Pith. sign in

REVIEW 3 major objections 4 minor 51 references

Mass-Zero constrained molecular dynamics for electrostatic interactions

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Poisson MaZe computes electrostatic forces by evolving the grid potential as zero-inertia auxiliary variables constrained by the Poisson equation, reducing the multigrid iterations needed and giving linear-scaling real-space electrostatics.

desk verdict The MaZe-for-Poisson formulation is a real idea, but the reported multigrid speedup looks like a residual-normalization artifact, so the paper needs a corrected performance comparison before the main claim can be trusted. read the letter →

arxiv 2505.20418 v2 pith:W5DK33HL submitted 2025-05-26 physics.comp-ph

Federica Troni , Davide Grassano , Jayashree Narayan , Benoît Roux , Sara Bonella This is my paper · ORCID
classification physics.comp-ph
keywords Poissonequationelectrostaticinteractionsmoleculardynamicsmass-zeroconstrainedmultigridreal-spacesolverextendedLagrangianmoltenNaCl
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

Poisson MaZe is a real-space method for computing electrostatic forces in molecular dynamics that avoids reciprocal-space Fast Fourier transforms. The paper argues that the value of the electrostatic potential at every grid point can be treated as a zero-mass auxiliary coordinate, with the discretized Poisson equation imposed as a dynamical constraint; the constraint update requires solving a linear system whose right-hand side is the mismatch between a Verlet-predicted potential and the ionic charge density. In a molten NaCl simulation the method reproduces the reference radial distribution functions and diffusion coefficients, and the multigrid solution of the constraint system converges in fewer cycles than solving the Poisson equation directly. The authors test the method as a complete electrostatic solver and demonstrate linear scaling with system size.

What carries the argument

The carrying object is the extended MaZe Lagrangian, in which the $N$ grid values of the electrostatic potential $\phi_n$ enter as auxiliary variables with zero inertia and the constraint $\sigma=\frac{h}{4\pi}M\phi+q=0$ is enforced by Lagrange multipliers. A Verlet step proposes a trial potential $\phi_p^{k+1}=2\phi^k-\phi^{k-1}$; requiring the corrected potential to satisfy the constraint turns the SHAKE step into the linear system $M y=\sigma_p^{k+1}$ (Equation 15), where $y$ is the rescaled multiplier vector and $\sigma_p$ is the constraint evaluated at the trial potential. This system shares the Laplacian matrix $M$ with the standard Poisson equation but has a different unknown and a right-hand side that measures how far the trial potential is from satisfying the Poisson equation, and it is solved iteratively with multigrid, initialized with the previous step's solution, which is the source of the faster convergence.

What would settle it

Run a moderate-size electrolyte or molten-salt simulation with multigrid tolerances of $10^{-4}$ and $10^{-10}$ and compare the resulting radial distribution functions and diffusion coefficients with a high-accuracy particle-mesh Ewald reference; any statistically significant difference between the two tolerance runs would show that the finite solver tolerance contaminates the sampled ensemble.

Watch

Extended reading notes

Core claim

The central claim is that embedding the Poisson problem in a mass-zero constrained extended dynamics yields a new linear system, $M y=\sigma_p$, whose multigrid solution reaches the electrostatic potential faster than solving the standard discretized Poisson equation $M\phi=-4\pi q/h$. At a tolerance of $10^{-7}$, Poisson MaZe converges in about four multigrid cycles, whereas a time-propagated direct multigrid solve needs roughly twice as many and a static-start direct solve roughly seven times as many, at essentially the same cost per cycle. The resulting forces, computed with a central-difference gradient of the grid potential and B-spline charge assignment, conserve total momentum and, when combined with a work-based energy estimator, keep the total energy stable at a relative fluctuation of $3.2\times 10^{-7}$. The validation on 250-ion molten NaCl matches the reference structural and transport observables, and the measured scaling of total CPU time with particle number is linear for all methods, with Poisson MaZe lowest on every tested size.

Load-bearing premise

The result depends on the zero-inertia limit of the constrained extended dynamics being exactly unbiased for the physical degrees of freedom, and on the finite multigrid tolerance not biasing the sampled configurations; if either fails, the dynamics will depart from the intended Boltzmann statistics even though the potential is correct.

