Pith. sign in

REVIEW 3 major objections 5 minor 47 references

Mesh-Based Filtering to Alleviate Time-Step Restrictions in Runge--Kutta Discontinuous Galerkin Methods in Spherical-polar Coordinates: Application to the Euler Equations

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read This paper proves that a projection–prolongation filter applied after every Runge–Kutta stage makes a fine-mesh discontinuous Galerkin update exactly equivalent to evolving on a coarser merged mesh, letting simulations take the merged…

desk verdict Sound 1D equivalence theorem for a DG cell-merging filter; the 2D/3D stability claims are empirical and should be labeled as such. read the letter →

arxiv 2608.07345 v1 pith:PUNYR4IY submitted 2026-08-07 math.NA cs.NAphysics.comp-ph

Joseph Hunter , Eirik Endeve , Yulong Xing This is my paper · ORCID
classification math.NAcs.NAphysics.comp-ph MSC 65M6065M1235L65
keywords discontinuousGalerkinmethodsRunge–Kuttatimeintegrationspherical-polarcoordinatesCFLtime-steprestrictionmeshmergingprojection-prolongationfilterEulerequationscellanisotropy
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

The paper proposes a post-processing filter for explicit Runge–Kutta discontinuous Galerkin (RKDG) methods on spherical-polar meshes, where converging coordinate lines create extremely small cells near the origin and poles and force tiny stable time steps. The filter projects the fine-mesh solution onto an auxiliary "merged mesh" with larger cells, evolves on the fine mesh using the merged mesh's larger time step, and prolongs back after each RK stage. In one dimension the paper proves that this filtered update is exactly equivalent to evolving the RKDG discretization directly on the merged mesh, so the filtered method inherits the merged mesh's accuracy and stability. In two and three dimensions the same idea is implemented with metric-weighted projections, and numerical tests on the Euler equations show accurate solutions with time steps independent of angular resolution and up to roughly 100 times larger (with a projected factor of about 400 in three dimensions).

What carries the argument

The load-bearing object is the projection–prolongation filter: an $L^2$ projection of the fine-mesh DG solution onto the merged mesh, followed by restriction of the merged polynomial back onto the fine cells. In nodal form it is assembled from cell merging weights $c^{(m)}_{i,\alpha} = \hat\ell^{(p)}_i(x^{(m)}_\alpha)$, which express each merged-cell basis function as a linear combination of fine-cell basis functions, so the update on any fine cell is a weighted sum of nodal values from all fine cells in the merged cell. The proof of Theorem 1 uses the fact that the merged basis functions are valid test functions on the fine cells and that, summed over a merged cell, the single-valued numerical fluxes at internal interfaces cancel, leaving only the merged-cell boundary terms. In two and three dimensions the projection is weighted by the metric volume factor $\sqrt{\gamma}$, with a merged representation $\sqrt{\hat\gamma}$ matched by moments, to preserve mass and energy.

What would settle it

Take the one-dimensional Burgers equation with quadratic DG and SSPRK3, and at each stage apply a slope limiter on the fine mesh before the projection–prolongation step; compare the resulting nodal values with a run that applies the same limiter on the merged mesh before prolongation. Any difference shows the practical limiter ordering used in two and three dimensions breaks the Theorem 1 equivalence, pinpointing the mechanism behind the reported shock overshoot.

Watch

Extended reading notes

Core claim

The central discovery is Theorem 1: if at time $t^n$ the fine-mesh DG solution is the prolongation of a merged-mesh solution, then one Forward Euler step on the fine mesh with the merged-mesh time step, followed by projection onto the merged mesh, produces exactly the same solution as one Forward Euler step on the merged mesh. Because the merged basis functions are valid test functions on the fine cells, and because single-valued numerical fluxes at internal fine-cell interfaces cancel when summed over a merged cell, the projected fine-mesh residual equals the merged-mesh residual. The argument extends stage by stage to any explicit Runge–Kutta method, provided the projection–prolongation is applied at each inner stage. Consequently the filtered fine-mesh method is equivalent to the RKDG method on the merged mesh and inherits its stability and accuracy properties; in particular, evolving with the merged-mesh time step is stable even though that time step exceeds the fine-mesh CFL limit.

