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

mom_step benchmark #47

Merged
merged 4 commits into from
Apr 17, 2023
Merged

mom_step benchmark #47

merged 4 commits into from
Apr 17, 2023

Conversation

b-fg
Copy link
Member

@b-fg b-fg commented Apr 16, 2023

Added mom_step.jl benchmark, also for the master branch. Remember to set threads to 1 when running on master for the baseline.

Added mom_step.jl benchmark, also for the master branch.
Remember to set threads to 1 when running on master for the baseline.
The mom_step.jl benchmark was to trivial (flow was mostly 0-filled arrays) so the compiler was probably doing some tricks to skip operations.
Adding the TGV initialisation provides higher speedups, as it is a non-trivial flow.

The new tgv.jl benchmark quantifies the full solver execution for the TGV test case during 0.1 CTU.
The speedups are aligned to what we observe when breaking down mom_step.
Added a the donut benchmark as a flow + body test case, and the master version for the baseline.
Plots are configured to produce the same figure as in the abstract (and commented out).
@b-fg
Copy link
Member Author

b-fg commented Apr 16, 2023

I added the donut benchmark and cleaned up a bit. Now we have benchmarks for TGV, TGV mom_step! break down, and donut for both CUDA and master branches.

@weymouth weymouth merged commit 4b8acfd into WaterLily-jl:CUDA Apr 17, 2023
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