Editorial extensions

If this is right

  • If the claim is right, long-range electrostatics can be computed entirely in real space with linear scaling, removing the global communication pattern of FFT-based particle-mesh methods.
  • At a multigrid tolerance of $10^{-7}$, Poisson MaZe converges in about four cycles per step, versus roughly seven cycles for a time-propagated direct solve and roughly 28 for a static-start direct solve, cutting total CPU time by factors of about 1.9 and 3.4 in the tested systems.
  • The molten NaCl runs reproduce the reference first-peak positions and heights for all three radial distribution functions and give diffusion coefficients $D_{\text{Na}}=0.16\times 10^{-3}$ cm$^2$/s and $D_{\text{Cl}}=0.14\times 10^{-3}$ cm$^2$/s, so the method yields correct structural and transport properties in a realistic liquid.
  • The same convergence advantage appears with a preconditioned conjugate-gradient solver, indicating that the benefit comes from the reformulated linear system rather than from a property specific to multigrid.
  • The work-based electrostatic energy estimator provides a stable conserved quantity for monitoring simulations, with relative total-energy fluctuations of $3.2\times 10^{-7}$, while the naive discretized energy oscillates badly due to grid self-energy terms.

Reading between the lines

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

  • Beyond the paper, the same constraint-embedding trick should transfer to other elliptic equations solved inside a molecular dynamics loop, such as generalized Poisson or linearized Poisson-Boltzmann problems, whenever the previous time step supplies a good initial guess for the iterative solve.
  • If the per-cycle cost remains flat at high core counts, replacing the global FFT communication of particle-mesh Ewald with local multigrid smoothing could make real-space electrostatics competitive at very large processor counts; the paper only measures scaling on 8 processors, so this is an extrapolation rather than a demonstrated result.
  • Because the trial potential is a second-order time prediction, the convergence gain should grow for slowly moving ions and small time steps; the paper's solid-versus-molten comparison (slowdown factor about 1.8 versus about 2.9 to 3.7 on melting) is consistent with that expectation.
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. The manuscript introduces Poisson MaZe, an extended-Lagrangian method for computing electrostatic interactions in molecular dynamics. The values of the discretized electrostatic potential are treated as zero-inertia auxiliary degrees of freedom, and the discretized Poisson equation is imposed as a constraint. The constraint is solved via SHAKE, leading to a linear system (Eq. 15) that differs from the direct Poisson system (Eq. 4). The authors validate the method on molten NaCl, reporting agreement with reference structural and transport properties, and demonstrate linear scaling with a multigrid solver. They also report that Poisson MaZe requires fewer solver iterations than direct multigrid solutions of the Poisson equation, and they discuss energy, momentum, time-reversibility, and stationarity properties. Appendices provide a toy-model validation and a conjugate-gradient variant.

Significance. If the performance claims are substantiated, Poisson MaZe would be a useful real-space alternative to FFT-based methods for large-scale simulations. The method is parameter-free, built on a clearly stated Lagrangian, and validated against external benchmarks (CHARMM SOR and published molten-salt data). The convergence and stationarity tests are valuable, and the linear-scaling property inherited from the multigrid solver is a strength. However, the central acceleration claim is not currently established because the residual comparison in Section III A 1 is not accuracy-equivalent, and the conditioning explanation in Appendix B is incorrect.