Load-bearing premise

The practical two- and three-dimensional implementation applies a slope limiter on the fine mesh before the filter, an ordering the authors explicitly note falls outside the conditions of Theorem 1, so the stability observed in those tests depends on the unproven assumption that the one-dimensional equivalence extends to multiple dimensions and survives that limiting order.

Editorial extensions

If this is right

  • The filtered RKDG method inherits the merged mesh's accuracy and stability in one dimension: with $N$ fine cells it reproduces the error of an $N/2$-cell merged computation while using a time step twice as large for pairwise merging.
  • Angular refinement no longer reduces the stable time step: in the two-dimensional Riemann test the number of time steps stays roughly constant as $N_\theta$ increases from 16 to 64, and the filtered $128\times64$ run uses about $3.6\times10^3$ steps versus $6.3\times10^4$ for the unfiltered mesh.
  • Conservation is preserved: relative changes in total mass and energy remain at $O(10^{-13})$ in the shock–bubble test, matching the unfiltered RKDG method.
  • In three dimensions the benefit is larger: filtered time steps are essentially independent of angular resolution, about 25 times larger on the $64\times4\times4$ mesh and 100 times on the $64\times8\times8$ mesh, with a projected factor of about 400 on the $64\times16\times16$ mesh.
  • The merging criterion can be relaxed to trade time-step size for angular resolution: the relaxed $128\times64$ mesh uses more steps than the standard merged mesh but has error closer to the unfiltered solution.

Reading between the lines

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

  • The equivalence argument is coordinate-agnostic, since it only uses element-wise polynomial spaces and single-valued fluxes, so the same filter should transfer to other structured curvilinear meshes with geometric anisotropy, such as cylindrical grids, provided a metric-weighted projection is used.
  • A testable extension: apply the slope limiter on the merged mesh before prolongation, as Remark 5 suggests, and compare the resulting 1D shock profiles with the fine-mesh-limited ordering used in the paper's 2D/3D tests; this would show whether the observed shock overshoot is an artifact of the limiter ordering.
  • Because the filtered time step becomes set by radial resolution rather than angular resolution, the merging criterion can be calibrated problem-by-problem, and the relaxed-criterion results in the paper indicate how to measure that accuracy-versus-speed tradeoff.
  • An implicit–explicit extension is natural: run explicit stages with the merged-mesh time step while performing implicit solves on the merged mesh, which would reduce cost for stiff source terms such as neutrino kinetics in core-collapse supernova simulations.
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.

Referee Report

3 major / 5 minor

Summary. The paper proposes a mesh-based filtering procedure for explicit Runge-Kutta discontinuous Galerkin (RKDG) methods on spherical-polar meshes. The method constructs an auxiliary merged mesh whose elements combine fine cells so that extreme angular-to-radial aspect ratios are removed, then evolves the solution on the original fine mesh using the larger time step allowed by the merged mesh, applying a projection-prolongation filter after each Runge-Kutta stage. In one dimension, Theorem 1 proves that, under exact quadrature and without limiters, one Forward Euler step on the fine mesh followed by projection equals one Forward Euler step on the merged mesh; Remark 3 extends this stage-by-stage to explicit Runge-Kutta methods. Numerical tests for linear transport and Burgers' equation confirm the expected convergence orders and show that the unfiltered fine-mesh update with the large time step blows up. In two and three dimensions, the authors introduce angular merging criteria (Eqs. 37 and 45) and a metric-weighted projection using sqrt(gamma) (Eqs. 38-40 and 46-48), and they demonstrate significantly larger stable time steps on a two-dimensional Riemann problem, a shock-bubble interaction, and a three-dimensional Sedov-Taylor blast wave. Section 4 explicitly discloses that the slope limiter is applied on the fine mesh before the filter, which the authors state falls outside the conditions of Theorem 1.

