, so this choice shapes the method's practical cost.
  • t0 (GEVP reference time) = 13 lattice units
    Chosen to satisfy the GEVP condition t0 >= t/2 over the plateau range; a standard analysis choice that affects the extracted energies.
assumptions (4)
  • ad hoc to paper The randomized sparse-grid estimator in Eq. (10) is unbiased; splitting the ensemble average into gauge and subspace parts ('one can show') is sufficient.
    Load-bearing lemma for the method, asserted rather than derived in Section 3.
  • domain assumption To keep the physical smearing radius constant, N must be scaled with physical volume, as cited from [10].
    Drives the volume-scaling comparison in Sections 2 and 3; not re-derived here.
  • domain assumption The GEVP energy extraction is valid when t0 >= t/2, following [17]; used to set t0 = 13.
    Standard lattice spectroscopy assumption invoked in Section 4.4.
  • domain assumption Effective-mass plateaus at late times are dominated by the lowest state.
    Used for the Nsep dependence fits and Tcc energy extraction; the authors note the tetraquark operators are not optimized and multiparticle levels are closely spaced.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Distillation and position-space sampling for local multiquark interpolators." pith.science (2026). https://pith.science/paper/TX4RNB2O

@misc{pith2026241209246,
  author       = {Pith},
  title        = {Pith review of: Distillation and position-space sampling for local multiquark interpolators},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TX4RNB2O}},
  note         = {Machine review of arXiv:2412.09246}
}
abstract

Distillation in lattice QCD is a smearing method that uses the lowest eigenvectors of the spatial Laplacian to construct a subspace in which the Dirac operator can be fully inverted. However, local multiquark interpolators are expensive in this framework because the cost of the contractions scales with a high power of the number of Laplacian eigenvectors. To address this, a position-space sampling method within distillation is presented that avoids this cost scaling. Our simulations show that this method works well for meson operators, but also for local tetraquark operators. In a preliminary study, we investigate the relevance of the latter for the ground state energy of the $T_{cc}(3875)^+$ tetraquark. There we find a downward shift in the lowest level when local tetraquark operators are added to a basis of bilocal scattering operators in the variational method. However, this shift is small compared to the error of the energy.

Figures

Figures reproduced from arXiv: 2412.09246 by the authors.

Figure 1
Figure 1. Pion (left) and 𝐷 meson (right) energy for zero momentum. They are plotted as a function of the point separation 𝑁sep in the sparse grids used for position-space sampling. which corresponds to using only 2 points in each spatial direction. We repeated this calculation for momentum 𝒑 = 2𝜋 𝐿 (1, 0, 0) and also found no increase in error up to 𝑁sep = 8, but a small increase for 𝑁sep = 16. In all cases, the energy was c… view at source ↗
Figure 2
Figure 2. Plateau values obtained from a constant fit to the effective masses of local 𝐷𝐷∗ (left) and diquark￾antidiquark (right) two-point functions with zero momentum. They are plotted as a function of the point separation 𝑁sep in the sparse grids used for position-space sampling. computing these two-point functions scales as 𝑁sep −6 , and since the contractions are more expensive than for single mesons, the smallest point … view at source ↗
Figure 3
Figure 3. Effective masses of the two lowest energy levels of the 𝑇𝑐𝑐 tetraquark. They were obtained using the variational method for the two bilocal operators (left) and when also including the two local operators (right). The red line shows the ground state energy extracted using a plateau fit and the orange dashed line shows the 𝑚𝐷 + 𝑚𝐷∗ value for this ensemble. shift in the lowest level, we see a larger shift in the first… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 8 canonical work pages

  1. [1]

    LHCb collaboration, Observation of an exotic narrow doubly charmed tetraquark,Nature Phys.18(2022) 751 [2109.01038]

    arXiv 2022

  2. [2]

    LHCb collaboration, Study of the doubly charmed tetraquark𝑇+ 𝑐𝑐,Nature Commun.13 (2022) 3351 [2109.01056]

    arXiv 2022

  3. [3]

    W ASA-at-COSY Collaborationcollaboration,Abashian-Booth-Crowe Effect in Basic Double-Pionic Fusion: A New Resonance?, Phys. Rev. Lett.106(2011) 242302

    work page 2011

  4. [4]

    Hadron Spectrumcollaboration,A Novel quark-field creation operator construction for hadronic physics in lattice QCD, Phys. Rev. D80(2009) 054506 [0905.2160]

    arXiv 2009

  5. [5]

    Padmanath and S

    M. Padmanath and S. Prelovsek,Signature of a Doubly Charm Tetraquark Pole in DD* Scattering on the Lattice,Phys. Rev. Lett.129 (2022) 032002 [2202.10110]

    arXiv 2022

  6. [6]

    Green, A.D

    J.R. Green, A.D. Hanlon, P.M. Junnarkar and H. Wittig,Weakly Bound𝐻 Dibaryon from SU(3)-Flavor-Symmetric QCD,Phys. Rev. Lett.127 (2021) 242003

    work page 2021

  7. [7]