major comments (3)
  1. [III A 1, Eqs. (22)-(23)] The comparison of iteration counts is not accuracy-equivalent. For any candidate potential φ = φ_p - (4π/h)y, the direct Poisson residual and the MaZe residual are related by r_direct = Mφ + (4π/h)q = -(4π/h)(My - σ_p) = -(4π/h) r_MaZe. With h = 0.172 Å this is a factor of about 73. Stopping both solvers at the same absolute tolerance 10^-7 therefore allows Poisson MaZe to terminate with a Poisson-equation residual about 73 times larger than the direct solver. The reported differences in iteration counts (4 vs 8/28 in Fig. 4) and the speedups (3.36±0.21 and 1.88±0.15 in Section III A 2) are consequently not comparisons at equal accuracy. The accompanying statement that the final potentials agree to ~10^-10–10^-14 is also inconsistent with this scaling: with tol=10^-7, the implied potential error for the MaZe solution is O(||M^{-1}|| × 7×10^-6), which is orders of magnitude larger than 10^-10. Please repeat the benchmark with equal Poisson-equation residual (or equal potential error) and report the resulting iteration counts and timings.
  2. [Appendix B] The claim that the MaZe acceleration "stems from an improved conditioning of the problem" is incorrect. The linear system solved by Poisson MaZe, Eq. (15), uses exactly the same matrix M as the direct Poisson system Eq. (4), so the condition number is identical. The faster convergence observed in Fig. 15 is plausibly due to the smaller right-hand side σ_p and the warm start, combined with the absolute-residual criterion discussed above; the conditioning explanation should be removed or revised.
  3. [II A and Conclusions] The force used in production, Eq. (18), is not the derivative of the extended-Lagrangian potential V(φ;{rβ}) appearing in Eq. (11). The derivative of V with respect to r_α at fixed φ is -Σ_n φ_n ∂q_n/∂r_α, whose discretized form is the alternative force Eq. (20); Eq. (18) differs by using the central difference of the potential rather than the gradient of the B-spline weights. Since the exact-sampling statement in the Conclusions cites Ref. 31 for the equations of motion (11), the authors should clarify whether the unbiasedness theorem applies to the implemented force Eq. (18) or only to the Lagrangian force Eq. (20). If it applies only to Eq. (20), the exact-sampling claim for the actual algorithm needs qualification.
minor comments (4)
  1. [II, Eq. (12)] The factor 1/2 in front of Δt^2 Mη is inconsistent with the equation of motion φ¨ = -Mη in Eq. (11), for which the Verlet update would contain -Δt^2 Mη. Since the algorithm solves for y directly, this does not change the final update rule, but the derivation should be corrected or the definition of η clarified.
  2. [Appendix B] The references to "arrows in Figure 4" and "the inset of Figure 4" should refer to Figure 15, which is the PLCG convergence figure.
  3. [III and Table I/Figure 3] The diffusion coefficients and g(r) peak parameters are quoted without statistical uncertainties; given the 25 ps trajectory, reporting error bars would allow a quantitative assessment of the agreement with Refs. 41 and 42.
  4. [Figure 5] The top-axis labels appear to have lost their formatting (e.g., "1043" rather than a readable grid-count notation); please reformat the axis labels.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the method is derived from an explicit Lagrangian and validated against external references; the one self-cited MaZe theorem is independent general support, not the target result.

full rationale

Poisson MaZe is constructed, not fitted: the extended Lagrangian (Eq. 5) uses V(phi;{r}) = (h/8*pi) phi^T M phi + phi^T q (Eq. 6), and the constraint sigma = grad_phi V = (h/4*pi) M phi + q = 0 (Eq. 8) is exactly the discretized Poisson equation (Eq. 4) rearranged. The SHAKE update algebraically transforms this into the auxiliary system M y = sigma_p (Eq. 15), with phi^{k+1} = phi_p - (4*pi/h) y (Eq. 16). Substitution shows that solving Eq. 15 is equivalent to solving Eq. 4 for the corrected potential, so the method is a reformulation rather than a circular derivation; the authors state the relationship between the matrices and right-hand sides explicitly. No parameter is fitted to the quantities later reported as results: the structural and transport validation compares against Galamba et al. and Tosi-Fumi reference data, and the potential is benchmarked against CHARMM's SOR solver. The only self-citation that carries theoretical weight is ref. 31 for the MaZe zero-inertia sampling theorem; it is a general, parameter-free theorem whose assumptions do not already contain the Poisson MaZe result, and the current paper additionally checks correctness against external data, so under the stated rules it is independent support rather than circularity. The skeptical concern about inconsistent residual scaling between Eq. 22 and Eq. 23 (r_direct = (4*pi/h) r_MaZe) is a benchmarking/correctness issue about the speedup claim, not an instance of a result reducing to its inputs by construction; it therefore does not raise the circularity score here.

