Pith. sign in

REVIEW 3 major objections 1 minor 7 cited by

Mol-R1: Towards Explicit Long-CoT Reasoning in Molecule Discovery

T0 review · 3 major / 1 minor · reviewed 2026-08-05 · deepseek-v4-flash

Pith's one-line read Mol-R1 makes molecule-generating LLMs reason out loud, and the paper claims this beats existing baselines.

desk verdict This is a mismatched submission: the claimed Mol-R1 manuscript is absent, replaced by an unrelated numerical-relativity paper, so the abstract's 'superior performance' claim is unverifiable. read the letter →

arxiv 2508.08401 v1 pith:AHPZ5NVZ submitted 2025-08-11 cs.CL

classification cs.CL
keywords largelanguagemodelschain-of-thoughtreasoningmoleculediscoverytext-basedgenerationdistillationreinforcementlearningexplainability
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 an R1-style large language model—one that produces a long, explicit chain of thought—can be trained to reason about molecule generation in a way that is both chemically valid and auditable. The proposed framework, Mol-R1, combines two mechanisms: PRID, a distillation method that produces paired reasoning traces guided by prior chemical regulations, and MoIA, an iterative training loop that alternates supervised fine-tuning with reinforced policy optimization. A sympathetic reader would care because if the claim is right, molecule discovery becomes a domain where a model can show its step-by-step reasoning instead of acting as a black box, and the training recipe can be reused for other knowledge-intensive fields.

What carries the argument

The load-bearing mechanism is the PRID+MoIA training loop. PRID (Prior Regulation via In-context Distillation) curates paired reasoning traces—each a molecule-generation task paired with a step-by-step solution that respects prior chemical regulations—so the model learns from grounded examples rather than fluent but ungrounded text. MoIA (Molecular Iterative Adaptation) then alternates supervised fine-tuning on those traces with reinforced policy optimization, where a reward signal reinforces chemically sound reasoning. Together they are what convert a generic long-CoT reasoner into a molecule-discovery reasoner.

What would settle it

Have trained chemists label the intermediate reasoning traces of Mol-R1 on a held-out set of molecule-generation prompts for structural validity and chemical plausibility, and compare the validity rate of the steps with that of the final molecules (and with a model trained without PRID/MoIA). A high rate of invalid valences, impossible reactions, or regulation violations in the traces—especially when final molecules still pass benchmark filters—would show the 'reasoning' is learned mimicry, not learned chemistry.

Watch

Extended reading notes

Core claim

Mol-R1's central claim is that explicit long-chain-of-thought reasoning can be made to work in the knowledge-intensive domain of text-based molecule generation. The paper argues that generic R1-style reasoning models fail there because they lack domain grounding, and that grounding can be supplied by PRID—distilling paired reasoning traces that follow prior chemical regulations to build a high-quality dataset—and then sharpened by MoIA, which iteratively combines supervised fine-tuning with reinforced policy optimization using a reward signal. The result, the paper reports, is superior performance against existing baselines on text-based molecule reasoning generation, with the practical bene

Load-bearing premise

The pipeline's gains depend on the 'prior regulations' used in distillation and the reward used in reinforcement learning actually encoding chemically correct ground truth; if those are wrong, the model learns fluent-sounding but invalid chemistry, and any benchmark improvement could reflect reward gaming.

Editorial extensions

If this is right

  • If Mol-R1 is right, long-CoT reasoning models can be adapted to knowledge-intensive scientific domains, not just commonsense and math.
  • Molecule generation gains an audit trail: each generated molecule is backed by explicit reasoning steps a chemist can inspect.
  • The PRID+MoIA recipe gives a reusable pattern for grounding reasoning LLMs in other regulated domains.
  • Iterative SFT-plus-reinforcement could become a standard recipe for domain-specific reasoning LLMs.