Significance. The one-dimensional equivalence theorem is a genuine and clearly presented contribution: it shows that projection-prolongation with a single-valued numerical flux reproduces exactly the merged-mesh DG update, so the filtered method inherits stability and accuracy in that idealized setting. The numerical validation is clean, including the blow-up of the unfiltered large-time-step update, and the paper is commendably explicit about the scope of its theorem. However, the practical two- and three-dimensional algorithm departs from the theorem in two load-bearing ways: the metric-weighted projection is not covered by the 1D analysis, and the production code applies the slope limiter on the fine mesh before filtering, an ordering the authors concede is outside the equivalence. Consequently, the abstract's and Section 5's claims that the filtered method 'inherits the accuracy and stability properties' are stronger than what the paper establishes for the tested 2D/3D method. The 2D/3D speedups in Figures 10 and 12 are valuable empirical results, but the mathematical bridge from Theorem 1 to those results remains incomplete.

major comments (3)
  1. [§4, first paragraph] The production implementation applies the slope limiter on T before the mesh-based filter, and the text states that this ordering 'falls outside the conditions for the equivalence established in Theorem 1; consequently, stability properties of the RKDG scheme on the merged mesh, such as the TVD property, are not guaranteed to carry over.' Because every 2D/3D result in Figures 8-12 is produced with this limiter ordering, the stability and robustness claims for the practical method are empirical only. Please either prove a stability statement for this ordering, modify the implementation so that the limiter is applied on the merged mesh before prolongation to T, or explicitly restrict the inheritance claim to the cases covered by Theorem 1 and Remark 5.
  2. [§3.1.2, Eqs. (38)-(40)] The metric-weighted projection is not covered by the one-dimensional analysis, and there is a concrete obstruction to the natural extension. Lemma 1 relies on the fact that the fine projection P annihilates u-Pu against every polynomial of degree at most k on a fine cell. In Eq. (39), sqrt(hat-gamma) is a degree-k polynomial on the merged cell, so for a merged test function phi of degree k, the integrand phi sqrt(hat-gamma) restricted to a fine cell has degree up to 2k. Consequently, bP[P[u]] = bP[u] is not guaranteed, and the induction base in Remark 2 fails unless additional conditions are imposed. For the same reason, the residual-sum argument in Theorem 1 does not automatically extend to the flux terms in Eq. (40), which involve sqrt(gamma) and sqrt(hat-gamma) on the two sides. The 2D/3D equivalence should either be proved under explicit hypotheses on the metric representation, or the claims should be weakened to empirical observations.
  3. [§3.2.2 and Eq. (28b)] The momentum equation in the Euler system contains a geometric source term (1/2) Pi^{ik} partial_j gamma_{ik}, and Remark 4 asserts that source terms can be included 'provided it is evaluated with the same quadrature and representation on the fine and merged meshes.' In the implementation, the fine-mesh and merged-mesh metric representations differ (sqrt(gamma) versus sqrt(hat-gamma) defined by moment matching), so the source-term evaluation on the fine mesh is not identical to the merged-mesh source-term evaluation. The equivalence for the momentum equation is therefore not established, and the conservation and accuracy statements for the 3D Euler tests should be presented as numerical evidence rather than as consequences of Theorem 1.
minor comments (5)
  1. [Figure 1 caption] The left-panel labels appear to list 'I5' twice in the fine-mesh sequence, which is likely a typo for I6.
  2. [§2.3.4, proof of Theorem 1] The notation x_{p-1/2} and x_{p+1/2} for the endpoints of a merged cell is used without definition; defining these endpoints would improve readability.
  3. [§3.2.1, Eq. (45)] The notation 'sin hat-theta' is ambiguous because hat-theta denotes a merged cell, not a specific angle; clarify how sin theta is evaluated on a merged cell, for example by using the cell midpoint or a quadrature rule.
  4. [§2.4.1] The sentence 'For the first order method, we include results when the fine-mesh solution is advanced with cDelta-t without applying the mesh-based filter' is followed by blow-up values in Table 1; stating explicitly that these values are included to demonstrate instability would help the reader.
  5. [General] The abstract contains the typo 'ans-stage' instead of 'an s-stage'; this should be corrected in the final version.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the filter analysis is a self-contained proof, with the admitted multi-dimensional gap being a limitation rather than an input-output identity.