    Hadron Spectrumcollaboration,Tetraquark operators in lattice QCD and exotic flavour states in the charm sector, JHEP11(2017) 033 [1709.01417]

    arXiv 2017

  8. [8]

    Ortiz-Pacheco, S

    E. Ortiz-Pacheco, S. Collins, L. Leskovec, M. Padmanath and S. Prelovsek,Doubly charmed tetraquark: isospin channels and diquark-antidiquark interpolators, 2312.13441

    arXiv

Show all 25 references
  1. [9]

    Lang, R.G

    N. Lang, R.G. Edwards and M.J. Peardon,Optimising stochastic algorithms for hadron correlation function computations in lattice QCD using a localised distillation basis, 2411.10395

    arXiv

  2. [10]

    Morningstar, J

    C. Morningstar, J. Bulava, J. Foley, K.J. Juge, D. Lenkner, M. Peardon et al.,Improved stochastic estimation of quark propagation with Laplacian Heaviside smearing in lattice QCD, Phys. Rev. D83(2011) 114505 [1104.3870]

    2011 arXiv

  3. [11]

    Detmold, D.J

    W. Detmold, D.J. Murphy, A.V. Pochinsky, M.J. Savage, P.E. Shanahan and M.L. Wagman, Sparsening algorithm for multihadron lattice QCD correlation functions, Phys. Rev. D104 (2021) 034502 [1908.07050]

    2021 arXiv

  4. [12]

    Li, S.-C

    Y. Li, S.-C. Xia, X. Feng, L.-C. Jin and C. Liu,Field sparsening for the construction of the correlation functions in lattice QCD, Phys. Rev. D103 (2021) 014514 [2009.01029]

    2021 arXiv

  5. [13]

    Wolff,Monte Carlo errors with less errors,Computer Physics Communications156 (2004) 143

    U. Wolff,Monte Carlo errors with less errors,Computer Physics Communications156 (2004) 143

    2004

  6. [14]

    Bruno et al.,Simulation of QCD with N𝑓 =2+1 flavors of non-perturbatively improved Wilson fermions,JHEP02 (2015) 043 [1411.3982]

    M. Bruno et al.,Simulation of QCD with N𝑓 =2+1 flavors of non-perturbatively improved Wilson fermions,JHEP02 (2015) 043 [1411.3982]

    2015 arXiv

  7. [15]

    Collins, A

    S. Collins, A. Nefediev, M. Padmanath and S. Prelovsek,Toward the quark mass dependence of Tcc+ from lattice QCD,Phys. Rev. D109(2024) 094509 [2402.14715]. 9 Distillation and position-space sampling for local multiquark interpolators Andres Stump

    2024 arXiv

  8. [16]

    Vujmilovic, S

    I. Vujmilovic, S. Collins, L. Leskovec, E. Ortiz-Pacheco, M. Padmanath and S. Prelovsek, 𝑇+ 𝑐𝑐 via the plane wave approach and including diquark-antidiquark operators, PoS LATTICE2024(2024) 112 [2411.08646]

    2024 arXiv

  9. [17]

    Blossier, M

    B. Blossier, M. Della Morte, G. von Hippel, T. Mendes and R. Sommer,On the generalized eigenvalue method for energies and matrix elements in lattice field theory, JHEP04(2009) 094 [0902.1265]

    2009 arXiv

  10. [18]

    Krause and P

    D. Krause and P. Thörnig,JURECA: Modular supercomputer at Jülich supercomputing centre. JLSRF 4,A132 10(2018)

    2018

  11. [19]

    Krause,JUWELS: Modular Tier-0/1 supercomputer at the Jülich supercomputing centre, Journal of large-scale research facilities JLSRF5(2019) A135

    D. Krause,JUWELS: Modular Tier-0/1 supercomputer at the Jülich supercomputing centre, Journal of large-scale research facilities JLSRF5(2019) A135

    2019

  12. [20]

    SciDAC, LHPC, UKQCDcollaboration, The Chroma software system for lattice QCD, Nucl. Phys. B Proc. Suppl.140 (2005) 832 [hep-lat/0409003]

    2005 arXiv

  13. [21]

    Stathopoulos and J.R

    A. Stathopoulos and J.R. McCombs,PRIMME: preconditioned iterative multimethod eigensolver—methods and software description,ACM Trans. Math. Softw.37(2010)

    2010

  14. [22]

    openQCD

    M. Lüscher and S. Schaefer, “openQCD.” https://luscher.web.cern.ch/luscher/openQCD/, 2012

    2012

  15. [23]

    Lukas Devos, Maarten Van Damme and contributors,TensorOperations.jl, 10, 2023

    J.H. Lukas Devos, Maarten Van Damme and contributors,TensorOperations.jl, 10, 2023. 10.5281/zenodo.3245496

    2023 doi

  16. [24]

    ADerrors.jl

    A. Ramos, “ADerrors.jl.”https://igit.ific.uv.es/alramos/aderrors.jl, 2020

    2020

  17. [25]

    Christ, D

    S. Christ, D. Schwabeneder, C. Rackauckas, M.K. Borregaard and T. Breloff,Plots.jl – a user extendable plotting API for the julia programming language, 2204.08775. 10

    arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.