Reading between the lines

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

  • The abstract does not define what the 'prior regulations' or the RPO reward are, or how trace correctness was validated; if those encode wrong chemistry, both SFT and RPO will amplify the error rather than correct it.
  • The same distillation-plus-iteration recipe could plausibly transfer to other structured generation tasks—retrosynthesis, materials design, or even legal or clinical reasoning—wherever checkable regulations exist.
  • A direct test would be to compare the chemical validity of intermediate reasoning steps against the final outputs; if steps are invalid more often than outputs, the claimed explainability is cosmetic.
  • The supplied full text is a different paper on numerical-relativity black-hole simulations, not the Mol-R1 manuscript; this summary is based solely on the Mol-R1 abstract.
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 / 1 minor

Summary. The submission is identified as arXiv:2508.08401 (cs.CL), titled "Mol-R1: Towards Explicit Long-CoT Reasoning in Molecule Discovery." The abstract claims a framework, Mol-R1, that combines PRID (Prior Regulation via In-context Distillation) and MoIA (Molecular Iterative Adaptation with SFT and RPO) to improve explainability and reasoning of R1-like long-CoT LLMs in text-based molecule generation, ending with an assertion of "superior performance against existing baselines." The full text provided, however, is an unrelated numerical-relativity paper: "Horizon tracking for asynchronous parallel black hole simulations" (arXiv:2508.08408v2). It contains no mention of Mol-R1, PRID, MoIA, molecule generation, reasoning traces, chemical validity, benchmarks, or any LLM. The central claims of the abstract are therefore entirely unsupported by the manuscript content.

Significance. If the claimed results were real and properly supported, a framework that enables auditable long-CoT reasoning for molecule generation would be a useful contribution to computational chemistry and LLM reasoning. The paper might also be notable for introducing PRID and MoIA as named training strategies. However, none of these contributions is present in the submitted full text. The actual manuscript is a numerical-relativity paper about asynchronous horizon tracking; whatever its own merits, it is not the paper described in the abstract. Consequently, there is no assessable scientific contribution to evaluate, and no evidence for the empirical claims.

major comments (3)
  1. [Abstract (evaluation claims)] The central claim of the abstract—that Mol-R1 with PRID and MoIA shows superior performance in text-based molecule generation—is unsupported because the full text is an entirely different paper, "Horizon tracking for asynchronous parallel black hole simulations." No section, equation, dataset, reward definition, baseline, or result related to Mol-R1 appears anywhere in the manuscript. This is a load-bearing, submission-level defect: the contribution asserted in the abstract is absent from the submitted artifact.
  2. [Abstract (evaluation claims)] Even taken in isolation, the abstract's claim of "superior performance against existing baselines" is unverifiable: it names no baselines, no metrics, no datasets, and no evaluation protocol. The PRID regulations and RPO reward are also unspecified, so the correctness of the reasoning traces and the possibility of reward gaming cannot be assessed. These omissions are independently serious for an empirical paper, and they compound the total absence of supporting content in the full text.
  3. [Full text (all sections)] The manuscript body addresses asynchronous parallelism, apparent-horizon finding, and feedback control in SpECTRE. None of these topics bears on molecule discovery, LLM reasoning, distillation, or reinforcement learning. The mismatch is not a local flaw in a derivation or a missing experiment; it is the complete absence of the claimed research object. For this reason, the paper cannot be checked for internal consistency, reproducibility, or the risk that RPO/trace data are circularly constructed from the evaluation benchmark.
minor comments (1)
  1. [Header/identification] The full text carries the arXiv identifier 2508.08408v2, which differs from the submission's stated identifier 2508.08401. This is consistent with the content mismatch and should be corrected if the authors resubmit the intended manuscript.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found because the claimed derivation chain is entirely absent from the full text.

full rationale

The abstract claims a contribution to molecule discovery via Mol-R1, PRID, and MoIA, with 'superior performance against existing baselines.' However, the supplied full text is an unrelated numerical-relativity paper, 'Horizon tracking for asynchronous parallel black hole simulations' (arXiv:2508.08408v2). It contains no mention of Mol-R1, PRID, MoIA, molecules, chemical validity, reasoning traces, LLMs, or any of the claimed experiments, equations, datasets, or reward definitions. Therefore, there is no derivation chain to walk and no specific equation or fitted parameter can be shown to reduce to an input by construction. The abstract's claims are unsupported by the full text, which is a severe submission-level defect, but it is not circularity under the definition provided: no self-definition, no fitted-input-called-prediction, no load-bearing self-citation, and no renaming of a known result can be identified without the manuscript's actual content. Thus the appropriate circularity score is 0, reflecting the absence of any circular step rather than endorsement of the paper's claims.

