Review of article: VERTEX-CFD

[Phoenixfire1081]

Here are my comments on the article,

• L62: High-fidelity simulation - why is LES included in this? As far as what I know, the quality of LES simulations depends on the underlying sub-grid scale models.
• L66-L67: Instead of "The main technique that reduces", it can be written as "One of the techniques that reduces.." since LES also reduces computational requirements.
• L126: Perhaps "for scalability" is better than "to scale".
• Figure 1 b, c, e - Legends and axis lables are too small to be readable.
• Since scalability appears to be one of the main features of the software, perhaps a plot showing it would be nice.

The paper is well-written and addresses requirements of the JOSS checklist quite well. I will report on my second set of comments focusing on the documentation and functionality part of the checklist.

[delcmo]

@Phoenixfire1081 thanks for providing comments. Below are my answers to your comments:

• L62: High-fidelity simulation - why is LES included in this? As far as what I know, the quality of LES simulations depends on the underlying sub-grid scale models.
  LES techniques are commonly used to generate high-fidelity simulations of flow in complex geometries. In the context of MHD flows this is still valid under the assumption that the LES methods is adapted to account for the MHD effects.

• L66-L67: Instead of "The main technique that reduces", it can be written as "One of the techniques that reduces.." since LES also reduces computational requirements.
  Thanks for the suggestion. We updated the text.

• L126: Perhaps "for scalability" is better than "to scale".
  Thanks for the suggestion. We updated the text.

• Figure 1 b, c, e - Legends and axis lables are too small to be readable.
  I will re-generate the plots and get back to you. Thanks.

• Since scalability appears to be one of the main features of the software, perhaps a plot showing it would be nice.
  Plot (c) in Figure 1 shows strong scaling of the VERTEX-CFD solver on Frontier.

The updated version can be found in here: https://github.com/ORNL/VERTEX-CFD/actions/runs/18129149340

Thanks,

Marco

[delcmo]

@Phoenixfire1081 I updated the figure and the latest version of the manuscript can be found here: https://github.com/ORNL/VERTEX-CFD/actions/runs/18138001609

Marco

[Phoenixfire1081]

Thanks for your efforts. I'll get back to you with my second set of comments soon.

[Phoenixfire1081]

Here are my second set of comments:

• The installation instructions are not very clear especially the Trilinos part - I wasn't even table to install the Trilinos-CPU. I tried installing the software on a Ubuntu 24.04 machine with NVIDIA RTX 6000 Ada Generation. It would be helpful to know what machines/OS the authors have tested the software on so that common issues during installation can be mentioned in the documentation itself. Due to this reason, I've been unable to test the functionality of the software. After the authors report the OS they've tested, I'll use a virtual machine with the same OS and test the functionality of the software.
• While the included example seems fine, additional examples in the documentation are needed so that the user has an easier time understanding the software.
• Did I miss it, or is there a documentation of the API?

[delcmo]

@Phoenixfire1081 thanks for the feedback.

• I will add some additional content to the installation instruction section.
• I will add one or two additional examples.
• What do you mean by API? An API to generate the XML input file?