{"id":"4f461e3c-2b26-423b-9448-c4f8c53950c5","arxiv_id":"1908.11673","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"openQ*D is an open-source lattice code that simulates QCD plus QED with C* boundary conditions and Wilson fermions, extending openQCD and NSPT with new algorithmic features.","lead":"This paper releases openQ*D, an open-source code for lattice simulations of QCD plus QED using C* boundary conditions and Wilson fermions. It matters because it gives a first-principles tool for computing isospin-breaking and radiative corrections in hadronic observables.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Phase-quenched approximation in Eq. (2.4) is the load-bearing risk: if the determinant/Pfaffian sign is not negligible at the parameters used, the generated ensembles do not represent QCD+QED, and none of the code tests probe this sign.","rationale":"The central claim is that openQ*D-1.0 can perform lattice simulations of QCD+QED. For that claim to hold, the simulated distribution must be (or be straightforwardly reweightable to) the target QCD+QED path integral. Equation (2.4) replaces the determinant/Pfaffian by its absolute value, and the paper's only justification is a citation to ref. [18] asserting mildness. This is a theoretical assertion, not a result of the code tests: the tests verify the integrator, solvers, and rational approximation, but none measure the sign. The Q*D1 run is at unphysical parameters (heavy quark masses, alpha_0 ~ 7 alpha_phys), so the sign could be more significant there. This concern is load-bearing because if the sign is not negligible, the ensembles are phase-quenched and all subsequent measurements are biased. It is, however, a known and disclosed limitation, and the code is open source, so a user could in principle implement sign reweighting; the verdict should remain CONDITIONAL rather than REJECT. I agree with the reader's identification of this as the weakest assumption.","tokens_in":33218,"tokens_out":9589,"duration_ms":83896,"concrete_test":"On the Q*D1 ensemble (or a small-lattice replica with identical bare parameters), evaluate the average sign of pf(CTD_f) using an exact Pfaffian computation for small volumes or the stochastic estimator of refs. [34,35] on a sample of 20-50 configurations. If the average sign is consistent with 1 within statistical errors of, say, 1% and does not fluctuate across configurations, the phase-quenched bias is negligible; if it deviates, the generated ensembles do not represent QCD+QED and the code's headline claim fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 2, Eq. (2.4) replaces det D_f and pf(CTD_f) by their absolute values, so the simulated distribution rho_sim in Eq. (2.6) is the phase-quenched theory, not the target rho_tar of Eqs. (2.2)-(2.3). The paper asserts this is a mild sign problem citing ref. [18] and that it becomes irrelevant close to the continuum, but no numerical evidence is provided. The test runs in Sec. 4 do not address the sign: Hamiltonian conservation (Fig. 4) tests the integrator, solver benchmarks (Figs. 6-7) test linear algebra, and the reweighting factors W_q in Fig. 8 quantify only the rational-approximation error, not the sign. If the sign is non-negligible at the parameters used (alpha_0 = 0.05 ~ 7 alpha_phys, m_PS ~ 660 MeV, a ~ 0.074 fm), the Q*D1 ensemble is biased with respect to the QCD+QED path integral, and the central claim that openQ*D can generate QCD+QED configurations is undermined. The paper itself defers this to future work, making it a live assumption.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents openQ*D-1.0, an open-source lattice QCD+QED simulation package built on openQCD-1.6 and NSPT-1.4. The code supports SU(3), U(1), and SU(3)×U(1) theories with O(a) improved Wilson fermions, periodic or C* spatial boundary conditions, several temporal boundary conditions, RHMC with rational approximations and frequency splitting, multiple deflation subspaces, and Fourier acceleration for the U(1) field. The manuscript gives the theoretical background for the actions and algorithms, a user guide for the dynamical simulation program iso1, and a set of tests: low-level code checks, Hamiltonian conservation tests with and without Fourier acceleration, solver benchmarks in an electroquenched setup, and a stable dynamical test run labeled Q*D1. The central claim is that openQ*D-1.0 is a stable, well-tested package that can generate QCD+QED configurations with C* boundary conditions.","tokens_in":33358,"tokens_out":6092,"duration_ms":61889,"significance":"If the claims hold, this is a valuable software contribution: it is the first open-source package of its kind for dynamical QCD+QED with C* boundary conditions, and it makes the framework of ref. [18] concretely usable. The strengths of the paper are that the code is publicly released, the low-level test suite covers C* boundary conditions, Dirac operators with generic electric charge, rational approximations and forces, and the Hamiltonian-conservation test reproduces the expected integrator orders. The authors also report a stable dynamical test run, which is a nontrivial algorithmic milestone. The main caveat is that the sign of the fermion determinant/Pfaffian is dropped in the simulated distribution, so the paper's claims must be read as claims about the phase-quenched theory unless the sign is shown to be negligible; this is acknowledged in the text but not quantified.","major_comments":[{"comment":"The treatment of the determinant/Pfaffian sign is a load-bearing assumption for the paper's central claim. Equation (2.4) replaces det D_f and pf(CT D_f) by their absolute values, so the simulated distribution rho_sim in Eq. (2.6) is the phase-quenched theory rather than the target rho_tar of Eqs. (2.2)-(2.3). The text says that the sign problem is mild, citing ref. [18], and that it becomes irrelevant sufficiently close to the continuum, but no numerical evidence is given at the parameters of the Q*D1 run (alpha_0 = 0.05 ~ 7 alpha_phys, m_PS ~ 660 MeV, a ~ 0.074 fm). The tests in Sec. 4 do not probe the sign: Hamiltonian conservation (Fig. 4) tests the integrator, the solver benchmarks (Figs. 6-7) test linear algebra, and the reweighting factors W_q in Fig. 8 quantify only the rational-approximation error, not the sign. If the sign is non-negligible at these parameters, the Q*D1 ensemble is biased with respect to the QCD+QED path integral, and the statement that openQ*D can generate QCD+QED configurations is not demonstrated. I ask the authors either to provide a quantitative estimate of the sign, for example by computing the stochastic estimate of the sign on a subset of generated configurations, or to explicitly and consistently recast the claims as applying to the phase-quenched theory, with the sign reweighting identified as a necessary additional ingredient that is not yet implemented.","section":"Sec. 2, Eq. (2.4); Sec. 4.5"},{"comment":"The Q*D1 run is presented as evidence that the code is stable for dynamical QCD+QED, but the run is explicitly an unphysical version of Nf=2+1 QCD+QED: the two down-type quarks are degenerate and the up-type quark is significantly heavier because the bare masses were taken equal. This is acknowledged in Sec. 4.5, and the run is described as sufficient to probe observables and performance. That is acceptable for a tools paper, but the wording of the abstract and Summary ('designed to perform lattice simulations of QCD+QED') should be qualified so that a reader does not infer that the package has been demonstrated to produce physical QCD+QED ensembles. The distinction between a code that can generate ensembles for the phase-quenched theory and a code that has been validated for the target QCD+QED distribution should be made prominent in the introduction or abstract.","section":"Sec. 4.5 and Summary"}],"minor_comments":[{"comment":"There is a typo: 'better rational appriximation' should read 'better rational approximation'.","section":"Sec. 4.5, text near Fig. 8"},{"comment":"The text uses 'Schwartz-Alternating-Procedure'; the standard spelling in lattice QCD is 'Schwarz' (as in Hermann Schwarz).","section":"Sec. 4.4"},{"comment":"The scaling plots have no error bars or repeated-run information, and the y-axis notation '16·t[16 cores]/t' is slightly ambiguous; a sentence stating how many runs were averaged and how representative the single-node baseline is would improve reproducibility.","section":"Fig. 3 and Sec. 4.1"},{"comment":"The instructions to modify 'lines 122-124 of main/Makefile' are tied to a specific line numbering that may change between versions; referencing the variable names (e.g., CFLAGS) would be more robust.","section":"Sec. 3.2.1"},{"comment":"The derivation of the small-determinant action S_sdet is only sketched; a short explanation of the identity det(D_f) = det(D_f,oo) det(hat D_f) in the text would help readers who are not already familiar with even-odd preconditioning.","section":"Eq. (2.5) and Sec. 2.3"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid software-description manuscript, and the code release is a genuine community resource. My main concern is that the sign-of-the-determinant issue is acknowledged but not addressed quantitatively, and the current wording allows readers to infer stronger validation than the tests actually provide. I believe this is fixable by either adding a sign estimate or carefully scoping all claims to the phase-quenched ensemble, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take on arXiv:1908.11673. The real content is the code release: openQ*D is the first open-source package that does dynamical QCD+QED with C* boundary conditions, and it extends its parent codes with multiple deflation subspaces (needed for quarks of different charge), Fourier acceleration for the U(1) field, rational approximations of arbitrary powers, and twisted-mass reweighting within RHMC. The paper documents the implementation carefully and ships a real test suite: low-level checks for C* boundary conditions, Dirac operators with generic charge, rational approximants, forces, plus Hamiltonian-conservation scalings that match the expected integrator orders and a stable dynamical run (Q*D1). That is genuine, reproducible work and should be credited.\n\nThe soft spots are in the testing, not the code. The phase-quenched approximation in Eq. (2.4) is the one that matters: the simulated distribution replaces determinant and Pfaffian by their absolute values, and the paper asserts the sign is mild citing [18], but nothing in Sec. 4 probes it. The Hamiltonian test tests the integrator; the solver benchmarks test linear algebra; the reweighting factors W_q quantify rational-approximation error, not the sign. At the test parameters (alpha_0 ~ 7 alpha_phys, m_PS ~ 660 MeV) that assumption is live. The paper itself says the sign will be investigated in future work, so it is not hiding the issue—but it means the central claim should be worded as 'generates ensembles for the phase-quenched theory' unless the sign is shown to be negligible. That is a fixable, addressable gap, not a fatal one.\n\nMinor: the benchmarks use one representative configuration without repeated-run error bars, and the raw data behind the figures are not shipped—annoying for reproducibility but not disqualifying. The scale setting for Q*D1 is admittedly rough. The citation pattern is fine; refs [18] and [20] are the authors' own theoretical framework, which is appropriate here since the code implements that framework.\n\nWho is this for? Any lattice group wanting to do QCD+QED with C* boundaries or with charged hadrons in finite volume. I'd take it seriously and would send it to a competent referee familiar with HMC and the C* formulation. It deserves peer review and, after the sign issue is either addressed or clearly scoped, likely publication as a code/tools paper. I'd cite it if I worked in this area.","headline":"A solid, genuinely useful code paper; the phase-quenched sign assumption is the one gap that should be explicitly scoped before acceptance.","tokens_in":34017,"tokens_out":1886,"would_cite":true,"duration_ms":17973,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["11.15.-q","11.15.Ha","12.20.-m","12.38.Gc","12.38.-t","02.70.-c","02.70.Uu"],"model":"deepseek-v4-flash","headline":"A new open-source code makes lattice QCD+QED simulations practical.","keywords":["lattice QCD","lattice QED","C* boundary conditions","rational hybrid Monte Carlo","Fourier acceleration","isospin breaking","open source code","Wilson fermions"],"falsifier":"Compute the average sign of the Pfaffian on the Q*D1 ensemble, or at lighter pseudoscalar masses, by evaluating $\\mathrm{pf}(C T D)$ on a sample of configurations; if the average sign departs from +1 by more than a few percent, or fluctuates strongly, the ensembles require sign reweighting and the code's stated assumption of a mild sign problem would fail at those parameters.","tokens_in":32922,"feed_emoji":"⚛️","tokens_out":12656,"duration_ms":105237,"temperature":0.7,"pith_summary":"This paper presents openQ*D-1.0, an open-source lattice simulation package designed to make QCD+QED calculations practical: it generates gauge-field configurations for quantum chromodynamics coupled to quantum electrodynamics, with or without C* boundary conditions. The payoff is a local, gauge-invariant way to compute isospin-breaking effects and QED radiative corrections to hadronic observables from first principles, avoiding the non-local actions that most existing approaches use to handle charged particles on a finite torus. The paper argues that the implementation of the C* orbifold, rational hybrid Monte Carlo with reweighting, multiple deflation subspaces, and Fourier acceleration for the U(1) field is correct, and it supports this with low-level tests, Hamiltonian-conservation checks, solver benchmarks, and a stable dynamical test run.","feed_headline":"Open-source code makes QCD+QED lattice simulations practical","feed_subtitle":"C* boundary conditions let charged hadrons be computed from first principles, enabling radiative corrections.","key_machinery":"The load-bearing mechanism is the C* orbifold: spatial directions are closed by C-parity boundary conditions, implemented by allocating a doubled lattice and mapping the mirror copy through shifted boundary conditions, which turns the fermion determinant into a Pfaffian and makes charged states accessible without non-local gauge fixing. Around this sit three algorithmic pieces: the rational hybrid Monte Carlo (RHMC) with a rational approximation of fractional powers of the Dirac operator and twisted-mass reweighting; multiple deflation subspaces so that quarks with different electric charges each receive their own low-mode subspace; and Fourier acceleration of the U(1) molecular-dynamics update, where the kinetic term uses an inverse Laplacian preconditioning. The package also provides O(a) improvement terms (Sheikholeslami-Wohlert) and a choice of SU(3) gauge actions.","core_discovery":"The central claim is that a single open-source program, openQ*D-1.0, can simulate QCD, QED, and QCD+QED with O(a)-improved Wilson fermions under periodic or C* boundary conditions, and that the implementation of the C* orbifold, the rational hybrid Monte Carlo with reweighting, multiple deflation subspaces, and Fourier acceleration for the U(1) field is correct and efficient. With C* boundary conditions, fields obey a charge-conjugation relation across the spatial boundary, implemented by simulating a doubled lattice; this yields the Pfaffian target distribution and provides a local gauge-invariant framework in which electrically charged hadrons can be studied on a finite volume. The paper reports that the dynamical test run Q*D1 is stable, and that the reweighting factors stay close to one, supporting the practical viability of the approach.","pith_inferences":["A direct test of the sign assumption would be to evaluate the Pfaffian sign on the Q*D1 ensembles and at lighter quark masses; a non-negligible average sign would require reweighting and would refine the paper's statement about the mildness of the sign problem.","The modular design, with the Dirac operator taking the electric charge as a parameter and the U(1) action kept separate, suggests the code could be extended to non-compact QED or arbitrary charge assignments without changing the core solver machinery.","If ensembles like Q*D1 are pushed to the physical point, the C*-boundary setup could become a standard first-principles route to isospin-breaking corrections, complementing methods that treat QED perturbatively."],"forward_implications":["Radiative corrections to hadronic observables can be computed from first principles on the lattice without non-local QED actions, because C* boundary conditions provide a local, gauge-invariant framework for charged states.","The code makes it practical to simulate QCD+QED with up- and down-type quarks of different electric charge, since each flavour can have its own deflation subspace and the solver performance stays charge-insensitive.","The demonstrated stability of the dynamical run Q*D1 and the smallness of the reweighting factors indicate that production ensembles at moderate quark masses are within reach.","Users can pick periodic, Schrödinger-functional, open, or open-SF boundary conditions in time, so the package can support scale-setting and step-scaling studies in addition to spectrum calculations."],"supporting_citations":[{"why":"Supplies the C* boundary-condition formulation and the Pfaffian target that the code implements.","marker":"[18]"},{"why":"Base code from which the Dirac operator, solvers, and HMC/RHMC machinery are inherited.","marker":"[2]"},{"why":"Source of the FFT module adapted for Fourier acceleration of the U(1) field.","marker":"[3]"},{"why":"Numerical evidence that charged-hadron masses can be extracted gauge-invariantly with C* boundary conditions.","marker":"[20]"},{"why":"Provides the rational-approximation method for fractional powers of the Dirac operator used in RHMC.","marker":"[37]"},{"why":"Supplies twisted-mass reweighting and frequency splitting used to control the RHMC.","marker":"[38,39]"},{"why":"Introduces the inexact deflation algorithm on which the locally deflated solver is based.","marker":"[22]"},{"why":"Establishes gauge-invariant Fourier acceleration, whose implementation is adapted for the U(1) field.","marker":"[23,24]"}],"fun_headline_variants":["OpenQ*D: open-source QCD+QED simulations","C* boundaries make charged hadrons computable","One package for QCD, QED, and QCD+QED","Versatile lattice tool: QCD and QED together"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the sign of the fermion determinant and Pfaffian is mild enough to be ignored at the simulated parameters; if that sign is not negligible, the generated ensembles would not represent the target QCD+QED path integral.","fun_headline_variants_meta":{"raw":{"variants":["OpenQ*D: open-source QCD+QED simulations","C* boundaries make charged hadrons computable","One package for QCD, QED, and QCD+QED","Versatile lattice tool: QCD and QED together"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000726,"raw_usage":{"total_tokens":3220,"prompt_tokens":880,"completion_tokens":2340,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":496,"completion_tokens_details":{"reasoning_tokens":2269}},"tokens_in":496,"tokens_out":2340,"duration_ms":17828,"temperature":1.0,"reasoning_tokens":2269,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:09:04.303031+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the average sign of the Pfaffian on the Q*D1 ensemble, or at lighter pseudoscalar masses, by evaluating $\\mathrm{pf}(C T D)$ on a sample of configurations; if the average sign departs from +1 by more than a few percent, or fluctuates strongly, the ensembles require sign reweighting and the code's stated assumption of a mild sign problem would fail at those parameters.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Base code from which the Dirac operator, solvers, and HMC/RHMC machinery are inherited."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Source of the FFT module adapted for Fourier acceleration of the U(1) field."},{"cited_title":"Gauge invariant determination of charged hadron masses","cited_arxiv_id":"1802.05474","evidence_quote":"Numerical evidence that charged-hadron masses can be extracted gauge-invariantly with C* boundary conditions."}],"review_version":1}