jaxlib v0.7.1

[regro-cf-autotick-bot]

It is very likely that the current package version for this feedstock is out of date.

Checklist before merging this PR:

• Dependencies have been updated if changed: see upstream
• Tests have passed
• Updated license if changed and license_file is packaged

Information about this PR:

1. Feel free to push to the bot's branch to update this PR if needed.
2. The bot will almost always only open one PR per version.
3. The bot will stop issuing PRs if more than 3 version bump PRs generated by the bot are open. If you don't want to package a particular version please close the PR.
4. If you want these PRs to be merged automatically, make an issue with @conda-forge-admin,please add bot automerge in the title and merge the resulting PR. This command will add our bot automerge feature to your feedstock.
5. If this PR was opened in error or needs to be updated please add the bot-rerun label to this PR. The bot will close this PR and schedule another one. If you do not have permissions to add this label, you can use the phrase @conda-forge-admin, please rerun bot in a PR comment to have the conda-forge-admin add it for you.

Pending Dependency Version Updates

Here is a list of all the pending dependency version updates for this repo. Please double check all dependencies before merging.

Name	Upstream Version	Current Version
cudnn	9.12.0.46

_{This PR was created by the regro-cf-autotick-bot. The regro-cf-autotick-bot is a service to automatically track the dependency graph, migrate packages, and propose package version updates for conda-forge. Feel free to drop us a line if there are any issues! This PR was generated by https://github.com/regro/cf-scripts/actions/runs/17107413748 - please use this URL for debugging.}

[conda-forge-admin]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe/meta.yaml) and found it was in an excellent condition.

I do have some suggestions for making it better though...

For recipe/meta.yaml:

• ℹ️ The recipe is not parsable by parser conda-souschef (grayskull). This parser is not currently used by conda-forge, but may be in the future. We are collecting information to see which recipes are compatible with grayskull.

_{This message was generated by GitHub Actions workflow run https://github.com/conda-forge/conda-forge-webservices/actions/runs/17881474427. Examine the logs at this URL for more detail.}

[lucascolley]

Hunk #1 FAILED at 22.

sounds like patches need to be regenerated?

[xhochy]

@h-vetinari May I ask you for a favour here? I fail at changing all compiler pins to clang 19. Could you have a look and help me?

[h-vetinari]

The switch to clang got lost somewhere in the bowels of bazel, which seems to be looking for GCC still...

[traversaro]

The switch to clang got lost somewhere in the bowels of bazel, which seems to be looking for GCC still...

I am not sure if it is related (I would need to check more in details) but while I was debugging another issue I remember seeing jax-ml/jax#29672 . I wanted to warn about this here, but then I forgot.

[xhochy]

The switch to clang got lost somewhere in the bowels of bazel, which seems to be looking for GCC still...

This depends on some changes in bazel-toolchain. I have them locally and will push them in the next days once everything runs smoothly.

[xhochy]

The switch to clang got lost somewhere in the bowels of bazel, which seems to be looking for GCC still...

I am not sure if it is related (I would need to check more in details) but while I was debugging another issue I remember seeing jax-ml/jax#29672 . I wanted to warn about this here, but then I forgot.

Good catch! The build is using the local toolchain for me but downloads a massive amount of LLVM binaries. So this is probably related.

[xhochy]

I have a local fix for osx-arm64 (trivial one). I don't understand the osx-64 error. Would we be OK with skipping it for now? Then I would push my fix and build the full matrix.

[h-vetinari]

I don't understand the osx-64 error.

It looks like a trivially missing header from onednn

In file included from external/xla/xla/service/cpu/runtime_matmul_c128.cc:22:
In file included from external/xla/xla/service/cpu/runtime_matmul_common.h:29:
external/xla/xla/tsl/framework/contraction/eigen_contraction_kernel.h:40:10: fatal error: 'dnnl.h' file not found
   40 | #include "dnnl.h"
      |          ^~~~~~~~
1 error generated.

[xhochy]

The error goes a bit deeper. It seems that the onednn archive is not unpacked at all. As simple as the error looks, the solution is not obvious to me.

[xhochy]

I have a script to bisect that in the XLA repository. This will take a while, but only CPU, not human, time.

[xhochy]

fe13dfb8ed08f5152a9e65f03bfe2f398e1d2927 is the first bad commit
commit fe13dfb8ed08f5152a9e65f03bfe2f398e1d2927
Author: Penporn Koanantakool <penporn@google.com>
Date:   Fri Aug 8 07:32:29 2025 -0700

    [xla:cpu] Add a unified `cc_library` target for oneDNN.

    The original `mkl_deps()` macro can't be used in a select statement because it also has a select statement inside. (Bazel doesn't support nested selects.)

    Users should depend on this new `cc_library` target instead.

    PiperOrigin-RevId: 792617157

 xla/backends/cpu/BUILD                | 15 ++-------------
 xla/backends/cpu/runtime/onednn/BUILD | 16 ++++------------
 xla/tsl/framework/contraction/BUILD   |  3 +--
 xla/tsl/mkl/BUILD.bazel               | 12 ++++++++----
 4 files changed, 15 insertions(+), 31 deletions(-)

[h-vetinari]

Since the linux builds passed previously and e9b7b9f is osx-specific, I'm cancelling the linux builds, due to heavy congestion on the opengpu server.

[h-vetinari]

Thanks so much for the huge effort here @xhochy! ❤️

[h-vetinari]

Since the OSX failures were just a question of timeouts where we'll have to play with restarts a bit (and we have some server capacity right now), I'm going to merge this as-is.