Pith. sign in

REVIEW 3 major objections 4 minor 38 references

Upwind embedded boundary SBP operators: New high order numerical schemes for arbitrarily shaped domains with Cartesian grids

T0 review · 3 major / 4 minor · reviewed 2026-08-02 · deepseek-v4-flash

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

Pith's one-line read This paper claims that new upwind SBP operators embed arbitrary boundaries at up to 9th interior order and demonstrate ~5.5-order convergence in 3D wave tests.

desk verdict New upwind/DRP embedded-boundary SBP operators are a real advance, but the 3D test injects exact solution near the excision boundary, so stability for genuine IBVPs remains unproven. read the letter →

arxiv 2607.14403 v2 pith:YTSQH4XR submitted 2026-07-15 gr-qc cs.NAmath-phmath.MPmath.NA

Conner Dailey This is my paper · ORCID
classification gr-qccs.NAmath-phmath.MPmath.NA MSC 65M0665M1283-08
keywords upwindsummationbypartsembeddedboundaryfinitedifferencehighorderaccuracyexcisionnumericalrelativitydispersionrelationpreservingdiagonalnorm
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 tries to establish that upwind summation-by-parts operators can be built for embedded boundaries—surfaces that cut through a Cartesian grid rather than aligning with grid lines—at interior accuracy orders p=2 through 9, with boundary closure accuracy up to order 4. The construction uses a parameterized ansatz for the derivative-pair (Q+, H), enforces the 2(b+1) accuracy conditions, and sets leftover free parameters by minimizing boundary error. If correct, these are the highest-order embedded-boundary finite-difference methods derived to date, and the 3D curvilinear wave-equation tests with a spherical excision show total L2 convergence near order 5.5. The payoff is that black-hole excision and worldtube coupling in numerical relativity could run on plain Cartesian grids without boundary-conforming coordinate transforms.

What carries the argument

The carrying object is the embedded-boundary upwind SBP pair (Q±, H): two derivative operators sharing a diagonal norm H and a boundary operator B = -e_l e_l^T + e_r e_r^T built from p-th-order accurate interpolation vectors. With Q+ = -(Q-)T, SBP holds automatically, and the scheme's dissipation is the matrix S = H(D+ - D-)/2, which is negative semi-definite and independent of the boundary position alpha. The construction enforces the accuracy conditions (Q± + 1/2 B) X^q - H X^{q-1} = 0 for q=1..b, then minimizes the total boundary error f over the free parameters to pin down a unique operator.

What would settle it

Take the D±_{9-4} operator and evolve the scalar wave equation with an excision sphere that is shallow relative to a grid line (so trapped points appear), or with a moving sphere whose velocity is comparable to the wave speed, or with characteristic data injected from a curved metric rather than a plane wave; if the L2 error grows without bound or the convergence rate drops below design as resolution increases, the stability claim fails for that regime.

Watch

Extended reading notes

Core claim

The central discovery is a family of diagonal-norm upwind SBP operators for embedded boundaries, labeled D±_{p-b} and DRP variants, for interior orders 2..9 and boundary orders b up to 4. The operators satisfy SBP by construction via the identity (Q± + 1/2 B) X^q = H X^{q-1} for q=1..b, and the remaining free parameters are fixed by minimizing the total boundary error f = sum over q=b+1..2b-1 of the L2 norms of the error vectors. The paper shows each operator has a positive-definite norm on a length-one interval of the boundary position parameter alpha, has S negative semi-definite (so it supplies numerical dissipation), and converges in the 3D excision test at the designed b+1.5/b+1 rates,

Load-bearing premise

The scheme's stability in 2D/3D with embedded boundaries is assumed rather than proven—energy stability is shown only in 1D, and the 3D robustness tests use a friendly sphere geometry with exact-solution injection at boundary points, avoiding the trapped-point and narrow-domain issues that can arise for moving or shallow boundaries.