full rationale

The central claim is Theorem 1, which is a constructive proof: given the projection/prolongation operators (16), one forward Euler step on the fine mesh with the merged-mesh time step followed by projection gives the same result as the merged-mesh DG update. The proof sums fine-cell weak forms, uses cancellation of internal numerical-flux terms, and invokes uniqueness of the L2 projection. No parameter is fitted to the target solution, and the merged-mesh construction and CFL choices are explicit geometric criteria (Equations (37), (45), (51)), not calibrated to the presented outputs. The paper's own Section 4 passage states that applying the slope limiter on the fine mesh before the filter 'falls outside the conditions for the equivalence established in Theorem 1; consequently, stability properties of the RKDG scheme on the merged mesh, such as the TVD property, are not guaranteed to carry over.' That is an honest, explicit limitation on the scope of the proven equivalence for the 2D/3D implementation, not a circular reduction. The metric-weighted projection (38)-(40) is likewise a stated generalization whose equivalence is not proven, but again this is a gap in proof coverage rather than an identity of inputs and outputs. Self-citations (e.g., thornado, prior DG applications) supply implementation context and background, not the load-bearing derivation. No step reduces by construction to its own inputs, and no fitted quantity is renamed as a prediction. No circularity is present.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central 1D claim rests only on standard DG assumptions plus the prolongation induction. The multi-dimensional claims add an unproven metric-weighted projection and a limiter ordering that the paper itself flags as outside the theorem.

free parameters (2)
  • Merging threshold ratio = 1 for Eq (37); 1/4 for Eq (51)
    The angular merge factor is chosen as the smallest power-of-2 integer making the merged proper length exceed the radial spacing, or one quarter of it in the relaxed test. This is a heuristic design constant, not fitted to the measured solutions.
  • CFL coefficient = 0.2 for Burgers; standard Cockburn-Shu values otherwise
    Stability limits taken from the literature; they set the time step but are not fitted to machine output and do not affect the equivalence theorem.
assumptions (4)
  • standard math Standard RKDG formulation with single-valued numerical flux
    Used throughout Section 2; the equivalence proof relies on cancellation of internal interface terms, which requires the numerical flux to be single-valued at each fine-cell interface.
  • standard math Exact quadrature or nodal-mass-consistent quadrature
    Assumed at the start of Section 2.3; needed so projection coefficients can be computed without mass-matrix inversion.
  • domain assumption Fine-mesh solution is the prolongation of the merged-mesh solution at each stage
    This is the inductive hypothesis of Theorem 1 and Remark 3; it holds for the initial condition by Lemma 1 but is broken when a slope limiter is applied on the fine mesh before filtering (Section 4).
  • ad hoc to paper Metric-weighted projection preserves the equivalence in 2D/3D
    The sqrt(gamma)-weighted projection in Eqs (38)-(40) and (46)-(48) was introduced after preliminary testing to maintain mass and energy conservation; no multi-dimensional analogue of Theorem 1 is proven.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Mesh-Based Filtering to Alleviate Time-Step Restrictions in Runge--Kutta Discontinuous Galerkin Methods in Spherical-polar Coordinates: Application to the Euler Equations." pith.science (2026). https://pith.science/paper/PUNYR4IY

@misc{pith2026260807345,
  author       = {Pith},
  title        = {Pith review of: Mesh-Based Filtering to Alleviate Time-Step Restrictions in Runge--Kutta Discontinuous Galerkin Methods in Spherical-polar Coordinates: Application to the Euler Equations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PUNYR4IY}},
  note         = {Machine review of arXiv:2608.07345}
}
abstract

We propose a mesh-based filtering approach to alleviate the severe timestep restrictions arising in explicit Runge--Kutta discontinuous Galerkin (RKDG) methods formulated in spherical-polar coordinates. The filter enables stable evolution on the original logically Cartesian mesh while using larger time steps associated with an auxiliary merged mesh constructed to eliminate the extreme cell anisotropies produced by converging coordinate lines near coordinate singularities. The filter is implemented as a sequence of post-processing operations applied within an $s$-stage RK time integrator, making it straightforward to incorporate into existing structured-mesh DG frameworks. We analyze the filter in one spatial dimension and prove that the filtered RKDG method is equivalent to evolving the RKDG discretization on a nonuniform mesh obtained by merging selected elements of the underlying uniform mesh. This equivalence implies that the filtered method inherits the accuracy and stability properties of the corresponding RKDG discretization on the merged mesh. We apply the mesh-based filter to an existing RKDG method for the Euler equations in spherical-polar coordinates and demonstrate, through selected two- and three-dimensional examples, its effectiveness in accelerating simulations through significantly larger stable timesteps.

