Pith. sign in

REVIEW 3 major objections 5 minor 42 references

This paper supplies the missing characteristic decomposition for GRMHD simulations, giving explicit eigenvalues and eigenvectors for the divergence-cleaning conservative formulation.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-03 21:42 UTC pith:PXWWTHW2

load-bearing objection Plausible and useful result, but the left-eigenvector machinery depends on an unpublished companion paper; referee it, but demand the missing theorem. the 3 major comments →

arxiv 2511.13837 v2 pith:PXWWTHW2 submitted 2025-11-17 gr-qc astro-ph.HE

Characteristic Decomposition for Relativistic Numerical Simulations: II. Magnetohydrodynamics

Saul A. Teukolsky This is my paper
classification gr-qc astro-ph.HE
keywords general-relativistic magnetohydrodynamicscharacteristic decompositioneigenvalueseigenvectorsdivergence cleaningquasi-invertible transformationconserved variablesRiemann solvers
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

General-relativistic magnetohydrodynamics (GRMHD) simulations of neutron-star mergers, black-hole accretion, and jets rely on numerical schemes that need to know how the fluid splits into independent waves, each with its own speed. That wave decomposition was known only in the fluid's own comoving frame, not in the coordinate frame and conserved variables that simulation codes actually evolve. This paper derives the complete decomposition for the divergence-cleaning conservative formulation that is known to be strongly hyperbolic, producing explicit eigenvalues and right/left eigenvectors. The key is a 'quasi-invertible transformation' that handles the magnetic field as well as the velocity, and the resulting expressions are simpler than earlier attempts. If correct, the equations give numerical codes the ability to use full-wave Riemann solvers and characteristic boundary conditions.

Core claim

The paper supplies the missing complete characteristic decomposition for GRMHD in the 3+1 coordinate frame and in the conserved variables used by numerical codes: explicit eigenvalues and right/left eigenvectors for the divergence-cleaning conservative formulation. Building on the quasi-invertible transformation method introduced in Paper I, the author extends the technique to the magnetic field, transforming the known comoving-frame Anile eigenvectors to the Eulerian frame and then to the conserved variables, while showing that the transformation yields the correct left eigenvectors despite the non-invertibility. The paper shows that the divergence-cleaning formulation's eigenvalues are alm

What carries the argument

The quasi-invertible transformation — a method for transforming characteristic eigenvectors between frames or variable sets when the transformation matrix is not invertible in the usual sense, because the constraint u^a u_a = -1 makes the Jacobian from comoving to Eulerian variables rank-deficient. The method constructs a one-sided inverse, fixes its non-uniqueness by an additional 'inverse' requirement, and uses a new transformation law for left eigenvectors (Eq. 2.82) to obtain correct left eigenvectors even when a weaker condition holds. The paper's main technical step is extending this machinery to the magnetic-field variables, deriving the transformation matrices between (u^a, b^a) and

Load-bearing premise

The derivation's correctness depends on a theorem stated in the companion hydrodynamics paper but not restated or proven here: that the quasi-invertible transformation, which satisfies only a weaker condition than full invertibility, nevertheless yields the correct left eigenvectors; if that theorem does not apply to the ideal-MHD system, the transformed eigenvectors would be wrong even though the comoving Anile eigenvectors are correct.

What would settle it

At a generic point, form the characteristic matrix A^a q_a of the divergence-cleaning conservative formulation (Eq. 4.12) contracted with a representative direction q_a, compute its eigenvalues and right/left eigenvectors numerically, and compare them to the closed-form expressions Eqs. (4.13), (4.32), (4.34)-(4.35) over a random sample of primitive states. Any disagreement would refute the paper's central claim.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • GRMHD codes can now implement full-wave Riemann solvers, the most accurate known flux prescriptions, rather than relying on approximate solvers.
  • Characteristic boundary conditions can be imposed accurately in GRMHD simulations, improving treatment of inflows, outflows, and shocks.
  • The decomposition applies to the divergence-cleaning conservative formulation, which is strongly hyperbolic and therefore well-posed for numerical evolution.
  • The relatively simple algebraic form of the eigenvectors makes them practical to evaluate at every grid point in a simulation.
  • The same frame-transformation method yields the Anile-formulation eigenvectors in conserved variables for comparison, connecting the new results to existing literature.