Editorial extensions

If this is right

  • Black-hole excision in numerical relativity could use a stationary Cartesian grid, eliminating the need for dual-frame coordinate evolution when the excision boundary moves.
  • Worldtube excision—coupling an analytical interior to a numerical exterior through characteristic boundary conditions—becomes a high-order option on Cartesian grids.
  • The upwind dissipation S is built into the operators and independent of alpha, so no separate artificial-dissipation operator is needed for stability in the tested regime.
  • Odd-order operators have better spectral radii than their even-order counterparts at the same interior stencil width, meaning larger stable time steps and equal memory loads, making them attractive for memory-bound codes.
  • Setting alpha = 0 recovers traditional boundary-conforming SBP operators, so the new operators generalize standard codes rather than replacing them.

Reading between the lines

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

  • The stability evidence in 3D rests on a single benign geometry: a convex sphere with no trapped points, no slender gaps, and exact time derivatives injected at near-boundary points. A moving or shallow-angle excision boundary could excite embedded-boundary modes that the fixed dissipation may not damp, so the robustness claim is narrower than the operator existence claim.
  • Since the dissipation matrix is alpha-independent, implementers could tabulate operator coefficients once per alpha range and interpolate, avoiding repeated symbolic derivation in production codes—an optimization the paper does not pursue.
  • The DRP operators sacrifice accuracy and time-step for better phase-velocity fidelity; for long-distance gravitational-wave propagation they might outperform the maximal-order operators, but this needs a dedicated study with distant observers.
  • The reported convergence order ~5.5 for the 9th-order interior operator is limited by the 4th-order boundary closure; pushing boundary order b to 5 or higher would likely raise the asymptotic rate, but the paper does not show such operators.
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

3 major / 4 minor

Summary. This paper constructs new upwind summation-by-parts (SBP) finite-difference operators for embedded boundaries on Cartesian grids. The construction in Section III builds a parameterized ansatz for (Q+,H), imposes polynomial accuracy conditions (Eqs. 18-19), minimizes a boundary-truncation-error function f (Eq. 21), and verifies positive-definiteness of H and negative semi-definiteness of S. This yields minimal-width operators D±_{p-b} for interior orders p=2..9 with boundary order up to 4, plus dispersion-relation-preserving (DRP) variants. The operators are then tested in Section IV on the 3D curvilinear scalar wave equation in a multiblock cubed-sphere grid with an inner spherical excision, with convergence rates reported for the total L2, inner-block L2, and L∞ norms. The paper claims these are the highest-order embedded-boundary finite-differencing operators derived to date and argues that they are suitable for black-hole and worldtube excision on Cartesian grids.

Significance. If the operator construction and the accompanying claims are accepted, this is a useful and nontrivial contribution: it extends embedded-boundary SBP from the previous order range to interior order 9 and boundary order 4, introduces upwind and DRP variants, and provides explicit operators with positive norms and dissipative S. The supplementary materials with polynomial coefficients and a Mathematica notebook are a strength, as is the fact that the closure parameters are determined by an internal error-minimization criterion rather than fitted to convergence data. However, the numerical demonstration in Section IV does not, as designed, exercise the embedded-boundary closure as a genuine initial-boundary-value problem, because active points just inside the excision sphere are evolved with exact time derivatives of the reference solution. This weakens the paper's central robustness claim for realistic excision applications, although the operator-existence and accuracy claims appear defensible.

