Chrono::Ray: A Distributed Framework for High-Throughput Simulation-Based Analysis of Multibody Systems

Dan Negrut; Khailanii Slaton

arxiv: 2605.13767 · v2 · pith:MCMJAZAAnew · submitted 2026-05-13 · 💻 cs.CE

Chrono::Ray: A Distributed Framework for High-Throughput Simulation-Based Analysis of Multibody Systems

Khailanii Slaton , Dan Negrut This is my paper

Pith reviewed 2026-05-19 13:49 UTC · model grok-4.3

classification 💻 cs.CE

keywords distributed simulationmultibody dynamicsRay frameworkChrono enginehigh-throughput computingparameter identificationterramechanicsworkflow orchestration

0 comments

The pith

Chrono::Ray integrates the Chrono simulation engine with Ray to support large-scale multibody analysis through distributed computing.

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

This paper presents Chrono::Ray as a framework that combines high-fidelity multibody simulation with distributed computing capabilities. It offers simple interfaces so that users can run many simulation trials in parallel for engineering studies. The approach avoids the need for users to set up and manage complex distributed systems themselves. Two examples illustrate its use: recovering parameters in a lunar lander model and designing experiments for a terramechanics model. The framework is open source and part of the Project Chrono ecosystem.

Core claim

The central claim is that by linking Chrono with Ray, the resulting framework provides modular and user-friendly abstractions that enable scalable orchestration of large ensembles of multibody simulations without requiring direct management of the underlying distributed infrastructure.

What carries the argument

Chrono::Ray, the modular workflow framework resulting from the integration of the Chrono multibody dynamics engine and the Ray distributed computing platform.

Load-bearing premise

The integration maintains the original simulation accuracy and provides meaningful performance benefits for practical large-scale engineering tasks.

What would settle it

Running the same simulation ensembles with and without Chrono::Ray and observing whether the results match exactly and whether runtime scales as expected with added compute resources.

Figures

Figures reproduced from arXiv: 2605.13767 by Dan Negrut, Khailanii Slaton.

**Figure 2.** Figure 2: Parameter estimation results for the multibody lunar lander example. Individual [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗

**Figure 3.** Figure 3: Representative snapshots from the continuum terramechanics design of experi [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗

read the original abstract

Large-scale simulation studies can provide invaluable insights across computational engineering efforts, but they are often computationally demanding, requiring the use of distributed computing, which is itself not a simple task. Chrono::Ray addresses this challenge by integrating the high-fidelity multibody dynamics simulation engine Chrono with the open-source distributed computing platform Ray. The result is a modular workflow framework providing user-friendly abstractions for large-scale engineering simulation studies, supporting scalable orchestration of large ensembles of simulation trials without requiring users to directly manage distributed infrastructure. The current capabilities of the framework are demonstrated through two representative examples: parameter recovery for a multibody lunar lander model, and design of experiments for parameters of a continuum terramechanics model. Chrono::Ray is a part of the larger Project Chrono ecosystem and is released as an open-source software package, with source code available at https://github.com/uwsbel/chrono-ray.git.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Chrono::Ray wraps Ray around Chrono for easier ensemble runs of multibody simulations, but the two examples give no scaling numbers or overhead data.

read the letter

This paper presents Chrono::Ray as a way to run large numbers of Chrono multibody simulations using Ray for distribution. The key idea is to give users simple interfaces so they can launch ensembles without managing the underlying compute resources themselves. The integration is new in the sense that it targets multibody dynamics workflows specifically, with two examples to show it in action: recovering parameters for a lunar lander model and running a design of experiments on a terramechanics model. Releasing the code openly on GitHub is a clear positive, as it lets others try it directly. The framework does a good job of packaging the distributed aspects into something usable for engineering studies. For someone already working with Chrono, this could save time on setting up parallel runs. Where it falls short is in the evaluation. The description relies on those two cases without providing metrics like how performance scales with more simulations, what the added overhead is, or any analysis of potential issues like load balancing or errors in distributed mode. That leaves the claim of practical scalability somewhat untested. Readers who do computational work in vehicle dynamics or terramechanics would find this relevant, especially if they need to explore parameter spaces at scale. It is the kind of software paper that deserves a serious look from referees, as the core contribution is a working tool rather than a new theory. I would recommend putting it through peer review.