Where Pith is reading between the lines

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

  • If the decomposition is correct, it should also enable characteristic-based shock detection and limiting for GRMHD, extending techniques long used in hydrodynamics but previously unavailable for magnetized relativistic flows.
  • The quasi-invertible transformation may be applicable to other constrained hyperbolic systems where rank-deficient Jacobians arise, such as relativistic viscous hydrodynamics or multifluid plasmas.
  • A natural test of the paper's premise is to compare simulations using this decomposition against existing constrained-transport codes: if the divergence constraint is maintained to machine precision, the weakly hyperbolic Valencia formulation might behave as well as the strongly hyperbolic one, as the paper speculates.
  • The deferred degeneracy analysis could be tackled by a perturbative regularization at degenerate points, or by using the characteristic matrix directly to compute eigenvectors numerically, which the paper suggests as a temporary measure.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper derives the characteristic decomposition of the ideal GRMHD system in the conservative variables used by numerical codes, with a divergence-cleaning scalar field. It starts from the comoving Anile formulation, obtains the eigenvalues and right/left eigenvectors, then uses the 'quasi-invertible transformation' method of the companion Paper I to pass to Eulerian variables and to conserved variables. The central new results are the scalar-field eigenvalues y = ±1 and the conservative-formulation eigenvectors, Eqs. (4.13), (4.32), (4.34)-(4.35). The paper also gives a detailed determinant evaluation in Appendix B and compares with the results of Anile, Antón et al., Schoepe et al., and Hilditch & Schoepe.

Significance. If correct, this fills a known gap: full-wave Riemann solvers for GRMHD require eigenvectors in evolved variables, which were previously unavailable. The paper's strengths are its explicit determinant calculations, several cross-checks against published comoving and Eulerian results, and the relative simplicity of the final expressions. However, the central left-eigenvector transformation law is imported from an unavailable companion paper, and the magnetic-field extension is justified only by a claim about a residual term. The ultimate correctness of the novel conservative left eigenvectors therefore cannot be checked from the manuscript as submitted.

major comments (3)
  1. [§II.B, Eq. (2.82)] The transformation law for left eigenvectors, L = L_old (A^a_old n_a) ∂U_old/∂U, is taken from Theorem 7.13 of Paper I, cited as arXiv:xxxx.yyyyy. This theorem is not restated or proved, and the companion paper is not available. Because all Eulerian and conserved left eigenvectors in §§II.B.4, III.E, and IV.D.2 rely on this step, the central claim is not independently verifiable. Please include the theorem statement with its hypotheses, or a self-contained proof tailored to the MHD system, or a companion manuscript that can be inspected.
  2. [§II.B, Eqs. (2.61)-(2.65)] The magnetic-field extension explicitly fails the inverse condition (2.61); the residual is (2.65). The text asserts that this is acceptable because A^a_old n_a annihilates the residual when multiplied on the right, but no calculation is shown. This is the load-bearing step that distinguishes the present method from a simple variable transformation. A detailed verification is needed: show that for the constructed matrices (2.67)-(2.68), (A^a_old n_a) times the residual term in (2.65) vanishes identically, and that the weakened quasi-invertibility is sufficient for Eq. (2.82).
  3. [§III.E, Eq. (3.42); §IV.D.2, Eqs. (4.34)-(4.35)] The final left eigenvectors in conserved variables are new and are not cross-checked against any published result. The text says computer algebra was used (§II.B.4), but no notebook or output is provided. Because an error in any single component would render the decomposition unusable, please supply a machine-checkable verification — for example, a script that checks L(A^a q_a)=0 and L R diagonal for a random sample of parameters in both the primitive and conservative forms.
minor comments (5)
  1. [References] Update Ref. [11] to a real arXiv number or DOI; 'arXiv:xxxx.yyyyy' is not citable.
  2. [Throughout] Typos: 'nonnonservative' in §II introduction, 'Jaocobian' in §III introduction, 'decompostion' in §I, 'amd' in Appendix A.
  3. [Eq. (2.94)] The xentropy display appears to omit the final +1 component in the (ρ,ε) variables; check the typesetting.
  4. [§V] The paper defers degeneracies to future work. Since denominators such as r1 in Eq. (2.76) and 1/G in Eq. (4.34) appear, a brief catalog of the degenerate limits would help implementers.
  5. [§IV.E] The discussion of constrained transport and weak hyperbolicity is speculative; consider labeling it explicitly as a conjecture.

