Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

How to profile/tune hypre BoomerAMG? #1235

Open
jczhang07 opened this issue Feb 12, 2025 · 4 comments
Open

How to profile/tune hypre BoomerAMG? #1235

jczhang07 opened this issue Feb 12, 2025 · 4 comments

Comments

@jczhang07
Copy link

Hello,
In an openfoam application, I want to use BoomerAMG (via petsc, on CPU for the time being). How do I know the best parameters to apply? Is there any documentation on that?

Thanks!

-Junchao

@victorapm
Copy link
Contributor

Hi Junchao,

The optimal parameters for BoomerAMG depend heavily on the specific problem setup, including the PDE, mesh type, and mesh size. We have some general guidelines available here, which may serve as a useful starting point.

However, fine-tuning requires deeper knowledge of the problem at hand. If you can provide more details about your setup, such as the type of equations you're solving, the mesh characteristics, and any solver settings you're using in PETSc, we can help suggest parameter choices that may improve performance.

@victorapm
Copy link
Contributor

For fine-grained performance profiling, you may use dedicated tools such as Caliper (--with-caliper), HPCToolkit, TAU, Intel VTune, NVIDIA Systems (for CUDA runs on NVIDIA GPUs) or ROCm Systems (for HIP runs on AMD GPUs) depending on your specific needs and platform.

@victorapm
Copy link
Contributor

@jczhang07 is this resolved? Thank you

@jczhang07
Copy link
Author

No. I don't have a good understanding of the problem (equations and meshes) on hand. Once I know more, I will give update. Thanks for the patience.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants