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

add longrun gpu pipeline #2441

Merged
merged 1 commit into from
Dec 18, 2023
Merged

add longrun gpu pipeline #2441

merged 1 commit into from
Dec 18, 2023

Conversation

szy21
Copy link
Member

@szy21 szy21 commented Dec 15, 2023

Purpose

Adds a new pipeline for GPU longruns. For now I only moved dycore simulations (dry and moist baroclinic wave, dry and moist held suarez). I'll move the ones with radiation when the performance issue is solved.

The build is here: https://buildkite.com/clima/climaatmos-gpulongruns/builds/7

To-do

Content


  • I have read and checked the items on the review checklist.

@szy21 szy21 marked this pull request as ready for review December 16, 2023 22:35
@szy21 szy21 force-pushed the zs/ci_longrun_gpu branch from b98f4c9 to 844ebd8 Compare December 17, 2023 19:33
@szy21 szy21 requested a review from simonbyrne December 17, 2023 19:35
@@ -6,7 +6,7 @@ z_elem: 63
dz_bottom: 30.0
dz_top: 3000.0
z_max: 55000.0
kappa_4: 2.0e16
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This run is stable on CPU but unstable on GPU. I decreased the hyperdiffusivity and it became stable. I think we are probably just close to the unstable regime.

@szy21 szy21 force-pushed the zs/ci_longrun_gpu branch from 844ebd8 to 791f033 Compare December 17, 2023 19:38
@szy21 szy21 force-pushed the zs/ci_longrun_gpu branch from 791f033 to bf6c69a Compare December 17, 2023 19:38
@szy21 szy21 added this pull request to the merge queue Dec 18, 2023
@simonbyrne
Copy link
Member

@szy21 out of curiousity, what are the corresponding GPU times?

@szy21
Copy link
Member Author

szy21 commented Dec 18, 2023

@szy21 out of curiousity, what are the corresponding GPU times?

For the same moist held suarez run (h_elem=16, z_elem=63, t_end=300days), it takes 3.5 hours on 1 GPU and 14 hours on 64 CPUs. I think it's pretty good:)

@szy21
Copy link
Member Author

szy21 commented Dec 18, 2023

I'll probably extend t_end for the GPU pipeline later. Hope it will be stable!

@szy21 szy21 mentioned this pull request Dec 18, 2023
1 task
Merged via the queue into main with commit d034af8 Dec 18, 2023
10 checks passed
@szy21 szy21 deleted the zs/ci_longrun_gpu branch December 18, 2023 20:16
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

Successfully merging this pull request may close these issues.

2 participants