Circularity Check

1 steps flagged

Load-bearing left-eigenvector transformation is imported from the author's own unpublished Paper I; otherwise the derivation is largely self-contained.

specific steps
  1. self citation load bearing [§II B 1, Eqs. (2.61)–(2.65); §II B 4, Eq. (2.82); Ref. [11]]
    "The question mark over the equals sign in Eq. (2.61) is there because we will find that we cannot in fact accomplish this requirement. Instead, we will find a weaker requirement that will still give a quasi-invertible transformation. ... As shown in Paper I, the transformation is given by L=L old (Aa old na) ∂Uold/∂U (2.82) (Eq. 7.13 of Paper I)."

    The paper's advertised complete decomposition in conserved variables (Eqs. (4.13), (4.32), (4.34)–(4.35)) depends on the left-eigenvector transformation rule (2.82), which is taken verbatim from the author's own Paper I [11], cited only as arXiv:xxxx.yyyyy and not proved or restated here. In extending this to MHD, the paper explicitly fails the inverse condition (2.61) and replaces it with the weaker nonzero residue (2.65); no argument is given that Paper I's theorem still applies to the 11-component divergence-cleaning system. Thus a load-bearing step is justified solely by self-citation to an unpublished, unavailable prior paper rather than by derivation in this text. Independent checks against Refs. [10], [12], and [13] and the first-principles determinant calculation keep the central r

full rationale

This is a derivation paper, not a fit-labeled-as-prediction paper, so there is no fitted-input circularity. The comoving eigensystem is derived directly from the characteristic matrix (Eqs. (2.26), (2.32), (4.14)), and the determinant is evaluated from first principles in Appendix B. Agreement with the comoving Anile eigenvectors, with Ref. [5], with Ref. [10], and with Refs. [12]–[13] provides external anchoring. However, the transformation of left eigenvectors from the comoving to the Eulerian/conserved frame relies on Eq. (2.82), stated as Eq. (7.13) of the author's own Paper I, whose argument ID is a placeholder and whose theorem is not restated. Moreover, the MHD generalization explicitly does not satisfy the inverse relation (2.61) and settles for a weaker condition (2.65), with no proof that Paper I's quasi-invertibility theorem extends to this case. This makes the central claim—that Eqs. (4.32), (4.34)–(4.35) give the correct conserved-variable left eigenvectors for divergence-cleaning GRMHD—partly dependent on an unverified self-citation. Score 4 reflects that some self-citation is load-bearing while substantial independent derivation and external checks remain.

Axiom & Free-Parameter Ledger

0 free parameters · 6 axioms · 0 invented entities

No free parameters were fitted; the paper is an analytical derivation. Its load-bearing axioms are the companion-paper quasi-invertibility theorem (unpublished), the ideal-MHD/divergence-cleaning modeling assumptions, the standard thermodynamic closure, and the assumption of non-degenerate generic configurations.

axioms (6)
  • ad hoc to paper Quasi-invertible transformation theorem of Paper I correctly gives left and right eigenvectors when the inverse transformation is only one-sided.
    The entire derivation imports this theorem from Ref [11], cited as arXiv:xxxx.yyyyy and not available. §IIB and Eq. (2.82) rely on it; a weaker condition replaces Eq. (2.61), so the method's validity for MHD is assumed rather than proven here.
  • domain assumption Ideal MHD / perfect conductor: the electromagnetic field is fully determined by the comoving magnetic field b^a = -*F^{ab} u_b.
    Eq. (2.1) and throughout. This is standard for GRMHD codes but restricts the scope to ideal MHD.
  • domain assumption The divergence-cleaning formulation of Hilditch & Schoepe [10] is equivalent to the flux-conservative equations used in numerical codes.
    §IV: 'The derivation below is patterned after that work, which shows that these equations are linear combinations of the flux-conservative ones.' The eigenvectors are derived for this formulation, not for the weakly hyperbolic Valencia formulation.
  • standard math The characteristic determinant identities in Appendix B are correct.
    The determinant evaluations in Appendix B and the generalized matrix determinant lemma (B9) are proven in the text; they are standard linear algebra, though the MHD-specific contractions are complex.
  • domain assumption Thermodynamic closure p = p(ρ, ε) with χ, κ, and sound speed c_s.
    Eqs. (2.10)-(2.12). Required for the primitive/conserved transformations in §III.
  • domain assumption Non-degenerate field orientations: the listed eigenvectors form a complete basis away from degeneracies.
    Degenerate cases are not treated; §V says 'this work will have to be redone for the conservative case' and proposes a numerical fallback.