Referee Report

2 major / 2 minor

Summary. The manuscript introduces Chrono::Ray, a framework integrating the Chrono multibody dynamics simulator with the Ray distributed computing platform. It provides modular, user-friendly abstractions for orchestrating large ensembles of simulations on distributed infrastructure without requiring users to manage low-level details, with capabilities shown through two examples: parameter recovery for a lunar lander multibody model and design of experiments for a continuum terramechanics model.

Significance. If the integration preserves simulation fidelity and delivers practical scalability, the work would offer a useful addition to the Project Chrono ecosystem for engineers performing high-throughput parametric studies and uncertainty quantification. The open-source release strengthens accessibility. However, the absence of quantitative performance data in the demonstrations limits evaluation of its advantages over direct use of Ray or other orchestration tools.

major comments (2)

[Demonstration sections] The two demonstration cases (lunar lander parameter recovery and terramechanics DoE) describe the workflow but report no quantitative metrics on wall-clock scaling, parallel efficiency, communication overhead, or failure rates as ensemble size increases. This directly undercuts the central claim of 'scalable orchestration' and 'practical scalability for real engineering workloads' without additional evidence.
[Framework architecture] The architecture description does not address how simulation fidelity is verified after distribution or how numerical errors might accumulate in the Ray-managed ensemble execution, which is load-bearing for claims that the framework supports high-fidelity multibody analysis.

minor comments (2)

[Abstract and Introduction] The abstract and introduction could clarify the specific Ray primitives wrapped by the abstractions to help readers assess novelty relative to existing Ray integrations in scientific computing.
[Figures and code examples] Figure captions and code snippets would benefit from explicit statements of the number of trials, hardware used, and any observed overhead to improve reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive review and for recognizing the potential utility of Chrono::Ray within the Project Chrono ecosystem. We have carefully considered the major comments and revised the manuscript accordingly to strengthen the evidence for scalability and to clarify fidelity preservation.

read point-by-point responses

Referee: [Demonstration sections] The two demonstration cases (lunar lander parameter recovery and terramechanics DoE) describe the workflow but report no quantitative metrics on wall-clock scaling, parallel efficiency, communication overhead, or failure rates as ensemble size increases. This directly undercuts the central claim of 'scalable orchestration' and 'practical scalability for real engineering workloads' without additional evidence.

Authors: We agree that quantitative performance metrics are necessary to support the claims of scalable orchestration. In the revised manuscript, we have expanded the demonstration sections with new results that include wall-clock scaling curves, parallel efficiency measurements, communication overhead estimates, and observed failure rates for ensemble sizes ranging from 10 to 1000 simulations. These data were obtained on a cluster with up to 64 nodes and are presented in new figures and tables that directly address the practical scalability for engineering workloads. revision: yes
Referee: [Framework architecture] The architecture description does not address how simulation fidelity is verified after distribution or how numerical errors might accumulate in the Ray-managed ensemble execution, which is load-bearing for claims that the framework supports high-fidelity multibody analysis.

Authors: We appreciate this observation. Chrono::Ray launches each multibody simulation as an independent Chrono instance on the allocated Ray workers; the distribution layer only manages task submission, input parameter passing, and result collection without modifying the underlying dynamics solver or introducing any approximation or coupling between simulations. Consequently, the numerical behavior of each run is identical to a local Chrono execution, and no additional numerical errors accumulate. We have added a dedicated paragraph in the architecture section that explains this design choice and reports a verification procedure in which selected ensemble members were re-run locally to confirm bit-for-bit agreement in output trajectories. revision: yes

Circularity Check

0 steps flagged

No circularity: software framework paper with no derivations or predictions

full rationale

The manuscript presents a software integration framework (Chrono::Ray) that combines the Chrono multibody simulator with the Ray distributed computing platform. It contains no equations, fitted parameters, predictions, uniqueness theorems, or ansatzes. The central contribution is a set of user-friendly abstractions for orchestrating simulation ensembles, illustrated only by two descriptive examples (lunar lander parameter recovery and terramechanics DoE). No derivation chain exists that could reduce to its own inputs by construction, and no self-citations are invoked to justify load-bearing mathematical claims. The paper is therefore self-contained as a software-engineering contribution whose value is assessed externally through usage and benchmarks rather than internal logical closure.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

