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

Replace if-Conditions by Ternary Operators in Nabla #1798

Merged
merged 4 commits into from
Aug 12, 2024

Conversation

fthaler
Copy link
Contributor

@fthaler fthaler commented Jul 24, 2024

The unconditional shifts (unconditional loads from neighbor tables) seem to help NVCC’s load hoisting and loop-invariant code motion in the case of k-blocking. Increases performance by up to 8% on double precision and up to 2.5% on single precision fused nabla implementations.

@gridtoolsjenkins
Copy link
Collaborator

Hi there, this is jenkins continuous integration...
Do you want me to verify this patch?

@fthaler
Copy link
Contributor Author

fthaler commented Jul 24, 2024

launch perftest

@fthaler
Copy link
Contributor Author

fthaler commented Jul 24, 2024

launch jenkins

Copy link
Contributor

@havogt havogt left a comment

Choose a reason for hiding this comment

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

we should look into gt4py codegen if these changes are relevant there

@fthaler
Copy link
Contributor Author

fthaler commented Aug 5, 2024

launch perftest

@fthaler
Copy link
Contributor Author

fthaler commented Aug 6, 2024

launch jenkins

@fthaler
Copy link
Contributor Author

fthaler commented Aug 6, 2024

launch perftest

@fthaler fthaler merged commit daf2892 into GridTools:master Aug 12, 2024
68 checks passed
@fthaler fthaler deleted the nabla-no-conditionals branch August 12, 2024 06:48
havogt pushed a commit that referenced this pull request Sep 30, 2024
The unconditional shifts (unconditional loads from neighbor tables) seem
to help NVCC’s load hoisting and loop-invariant code motion in the case
of k-blocking. Increases performance by up to 8% on double precision and
up to 2.5% on single precision fused nabla implementations.
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.

3 participants