major comments (3)
  1. [III.A, Eqs. (18)-(19)] The text calls Eqs. (18)-(19) the '2(b+1) accuracy conditions', but the conditions are displayed only for q∈[1,b], which is 2b conditions. More importantly, starting at q=1 omits the zero-th order consistency condition (Q± + (1/2)B)X^0 = 0 (with X^{-1} understood as zero), i.e. that the derivative of a constant vanishes. Without this condition the boundary row-sum constraints are not enforced, and the stated uniqueness and free-parameter counts in Section III.A and Table I do not follow from the equations as written. The authors should correct the stated q-range (probably q=0,...,b) and confirm that the actual derivation and the supplementary materials impose the q=0 condition.
  2. [IV.A and IV.B] The 3D convergence/robustness test does not exercise the embedded-boundary closure as a genuine IBVP. Section IV.A states that points just inside the excision boundary are active but 'evolved via the exact time derivatives obtained by differentiating the reference solution', and both the inner and outer boundary data are injected as exact incoming characteristics. Thus the stencils that include these boundary-layer points are fed with analytically prescribed values rather than values generated by the PDE, so any unstable boundary mode is suppressed by construction. The convergence rates in Figures 4-15 validate the interior stencils, multiblock interfaces, and boundary interpolation against exact data, but they do not demonstrate stability of the embedded-boundary operator when coupled with characteristic data derived from the interior solution. Please add a test in which all active poin
  3. [III.B and IV.A] The paper itself concedes in Section III.B that 2D/3D stability with embedded boundaries 'cannot be proven' and relies on 'the hope that with appropriate numerical dissipation, the scheme can be stabilized.' Section IV.A also states that the test geometry avoids trapped points, slender geometries, and narrow domains. Given these limitations, the conclusion in Section V that the operators 'can be used to evolve hyperbolic problems in first order form with arbitrarily shaped boundaries' is broader than the evidence supports. Either provide additional stability evidence for a representative non-benign configuration (for example an inclined or tilted boundary, or a domain requiring the derivative-extrapolation treatment for trapped points), or narrow the stated claims to the tested spherical-excision setting with injected inner solution.
minor comments (4)
  1. [III.B] Typo: 'dissapative' should be 'dissipative' in the phrase 'asymmetric dissapative scheme'.
  2. [IV.C] Typo: 'nontheless' should be 'nonetheless'.
  3. [III.A.2 / Table I] The distinction between the minimal-width operators and the DRP operators is hard to follow in plain text because the two families appear with nearly identical notation. Please make the notational difference explicit in all table captions, figure captions, and the text.
  4. [IV.A] The sentence describing active points inside the excision boundary should be more prominent: it is the key reason the subsequent convergence test is not a genuine IBVP test for the embedded boundary closure. At minimum, the authors should state what is injected and why this does not test feedback from the interior solution.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the operator parameters are fixed by internal accuracy and error-minimization criteria, and the numerical convergence rates are genuine predictions, not fitted outcomes.

full rationale

The derivation chain is self-contained. The novel operators are defined by a parameterized ansatz for (Q+,H) (Eq. 17), with the free parameters constrained by the 2(b+1) accuracy conditions (Eqs. 18-19) and then set by minimizing the internal boundary error function f (Eq. 21), together with positive-definiteness of H and negative semi-definiteness of S. No parameter is fitted to the convergence data, and the observed rates (b+1.5 and b+1) are compared with independent theoretical expectations for upwind SBP operators. The self-citation [14] is used only to describe a prior application of the embedded-boundary framework and is not load-bearing for the operator derivations presented here. The manuscript itself identifies the principal caveat: 2D/3D stability with embedded boundaries is not proven (Sec. III.B), and the 3D test injects exact time derivatives and characteristic data near the excision boundary (Sec. IV.A). That is an evidentiary limitation of the robustness demonstration, not a circularity: the test does not encode the claimed convergence rates into the operator design. No step in the derivation reduces by construction to its own inputs.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claim rests on: (1) standard SBP identities, satisfied by construction (Eqs. 7-10); (2) the [13] embedded-boundary framework — interpolation vectors e_l,r (Eq. 15) and boundary operator (Eq. 16) — adopted wholesale; (3) the [15]/[16] interior stencils, taken as given; (4) the [20] prediction that upwind SBP converges at b+1.5/b+1, assumed to transfer to embedded closures; (5) the unproven premise that fixed dissipation stabilizes the 3D embedded scheme (Sec. III B). The derived operators are concrete constructed matrices; no new physical entities are postulated.