pith-pipeline@v1.3.0-alltime-deepseek · 27583 in / 12449 out tokens · 122579 ms · 2026-08-03T21:42:17.068389+00:00 · methodology

0 comments
read the original abstract

The characteristic decomposition for GRMHD in the comoving frame of the fluid has been known for a long time. However, it has not been known in the coordinate frame of the simulation and in terms of the conserved variables evolved in typical numerical simulations. This paper applies the method of quasi-invertible transformations developed in Paper I to derive this decomposition. Among other benefits, this will allow us to use the most accurate known computational methods, such as full-wave Riemann solvers. The results turn out to be simpler than expected based on earlier attempts.

discussion (0)

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

Reference graph

Works this paper leans on

42 extracted references · 1 canonical work pages

  1. [1]

    We then set to zero the determinant of the characteristic matrix Aaqa

    Characteristic speeds As explained in Section II of Paper I, to find the char- acteristic speeds in the comoving frame, we introduce a vector qa =y uua +ζ a,(2.27) where yu is the comoving eigenvalue (characteristic speed) and ζa is a unit vector orthogonal to ua defining the direction of the decomposition. We then set to zero the determinant of the chara...

  2. [2]

    tangential

    Right Eigenvectors in the Comoving Frame The right eigenvectors follow from solving (Aaqa)X = 0. TakeXto be of the form X=   Xb Yb X7 X8   .(2.31) 5 Here the unknown vectors Xb and Yb can be taken to be orthogonal to ub. Using the explicit form (2.26) of Aa and the definitions (2.29), we get the equations aXc + 1 ρh∗ bbYb hacqa −BY c +qahacX7 + ...

  3. [3]

    TakeLto be of the form L= Lb Yb L7 L8 ,(2.39) with Lb and Yb orthogonal to ub

    Left Comoving Eigenvectors The left eigenvectors follow from solving L(Aaqa) = 0 analogously to the steps to find the right eigenvectors in §II A 2. TakeLto be of the form L= Lb Yb L7 L8 ,(2.39) with Lb and Yb orthogonal to ub. Using the explicit form (2.26) of Aa and the definitions (2.29), we get the equations aLc −BY c +h caqa(bbYb +ρhc 2 sL7 +pL 8/ρ) ...

  4. [4]

    Transformation between(v a, Ba)and(u a, ba) As shown in Paper I, to get both the left and right transformed eigenvectors, we need the Jacobian matrix for the transformation from ( ua, ba) to ( va, Ba) as well as the matrix for the inverse transformation. Recall the detailed discussion there for the case without a magnetic field: Because of the constraint ...

  5. [5]

    (2.27) becomes qa =yn a +s a,(2.69) where sa is a unit spatial vector in the direction of the de- composition

    Eigenvalues for the Eulerian frame In the Eulerian frame, the vectorqa defined in Eq. (2.27) becomes qa =yn a +s a,(2.69) where sa is a unit spatial vector in the direction of the de- composition. Here y is now the eigenvalue in the Eulerian frame. The degenerate eigenvalues corresponding to a = qaua = 0 have the value in the Eulerian frame y=v asa.(2.70)...

  6. [6]

    The expressions of the eigenvectors have been obtained after tedious algebraic manipulations

    Right Eigenvectors in the Eulerian frame The transformation from the comoving to the Eule- rian right eigenvectors is given by the Jacobian matrix ∂(va, Ba)/∂(ub, bb). This matrix acts on the vector part of each eigenvector. The scalar part of the eigenvectors is unchanged. a. Degenerate EigenvectorsApplying this transfor- mation to the degenerate eigenve...