Assumptions & free parameters 2 free parameters · 3 assumptions · 2 invented entities

The ledger is necessarily incomplete because only the abstract is available. The two named procedures (PRID, MoIA) are the paper's invented machinery, and the untested assumptions are trace correctness, transferability of RL recipes, and benchmark validity. The free parameters are unstated in the abstract: the regulation set and the RPO reward are the two most consequential choices, and neither is described.

free parameters (2)
  • Prior regulations set for PRID
    The abstract says distillation traces are 'guided by prior regulations' but does not define these rules. If they are hand-authored chemical constraints, they are chosen by hand and directly shape all training data; this is a load-bearing design choice for trace quality.
  • RPO reward design and RL hyperparameters
    Reinforced Policy Optimization requires a reward signal for molecule outputs (validity, similarity, property correctness) plus RL hyperparameters (KL coefficients, iteration counts, sampling budgets). None are specified in the abstract, yet they determine whether the RL loop improves or merely games the metric.
assumptions (3)
  • domain assumption Distilled reasoning traces generated under PRID are chemically correct enough to serve as training supervision
    The entire SFT/RPO loop inherits errors from the traces. The abstract asserts a 'high-quality reasoning dataset curated through PRID' but provides no verification protocol for trace correctness.
  • domain assumption Long-CoT RL recipes that work for math and commonsense transfer to molecule generation
    The paper's motivation states that such transfer is non-obvious ('often criticized for limited ability in knowledge-intensive domains'), yet MoIA assumes the R1-style iterative SFT+RL recipe is the right vehicle for this domain.
  • domain assumption Evaluation benchmark metrics reflect real molecule generation quality
    The claim of 'superior performance' rests on benchmark metrics that the abstract does not name and does not validate against wet-lab or quantum-chemistry ground truth. If the metric rewards format over chemistry, the central claim is weaker than stated.
invented entities (2)
  • PRID (Prior Regulation via In-context Distillation)
    purpose: Distillation strategy to generate paired reasoning traces guided by prior regulations
    A new training-data generation procedure. It is a methodological component, not a physical entity, but it is an invented construct whose only evidence is the benchmark improvement claimed in the abstract; it has no external falsifiable handle described.
  • MoIA (Molecular Iterative Adaptation)
    purpose: Iterative combination of Supervised Fine-tuning (SFT) with Reinforced Policy Optimization (RPO) to boost reasoning performance
    A new training strategy. Its only claimed evidence is the benchmark gain in the abstract; whether it outperforms simpler one-shot SFT+RL is not established by anything in the abstract.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Mol-R1: Towards Explicit Long-CoT Reasoning in Molecule Discovery." pith.science (2026). https://pith.science/paper/AHPZ5NVZ

@misc{pith2026250808401,
  author       = {Pith},
  title        = {Pith review of: Mol-R1: Towards Explicit Long-CoT Reasoning in Molecule Discovery},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AHPZ5NVZ}},
  note         = {Machine review of arXiv:2508.08401}
}
read the original abstract

Large language models (LLMs), especially Explicit Long Chain-of-Thought (CoT) reasoning models like DeepSeek-R1 and QWQ, have demonstrated powerful reasoning capabilities, achieving impressive performance in commonsense reasoning and mathematical inference. Despite their effectiveness, Long-CoT reasoning models are often criticized for their limited ability and low efficiency in knowledge-intensive domains such as molecule discovery. Success in this field requires a precise understanding of domain knowledge, including molecular structures and chemical principles, which is challenging due to the inherent complexity of molecular data and the scarcity of high-quality expert annotations. To bridge this gap, we introduce Mol-R1, a novel framework designed to improve explainability and reasoning performance of R1-like Explicit Long-CoT reasoning LLMs in text-based molecule generation. Our approach begins with a high-quality reasoning dataset curated through Prior Regulation via In-context Distillation (PRID), a dedicated distillation strategy to effectively generate paired reasoning traces guided by prior regulations. Building upon this, we introduce MoIA, Molecular Iterative Adaptation, a sophisticated training strategy that iteratively combines Supervised Fine-tuning (SFT) with Reinforced Policy Optimization (RPO), tailored to boost the reasoning performance of R1-like reasoning models for molecule discovery. Finally, we examine the performance of Mol-R1 in the text-based molecule reasoning generation task, showing superior performance against existing baselines.