free parameters (3)
  • Boundary-closure free parameters (1/4/9 per operator family) = Not printed in text; promised in supplementary coefficient tables
    D±4−2/D±5−2 have 1 closure parameter, D±6−3/D±7−3 have 4, D±8−4/D±9−4 have 9 (Sec. III A). Set by minimizing the operator's own boundary error function f (Eq. 21) — an internal truncation-error criterion independent of the wave-equation test, so the convergences measured in Sec. IV are predictions, not fits.
  • Boundary position range [α_min, α_max) = [0,1) for p=2; [−1/2,1/2) for p=4,5; [−3/5,2/5) for p=6,7; [−2/3,1/3) for p=8,9
    Construction step 6 (Sec. III A): the unit-length interval is selected per operator so the norm H is positive definite and the maximum spectral radius is minimized (Table I). This design input controls the stable time step and excludes extrapolative α values; it is not fitted to the test data.
  • Dissipation weight on H^{-1}S term = 1
    The asymmetric dissipative scheme (Eq. 29) adds I2⊗(H^{-1}S) with unit weight. Sec. III B notes a weight >1 is possible but 'was not found to be necessary' — the demonstrated 3D stability depends on this untuned default.
assumptions (5)
  • domain assumption The interior upwind stencils of [15] and the DRP stencils of [16] have the accuracy and dispersion properties claimed there and transfer unchanged into the embedded-boundary ansatz.
    Section III A builds (Q+,H) from these interior stencils without re-deriving them. If the stencil coefficients were wrong, the operator orders and the abstract's dispersion claims would fail.
  • domain assumption The embedded-boundary representation B = −el el^T + er er^T with b-th-order interpolation vectors (Eqs. 14-16) from [13] satisfies the generalized SBP property in multi-dimensions via the line-by-line tensor-product construction.
    Section II A adopts [13]'s framework wholesale, including the treatment of trapped points by derivative extrapolation, none of which is tested in this work.
  • domain assumption In 2D/3D embedded-boundary use, the added S-based dissipation stabilizes the scheme even though the energy-method proof does not go through.
    Section III B: 'This cannot be proven however in the case of 2D or 3D in the presence of embedded boundaries... the hope is that with appropriate numerical dissipation, the scheme can be stabilized.' This is the load-bearing unproven premise for the robustness claim.
  • domain assumption The predicted convergence orders b+1.5 (total L2) and b+1 (central-block/L∞) for upwind SBP operators [20] hold for the embedded-boundary generalizations.
    Used to interpret Figures 4-17 and to label results as matching or exceeding expectations; the paper supplies no error analysis of its own for the embedded case.
  • domain assumption H positive-definiteness and S negative semi-definiteness over the quoted α ranges are computed correctly.
    Asserted per operator in Section III A ('the norm is positive definite in the range...', 'S is naturally negative semi-definite') without proof; the reader cannot check these without the promised operator coefficients.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Upwind embedded boundary SBP operators: New high order numerical schemes for arbitrarily shaped domains with Cartesian grids." pith.science (2026). https://pith.science/paper/YTSQH4XR

@misc{pith2026260714403,
  author       = {Pith},
  title        = {Pith review of: Upwind embedded boundary SBP operators: New high order numerical schemes for arbitrarily shaped domains with Cartesian grids},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YTSQH4XR}},
  note         = {Machine review of arXiv:2607.14403}
}
read the original abstract

Embedded boundary summation by parts (SBP) methods define finite differencing based derivative operators with the added feature that the boundary need not coincide with a grid cell, allowing a boundary to be embedded on a regular Cartesian grid. This is achieved by the introduction of interpolation/extrapolation operators that match the accuracy of the boundary closure. These methods have been used to perform black hole excision simulations on a domain with a spherical boundary embedded in a regular Cartesian grid, demonstrating their usefulness for nonlinear problems. In this work, new operators are derived using this embedded boundary framework to increase the order of accuracy of the interior and boundary closure while minimizing the boundary error. Additionally, these novel operators improve the spectral properties on the grid by generalizing to an upwind scheme that has better dispersion relation preserving properties compared to traditional SBP schemes for wave equations. These operators are tested with the curvilinear scalar wave equation on a 3D multiblock grid with an excision sphere embedded in the center block to demonstrate the robustness and accuracy of these novel embedded operators.

Figures

Figures reproduced from arXiv: 2607.14403 by the authors.

Figure 1
Figure 1. FIG. 1. Equatorial slice of the test domain grid for a resolu [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Spectral radii [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Spectral radii [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: FIG. 4. Convergence results for the [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 6
Figure 6. Figure 6: FIG. 6. Convergence results for the [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 8
Figure 8. Figure 8: FIG. 8. Convergence results for the [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 10
Figure 10. Figure 10: FIG. 10. Convergence results for the [PITH_FULL_IMAGE:figures/full_fig_p010_10.png]
Figure 12
Figure 12. Figure 12: FIG. 12. Convergence results for the [PITH_FULL_IMAGE:figures/full_fig_p011_12.png]
Figure 14
Figure 14. Figure 14: FIG. 14. Convergence results for the [PITH_FULL_IMAGE:figures/full_fig_p012_14.png]
Figure 16
Figure 16. Figure 16: FIG. 16. Comparison of the convergence of the [PITH_FULL_IMAGE:figures/full_fig_p012_16.png]
Figure 17
Figure 17. Figure 17: FIG. 17. Comparison of the convergence of the [PITH_FULL_IMAGE:figures/full_fig_p013_17.png]
Figure 14
Figure 14. Figure 14: Notably absent is a higher than expected con [PITH_FULL_IMAGE:figures/full_fig_p013_14.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references