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

The method introduces no data-fitted constants. Its validity depends on the prior MaZe theorem for the zero-inertia limit, the accuracy of the standard force discretization, the reference data for molten NaCl, and the representative behavior of the chosen multigrid operators.

assumptions (4)
  • standard math The zero-inertia limit of the MaZe extended dynamics yields an unbiased probability density for the physical degrees of freedom.
    Invoked in Sections II (after Eq 11) and IV; the entire method assumes the ionic dynamics is correct when the potential is obtained by constraint.
  • domain assumption The central-difference force estimator (Eq 17) with cubic B-spline charge assignment conserves total momentum and cancels self-energy terms.
    Section II A and Eq 18; standard result from Hockney and Eastwood (ref 21), adopted without derivation.
  • domain assumption The Tosi-Fumi force field and the reference data of Galamba et al. provide accurate benchmarks for molten NaCl at 1550 K.
    Section III; the validation of g(r) and diffusion coefficients relies on these external references.
  • domain assumption The multigrid operator set (Red-Black Gauss-Seidel, 27-point full-weighting, trilinear prolongation) converges linearly and its convergence behavior is representative.
    Section III A; operators from Sagui and Darden (ref 23).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Mass-Zero constrained molecular dynamics for electrostatic interactions." pith.science (2026). https://pith.science/paper/W5DK33HL

@misc{pith2026250520418,
  author       = {Pith},
  title        = {Pith review of: Mass-Zero constrained molecular dynamics for electrostatic interactions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/W5DK33HL}},
  note         = {Machine review of arXiv:2505.20418}
}
read the original abstract

Optimal exploitation of supercomputing resources for the evaluation of electrostatic forces remains a challenge in molecular dynamics simulations of very large systems. The most efficient methods are currently based on particle-mesh Ewald sums and achieve semi-logarithmic scaling in the number of particles. These methods solve the problem in reciprocal space, requiring extensive use of Fast Fourier transforms (FFTs). While highly efficient in many contexts, FFTs may encounter scalability challenges at very large processor counts due to their communication requirements. To mitigate these problems, the development and scalable coding of real-space approaches to solve the Poisson equation on a grid is an active field of research. In this work, we introduce a novel real-space approach that provides some advantages over alternatives. Our method exploits an extended Lagrangian in which the values of the field at the grid points are treated as auxiliary variables of zero inertia and the discretized Poisson equation is enforced as a dynamical constraint. The solution of the constraints leads to a linear system - different from those appearing in other real-space approaches - that can be efficiently solved via state-of-the-art iterative methods. The method inherits the numerical scaling of the adopted iterative solver, e.g. linear with a multigrid (MG) approach, but converges with fewer cycles. We analyze this approach considering realistic simulations of molten NaCl that validate its ability to reproduce structural and transport properties. Using this non-trivial benchmark, we demonstrate linear scaling and illustrate some features of our algorithm.

Figures

Figures reproduced from arXiv: 2505.20418 by the authors.

Figure 1
Figure 1. FIG. 1. Pseudocode for the Poisson MaZe algorithm imple [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Normalized VACFs for the two species (main panel) [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Radial pair distribution function obtained with a [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: FIG. 4. Convergence of Poisson MaZe (red), plain MG with [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5. Scaling of the total CPU time per step as a function [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6. Comparison of electrostatic and total energies. The [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 8
Figure 8. Figure 8: FIG. 8. Normalized histograms of the mean force error for [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 7
Figure 7. Figure 7: FIG. 7. Time evolution of kinetic energy (top panel), total [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 9
Figure 9. Figure 9: FIG. 9. Conservation of total momentum for two different [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]
Figure 11
Figure 11. Figure 11: FIG. 11. Average particle positions over time, grouped into [PITH_FULL_IMAGE:figures/full_fig_p011_11.png]
Figure 12
Figure 12. Figure 12: FIG. 12. Section of the electrostatic potential generated by [PITH_FULL_IMAGE:figures/full_fig_p012_12.png]
Figure 13
Figure 13. Figure 13: FIG. 13. Comparison between the different expressions for [PITH_FULL_IMAGE:figures/full_fig_p012_13.png]
Figure 14
Figure 14. Figure 14: FIG. 14. Comparison between the different expressions for [PITH_FULL_IMAGE:figures/full_fig_p013_14.png]
Figure 16
Figure 16. Figure 16: FIG. 16. Scaling analysis of Poisson MaZe as a function of the [PITH_FULL_IMAGE:figures/full_fig_p014_16.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