Discussion (0). Sign in to comment.

Forward citations

Cited by 7 Pith papers

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

  1. Chemical Chain-of-Thought Functions as a Hallucination-Prone Molecular Scratchpad

    cs.CE 2026-07 conditional novelty 7.0 of 10

    Chemical chain-of-thought traces from LLMs are not faithful explanations: they frequently contain fabricated functional-group claims even on correct answers, yet partial SMILES drafts within the trace can still causal...

  2. PolyReal: A Benchmark for Real-World Polymer Science Workflows

    cs.CV 2026-04 unverdicted novelty 7.0 of 10

    PolyReal benchmark shows leading MLLMs perform well on polymer knowledge reasoning but drop sharply on practical tasks like lab safety analysis and raw data extraction.

  3. Speak-to-Structure: Evaluating LLMs in Open-domain Natural Language-Driven Molecule Generation

    cs.CL 2024-12 unverdicted novelty 7.0 of 10

    S^2-Bench is a new one-to-many benchmark for natural language-driven molecule generation with three tasks, and OpenMolIns is an instruction dataset enabling Llama3.1-8B to outperform GPT-4o and Claude-3.5 on it.

  4. MolGVR: A Chemistry-Grounded Framework for Text-to-Molecule Generation

    cs.LG 2026-07 conditional novelty 6.0 of 10

    Adding an LLM-based verifier and a feedback-trained refiner to an already strong generator lifts exact-match accuracy by about one point on two benchmarks, with larger gains coming from the generator's MSR+RL training.

  5. Mol-Debate: Multi-Agent Debate Improves Structural Reasoning in Molecular Design

    cs.AI 2026-04 unverdicted novelty 6.0 of 10

    Mol-Debate applies multi-agent debate in an iterative loop with perspective orchestration to achieve state-of-the-art text-guided molecular design, scoring 59.82% exact match on ChEBI-20 and 50.52% weighted success on...

  6. ChemVLR: Prioritizing Reasoning in Perception for Chemical Vision-Language Understanding

    cs.CL 2026-04 unverdicted novelty 6.0 of 10

    ChemVLR prioritizes reasoning in perception for chemical VLMs by identifying descriptors such as functional groups before generating answers, using a 760k curated dataset and three-stage training to reach SOTA performance.

  7. STRIDE: Post-Training LLMs to Reason and Refine Bio-Sequences via Edit Trajectories

    cs.CE 2026-03 conditional novelty 6.0 of 10

    Training LLMs to emit executable edit trajectories (INSERT/DELETE/REPLACE) from Levenshtein alignments plus policy optimization improves oracle-scored bio-sequence optimization success and novelty.

Reference graph

Works this paper leans on

55 extracted references · 51 canonical work pages · cited by 7 Pith papers

  1. [1]

    Aasi et al

    J. Aasi et al. Advanced LIGO. Class. Quant. Grav., 32:074001, 2015

  2. [2]

    Acernese et al

    F. Acernese et al. Advanced Virgo: a second-generation interferometric gravitational wave detector. Class. Quant. Grav., 32(2):024001, 2015

  3. [3]

    Akutsu et al

    T. Akutsu et al. Overview of KAGRA: Detector design and construction history. Progress of Theoretical and Ex- perimental Physics, 2021(5):05A101, May 2021

  4. [4]

    K. L. Dooley et al. GEO 600 and the GEO-HF upgrade program: successes and challenges. Class. Quant. Grav., 33:075009, 2016

  5. [5]

    B. P. Abbott et al. Observation of gravitational waves from a binary black hole merger. Phys. Rev. Lett., 116:061102, Feb 2016

  6. [6]

    Capote et al

    E. Capote et al. Advanced LIGO detector performance in the fourth observing run. Phys. Rev. D, 111(6):062002, 2025

  7. [7]

    Mroue et al

    Abdul H. Mroue et al. Catalog of 174 Binary Black Hole Simulations for Gravitational Wave Astronomy. Phys. Rev. Lett., 111(24):241104, 2013

  8. [8]