  1. [13]

    Instruc- tion roofline: An insightful visual performance model for gpus.Concurrency and Computation: Practice and Ex- perience, 34(20):e6591, 2022

    Nan Ding, Muaaz Awan, and Samuel Williams. Instruc- tion roofline: An insightful visual performance model for gpus.Concurrency and Computation: Practice and Ex- perience, 34(20):e6591, 2022

    2022

  2. [1]

    Construct a parameterized ansatz for (Q +, H) given an upwind interior stencil

  3. [2]

    ConstructBwithb th order accuratee l,r vectors

  4. [3]

    Find a solution of (Q+, H) that satisfies the 2(b+1) accuracy conditions

  5. [4]

    Minimize the total error functionfwith respect to the remaining free parameters

  6. [5]

    Properties of all derived operators

    Ensure that the matrixSis negative semi-definite Operator [αmin, αmax) ρ(1) max ρ(2) max ρ(3) max D± 2−1 [ 0,1) 2.70 3.86 5.55 D± 3−1 [−1/5,4/5) 1.21 1.45 1.92 D± 4−2 [−1/2,1/2) 5.04 4.63 4.92 D± 5−2 [−1/2,1/2) 1.48 1.58 1.59 D± 6−3 [−3/5,2/5) 1.89 2.18 2.80 D± 7−3 [−3/5,2/5) 1.66 1.69 1.75 D± 8−4 [−2/3,1/3) 2.02 2.40 2.54 D± 9−4 [−2/3,1/3) 2.04 2.05 2.08...

  7. [6]

    Solutions for the DRP interior stencils of [16] are labeledD ± p−b

    Select a range [α min, αmax) forα max −α min = 1 for which the norm is positive definite and the maxi- mum spectral radius is minimized In this work, solutions were found for interior orders p= 2 throughp= 9 using the interior stencils of [15] and these operators are labeled asD ± p−b for interior orderp and boundary orderb. Solutions for the DRP interior...

  8. [7]

    Multi- block simulations in general relativity: high-order dis- cretizations, numerical stability and applications.Clas- sical and Quantum Gravity, 22(24):5283–5321, December 2005

    Luis Lehner, Oscar Reula, and Manuel Tiglio. Multi- block simulations in general relativity: high-order dis- cretizations, numerical stability and applications.Clas- sical and Quantum Gravity, 22(24):5283–5321, December 2005

    2005

Show all 38 references
  1. [8]

    The norm of theD ± 3−1 operator is positive definite in the range −0.6≲α l,r ≲1.4, and the range−1/5≤α l,r <4/5 min- imizes the spectral radius

    Minimal Width Upwind Operators TheD ± 2−1 andD ± 3−1 operators are unique in the minimal boundary closure, but surprisingly the norm forD ± 2−1 is only positive definite in the range −0.3≲α l,r ≲1.6, and the spectral radius is best in the purely extrapolating range 0≤α l,r <1....

  2. [9]

    The free param- eter can be used to minimize the total error function, and this results in a unique minimum andSis naturally negative semi-definite for this minimum

    Dispersion Relation Preserving Operators TheD ± 4−2 andD ± 5−2 operators have one free param- eter each, the norm is positive definite in the range −0.6≲α l,r ≲0.8, and the spectral radius is best in the range−1/2≤α l,r <1/2 for both cases. The free param- eter can be used to ...

  3. [10]

    Minimal Width Operators TheD ± 2−1 OperatorsThep= 2 case is the lowest order option possible and it requires an interior stencil width of 5. In a departure from the general trend of traditional SBP operators, thep= 2 case has the worst spectral radius properties by far, requir...