Figures

Figures reproduced from arXiv: 2608.07345 by the authors.

Figure 1
Figure 1. The left panel depicts a one-dimensional fine mesh of eight cells and its merging region consisting of two merged cells. Blue vertical lines indicate the boundaries of the two merged cells. Dashed black lines indicate the boundaries of fine cells. The fine and merged cells have been labeled on the bottom and top axes, respectively. The right panel depicts a two-dimensional spherical-polar fine mesh (black) and its (… view at source ↗
Figure 2
Figure 2. Top row: example of local linear bases (left) on two cells (red and blue) and the basis for the two cells when merged together (green), and the merged basis functions (right) expressed locally on each cell. The nodal values are the cell merging weights as defined in Section 2.3.2. Bottom row: example of a linear DG solution across two cells (left), and the same solution (right) after the projection–prolongation step… view at source ↗
Figure 3
Figure 3. A flowchart describing the mesh-based filtering process using Forward Euler. In the next subsection we show, under the assumptions stated there, that advancing the fine-mesh DG solution by one Forward Euler step with time step ∆ct, followed by the projection–prolongation update (17), is equivalent to advancing the DG solution directly on the merged mesh Tb with the same time step. 2.3. Projection–Prolongation Filter… view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: The left panel shows the linear transport solution on a subset of Ω after one mesh-based filtering step, using explicit midpoint time stepping and linear polynomials. The black line denotes the exact solution, the red lines the solutions advanced with time step ∆ct, an…
Figure 5
Figure 5. Figure 5: The left panel plots the solution without mesh-based filtering (red) and the solution with mesh-based filtering (blue) to the Burgers’ equation at t = 2 using N = 80 cells. The right panel zooms in on the shock region of the solution in the left panel. The markers are …
Figure 6
Figure 6. Figure 6: Top row: the result of applying Equation (37) to a 32 × 32 mesh in coordinate space (left) and the same mesh transformed into the physical space (right). Bottom row: depicted are θ-ϕ coordinate meshes for four different slices of r for a 64 × 16 × 32 mesh. The slices o…
Figure 7
Figure 7. Figure 7: The upper half (0 ≤ θ ≤ π/2) of the merging region, Ω2, of all four Tb meshes used for the two-dimensional Riemann problem. Elements in which two or more fine cells have been merged together are shaded. The first three meshes are merged according to Equation (37), whil…
Figure 8
Figure 8. Figure 8: The four panels on each row plot radial lineouts of the density, pressure, θ-velocity, and r-velocity for the two-dimensional Riemann problem. The first row is the 128 × 64 fine mesh solution, the second row is the 128 × 64 mesh-filtering solution, and the last row is …
Figure 9
Figure 9. Figure 9: Weighted L p error over time for the two-dimensional Riemann problem for fine-mesh (solid lines) and mesh-based filtering solutions (dashed lines) over time on Ω (left column) and Ω1 (right column) for all four meshes. Recall Ω1 is the union of all elements of Tb that …
Figure 10
Figure 10. Figure 10: Results for the two-dimensional Riemann problem. The left panel plots the total number of time steps for 128× Nθ meshes. The blue line is for mesh-filtering solutions, while the red line is for fine mesh solutions. The green dots are for the 128 × 64 mesh with relaxed…
Figure 11
Figure 11. Figure 11: Time series Schlieren plots for the shock–bubble interaction problem on a 256 × 256 mesh. The left half of each panel shows the result with mesh-based filtering, while the right half is the fine mesh result. The plotted quantity is |∇ρ|/ρ, which provides a visualizati…
Figure 12
Figure 12. Figure 12: Top row: radial profiles of the mass density (left) and internal energy density (right) for the 3D Sedov–Taylor blast wave at time t = 1. The radial extent of the plots is restricted to a region around the wavefront. The solid black line is a 1D reference solution wit…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 24 canonical work pages

  1. [1]

    2017, Journal of Computational Physics, 341, 377, doi: https://doi.org/10.1016/j.jcp.2017.04.025

    Asaithambi, R., & Mahesh, K. 2017, Journal of Computational Physics, 341, 377, doi: https://doi.org/10.1016/j.jcp.2017.04.025

    doi:10.1016/j.jcp.2017.04.025 2017

  2. [2]