51 extracted references · 45 canonical work pages

  1. [1]

    merlin.mbs aapmrev4-1.bst 2010-07-25 4.21a (PWD, AO, DPC) hacked

    FUNCTION id.bst "merlin.mbs aapmrev4-1.bst 2010-07-25 4.21a (PWD, AO, DPC) hacked" ENTRY address archive archivePrefix author bookaddress booktitle chapter collaboration doi edition editor eid eprint howpublished institution isbn issn journal key language month note number organization pages primaryClass publisher school SLACcitation series title translat...

    2010

  2. [2]

    merlin.mbs aipauth4-1.bst 2010-07-25 4.21a (PWD, AO, DPC) hacked

    FUNCTION id.bst "merlin.mbs aipauth4-1.bst 2010-07-25 4.21a (PWD, AO, DPC) hacked" ENTRY address archive archivePrefix author bookaddress booktitle chapter collaboration doi edition editor eid eprint howpublished institution isbn issn journal key language month note number organization pages primaryClass publisher school SLACcitation series title translat...

    2010

  3. [3]

    merlin.mbs aipnum4-1.bst 2010-07-25 4.21a (PWD, AO, DPC) hacked

    FUNCTION id.bst "merlin.mbs aipnum4-1.bst 2010-07-25 4.21a (PWD, AO, DPC) hacked" ENTRY address archive archivePrefix author bookaddress booktitle chapter collaboration doi edition editor eid eprint howpublished institution isbn issn journal key language month note number organization pages primaryClass publisher school SLACcitation series title translati...

    2010

  4. [4]

    merlin.mbs apsrev4-1.bst 2010-07-25 4.21a (PWD, AO, DPC) hacked

    FUNCTION id.bst "merlin.mbs apsrev4-1.bst 2010-07-25 4.21a (PWD, AO, DPC) hacked" ENTRY address archive archivePrefix author bookaddress booktitle chapter collaboration doi edition editor eid eprint howpublished institution isbn issn journal key language month note number organization pages primaryClass publisher school SLACcitation series title translati...

    2010

  5. [5]

    merlin.mbs apsrmp4-1.bst 2010-07-25 4.21a (PWD, AO, DPC) hacked

    FUNCTION id.bst "merlin.mbs apsrmp4-1.bst 2010-07-25 4.21a (PWD, AO, DPC) hacked" ENTRY address archive archivePrefix author bookaddress booktitle chapter collaboration doi edition editor eid eprint howpublished institution isbn issn journal key language month note number organization pages primaryClass publisher school SLACcitation series title translati...

    2010

  6. [6]

    Allen \ and\ author D

    author author M. Allen \ and\ author D. Tildesley ,\ @noop title Computer Simulation of Liquids \ ( publisher Oxford Science Publications, Clarendon Press ,\ address Oxford ,\ year 1989 ) NoStop

    work page 1989

  7. [7]

    Darden , author D

    author author T. Darden , author D. York , \ and\ author L. Pedersen ,\ @noop journal journal The Journal of Chemical Physics \ volume 98 ,\ pages 10089 ( year 1993 ) NoStop

    work page 1993

  8. [8]

    Essmann , author L

    author author U. Essmann , author L. Perera , author M. L. \ Berkowitz , author T. Darden , author H. Lee , \ and\ author L. G. \ Pedersen ,\ @noop journal journal The Journal of Chemical Physics \ volume 103 ,\ pages 8577 ( year 1995 ) NoStop

    work page 1995

Show all 51 references
  1. [9]

    Darden , author A

    author author T. Darden , author A. Toukmaji , \ and\ author L. Pedersen ,\ @noop journal journal Journal de Chimie Physique et de Physico-Chimie Biologique \ volume 94 ,\ pages 1346 ( year 1997 ) NoStop

    1997

  2. [10]

    Hwang , author S

    author author W. Hwang , author S. L. \ Austin , author A. Blondel , author E. D. \ Boittier , author S. Boresch , author M. Buck , author J. Buckner , author A. Caflisch , author H.-T. \ Chang , author X. Cheng , et al. ,\ @noop journal journal The Journal of Physical Chemist...

