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.

Sign up for GitHub

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

Jump to bottom

Timing of new MPI communication (comm only send faces, not entire array) #96

Open
kaschau opened this issue Jan 21, 2022 · 0 comments
Open

Timing of new MPI communication (comm only send faces, not entire array) #96

kaschau opened this issue Jan 21, 2022 · 0 comments

Comments

@kaschau
Copy link
Owner

kaschau commented Jan 21, 2022

Using the threeDTaylorVortex example (results in seconds)

Nx Entire Array Just sends Speed Up
65 1170 1050 10%
50 554 550 0%
35 364 478 -20%

I think the function calling in python, mixed with the overhead of an individual copy call, is responsible for the speed up.

Not really sure what to do, but its clear for lots of small blocks, especially when they are shared by a process, the individual calls are a killer.
kaschau pushed a commit that referenced this issue Jan 21, 2022
Adding back the communicateWholeView routine as the Newark
e95c8b0 
method is not clearly better in all circumstances. Also added
a real gas eos example script. Fixes #92 and addresses #96
kaschau added a commit that referenced this issue Dec 16, 2023
@kaschau
Adding back the communicateWholeView routine as the Newark
16b39c2 
method is not clearly better in all circumstances. Also added
a real gas eos example script. Fixes #92 and addresses #96
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@kaschau