    1997, Applied Numerical Mathematics, 25, 151, doi: 10.1016/S0168-9274(97)00056-1

    Ascher, U., Ruuth, S., & Spiteri, R. 1997, Applied Numerical Mathematics, 25, 151, doi: 10.1016/S0168-9274(97)00056-1

    doi:10.1016/s0168-9274(97)00056-1 1997

  3. [3]

    W., Montero, P

    Baumgarte, T. W., Montero, P. J., Cordero-Carri´ on, I., & M¨ uller, E. 2013, Phys. Rev. D, 87, 044026, doi: 10.1103/PhysRevD.87.044026

    work page doi:10.1103/physrevd.87.044026 2013

  4. [4]

    1989, Journal of Computational Physics, 82, 64, doi: https://doi.org/10.1016/0021-9991(89)90035-1 27

    Berger, M., & Colella, P. 1989, Journal of Computational Physics, 82, 64, doi: https://doi.org/10.1016/0021-9991(89)90035-1 27

    doi:10.1016/0021-9991(89)90035-1 1989

  5. [5]

    M., & Lufkin, E

    Blondin, J. M., & Lufkin, E. A. 1993, The Astrophysical Journal Supplement Series, 88, 589, doi: 10.1086/191834

    work page doi:10.1086/191834 1993

  6. [6]

    2026, The Astrophysical Journal Supplement Series, 285, 48, doi: 10.3847/1538-4365/ae825b

    Endeve, E. 2026, The Astrophysical Journal Supplement Series, 285, 48, doi: 10.3847/1538-4365/ae825b

    work page doi:10.3847/1538-4365/ae825b 2026

  7. [7]

    1967, An Introduction to the Study of Stellar Structure (New York, NY: Dover Publications)

    Chandrasekhar, S. 1967, An Introduction to the Study of Stellar Structure (New York, NY: Dover Publications)

    work page 1967

  8. [8]

    1998, SIAM Journal on Numerical Analysis, 35, 2440, doi: https://doi.org/10.1137/S0036142997316712

    Cockburn, B., & Shu, C.-W. 1998, SIAM Journal on Numerical Analysis, 35, 2440, doi: https://doi.org/10.1137/S0036142997316712

    doi:10.1137/s0036142997316712 1998

Show all 47 references
  1. [9]

    Cockburn, B., & Shu, C.-W. 2001, J. Sci. Comput., 16, 173, doi: https://doi.org/10.1023/A:1012873910884

    2001 doi

  2. [10]

    1928, Mathematische Annalen, 100, 32, doi: 10.1007/BF01448839

    Courant, R., Friedrichs, K., & Lewy, H. 1928, Mathematische Annalen, 100, 32, doi: 10.1007/BF01448839

    1928 doi

  3. [11]

    J., Endeve, E., Mezzacappa, A., et al

    Dunham, S. J., Endeve, E., Mezzacappa, A., et al. 2024, The Astrophysical Journal, 964, 38, doi: 10.3847/1538-4357/ad206c

    2024 doi

  4. [12]

    J., et al

    Endeve, E., Buffaloe, J., Dunham, S. J., et al. 2019, Journal of Physics: Conference Series, 1225, 012014, doi: 10.1088/1742-6596/1225/1/012014

    2019 doi

  5. [13]

    A., et al

    Endeve, E., Mewes, V., Harris, J. A., et al. 2026, The Astrophysical Journal Supplement Series, 284, 40, doi: 10.3847/1538-4365/ae57ad

    2026 doi

  6. [14]

    2018, Monthly Notices of the Royal Astronomical Society, 477, 4543, doi: 10.1093/mnras/sty734

    Zanotti, O. 2018, Monthly Notices of the Royal Astronomical Society, 477, 4543, doi: 10.1093/mnras/sty734

    2018 doi

  7. [15]

    S., & Meyer, C

    Florinski, V., Guo, X., Balsara, D. S., & Meyer, C. 2013, The Astrophysical Journal Supplement Series, 205, 19, doi: 10.1088/0067-0049/205/2/19

    2013 doi

  8. [16]

    2000, The Astrophysical Journal Supplement Series, 131, 273, doi: 10.1086/317361

    Fryxell, B., Olson, K., Ricker, P., et al. 2000, The Astrophysical Journal Supplement Series, 131, 273, doi: 10.1086/317361