    2024

  3. [11]

    Salomon-Ferrer , author D

    author author R. Salomon-Ferrer , author D. A. \ Case , \ and\ author R. C. \ Walker ,\ @noop journal journal Wiley Interdisciplinary Reviews: Computational Molecular Science \ volume 3 ,\ pages 198 ( year 2013 ) NoStop

    2013

  4. [12]

    author author M. J. \ Abraham , author T. Murtola , author R. Schulz , author S. P \'a ll , author J. C. \ Smith , author B. Hess , \ and\ author E. Lindahl ,\ @noop journal journal SoftwareX \ volume 1 ,\ pages 19 ( year 2015 ) NoStop

    2015

  5. [13]

    author author J. C. \ Phillips , author D. J. \ Hardy , author J. D. C. \ Maia , author J. E. \ Stone , author J. V. \ Ribeiro , author R. C. \ Bernardi , author R. Buch , author G. Fiorin , author J. Hénin , author W. Jiang , author R. McGreevy , author M. C. R. \ Melo , auth...

    2020

  6. [14]

    author author A. P. \ Thompson , author H. M. \ Aktulga , author R. Berger , author D. S. \ Bolintineanu , author W. M. \ Brown , author P. S. \ Crozier , author P. J. \ In't Veld , author A. Kohlmeyer , author S. G. \ Moore , author T. D. \ Nguyen , et al. ,\ @noop journal jo...

    2022

  7. [15]

    Eastman , author J

    author author P. Eastman , author J. Swails , author J. D. \ Chodera , author R. T. \ McGibbon , author Y. Zhao , author K. A. \ Beauchamp , author L.-P. \ Wang , author A. C. \ Simmonett , author M. P. \ Harrigan , author C. D. \ Stern , et al. ,\ @noop journal journal PLoS c...

    2017

  8. [16]

    author author N. A. \ Gumerov \ and\ author R. Duraiswami ,\ @noop title Fast Multipole Methods for the Helmholtz Equation in Three Dimensions \ ( publisher Elsevier ,\ year 2004 ) NoStop

    2004

  9. [17]

    author author C. J. \ Fennell \ and\ author J. D. \ Gezelter ,\ @noop journal journal The Journal of Chemical Physics \ volume 124 ,\ pages 234104 ( year 2006 ) NoStop

    2006

  10. [18]

    author author J. M. \ Rodgers \ and\ author J. D. \ Weeks ,\ @noop journal journal Proceedings of the National Academy of Sciences \ volume 105 ,\ pages 19136 ( year 2008 ) NoStop

    2008

  11. [19]

    Dehnen ,\ @noop journal journal Journal of Computational Physics \ volume 179 ,\ pages 27 ( year 2002 ) NoStop

    author author W. Dehnen ,\ @noop journal journal Journal of Computational Physics \ volume 179 ,\ pages 27 ( year 2002 ) NoStop

    2002

  12. [20]

    author author H. A. \ Boateng ,\ @noop journal journal Journal of Chemical Theory and Computation \ volume 16 ,\ pages 7 ( year 2019 ) NoStop

    2019

  13. [21]

    author author R. D. \ Skeel , author I. Tezcan , \ and\ author D. J. \ Hardy ,\ @noop journal journal Journal of Computational Chemistry \ volume 23 ,\ pages 673 ( year 2002 ) NoStop

    2002

  14. [22]

    author author D. J. \ Hardy , author Z. Wu , author J. C. \ Phillips , author J. E. \ Stone , author R. D. \ Skeel , \ and\ author K. Schulten ,\ @noop journal journal Journal of Chemical Theory and Computation \ volume 11 ,\ pages 766 ( year 2015 ) NoStop

