Pith. sign in

REVIEW 2 major objections 4 minor 3 cited by

2N-storage Runge-Kutta methods: Order conditions, general properties and some analytic solutions

T0 review · 2 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A single quadratic equation now generates every 4- and 5-stage third-order 2N-storage Runge-Kutta method.

desk verdict The structural results (constraints, corrected Williamson formula, recursive parameterization) are solid, but the central closed-form quadratic Eq. (95) is contradicted by the paper's own Table 5 example, so the headline closed-form solution claim is not supported as written. read the letter →

arxiv 2506.07359 v1 pith:SEO2X3RJ submitted 2025-06-09 math.NA cs.NAhep-latphysics.comp-ph

Alexei Bazavov This is my paper
classification math.NAcs.NAhep-latphysics.comp-ph MSC 65L06
keywords low-storageRunge-Kuttamethods2N-storageWilliamsonformorderconditionsButchertableauexplicitthird-orderaccuracycoefficientconversion
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 works out, for the first time, the explicit algebraic constraints that a classical Runge-Kutta tableau must satisfy to be implementable in Williamson's 2N-storage (two-registers-per-variable) form. With those constraints in hand, it shows that for third-order accuracy with four or five stages the entire method is determined by the nodes and weights, and that choosing one weight reduces the whole design to solving one quadratic equation. It also finds and corrects a longstanding error in Williamson's conversion formula for special cases where some weights vanish. If the derivation is right, designing new low-storage integrators of this class becomes a matter of picking a few free parameters and solving a single quadratic, which is why the author can report over two million rational-coefficient 5-stage schemes from a simple search.

What carries the argument

The central object is the α-form of the method, an intermediate representation where each stage updates from the previous stage (Equations (23)-(25)). Lemma 2 gives explicit expressions for the α-coefficients in terms of the storage coefficients A_i, B_i, and the ratio identity (42) follows by taking ratios of α-coefficients. This identity is converted into the explicit a-form constraints (45)-(47), which are then solved to yield the recursive parametrization (71) of the tableau by nodes c_i and weights b_i. The quadratic equation for b_2, Equations (95) and (115) for the (4,3) and (5,3) cases, is what carries the reduction: after eliminating all other weights, one free parameter remains and satisfies a single quadratic, whose coefficients are explicit rational functions of the chosen free nodes and weights.

What would settle it

One could take the (5,3) scheme in Table 2 (which has b_4 = 0 in the paper) and convert its Butcher tableau to storage form using the old equation (12) and the new equation (70); if the old conversion actually produced the same tableau and the same order conditions, the claim that (12) is wrong for i > 2 would lose its key supporting example. More generally, solving the full set of 2N-storage order conditions for a randomly chosen free-parameter selection in (95) or (115) and checking the resulting tableau against the constraints (45)-(47) would either confirm or break the claimed parametrization.

Watch

Extended reading notes

Core claim

The paper claims that the 2N-storage property of an explicit Runge-Kutta method can be expressed directly as a set of explicit quadratic constraints on the standard Butcher tableau, namely Equations (45)-(47), derived here for the first time. In the generic case these constraints can be solved: every sub-diagonal coefficient a_ij is determined by the nodes c_i and weights b_i through the recursive relation (71), and the storage-coefficients A_i, B_i are likewise given by (81)-(82). Applying this to third order with four and five stages reduces the full set of order-plus-storage conditions to a single quadratic equation for one weight, b_2; solving it and substituting back reconstructs the whole tableau. The paper further claims that Williamson's special-case formula (12), used when a weight b_i vanishes for i>2, is incorrect, and supplies the corrected relation (70). These claims are supported by explicit four- and five-stage schemes, including examples with vanishing weights that demonstrate the failure of the old formula.

Load-bearing premise

The closed-form results hold only in the generic case where all the relevant storage coefficients A_i and denominators such as the accumulated weights b_1+...+b_k minus a node c_j stay nonzero; special cases like repeated nodes or zero weights need to be handled separately, and for five stages they are not worked out here.