  4. [11]

    This suggests that around this order of accuracy, dispersion error is a significant driver in the overallL 2 error

    Dispersion Relation Preserving Operators TheD ± 4−2 OperatorsRequiring an interior stencil width of 9, the DRPp= 4 scheme shows higher than expected convergence, better even than theD ± 4−2 opera- tor as shown in Figure 12, at least in the pre-asymptotic regime. This suggests ...

    2000

  5. [12]

    Roofline: an insightful visual performance model for multicore architectures.Commun

    Samuel Williams, Andrew Waterman, and David Pat- terson. Roofline: an insightful visual performance model for multicore architectures.Commun. ACM, 52(4):65–76, April 2009

    2009

  6. [14]

    High-performance xpu stencil computations in julia, 2022

    Samuel Omlin and Ludovic R¨ ass. High-performance xpu stencil computations in julia, 2022

    2022

  7. [15]

    Duez, Lawrence E

    Enrique Pazos, Manuel Tiglio, Matthew D. Duez, Lawrence E. Kidder, and Saul A. Teukolsky. Orbiting binary black hole evolutions with a multipatch high or- der finite-difference approach.Physical Review D, 80(2), 2009

    2009

  8. [16]

    Dis- tributed parallelization of xpu stencil computations in ju- lia.Proceedings of the JuliaCon Conferences, 6(65):137, 2024

    Samuel Omlin, Ludovic R¨ ass, and Ivan Utkin. Dis- tributed parallelization of xpu stencil computations in ju- lia.Proceedings of the JuliaCon Conferences, 6(65):137, 2024

    2024

  9. [17]

    Traditionally, the interior stencil is taken to be of min- imal width for a given order of interior accuracy

    that some operators do not support these spurious modes in the asymmetric scheme, and all of the operators in that work lessen their existence compared to centered schemes, while [16] eliminates them almost entirely. Traditionally, the interior stencil is taken to be of min- i...

  10. [18]

    Optimized high-order derivative and dis- sipation operators satisfying summation by parts, and applications in three-dimensional multi-block evolutions

    Peter Diener, Ernst Nils Dorband, Erik Schnetter, and Manuel Tiglio. Optimized high-order derivative and dis- sipation operators satisfying summation by parts, and applications in three-dimensional multi-block evolutions. Journal of Scientific Computing, 32(1):109–145, mar 2007

    2007

  11. [19]

    A multi-block infrastructure for three- dimensional time-dependent numerical relativity.Classi- cal and Quantum Gravity, 23(16):S553–S578, 2006

    Erik Schnetter, Peter Diener, Ernst Nils Dorband, and Manuel Tiglio. A multi-block infrastructure for three- dimensional time-dependent numerical relativity.Classi- cal and Quantum Gravity, 23(16):S553–S578, 2006

    2006

  12. [20]

    Scheel, Harald P

    Mark A. Scheel, Harald P. Pfeiffer, Lee Lindblom, Lawrence E. Kidder, Oliver Rinne, and Saul A. Teukol- sky. Solving einstein’s equations with dual coordinate frames.Physical Review D, 74(10), November 2006

    2006

  13. [21]

    Wittek, Mekhi Dhesi, Leor Barack, Harald P

    Nikolas A. Wittek, Mekhi Dhesi, Leor Barack, Harald P. Pfeiffer, Adam Pound, Hannes R. R¨ uter, Marceline S. Bonilla, Nils Deppe, Lawrence E. Kidder, Prayush Ku- mar, Mark A. Scheel, William Throwe, and Nils L. Vu. 16 Worldtube excision method for intermediate-mass-ratio inspi...