    2000 doi

  9. [17]

    J., & Winters, A

    Gassner, G. J., & Winters, A. R. 2021, Frontiers in Physics, Volume 8 - 2020, doi: 10.3389/fphy.2020.500690

    2021

  10. [18]

    S., & Bell, J

    Gulizzi, V., Almgren, A. S., & Bell, J. B. 2022, Journal of Computational Physics, 450, 110861, doi: 10.1016/j.jcp.2021.110861

    2022

  11. [19]

    P., & Xing, Y

    Hunter, J., Endeve, E., Laiu, M. P., & Xing, Y. 2025, Phys. Rev. D, 111, 043001, doi: 10.1103/PhysRevD.111.043001

    2025 doi

  12. [20]

    2023, Phys

    Ji, L., Mewes, V., Zlochower, Y., et al. 2023, Phys. Rev. D, 108, 104005, doi: 10.1103/PhysRevD.108.104005

    2023 doi

  13. [21]

    2004, Geochemistry, Geophysics, Geosystems, 5, doi: https://doi.org/10.1029/2004GC000734

    Kageyama, A., & Sato, T. 2004, Geochemistry, Geophysics, Geosystems, 5, doi: https://doi.org/10.1029/2004GC000734

    2004 doi

  14. [22]

    E., Field, S

    Kidder, L. E., Field, S. E., Foucart, F., et al. 2017, Journal of Computational Physics, 335, 84, doi: https://doi.org/10.1016/j.jcp.2016.12.059

    2017 doi

  15. [23]

    D., & Lifshitz, E

    Landau, L. D., & Lifshitz, E. M. 1987, Course of Theoretical Physics, Vol. 6, Fluid Mechanics, 2nd edn. (Pergamon Press)

    1987

  16. [24]

    2020, Phys

    Mewes, V., Zlochower, Y., Campanelli, M., et al. 2020, Phys. Rev. D, 101, 104007, doi: 10.1103/PhysRevD.101.104007

    2020 doi

  17. [25]

    Mezzacappa, A., Endeve, E., Messer, O. E. B., & Bruenn, S. W. 2020, Living Reviews in Computational Astrophysics, 6, 4, doi: 10.1007/s41115-020-00010-8

    2020 doi

  18. [26]

    M., & Schnetter, E

    Miller, J. M., & Schnetter, E. 2017, Classical and Quantum Gravity, 34, 015003, doi: 10.1088/1361-6382/34/1/015003 M¨ onchmeyer, R., & M¨ uller, E. 1989, A&A, 217, 351 M¨ uller, B. 2015, Monthly Notices of the Royal Astronomical Society, 453, 287, doi: 10.1093/mnras/stv1611 M¨...

    2017 doi

  19. [27]

    2019, Publications of the Astronomical Society of Japan, 71, 98, doi: 10.1093/pasj/psz080

    Nakamura, K., Takiwaki, T., & Kotake, K. 2019, Publications of the Astronomical Society of Japan, 71, 98, doi: 10.1093/pasj/psz080

    2019 doi

  20. [28]

    2006, Journal of Computational Physics, 213, 391, doi: 10.1016/j.jcp.2005.08.023

    Omang, M., Børve, S., & Trulsen, J. 2006, Journal of Computational Physics, 213, 391, doi: 10.1016/j.jcp.2005.08.023

    2006 doi

  21. [29]

    D., Abdikamalov, E., O’Connor, E., et al

    Ott, C. D., Abdikamalov, E., O’Connor, E., et al. 2012, Phys. Rev. D, 86, 024026, doi: 10.1103/PhysRevD.86.024026

    2012 doi

  22. [30]

    D., Abdikamalov, E., M¨ osta, P., et al

    Ott, C. D., Abdikamalov, E., M¨ osta, P., et al. 2013, The Astrophysical Journal, 768, 115, doi: 10.1088/0004-637X/768/2/115

    2013 doi

  23. [31]

    2005, Journal of Scientific Computing, 25, 129, doi: 10.1007/s10915-004-4636-4

    Pareschi, L., & Russo, G. 2005, Journal of Scientific Computing, 25, 129, doi: 10.1007/s10915-004-4636-4

    2005 doi

  24. [32]