    Clark, Lionel Lon- don, Pablo Laguna, and Deirdre Shoemaker

    Karan Jani, James Healy, James A. Clark, Lionel Lon- don, Pablo Laguna, and Deirdre Shoemaker. Georgia Tech Catalog of Gravitational Waveforms. Class. Quant. Grav., 33(20):204001, 2016

Show all 55 references
  1. [9]

    Lousto, Yosef Zlochower, and Manuela Campanelli

    James Healy, Carlos O. Lousto, Yosef Zlochower, and Manuela Campanelli. The RIT binary black hole simula- tions catalog. Class. Quant. Grav., 34(22):224001, 2017

  2. [10]

    Lousto, Jacob Lange, Richard O’Shaughnessy, Yosef Zlochower, and Manuela Campan- elli

    James Healy, Carlos O. Lousto, Jacob Lange, Richard O’Shaughnessy, Yosef Zlochower, and Manuela Campan- elli. Second RIT binary black hole simulations catalog and its application to gravitational waves parameter es- timation. Phys. Rev. D, 100(2):024021, 2019

  3. [11]

    The SXS Collaboration catalog of binary black hole simulations

    Michael Boyle et al. The SXS Collaboration catalog of binary black hole simulations. Class. Quant. Grav., 36(19):195006, 2019

  4. [12]

    Scheel et al

    Mark A. Scheel et al. The sxs collaboration’s third cat- alog of binary black hole simulations, 2025

  5. [13]

    James Healy and Carlos O. Lousto. Third RIT bi- nary black hole simulations catalog. Phys. Rev. D , 102(10):104018, 2020

  6. [14]

    James Healy and Carlos O. Lousto. Fourth RIT binary black hole simulations catalog: Extension to eccentric orbits. Phys. Rev. D, 105(12):124010, 2022

  7. [15]

    Second MAYA Catalog of Binary Black Hole Numerical Relativity Waveforms

    Deborah Ferguson et al. Second MAYA Catalog of Binary Black Hole Numerical Relativity Waveforms. arXiv e- prints, page arXiv:2309.00262, 9 2023

  8. [16]

    Laser Interferometer Space Antenna

    Pau others Amaro-Seoane. Laser Interferometer Space Antenna. arXiv e-prints, page arXiv:1702.00786, Febru- ary 2017

  9. [17]

    Punturo et al

    M. Punturo et al. The Einstein Telescope: a third- generation gravitational wave observatory. Class. Quan- tum Grav., 27(19):194002, October 2010

  10. [18]

    A Horizon Study for Cosmic Ex- plorer: Science, Observatories, and Community

    Matthew Evans et al. A Horizon Study for Cosmic Ex- plorer: Science, Observatories, and Community. arXiv e-prints, page arXiv:2109.09882, 9 2021

  11. [19]

    Gravitational waveform accuracy requirements for future ground-based detectors

    Michael P¨ urrer and Carl-Johan Haster. Gravitational waveform accuracy requirements for future ground-based detectors. Phys. Rev. Res., 2(2):023151, 2020

  12. [20]

    Assessing the readiness of numer- ical relativity for LISA and 3G detectors

    Deborah Ferguson, Karan Jani, Pablo Laguna, and Deirdre Shoemaker. Assessing the readiness of numer- ical relativity for LISA and 3G detectors. Phys. Rev. D, 104(4):044037, 2021

  13. [21]

    Accuracy limita- tions of existing numerical relativity waveforms on the data analysis of current and future ground-based detec- tors

    Aasim Jan, Deborah Ferguson, Jacob Lange, Deirdre Shoemaker, and Aaron Zimmerman. Accuracy limita- tions of existing numerical relativity waveforms on the data analysis of current and future ground-based detec- tors. Phys. Rev. D, 110(2):024023, 2024

  14. [22]

    https://www.black-holes.org/SpEC.html

  15. [23]

    Kidder, Nils L

    Nils Deppe, William Throwe, Lawrence E. Kidder, Nils L. Vu, Kyle C. Nelli, Crist´ obal Armaza, Marceline S. Bonilla, Fran¸ cois H´ ebert, Yoonsoo Kim, Prayush Kumar, Geoffrey Lovelace, Alexandra Macedo, Jordan Moxon, Eamonn O’Shea, Harald P. Pfeiffer, Mark A. Scheel, Saul A. T...