  7. [7]

    The reason is that the transformation is not invertible, but only quasi-invertible

    Left Eigenvectors in the Eulerian frame The transformation from comoving to Eulerian frames for the left eigenvectors is more complicated than the one for the right eigenvectors. The reason is that the transformation is not invertible, but only quasi-invertible. As shown in Paper I, the transformation is given by L=L old (Aa oldna) ∂Uold ∂U (2.82) 9 (Eq. ...

  8. [8]

    Valencia

    Transformation to(ρ, ϵ) Since most numerical codes use the pair ( ρ, ϵ) as primi- tive variables rather than (p, ϵ), we transform the Eulerian eigenvectors determined above to these variables. The transformation is invertible: dUold dU = ∂(p, ϵ) ∂(ρ, ϵ)= χ κ 0 1 ,(2.92) dU dUold = ∂(ρ, ϵ) ∂(p, ϵ)= 1/χ−κ/χ 0 1 .(2.93) The part of the transformation acting ...

  9. [9]

    Take X to be of the form (2.31) with an extra component X9 at the end

    Right Eigenvectors in the Comoving Frame The right eigenvectors follow from solving (Aaqa)X = 0. Take X to be of the form (2.31) with an extra component X9 at the end. Then the analog of Eqs. (2.32) is aXc + 1 ρh∗ bbYb hacqa −BY c +qahacX7 + Bbc ρh X7 − h∗bc h qbYb = 0,(4.14a) BXc −b c qaXa −aY c −h acqaX9 = 0,(4.14b) ρhc2 sqaXa +aX 7 − κB ρ X9 = 0,(4.14c...

  10. [10]

    There is an additional scalar L9 in the equation corresponding to (2.39)

    Left Eigenvectors in the Comoving Frame The equations L(Aaqa) = 0 for the left eigenvectors are very similar to the case of the Anile formulation treated in§II A 3. There is an additional scalar L9 in the equation corresponding to (2.39). The equations are the same as those in Eq. (2.40) except that there is an additional term in Eq. (2.40b) of the form h...

  11. [11]

    This matrix acts on the vector part of each eigenvector

    Right Eigenvectors in the Eulerian frame As for the non-conservative formulation, the transforma- tion from the comoving to the Eulerian right eigenvectors is given by the Jacobian matrix ∂(va, Ba)/∂(ub, bb). This matrix acts on the vector part of each eigenvector. The scalar part of the eigenvectors is unchanged. Since all the comoving right eigenvectors...

  12. [12]

    First, expand (2.84) to a 5 × 5 matrix by padding it with zeros

    Left Eigenvectors in the Eulerian frame The transformation (2.84) for the divergence cleaning case has some additional terms. First, expand (2.84) to a 5 × 5 matrix by padding it with zeros. Then add the following terms:   0W b avb/(ρh) 0 0 0 0 0 0 0−W H a 0 0 0 0−κ(b ana)/ρ 0 0 0 0−(b ana)/ρ 0−W v b 0 0−W   (4.27) Applying this transformati...

  13. [13]

    Since the transformation matrix (3.5) is the same, the eigenvectors for the conserved variables are the same as in Eq

    Right Eigenvectors for Conserved Variables As mentioned already in§IV C 1, with the exception of the scalar eigenvectors, the right eigenvectors are the 17 same as for the Anile case with a zero appended. Since the transformation matrix (3.5) is the same, the eigenvectors for the conserved variables are the same as in Eq. (3.39), with the exception of Rco...

  14. [14]

    Thus, the inverse transformation matrix derived in§III C simply gets an extra row and column of zeros, with a 1 in the lower- right corner

    Left Eigenvectors for Conserved Variables For the divergence cleaning case, the field ϕ is both a primitive and a conservative variable. Thus, the inverse transformation matrix derived in§III C simply gets an extra row and column of zeros, with a 1 in the lower- right corner. Applying the transformation to the left eigenvectors given in§IV C 2, we get Len...

  15. [15]

    The Method of Anton et al. In Ref. [12], Ant´ on et al. found the characteristic decom- position for MHD in special relativity by a transformation method. They start with the covariant Anile eigenvectors for the 10 variables Ucov = (ua, ba, p, s). The particular choice of the two thermodynamic variables is not rele- vant for the discussion here. They use ...