Editorial extensions

If this is right

  • If a user wants a 4- or 5-stage third-order 2N-storage scheme, they no longer need to solve a large nonlinear system: choosing free nodes/weights, solving one quadratic, and substituting into (71) produces the full tableau.
  • Because the final equation is quadratic, rational-coefficient schemes are easy to find; the paper demonstrates this with over two million 5-stage rational schemes from a grid search.
  • The corrected conversion formulas (61)-(63) and (70) replace Williamson's original branching formulas, removing the special-case trap for vanishing weights b_i with i>2.
  • For higher-order methods the explicit constraints (45)-(47) give a new way to formulate the design problem directly in standard Butcher-tableau parameters, rather than in storage parameters A_i, B_i.
  • The structure-preserving (Lie group) property observed for 2N-storage methods becomes easier to study now that all tableaux in the (4,3) and (5,3) families are parametrized explicitly.

Reading between the lines

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

  • Since the parametrization (71) depends only on the storage form and not on the order conditions, the same recursive construction may extend to families with more stages, and the author's suggestion that the final equation stays low-degree for any s with third order looks testable by computer algebra.
  • The corrected conversion formulas may change how previously published low-storage schemes with vanishing weights are interpreted; any implementation relying on Williamson's equation (12) for b_i = 0, i > 2, would silently compute a different (and lower-order) scheme.
  • The availability of millions of rational (5,3) schemes means the search for schemes with additional goals — larger stability regions, embedded error estimates, fourth-order accuracy on linear problems — becomes a filtering problem over an explicit rational family, which is exactly how the paper's illustrative schemes appear to have been selected.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 4 minor

Summary. The paper develops a structural theory of 2N-storage (Williamson-type) Runge-Kutta methods in the standard Butcher-tableau representation. It derives explicit 2N-storage constraints in both the alpha-form, Eq. (42), and the a-form, Eqs. (45)-(47); shows that, in a general nondegenerate case, all tableau entries a_ij and all low-storage coefficients A_i, B_i can be expressed through the nodes c_i and weights b_i, Eqs. (71), (81), (82); identifies and corrects an error in Williamson's special-case conversion formula, Eqs. (12) versus (70); and then claims closed-form general solutions for the families of four- and five-stage third-order 2N-storage methods, with the final condition reduced to a quadratic in b_2, Eqs. (95) and (115). Several rational schemes are exhibited and tested numerically.

Significance. If the central derivations are correct, the paper would make a useful contribution: the explicit 2N-storage constraints in the original Runge-Kutta parameters are new and potentially convenient for constructing low-storage methods; the correction of Williamson's formula is a concrete, checkable improvement; and a closed-form parametrization of all (4,3) and (5,3) 2N-storage methods would be a valuable reference result. The paper also provides reproducible Matlab code for the conversion issue and shows that the claimed framework can generate many rational schemes. However, the main closed-form result for (4,3) methods is contradicted by the paper's own example, so the significance is contingent on a corrected derivation.

major comments (2)
  1. [Sec. 7, Eq. (95) and Table 5] The central quadratic for the (4,3) family is not satisfied by the paper's own (4,3)_1 scheme. Substituting c2=1/4, c3=7/12, c4=4/5, b2=1/6 into the definitions (96)-(101) gives z1=-1/120, z2=1/60, z3=11/360, z4=-1/1440, z5=-13/288, and x=b2*c2=1/24; the left-hand side of Eq. (95) then evaluates to exactly 1/25920, not zero. This tableau satisfies the standard order conditions (B.1)-(B.4) and the 2N-storage constraints (45)-(47), and the text states that the listed schemes were found by searching for rational solutions of Eq. (95). The displayed equation is therefore either misprinted or misderived, and the omitted 'tedious but straightforward algebra' preceding Eq. (95) contains an error. As written, Eq. (95) does not support the claimed closed-form general solution for (4,3) methods.
  2. [Sec. 8, Eqs. (103)-(115)] The analogous (5,3) result is presented as a long chain of intermediate variables with no derivation and with only the assertion that 'direct substitution shows C3 = 0'. Since the corresponding (4,3) derivation has just been shown to contain a concrete error, the (5,3) quadratic must be re-derived or independently verified with computer algebra before it can be relied upon. As written, Eq. (115) cannot support the claimed general solution for (5,3) methods or the reported search over rational coefficients. The paper should provide the missing algebra or, at minimum, a machine-checkable symbolic verification for both Eq. (95) and Eq. (115).
minor comments (4)
  1. [Sec. 3] The word 'auxilliary' in the first paragraph should be 'auxiliary'.
  2. [Sec. 5.1] There is a typo in 'thea-form', which should read 'the a-form'.
  3. [Sec. 6, Eq. (80)] The proof of Eq. (80) is omitted; since this non-recursive formula is not used in the rest of the paper, the omission is acceptable, but the statement should be marked as a direct but unproved substitution result.
  4. [General presentation] A few punctuation and formatting lapses occur, such as 'e.g, Eqs.' instead of 'e.g., Eqs.' and the use of 'improper' in the captions of Tables 3 and 4, where 'incorrect' would be clearer.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: 2N-storage constraints and closed-form solutions are derived from the method definitions and standard order conditions; author self-citations are motivational only.

