Skip to content

Parallel training via do concurrent - #202

Merged
rouson merged 16 commits into
mainfrom
parallel
Sep 14, 2024
Merged

Parallel training via do concurrent#202
rouson merged 16 commits into
mainfrom
parallel

Conversation

@rouson

Copy link
Copy Markdown
Contributor

No description provided.

rousonand others added 9 commits August 1, 2024 17:52
The program cloud-microphysics/app/train-cloud-microphysics.f90
has makes three subroutine calls, two of which have now been
tested in parallel:
1. get_command_line_arguments: multi-image execution works
2. create_or_append_plot_file: now happens only on image 1
3. read_train_write: (to do)
This commit fixes the `do concurrent` blocks in the training
subroutine so that the iterations can execute concurrently:
1. Add locality specifiers for compilers that support them.
2. Otherwise, define blocks with local variables/allocations
for each loop iteration and perform necessary explicit sums.
Co-authored-by: Dan Bonachea <dobonachea@lbl.gov>
@rouson
rouson marked this pull request as ready for review September 13, 2024 17:10
Comment threadsrc/inference_engine/trainable_engine_s.F90

@bonacheabonachea left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

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

Probably also need to update the status of do concurrent support and usage in the top-level README

@rouson

Copy link
Copy Markdown
ContributorAuthor

@bonachea I think all the changes we've discussed have been made so I'll merge this.

@rouson
rouson merged commit 752d7a9 into mainSep 14, 2024
Sign up for freeto 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

@rouson@bonachea