  16. [24]

    Performance- Portable Numerical Relativity with AthenaK

    Hengrui Zhu, Jacob Fields, Francesco Zappa, David Radice, James Stone, Alireza Rashti, William Cook, Se- bastiano Bernuzzi, and Boris Daszuta. Performance- Portable Numerical Relativity with AthenaK. arXiv e- prints, page arXiv:2409.10383, 9 2024

  17. [25]

    Brandt, Roland Haas, Erik Schnetter, and Yannick de Graaf

    Swapnil Shankar, Philipp M¨ osta, Steven R. Brandt, Roland Haas, Erik Schnetter, and Yannick de Graaf. GRaM-X: a new GPU-accelerated dynamical spacetime GRMHD code for Exascale computing with the Einstein Toolkit. Class. Quant. Grav., 40(20):205009, 2023

  18. [26]

    Kalinani et al

    Jay V. Kalinani et al. AsterX: a new open-source GPU- accelerated GRMHD code for dynamical spacetimes. Class. Quant. Grav., 42(2):025016, 2025

  19. [27]

    Amrex: Block-structured adaptive mesh refinement for multiphysics applications

    Weiqun Zhang, Andrew Myers, Kevin Gott, Ann Alm- gren, and John Bell. Amrex: Block-structured adaptive mesh refinement for multiphysics applications. The In- ternational Journal of High Performance Computing Ap- plications, 35(6):508–526, 2021

  20. [28]

    Trott et al

    Christian R. Trott et al. Kokkos 3: Programming model extensions for the exascale era. IEEE Transactions on Parallel and Distributed Systems, 33(4):805–817, 2022

  21. [29]

    The new discontinuous Galerkin methods based numerical relativity program Nmesh

    Wolfgang Tichy, Liwei Ji, Ananya Adhikari, Alireza Rashti, and Michal Pirog. The new discontinuous Galerkin methods based numerical relativity program Nmesh. Class. Quant. Grav., 40(2):025004, 2023

  22. [30]

    Davison de St

    J. Davison de St. Germain, John McCorquodale, Steven G. Parker, and Christopher R. Johnson. Uin- tah: A massively parallel problem solving environment. In Proceedings of the 9th IEEE International Symposium on High Performance Distributed Computing, HPDC ’00, page 33, USA, 200...

  23. [31]

    Chapel: a versatile language for teaching parallel programming: conference workshop

    Kyle Burke. Chapel: a versatile language for teaching parallel programming: conference workshop. J. Comput. Sci. Coll., 30(6):16, June 2015. 15

  24. [32]

    Uiuc-ppl/charm: Charm++ ver- sion 6.10.2

    Laxmikant Kale et al. Uiuc-ppl/charm: Charm++ ver- sion 6.10.2. 10.5281/zenodo.3972617, August 2020

  25. [33]

    Bennett, David Hollman, Sean Treichler, Patrick S

    Philippe P´ eba¨ y, Janine C. Bennett, David Hollman, Sean Treichler, Patrick S. McCormick, Christine M. Sweeney, Hemanth Kolla, and Alex Aiken. Towards asynchronous many-task in situ data analysis using legion. In 2016 IEEE International Parallel and Distributed Process- ing ...

  26. [34]

    Protocol for asynchronous, reliable, secure and efficient consensus (parsec) version 2.0, 2019

    Pierre Chevalier, Bartlomiej Kaminski, Fraser Hutchison, Qi Ma, Spandan Sharma, Andreas Fackler, and William J Buchanan. Protocol for asynchronous, reliable, secure and efficient consensus (parsec) version 2.0, 2019

  27. [35]

    MPI: A Message- Passing Interface Standard Version 5.0, June 2025

    Message Passing Interface Forum. MPI: A Message- Passing Interface Standard Version 5.0, June 2025

  28. [36]

    Pseudospectral apparent horizon finders: An Efficient new algorithm

    Carsten Gundlach. Pseudospectral apparent horizon finders: An Efficient new algorithm. Phys. Rev. D, 57:863–875, 1998

  29. [37]

    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. Phys. Rev., D74:104006, 2006

  30. [38]