  16. [16]

    dual frames

    The Method of Schoepe et al. Schoepe et al. [ 5] introduced an algorithm for charac- teristic decomposition in relativity that relies on a dual frames approach. The two frames in the algorithm are again the comoving and the Eulerian frames of the fluid. The formalism is derived from the dual foliation formal- ism [23], but because the comoving frame does ...

  17. [17]

    Now it is easy to show that

    Non-Conservative F ormulation In the usual comoving approach in the literature, one typically deals with a 10 × 10 characteristic determinant corresponding to the variables (ua, ba) and two thermody- namic variables. The standard form of the determinant is given in [ 3], with the remark “Now it is easy to show that . . . ” and the value of the determinant...

    1966

  18. [18]

    Divergence Cleaning F ormulation Equation (4.12) gives in the comoving frame det(Aaqa) = det   aδij (−Bδ ij +b iqj)/(ρh∗)−q ibj/(ρh) (Bb j +ρhq j)/(ρ2hh∗) 0 0 −Bδ ij +q ibj aδij 0 0q j ρhc2 sqi 0a0−κB/ρ (p/ρ)qi 0 0a−B/ρ 0q i 0 0a   .(B14) Multiply the last column by qi/a and subtract from the second column to eliminate qi in the bottom row. ...

  19. [19]

    Bruhat, Etude des equations des fluides charg´ es rela- tivistes inductifs et conducteurs, Comm

    Y. Bruhat, Etude des equations des fluides charg´ es rela- tivistes inductifs et conducteurs, Comm. Math. Phys.3, 334 (1966)

    1966

  20. [20]

    A. M. Anile and S. Pennisi, On the mathematical struc- ture of test relativistic magnetofluiddynamics, Annales de l’I.H.P. Physique th´ eorique46, 27 (1987), avail- able at http://www.numdam.org/item/AIHPA_1987__46_ 1_27_0

    1987

  21. [21]

    A. M. Anile,Relativistic fluids and magneto-fluids : with applications in astrophysics and plasma physics(Cam- bridge University Press, New York, 1989)

    1989

  22. [22]

    M. H. P. M. van Putten, Maxwell’s equations in divergence form for general media with applications to MHD, Comm. Math. Phys.141, 63 (1991)

    1991

  23. [23]

    Schoepe, D

    A. Schoepe, D. Hilditch, and M. Bugner, Revisiting hy- perbolicity of relativistic fluids, Phys. Rev. D97, 123009 (2018), arXiv:1712.09837 [gr-qc]

    Pith/arXiv arXiv 2018

  24. [24]

    J. A. Font, Numerical hydrodynamics and magnetohydro- dynamics in general relativity, Liv. Rev. Relativity11, 7 (2008)

    2008

  25. [25]

    M¨ osta, B

    P. M¨ osta, B. C. Mundim, J. A. Faber, R. Haas, S. C. No- ble, T. Bode, F. L¨ offler, C. D. Ott, C. Reisswig, and E. Schnetter, GRHydro: a new open-source general- relativistic magnetohydrodynamics code for the Ein- stein toolkit, Class. Quantum Grav.31, 015005 (2014), arXiv:1304.5544 [gr-qc]

    Pith/arXiv arXiv 2014

  26. [26]

    S. L. Liebling, L. Lehner, D. Neilsen, and C. Palenzuela, Evolutions of magnetized and rotating neutron stars, Phys. Rev. D81, 124023 (2010), arXiv:1001.0575 [gr-qc]

    Pith/arXiv arXiv 2010

  27. [27]

    A. J. Penner, General relativistic magnetohydrodynamic Bondi-Hoyle accretion, Mon. Not. R. Astron. Soc.414, 1467 (2011), arXiv:1011.2976 [astro-ph.HE]

    Pith/arXiv arXiv 2011

  28. [28]

    Hilditch and A

    D. Hilditch and A. Schoepe, Hyperbolicity of divergence cleaning and vector potential formulations of general rela- tivistic magnetohydrodynamics, Phys. Rev. D99, 104034 (2019)

    2019

  29. [29]

    S. A. Teukolsky, Characteristic decomposition for rela- tivistic numerical simulations: I. Hydrodynamics, arXiv e-prints , arXiv:xxxx.yyyyy (2025), arXiv:xxxx.yyyyy [gr- qc]

    2025

  30. [30]

