Skip to content

[MRG] DOC More details about parallelism (joblib, openMP, MKL...) #15116

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
Merged
ogrisel merged 13 commits into from
Oct 15, 2019
Merged

[MRG] DOC More details about parallelism (joblib, openMP, MKL...) #15116

ogrisel merged 13 commits into from
Oct 15, 2019
+171 −66

Conversation

@NicolasHug
Copy link
Member

@NicolasHug NicolasHug commented Oct 1, 2019

Follow up on #14265

ping @ogrisel @jeremiedbb please make sure I didn't write any nonsense :)

albertcthomas
albertcthomas reviewed Oct 2, 2019
View reviewed changes
Copy link
Contributor

@albertcthomas albertcthomas left a comment

Choose a reason for hiding this comment

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

I cannot really assess the correctness but I find it very clear. I also like the small example with the 4 * 4 threads in total. Thanks!

rth
rth reviewed Oct 2, 2019
View reviewed changes
Copy link
Member

@rth rth left a comment

Choose a reason for hiding this comment

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

Thanks @NicolasHug ! A few comments

@NicolasHug NicolasHug changed the title [MRG] DOC for controlling number of OpenMP threads [MRG] DOC More details about parallelism (joblib, openMP, MKL...) Oct 2, 2019
@NicolasHug
Copy link
Member Author

NicolasHug commented Oct 2, 2019

I made a big update.

Tried to trim down the n_jobs glossary description and added more details in the UG.

That's pretty much all I know about this... Hope that's useful, feedback welcome

albertcthomas
albertcthomas reviewed Oct 3, 2019
View reviewed changes
jeremiedbb
jeremiedbb reviewed Oct 3, 2019
View reviewed changes
Copy link
Member

@jeremiedbb jeremiedbb left a comment

Choose a reason for hiding this comment

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

Look good overall. I find it very clear, thanks !

Numpy recently added the possibility to link against BLIS (another multi-threaded BLAS implementation). For completeness you could add BLIS everywhere you mention MKL and OpenBLAS :)
It also has a env var: BLIS_NUM_THREADS.

@NicolasHug
Copy link
Member Author

NicolasHug commented Oct 3, 2019

Thanks a lot for the comments. @jeremiedbb joblib doesn't protect from oversubscription from BLIS for now, right?

@jeremiedbb
Copy link
Member

jeremiedbb commented Oct 3, 2019 •
edited
Loading

joblib doesn't protect from oversubscription from BLIS for now, right?

Actually it does because BLIS looks for OMP_NUM_THREADS as well.

(That might change in the future:
Note: We highly discourage use of the OMP_NUM_THREADS environment variable and may remove support for it in the future. If you wish to set parallelism globally via environment variables, please use BLIS_NUM_THREADS.)

jeremiedbb
jeremiedbb reviewed Oct 3, 2019
View reviewed changes
jeremiedbb
jeremiedbb approved these changes Oct 3, 2019
View reviewed changes
Copy link
Member

@jeremiedbb jeremiedbb left a comment

Choose a reason for hiding this comment

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

LGTM !

tomMoral
tomMoral reviewed Oct 3, 2019
View reviewed changes
Copy link
Contributor

@tomMoral tomMoral left a comment

Choose a reason for hiding this comment

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

This is very clear and synthetic! Thanks @NicolasHug for doing this.
2/3 comments on some specific points.

tomMoral
tomMoral reviewed Oct 3, 2019
View reviewed changes
jeremiedbb
jeremiedbb reviewed Oct 5, 2019
View reviewed changes
@NicolasHug NicolasHug mentioned this pull request Oct 7, 2019
Merged
ogrisel
ogrisel approved these changes Oct 7, 2019
View reviewed changes
Copy link
Member

@ogrisel ogrisel left a comment

Choose a reason for hiding this comment

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

Overall, LGTM. Just a few improvement suggestions:

@ogrisel ogrisel merged commit 60ce68c into scikit-learn:master Oct 15, 2019
@ogrisel
Copy link
Member

ogrisel commented Oct 15, 2019

Merged! Thanks @NicolasHug.

@NicolasHug
Copy link
Member Author

NicolasHug commented Oct 15, 2019

This might need an update depending on the outcome of #14196

@NicolasHug NicolasHug mentioned this pull request Jan 9, 2020
Closed
@NicolasHug NicolasHug mentioned this pull request Jan 24, 2020
Merged
6 tasks
@cmarmo cmarmo mentioned this pull request Mar 31, 2021
Closed
8 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@ogrisel ogrisel ogrisel approved these changes

@thomasjpfan thomasjpfan thomasjpfan left review comments

@jeremiedbb jeremiedbb jeremiedbb approved these changes

+3 more reviewers

@rth rth rth left review comments

@tomMoral tomMoral tomMoral left review comments

@albertcthomas albertcthomas albertcthomas left review comments

Reviewers whose approvals may not affect merge requirements

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

7 participants

@NicolasHug @jeremiedbb @ogrisel @rth @tomMoral @thomasjpfan @albertcthomas