    Hemberger, Mark A

    Daniel A. Hemberger, Mark A. Scheel, Lawrence E. Kid- der, B´ ela Szil´ agyi, Geoffrey Lovelace, Nicholas W. Taylor, and Saul A. Teukolsky. Dynamical excision boundaries in spectral evolutions of binary black hole spacetimes. Class. Quant. Grav., 30:115001, 2013

  31. [39]

    Scheel, Lawrence E

    Lee Lindblom, Mark A. Scheel, Lawrence E. Kidder, Robert Owen, and Oliver Rinne. A New generalized har- monic evolution system. Class. Quant. Grav., 23:S447– S462, 2006

  32. [40]

    Bela Szilagyi, Lee Lindblom, and Mark A. Scheel. Sim- ulations of Binary Black Hole Mergers Using Spectral Methods. Phys. Rev. D, 80:124010, 2009

  33. [41]

    Kidder, Mark A

    Nils Deppe, Lawrence E. Kidder, Mark A. Scheel, and Saul A. Teukolsky. Critical behavior in 3D gravita- tional collapse of massless scalar fields. Phys. Rev., D99(2):024018, 2019

  34. [42]

    Simulating binary black hole mergers using discontinuous Galerkin methods

    Geoffrey Lovelace et al. Simulating binary black hole mergers using discontinuous Galerkin methods. Class. Quant. Grav., 42(3):035001, 2025

  35. [43]

    Post-Newtonian theory for gravitational waves

    Luc Blanchet. Post-Newtonian theory for gravitational waves. Living Rev. Rel., 27(1):4, 2024

  36. [44]

    Fundamentals of parallel multicore archi- tecture

    Yan Solihin. Fundamentals of parallel multicore archi- tecture. Chapman & Hall/CRC Computational Science. Chapman & Hall/CRC, Philadelphia, PA, June 2020

  37. [45]

    Michael J. Quinn. Parallel computing (2nd ed.): theory and practice. McGraw-Hill, Inc., USA, 1994

  38. [46]

    Binary neutron star mergers using a discontinuous Galerkin-finite difference hybrid method

    Nils Deppe et al. Binary neutron star mergers using a discontinuous Galerkin-finite difference hybrid method. Class. Quant. Grav., 41(24):245002, 2024

  39. [47]

    and Zanotti, O

    Rezzolla, L. and Zanotti, O. Relativistic Hydrodynamics. Oxford University Press, sep 2013

  40. [48]

    Soham Mandal and Paul C. Duffell. Sprout: A moving- mesh hydro code using a uniformly expanding carte- sian grid. The Astrophysical Journal Supplement Series, 269(1):30, nov 2023

  41. [49]

    Sarah Habib et al. 2025. In preparation

  42. [50]

    Pfeiffer, Nikolas A

    Guillermo Lara, Harald P. Pfeiffer, Nikolas A. Wittek, Nils L. Vu, Kyle C. Nelli, Alexander Carpenter, Geoffrey Lovelace, Mark A. Scheel, and William Throwe. Scalar- ization of isolated black holes in scalar Gauss-Bonnet the- ory in the fixing-the-equations approach. Phys. Rev...

  43. [51]

    Wittek, Leor Barack, Harald P

    Nikolas A. Wittek, Leor Barack, Harald P. Pfeiffer, Adam Pound, Nils Deppe, Lawrence E. Kidder, Alexandra Macedo, Kyle C. Nelli, William Throwe, and Nils L. Vu. Relieving scale disparity in binary black hole simulations. 10 2024

  44. [52]

    Signatures from metastable oppositely-charged black hole binaries in scalar Gauss- Bonnet gravity

    Guillermo Lara et al. Signatures from metastable oppositely-charged black hole binaries in scalar Gauss- Bonnet gravity. 5 2025

  45. [53]

    T. Tantau. The tikz and pgf packages. github.com/pgf- tikz/pgf

  46. [54]

    The ParaView Guide: A Parallel Vi- sualization Application

    Utkarsh Ayachit. The ParaView Guide: A Parallel Vi- sualization Application. Kitware, Inc., Clifton Park, NY, USA, 2015

  47. [55]

    Ahrens, Berk Geveci, and C

    J. Ahrens, Berk Geveci, and C. Law. Paraview: An end- user tool for large-data visualization. In Visualization Handbook. Butterworth-Heinemann, Burlington, 2005

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.