This is a software framework description paper. No free parameters, mathematical axioms, or invented physical entities are introduced; the work relies on the existing Chrono and Ray codebases.

pith-pipeline@v0.9.0 · 5689 in / 988 out tokens · 37236 ms · 2026-05-19T13:49:59.014402+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Chrono::Ray is a modular software interface that bridges PyChrono... with the Ray distributed computing platform... workflow-oriented interfaces for large-scale simulation studies

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

6 extracted references · 6 canonical work pages · 1 internal anchor

[1]

thebibliography tasora2016

format for author-year citations. thebibliography tasora2016

work page
[2]

Chrono: An open source multi-physics dynamics engine

Tasora, A., Serban, R., Mazhar, H., Pazouki, A., Melanz, D., Fleischmann, J., Taylor, M., Sugiyama, H., and Negrut, D. Chrono: An open source multi-physics dynamics engine. In High Performance Computing in Science and Engineering, pages 19--49. Springer, 2016

work page 2016
[3]

I., and Stoica, I

Moritz, P., Nishihara, R., Wang, S., Tumanov, A., Liaw, R., Liang, E., Elibol, M., Yang, Z., Paul, W., Jordan, M. I., and Stoica, I. Ray: A distributed framework for emerging AI applications. In 13th USENIX Symposium on Operating Systems Design and Implementation (OSDI 18), pages 561--577, 2018

work page 2018
[4]

Tune: A Research Platform for Distributed Model Selection and Training

Liaw, R., Liang, E., Nishihara, R., Moritz, P., Gonzalez, J. E., and Stoica, I. Tune: A research platform for distributed model selection and training. arXiv preprint arXiv:1807.05118, 2018

work page internal anchor Pith review Pith/arXiv arXiv 2018
[5]

A physics-based continuum model for versatile, scalable, and fast terramechanics simulation

Unjhawala, H., et al. A physics-based continuum model for versatile, scalable, and fast terramechanics simulation. arXiv preprint arXiv:2507.05643, 2025

work page arXiv 2025
[6]

Variance based sensitivity analysis of model output: Design and estimator for the total sensitivity index

Saltelli, A., Annoni, P., Azzini, I., Campolongo, F., Ratto, M., and Tarantola, S. Variance based sensitivity analysis of model output: Design and estimator for the total sensitivity index. Computer Physics Communications, 181(2):259--270, 2010

work page 2010

[1] [1]

thebibliography tasora2016

format for author-year citations. thebibliography tasora2016

work page

[2] [2]

Chrono: An open source multi-physics dynamics engine

Tasora, A., Serban, R., Mazhar, H., Pazouki, A., Melanz, D., Fleischmann, J., Taylor, M., Sugiyama, H., and Negrut, D. Chrono: An open source multi-physics dynamics engine. In High Performance Computing in Science and Engineering, pages 19--49. Springer, 2016

work page 2016

[3] [3]

I., and Stoica, I

Moritz, P., Nishihara, R., Wang, S., Tumanov, A., Liaw, R., Liang, E., Elibol, M., Yang, Z., Paul, W., Jordan, M. I., and Stoica, I. Ray: A distributed framework for emerging AI applications. In 13th USENIX Symposium on Operating Systems Design and Implementation (OSDI 18), pages 561--577, 2018

work page 2018

[4] [4]

Tune: A Research Platform for Distributed Model Selection and Training

Liaw, R., Liang, E., Nishihara, R., Moritz, P., Gonzalez, J. E., and Stoica, I. Tune: A research platform for distributed model selection and training. arXiv preprint arXiv:1807.05118, 2018

work page internal anchor Pith review Pith/arXiv arXiv 2018

[5] [5]

A physics-based continuum model for versatile, scalable, and fast terramechanics simulation

Unjhawala, H., et al. A physics-based continuum model for versatile, scalable, and fast terramechanics simulation. arXiv preprint arXiv:2507.05643, 2025

work page arXiv 2025

[6] [6]

Variance based sensitivity analysis of model output: Design and estimator for the total sensitivity index

Saltelli, A., Annoni, P., Azzini, I., Campolongo, F., Ratto, M., and Tarantola, S. Variance based sensitivity analysis of model output: Design and estimator for the total sensitivity index. Computer Physics Communications, 181(2):259--270, 2010

work page 2010