full rationale

The derivation chain is self-contained. Lemma 2 obtains the α-form coefficients from the A-form definition, Eq. (42) follows from ratios of those products, and Theorem 2 converts these to the explicit a-form constraints, Eqs. (45)–(47). Theorem 4 proves Eq. (71) by induction from those constraints plus the row-sum condition, Eq. (6), without assuming any target tableau; Theorem 5 then derives the A-form coefficients (81)–(82) from Eq. (71). The (4,3) and (5,3) derivations are algebraic eliminations of order conditions (B.1)–(B.4) after substituting Eq. (71), leading to quadratic equations in b2; the presented tableaus are obtained by solving these equations and then independently satisfy the order and 2N-storage constraints, so no fitted parameter is renamed as a prediction. The author's prior work [13] is cited only as motivation for studying the structure and as a conjecture about Lie-group preservation, not as a premise for the derived constraints or closed-form solutions. The apparent algebraic discrepancy in Eq. (95) raised by numerical substitution is a correctness matter, not a circularity: the formula is claimed to follow from algebra, not assumed as an input. Accordingly, no circular step is present.

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

The paper introduces no new physical entities. It uses the standard mathematical framework of Runge-Kutta methods, with the free parameters being the nodes and weights of the method family. The main assumptions are the non-degeneracy conditions for the general closed-form solutions.

free parameters (3)
  • nodes c2, c3, c4 (for (4,3) family)
    In the (4,3) closed-form solution, c2, c3, c4 are free parameters of the method family; the paper treats them as arbitrary subject to non-degeneracy conditions.
  • nodes c2, c3, c4, c5 and weight b5 (for (5,3) family)
    In the (5,3) solution, these five parameters are free; the paper chooses convenient values to produce rational-coefficient examples.
  • chosen nodes and weights in example schemes = e.g., (4,3)1: c2=1/4, c3=2/3, c4=4/5; (5,3)1: various
    The specific schemes in Tables 5-10 are constructed by selecting particular values of the free parameters and solving the quadratic. These are design choices, not fits to data.
assumptions (4)
  • standard math Standard order conditions (B.1)-(B.4) are necessary and sufficient for global third-order accuracy.
    Invoked in Secs. 7 and 8 to define the system the closed-form solution solves.
  • domain assumption The method is explicit and self-starting, so A1=0 and aij=0 for i<=j.
    Assumed throughout the paper, stated in Sec. 2.
  • domain assumption All A_i (i=2..s) are nonzero and no denominators vanish in Eqs. (71), (81), (82).
    Remark 2 and the statement of Theorem 4 restrict the main results to the 'general case'; special cases are handled separately only for (4,3).
  • domain assumption The 2N-storage form (Eqs. 7-9) is equivalent to a subset of standard explicit Runge-Kutta methods.
    This is the standard definition from Williamson [3] and is used to derive the constraints.

how reviews work

0 comments
Cite this review

Pith. "Pith review of 2N-storage Runge-Kutta methods: Order conditions, general properties and some analytic solutions." pith.science (2026). https://pith.science/paper/SEO2X3RJ

@misc{pith2026250607359,
  author       = {Pith},
  title        = {Pith review of: 2N-storage Runge-Kutta methods: Order conditions, general properties and some analytic solutions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SEO2X3RJ}},
  note         = {Machine review of arXiv:2506.07359}
}
read the original abstract

Low-storage Runge-Kutta schemes of Williamson's type, so-called 2N-storage schemes, are examined. Explicit 2N-storage constraints are derived for the first time and used to establish new relations between the entries of the Butcher tableau. An error in the Williamson's formula for converting coefficients between the standard and 2N-storage formats in the special case is pointed out and corrected. The new relations are used to derive a closed-form solution for four- and five-stage 2N-storage methods with the third order of global accuracy. Several new four- and five-stage schemes with rational coefficients are presented and numerically examined for illustration.