    2015

  15. [23]

    Sagui \ and\ author T

    author author C. Sagui \ and\ author T. Darden ,\ @noop journal journal The Journal of Chemical Physics \ volume 114 ,\ pages 6578 ( year 2001 ) NoStop

    2001

  16. [24]

    Rottler \ and\ author A

    author author J. Rottler \ and\ author A. C. \ Maggs ,\ @noop journal journal Phys. Rev. Lett. \ volume 93 ,\ pages 170201 ( year 2004 a ) NoStop

    2004

  17. [25]

    Rottler \ and\ author A

    author author J. Rottler \ and\ author A. C. \ Maggs ,\ @noop journal journal The Journal of Chemical Physics \ volume 120 ,\ pages 3119 ( year 2004 b ) NoStop

    2004

  18. [26]

    \ Ryckaert , author A

    author author J.-P. \ Ryckaert , author A. Bellemans , \ and\ author G. C. \ and ,\ 10.1080/00268978100102931 journal journal Molecular Physics \ volume 44 ,\ pages 979 ( year 1981 ) NoStop

    1981 doi

  19. [27]

    Coretti , author S

    author author A. Coretti , author S. Bonella , \ and\ author G. Ciccotti ,\ @noop journal journal The Journal of Chemical Physics \ volume 149 ,\ pages 191102 ( year 2018 ) NoStop

    2018

  20. [28]

    Bonella , author A

    author author S. Bonella , author A. Coretti , author R. Vuilleumier , \ and\ author G. Ciccotti ,\ @noop journal journal Phys. Chem. Chem. Phys. \ volume 22 ,\ pages 10775 ( year 2020 ) NoStop

    2020

  21. [29]

    Girardier , author A

    author author D. Girardier , author A. Coretti , author G. Ciccotti , \ and\ author S. Bonella ,\ @noop journal journal The European Physical Journal B \ volume 94 ,\ pages 158 ( year 2021 ) NoStop

    2021

  22. [30]

    Coretti , author L

    author author A. Coretti , author L. Scalfi , author C. Bacon , author B. Rotenberg , author R. Vuilleumier , author G. Ciccotti , author M. Salanne , \ and\ author S. Bonella ,\ @noop journal journal The Journal of Chemical Physics \ volume 152 ,\ pages 194701 ( year 2020 ) NoStop

    2020

  23. [31]

    Coretti , author T

    author author A. Coretti , author T. Baird , author R. Vuilleumier , \ and\ author S. Bonella ,\ @noop journal journal The Journal of Chemical Physics \ volume 157 ( year 2022 ) NoStop

    2022

  24. [32]

    Im , author D

    author author W. Im , author D. Beglov , \ and\ author B. Roux ,\ @noop journal journal Computer Physics Communications \ volume 111 ,\ pages 59 ( year 1998 ) NoStop

    1998

  25. [33]

    \ Ryckaert , author G

    author author J.-P. \ Ryckaert , author G. Ciccotti , \ and\ author H. J. \ Berendsen ,\ @noop journal journal Journal of Computational Physics \ volume 23 ,\ pages 327 ( year 1977 ) NoStop

    1977

  26. [34]

    note Verlet is chosen for simplicity since only the values of the field (and not those of the field NoStop

  27. [35]

    author author M. R. \ Hestenes \ and\ author E. Stiefel ,\ @noop journal journal Journal of research of the National Bureau of Standards \ volume 49 ,\ pages 409 ( year 1952 ) NoStop

    1952

  28. [36]

    Elber , author A

    author author R. Elber , author A. P. \ Ruymgaart , \ and\ author B. Hess ,\ @noop journal journal The European Physical Journal Special Topics \ volume 200 ,\ pages 211 ( year 2011 ) NoStop

    2011

  29. [37]

    Note also that explicit storage of the matrix is not necessary NoStop

    note The fact that every row of M has only 7 non zero entries enables to perform the matrix-vector multiplication with a complexity of O ( N ) , strongly reducing the computational bottleneck of the brute-force approach that scales like O ( N ^2) . Note also that explicit stor...

