[MRG+2] use manylinux dev wheels for numpy / scipy #8536
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Use daily manylinux wheels for numpy and scipy, instead of soon-to-be-discontinued per-commit Precise wheels.
|
LGTM, thanks |
|
Why did you remove the atlas libraries? It seems to me that an optimized blas would be useful. Do we have others on the system? |
|
I removed the blas libraries thinking that you didn't need them anymore, now you are not using the numpy wheels that depend on them. Is that not correct? |
|
I am pretty sure that we use a blas when it's available, for instance in sklearn/utils/weight_vector.pyx. We carry around our own blas headers (in sklearn/src/cblas) but I would rather rely on a proper blas implementation for building wheels. What do you think? It seems to me that it would be better to have a good blas on the build system. Ideally openblas, though. |
|
I may have misunderstood, but as far as I know the wheel-building repo is https://github.com/MacPython/scikit-learn-wheels - and you are not building wheels for redistribution here. If you sometimes build with an external blas, and sometimes without, I guess the best option would be to have a build matrix entry for each option? |
|
but as far as I know the wheel-building repo is https://github.com/MacPython/scikit-learn-wheels - and you are not building wheels for redistribution here.
Good point. I got confused.
If you sometimes build with an external blas, and sometimes without, I guess the best option would be to have a build matrix entry for each option?
I'd rather avoid added one option to our build matrix, it's already big
enough, and consumming a lot of resources. The question is whether the
default should be with atlas or without.
If I compare the time of travis on your PR:
https://travis-ci.org/scikit-learn/scikit-learn/builds/207824847
to another recent build:
https://travis-ci.org/scikit-learn/scikit-learn
It seems to me that when there is atlas installed, the tests are
significantly faster (it could be something else, though). Hence, I am
more inclined to go for the atlas solution.
|
|
Hmm - looking at |
0aab63a to
ecdafc9
Compare
scikit-learn can link against BLAS libraries still at the same location as they were duing numpy build.
ecdafc9 to
1468380
Compare
|
I reverted the change - so ATLAS is always installed now, as it was before this PR. |
|
LGTM |
|
Thanks @matthew-brett 🎉 |
Closed
herilalaina
pushed a commit
to herilalaina/scikit-learn
that referenced
this pull request
Mar 26, 2017
* MAINT: use manylinux dev wheels for numpy / scipy Use daily manylinux wheels for numpy and scipy, instead of soon-to-be-discontinued per-commit Precise wheels. * BF: add back ATLAS install for ubuntu build entry scikit-learn can link against BLAS libraries still at the same location as they were duing numpy build.
massich
pushed a commit
to massich/scikit-learn
that referenced
this pull request
Apr 26, 2017
* MAINT: use manylinux dev wheels for numpy / scipy Use daily manylinux wheels for numpy and scipy, instead of soon-to-be-discontinued per-commit Precise wheels. * BF: add back ATLAS install for ubuntu build entry scikit-learn can link against BLAS libraries still at the same location as they were duing numpy build.
Sundrique
pushed a commit
to Sundrique/scikit-learn
that referenced
this pull request
Jun 14, 2017
* MAINT: use manylinux dev wheels for numpy / scipy Use daily manylinux wheels for numpy and scipy, instead of soon-to-be-discontinued per-commit Precise wheels. * BF: add back ATLAS install for ubuntu build entry scikit-learn can link against BLAS libraries still at the same location as they were duing numpy build.
NelleV
pushed a commit
to NelleV/scikit-learn
that referenced
this pull request
Aug 11, 2017
* MAINT: use manylinux dev wheels for numpy / scipy Use daily manylinux wheels for numpy and scipy, instead of soon-to-be-discontinued per-commit Precise wheels. * BF: add back ATLAS install for ubuntu build entry scikit-learn can link against BLAS libraries still at the same location as they were duing numpy build.
paulha
pushed a commit
to paulha/scikit-learn
that referenced
this pull request
Aug 19, 2017
* MAINT: use manylinux dev wheels for numpy / scipy Use daily manylinux wheels for numpy and scipy, instead of soon-to-be-discontinued per- 9338 commit Precise wheels. * BF: add back ATLAS install for ubuntu build entry scikit-learn can link against BLAS libraries still at the same location as they were duing numpy build.
maskani-moh
pushed a commit
to maskani-moh/scikit-learn
that referenced
this pull request
Nov 15, 2017
* MAINT: use manylinux dev wheels for numpy / scipy Use daily manylinux wheels for numpy and scipy, instead of soon-to-be-discontinued per-commit Precise wheels. * BF: add back ATLAS install for ubuntu build entry scikit-learn can link against BLAS libraries still at the same location as they were duing numpy build.
jwjohnson314
pushed a commit
to jwjohnson314/scikit-learn
that referenced
this pull request
Dec 18, 2017
* MAINT: use manylinux dev wheels for numpy / scipy Use daily manylinux wheels for numpy and scipy, instead of soon-to-be-discontinued per-commit Precise wheels. * BF: add back ATLAS install for ubuntu build entry scikit-learn can link against BLAS libraries still at the same location as they were duing numpy build.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Use daily manylinux wheels for numpy and scipy, instead of
soon-to-be-discontinued per-commit Precise wheels.