Figures

Figures reproduced from arXiv: 2506.07359 by the authors.

Figure 1
Figure 1. Scaling of the new (4,3)1 and (5,3)i, i = 1, . . . 4 methods, shown in color, and the five (4,3) methods of Ref. [4], shown in gray, for test problems 1 (left) and 2 (right). a (5,3) method across these three test problems, the (5,3)1 scheme in [PITH_FULL_IMAGE:figures/full_fig_p021_1.png] view at source ↗
Figure 2
Figure 2. Scaling of the new (4,3)1 and (5,3)i, i = 1, . . . 4 methods, shown in color, and the five (4,3) methods of Ref. [4], shown in gray, for test problem 3 (left). Stability regions for all (4,3) methods (identical due to satisfying Eq. (102)) and (5,3) methods considered (right). 10. Conclusion It appears that 2N-storage Runge-Kutta methods of Williamson [3] possess some interesting struc￾ture that was explored in this… view at source ↗

Discussion (0). Sign in to comment.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. D-splitting methods: 2N -storage embedded explicit pseudo-geometric Runge-Kutta methods using splitting methods

    math.NA 2026-04 unverdicted novelty 7.0 of 10

    D-splitting on the duplicated phase space produces arbitrary-order 2N-storage embedded explicit RK methods without a third register, with optional order elevation by averaging.

  2. CosmoLattice 2.0

    astro-ph.CO 2026-07 conditional novelty 6.0 of 10

    CosmoLattice v2.0 extends lattice cosmology simulations with non-minimal scalars, ALP–gauge couplings, defect networks, low-storage RK integrators, optimized GWs, and O(10) GPU speedups.

  3. 2N-storage Runge-Kutta methods: c-reflection symmetry and factorization of the Butcher tableau

    math.NA 2025-06 conditional novelty 6.0 of 10

    2N-storage Runge-Kutta methods of order four or lower come in c-reflected pairs, proved via a new factorization of the Butcher tableau into lower triangular matrices.

Reference graph

Works this paper leans on

18 extracted references · 4 canonical work pages · cited by 3 Pith papers

  1. [1]

    Butcher, Numerical Methods for Ordinary Differential Equations, 3rd Edition, Wiley, 2016

    J. Butcher, Numerical Methods for Ordinary Differential Equations, 3rd Edition, Wiley, 2016

    2016

  2. [2]

    Hairer, S

    E. Hairer, S. Nørsett, G. Wanner, Solving Ordinary Differential Equations I Nonstiff problems, 2nd Edition, Springer, Berlin, 2000

    2000

  3. [4]

    Carpenter, C

    M. Carpenter, C. Kennedy, Fourth-order 2N-storage Runge-Kutta schemes, Tech. Rep. NASA- TM-109112, NASA (1994)

    1994

  4. [5]

    Stanescu, W

    D. Stanescu, W. Habashi, 2N-storage low dissipation and dispersion Runge-Kutta schemes for computational acoustics, Journal of Computational Physics 143 (2) (1998) 674 – 681.doi:https: 31 //doi.org/10.1006/jcph.1998.5986. URL http://www.sciencedirect.com/science/article/pii/S0021999198959861

    work page arXiv 1998

  5. [6]

    Berland, C

    J. Berland, C. Bogey, C. Bailly, Low-dissipation and low-dispersion fourth-order Runge-Kutta algorithm, Computers and Fluids 35 (10) (2006) 1459 – 1463. doi:https://doi.org/10.1016/ j.compfluid.2005.04.003. URL http://www.sciencedirect.com/science/article/pii/S0045793005000575

    2006

  6. [7]

    Allampalli, R

    V. Allampalli, R. Hixon, M. Nallasamy, S. D. Sawyer, High-accuracy large-step explicit Runge- Kutta (HALE-RK) schemes for computational aeroacoustics, Journal of Computational Physics 228 (10) (2009) 3837 – 3850. doi:https://doi.org/10.1016/j.jcp.2009.02.015. URL http://www.sciencedirect.com/science/article/pii/S0021999109000849

    doi:10.1016/j.jcp.2009.02.015 2009

  7. [8]