  30. [38]

    author author R. W. \ Hockney \ and\ author J. W. \ Eastwood ,\ @noop title Computer Simulation Using Particles \ ( year 1981 ) NoStop

    1981

  31. [39]

    author author M. K. \ Gilson , author A. Rashin , author R. Fine , \ and\ author B. Honig ,\ @noop journal journal Journal of Molecular Biology \ volume 184 ,\ pages 503 ( year 1985 ) NoStop

    1985

  32. [40]

    author author M. K. \ Gilson , author M. E. \ Davis , author B. A. \ Luty , \ and\ author J. A. \ McCammon ,\ @noop journal journal The Journal of Physical Chemistry \ volume 97 ,\ pages 3591 ( year 1993 ) NoStop

    1993

  33. [41]

    Klapper , author R

    author author I. Klapper , author R. Hagstrom , author R. Fine , author K. Sharp , \ and\ author B. Honig ,\ @noop journal journal Proteins: Structure, Function, and Bioinformatics \ volume 1 ,\ pages 47 ( year 1986 ) NoStop

    1986

  34. [42]

    Nicholls \ and\ author B

    author author A. Nicholls \ and\ author B. Honig ,\ @noop journal journal Journal of Computational Chemistry \ volume 12 ,\ pages 435 ( year 1991 ) NoStop

    1991

  35. [43]

    Jo , author T

    author author S. Jo , author T. Kim , author V. G. \ Iyer , \ and\ author W. Im ,\ @noop journal journal Journal of Computational Chemistry \ volume 29 ,\ pages 1859 ( year 2008 ) NoStop

    2008

  36. [44]

    author author C. K. \ Birdsall \ and\ author D. Fuss ,\ @noop journal journal Journal of Computational Physics \ volume 3 ,\ pages 494 ( year 1969 ) NoStop

    1969

  37. [45]

    Galamba \ and\ author B

    author author N. Galamba \ and\ author B. J. \ Costa Cabral ,\ @noop journal journal The Journal of Chemical Physics \ volume 126 ,\ pages 124502 ( year 2007 ) NoStop

    2007

  38. [46]

    author author S. B. \ F. Mouhat \ and\ author C. Pierleoni ,\ https://doi.org/10.1080/00268976.2013.846486 journal journal Molecular Physics \ volume 111 ,\ pages 3651 ( year 2013 ) NoStop

    2013

  39. [47]

    Tosi \ and\ author F

    author author M. Tosi \ and\ author F. Fumi ,\ @noop journal journal Journal of Physics and Chemistry of Solids \ volume 25 ,\ pages 45 ( year 1964 ) NoStop

    1964

  40. [48]

    author author D. A. \ Sivak , author J. D. \ Chodera , \ and\ author G. E. \ Crooks ,\ @noop journal journal The Journal of Physical Chemistry. B \ volume 118 ,\ pages 6466 ( year 2013 ) NoStop

    2013

  41. [49]

    author author B. J. \ Leimkuhler ,\ in\ @noop booktitle Computational Molecular Dynamics: Challenges, Methods, Ideas: Proceedings of the 2nd International Symposium on Algorithms for Macromolecular Modelling, Berlin, May 21--24, 1997 \ ( organization Springer ,\ year 1999 )\ p...

    1997

  42. [50]

    Ayala , author S

    author author A. Ayala , author S. Tomov , author M. Stoyanov , author A. Haidar , \ and\ author J. Dongarra ,\ in\ @noop booktitle 2022 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW) \ ( year 2022 )\ pp.\ pages 372--381 NoStop

    2022

  43. [51]

    Zhou , author P

    author author Q. Zhou , author P. Kousha , author Q. Anthony , author K. Shafie Khorassani , author A. Shafi , author H. Subramoni , \ and\ author D. K. \ Panda ,\ in\ @noop booktitle High Performance Computing ,\ editor edited by\ editor A.-L. \ Varbanescu , editor A. Bhatele...

    2022

Pith tools

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