    L., Endeve, E., et al

    Pochik, D., Barker, B. L., Endeve, E., et al. 2021, The Astrophysical Journal Supplement Series, 253, 21, doi: 10.3847/1538-4365/abd700

    2021 doi

  25. [33]

    2011, Phys

    Radice, D., & Rezzolla, L. 2011, Phys. Rev. D, 84, 024010, doi: 10.1103/PhysRevD.84.024010

    2011 doi

  26. [34]

    E., & Shephard, M

    Remacle, J.-F., Flaherty, J. E., & Shephard, M. S. 2003, SIAM Review, 45, 53, doi: 10.1137/S00361445023830

    2003 doi

  27. [35]

    1996, Journal of Computational Physics, 124, 93, doi: https://doi.org/10.1006/jcph.1996.0047

    Ronchi, C., Iacono, R., & Paolucci, P. 1996, Journal of Computational Physics, 124, 93, doi: https://doi.org/10.1006/jcph.1996.0047

    1996

  28. [36]

    2015, Monthly Notices of the Royal Astronomical Society, 453, 4278, doi: 10.1093/mnras/stv1859

    Schaal, K., Bauer, A., Chandrashekar, P., et al. 2015, Monthly Notices of the Royal Astronomical Society, 453, 4278, doi: 10.1093/mnras/stv1859

    2015 doi

  29. [37]

    C., K¨ appeli, R., & Liebend¨ orfer, M

    Scheidegger, S., Whitehouse, S. C., K¨ appeli, R., & Liebend¨ orfer, M. 2010, Classical and Quantum Gravity, 27, 114101, doi: 10.1088/0264-9381/27/11/114101 28

    2010 doi

  30. [38]

    L., & Teukolsky, S

    Shapiro, S. L., & Teukolsky, S. A. 1983, Black Holes, White Dwarfs, and Neutron Stars: The Physics of Compact Objects (New York, NY: John Wiley & Sons)

    1983

  31. [39]

    2016, Journal of Computational Physics, 316, 598, doi: 10.1016/j.jcp.2016.04.030

    Shu, C.-W. 2016, Journal of Computational Physics, 316, 598, doi: 10.1016/j.jcp.2016.04.030

    2016 doi

  32. [40]

    1988, Journal of Computational Physics, 77, 439, doi: https://doi.org/10.1016/0021-9991(88)90177-5

    Shu, C.-W., & Osher, S. 1988, Journal of Computational Physics, 77, 439, doi: https://doi.org/10.1016/0021-9991(88)90177-5

    1988 doi

  33. [41]

    2019, The Astrophysical Journal Supplement Series, 241, 7, doi: 10.3847/1538-4365/ab007f

    Vartanyan, D. 2019, The Astrophysical Journal Supplement Series, 241, 7, doi: 10.3847/1538-4365/ab007f

    2019 doi

  34. [42]

    Sod, G. A. 1978, Journal of Computational Physics, 27, 1, doi: https://doi.org/10.1016/0021-9991(78)90023-2

    1978 doi

  35. [43]

    Teukolsky, S. A. 2016, Journal of Computational Physics, 312, 333, doi: https://doi.org/10.1016/j.jcp.2016.02.031

    2016 doi

  36. [44]

    J., & M¨ uller, E

    Wongwathanarat, A., Hammer, N. J., & M¨ uller, E. 2010, A&A, 514, A48, doi: 10.1051/0004-6361/200913435

    2010 doi

  37. [45]

    2017, Phys

    Wu, K. 2017, Phys. Rev. D, 95, 103001, doi: 10.1103/PhysRevD.95.103001

    2017 doi

  38. [46]

    2021, SIAM Journal on Scientific Computing, 43, A472, doi: https://doi.org/10.1137/20M133782X

    Wu, K., & Xing, Y. 2021, SIAM Journal on Scientific Computing, 43, A472, doi: https://doi.org/10.1137/20M133782X

    2021 doi

  39. [47]

    2011, Journal of Computational Physics, 230, 1238, doi: https://doi.org/10.1016/j.jcp.2010.10.036

    Zhang, X., & Shu, C.-W. 2011, Journal of Computational Physics, 230, 1238, doi: https://doi.org/10.1016/j.jcp.2010.10.036

    2011 doi

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.