    Bernardini, S

    M. Bernardini, S. Pirozzoli, A general strategy for the optimization of Runge-Kutta schemes for wave propagation phenomena, Journal of Computational Physics 228 (11) (2009) 4182 – 4199. doi:https://doi.org/10.1016/j.jcp.2009.02.032. URL http://www.sciencedirect.com/science/article/pii/S0021999109001077

    work page doi:10.1016/j.jcp.2009.02.032 2009

  8. [9]

    D. I. Ketcheson, Runge-Kutta methods with minimum storage implementations, Journal of Com- putational Physics 229 (5) (2010) 1763 – 1773. doi:https://doi.org/10.1016/j.jcp.2009. 11.006. URL http://www.sciencedirect.com/science/article/pii/S0021999109006251

    doi:10.1016/j.jcp.2009 2010

Show all 18 references
  1. [10]

    Toulorge, W

    T. Toulorge, W. Desmet, Optimal Runge-Kutta schemes for discontinuous Galerkin space dis- cretizations applied to wave propagation problems, Journal of Computational Physics 231 (4) (2012) 2067 – 2091. doi:https://doi.org/10.1016/j.jcp.2011.11.024. URL http://www.sciencedirect...

    2012 doi

  2. [11]

    Niegemann, R

    J. Niegemann, R. Diehl, K. Busch, Efficient low-storage Runge-Kutta schemes with optimized stability regions, Journal of Computational Physics 231 (2) (2012) 364 – 372. doi:https://doi. org/10.1016/j.jcp.2011.09.003. URL http://www.sciencedirect.com/science/article/pii/S002199...

    2012 doi

  3. [12]

    an Yan, Low-storage Runge-Kutta method for simulating time-dependent quantum dynamics, Chinese Journal of Chemical Physics 30 (3) (2017) 277 – 286

    Y. an Yan, Low-storage Runge-Kutta method for simulating time-dependent quantum dynamics, Chinese Journal of Chemical Physics 30 (3) (2017) 277 – 286

    2017

  4. [13]

    Bazavov, Commutator-free lie group methods with minimum storage requirements and reuse of exponentials, BIT Numerical Mathematics 62 (3) (2021) 745–771

    A. Bazavov, Commutator-free lie group methods with minimum storage requirements and reuse of exponentials, BIT Numerical Mathematics 62 (3) (2021) 745–771. doi:10.1007/ s10543-021-00892-x . URL http://dx.doi.org/10.1007/s10543-021-00892-x

    2021 doi

  5. [14]

    C.-W. Shu, S. Osher, Efficient implementation of essentially non-oscillatory shock-capturing schemes, Journal of Computational Physics 77 (2) (1988) 439–471. doi:https://doi.org/10. 1016/0021-9991(88)90177-5. URL https://www.sciencedirect.com/science/article/pii/0021999188901775

    1988

  6. [15]

    S. J. Ruuth, Global optimization of explicit strong-stability-preserving runge-kutta methods, Mathematics of Computation 75 (253) (2006) 183–207. URL http://www.jstor.org/stable/4100149

    2006

  7. [16]

    Berland, C

    J. Berland, C. Bogey, C. Bailly, Optimized Explicit Schemes: Matching and Boundary Schemes, and 4th-order Runge-Kutta Algorithm, 2012. arXiv:https://arc.aiaa.org/doi/pdf/10. 2514/6.2004-2814, doi:10.2514/6.2004-2814. URL https://arc.aiaa.org/doi/abs/10.2514/6.2004-2814 32

    2012 doi

  8. [17]

    Bazavov, T

    A. Bazavov, T. Chuna, Efficient integration of gradient flow in lattice gauge theory and properties of low-storage commutator-free lie group methods (2021). arXiv:2101.05320. URL https://arxiv.org/abs/2101.05320

    2021 arXiv

  9. [18]

    W. H. Press, W. T. Vetterling, S. A. Teukolsky, B. P. Flannery, Numerical Recipes in C++: the art of scientific computing, 2nd Edition, Cambridge University Press, USA, 2001

    2001

  10. [19]

    Fousse, G

    L. Fousse, G. Hanrot, V. Lef` evre, P. P´ elissier, P. Zimmermann, Mpfr: A multiple-precision binary floating-point library with correct rounding, ACM Trans. Math. Softw. 33 (2) (2007) 13–es. doi:10.1145/1236463.1236468. URL https://doi.org/10.1145/1236463.1236468 33

    2007

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.