    2023

  14. [22]

    Sandberg

    Mathieu Deuse and Richard D. Sandberg. Implementa- tion of a stable high-order overset grid method for high- fidelity simulations.Computers & Fluids, 211:104449, 2020

    2020

  15. [23]

    Computational Aeroacoustics

    Eric Manoha, St´ ephane Redonnet, and St´ ephane Caro. Computational Aeroacoustics. John Wiley & Sons, Ltd, 2010

    2010

  16. [24]

    A high-order ac- curate embedded boundary method for first order hy- perbolic equations.Journal of Computational Physics, 334:255–279, 2017

    Ken Mattsson and Martin Almquist. A high-order ac- curate embedded boundary method for first order hy- perbolic equations.Journal of Computational Physics, 334:255–279, 2017

    2017

  17. [25]

    Schnetter, and N

    Conner Dailey, E. Schnetter, and N. Afshordi. Formulat- ing the complete initial boundary value problem in nu- merical relativity to model black hole echoes.Classical and Quantum Gravity, 42(2):025002, 2024

    2024

  18. [26]

    Diagonal-norm upwind sbp operators

    Ken Mattsson. Diagonal-norm upwind sbp operators. Journal of Computational Physics, 335:283–310, 2017

    2017

  19. [27]

    Full-spectrum dispersion relation preserving summation-by-parts oper- ators.SIAM Journal on Numerical Analysis, 62(4):1565– 1588, 2024

    Christopher Williams and Kenneth Duru. Full-spectrum dispersion relation preserving summation-by-parts oper- ators.SIAM Journal on Numerical Analysis, 62(4):1565– 1588, 2024

    2024

  20. [28]

    Dual-pairing summation by parts finite dif- ference methods for large scale elastic wave simulations in 3d complex geometries.Journal of Computational Physics, 454:110966, 2022

    Kenneth Duru, Frederick Fung, and Christopher Williams. Dual-pairing summation by parts finite dif- ference methods for large scale elastic wave simulations in 3d complex geometries.Journal of Computational Physics, 454:110966, 2022

    2022

  21. [29]

    Spectrally-tuned compact finite- difference schemes with domain decomposition and appli- cations to numerical relativity.Journal of Computational Physics, 508:112958, 2024

    Boris Daszuta. Spectrally-tuned compact finite- difference schemes with domain decomposition and appli- cations to numerical relativity.Journal of Computational Physics, 508:112958, 2024

    2024

  22. [30]

    Ken Mattsson, David Niemel¨ a, and Andrew R. Winters. Optimal boundary closures for diagonal-norm upwind sbp operators, 2026

    2026

  23. [31]

    Upwind summation- by-parts finite differences: Error estimates and weno methodology.Journal of Scientific Computing, 100:75, 2024

    Yan Jiang and Siyang Wang. Upwind summation- by-parts finite differences: Error estimates and weno methodology.Journal of Scientific Computing, 100:75, 2024

    2024

  24. [32]

    High-accuracy finite-difference schemes for solving elastodynamic prob- lems in curvilinear coordinates within multiblock ap- proach.Applied Numerical Mathematics, 93:176–194,

    Leonid Dovgilovich and Ivan Sofronov. High-accuracy finite-difference schemes for solving elastodynamic prob- lems in curvilinear coordinates within multiblock ap- proach.Applied Numerical Mathematics, 93:176–194,

  25. [34]

    Parsani, David I

    M. Parsani, David I. Ketcheson, and W. Deconinck. Op- timized explicit runge–kutta schemes for the spectral dif- ference method applied to wave propagation problems. SIAM Journal on Scientific Computing, 35(2):A957– A986, 2013

    2013

  26. [35]

    R¨ ass, I

    L. R¨ ass, I. Utkin, T. Duretz, S. Omlin, and Y. Y. Pod- ladchikov. Assessing the robustness and scalability of the accelerated pseudo-transient method.Geoscientific Model Development, 15(14):5757–5786, 2022

    2022

  27. [36]

    Julia: A fresh approach to numerical com- puting.SIAM Review, 59(1):65–98, 2017

    Jeff Bezanson, Alan Edelman, Stefan Karpinski, and Vi- ral B Shah. Julia: A fresh approach to numerical com- puting.SIAM Review, 59(1):65–98, 2017

    2017

  28. [37]

    van Rees

    James Gabbard and Wim M. van Rees. A high-order finite difference method for moving immersed domain boundaries and material interfaces.Journal of Compu- tational Physics, 507:112979, 2024

    2024

  29. [38]

    Tensorial.jl: a Julia package for tensor operations, 2024

    Keita Nakamura. Tensorial.jl: a Julia package for tensor operations, 2024

    2024

  30. [2015]

    International Conference Difference Schemes and Applications in Honor of the 90-th Birthday of Professor V. S. Ryaben’kii

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.