    Ant´ on, J

    L. Ant´ on, J. A. Miralles, J. M. Mart ´ ı, J. M. Ib´ a˜ nez, M. A. Aloy, and P. Mimica, Relativistic magnetohydrodynam- ics: Renormalized eigenvectors and full wave decomposi- tion Riemann solver, Astrophys. J. Suppl.188, 1 (2010), arXiv:0912.4692 [astro-ph.IM]

    Pith/arXiv arXiv 2010

  31. [31]

    Ib´ a˜ nez, I

    J.-M. Ib´ a˜ nez, I. Cordero-Carri´ on, M.-´A. Aloy, J.-M. Mart ´ ı, and J.-A. Miralles, On the convexity of relativistic ideal magnetohydrodynamics, Class. Quantum Grav.32, 095007 (2015), arXiv:1503.00978 [gr-qc]

    Pith/arXiv arXiv 2015

  32. [32]

    C. W. Misner, K. S. Thorne, and J. A. Wheeler,Gravita- tion(W. H. Freeman, San Francisco, 1973)

    1973

  33. [33]

    Ant´ on,Magnetohidrodin´ amica relativista num´ erica: Aplicaciones en relatividad especial y general, Ph.D

    L. Ant´ on,Magnetohidrodin´ amica relativista num´ erica: Aplicaciones en relatividad especial y general, Ph.D. thesis, Universitat de Val` encia (2008), (in Spanish)

    2008

  34. [34]

    C. R. Evans and J. F. Hawley, Simulation of magne- tohydrodynamic flows: A constrained transport model, Astrophys. J.332, 659 (1988)

    1988

  35. [35]

    D. S. Balsara and D. S. Spicer, A staggered mesh al- gorithm using high order Godunov fluxes to ensure solenoidal magnetic fields in magnetohydrodynamic simu- lations, J. Comp. Phys.149, 270 (1999)

    1999

  36. [36]

    T´ oth, The ∇·B=0 constraint in shock-capturing magnetohydrodynamics codes, J

    G. T´ oth, The ∇·B=0 constraint in shock-capturing magnetohydrodynamics codes, J. Comp. Phys.161, 605 (2000)

    2000

  37. [37]

    Deppe, W

    N. Deppe, W. Throwe, L. E. Kidder, N. L. Vu, K. C. Nelli, C. Armaza, M. S. Bonilla, F. H´ ebert, Y. Kim, P. Kumar, G. Lovelace, A. Macedo, J. Moxon, E. O’Shea, H. P. Pfeiffer, M. A. Scheel, S. A. Teukolsky, N. A. Wittek, et al., SpECTRE v2025.08.19, 10.5281/zenodo.16906840 (2025)

    work page doi:10.5281/zenodo.16906840 2025

  38. [38]

    Donat and A

    R. Donat and A. Marquina, Capturing shock reflections: An improved flux formula, J. Comp. Phys.125, 42 (1996)

    1996

  39. [39]

    M. A. Aloy, J. M. Ib´ a˜ nez, J. M. Mart ´ ı, and E. M¨ uller, GENESIS: A high-resolution code for three-dimensional relativistic hydrodynamics, Astrophys. J. Suppl.122, 151 (1999), arXiv:astro-ph/9903352 [astro-ph]

    Pith/arXiv arXiv 1999

  40. [40]

    J. M. Martin-Garcia, xact: Efficient tensor computer algebra,https://josmar493.dreamhosters.com(2025)

    2025

  41. [41]

    Hilditch, Dual foliation formulations of general relativity, arXiv e-prints , arXiv:1509.02071 (2015), arXiv:1509.02071 [gr-qc]

    D. Hilditch, Dual foliation formulations of general relativity, arXiv e-prints , arXiv:1509.02071 (2015), arXiv:1509.02071 [gr-qc]

    Pith/arXiv arXiv 2015

  42. [42]

    M. A. Scheel, H. P. Pfeiffer, L. Lindblom, L. E. Kid- der, O. Rinne, and S. A. Teukolsky, Solving Einstein’s equations with dual coordinate frames, Phys. Rev. D74, 104006 (2006), arXiv:gr-qc/0607056 [gr-qc]. [25]http://www.black-holes.org/SpEC.html

    Pith/arXiv arXiv 2006