2.4.x kineto backport #372
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
Explicitly pass `-Wl,-z,noexecstack` to the linker, to ensure that `libpytorch_cpu.so` is compiled without an executable stack. This is necessary because the raw assembly in oneDNN triggers: ``` $BUILD_PREFIX/bin/../lib/gcc/x86_64-conda-linux-gnu/13.3.0/../../../../x86_64-conda-linux-gnu/bin/ld: warning: ittptmark64.S.o: missing .note.GNU-stack section implies executable stack ``` ...and glibc 2.41 no longer permits loading libraries with executable stack. Fixes conda-forge#350
When building with `mkl`, upstream build system uses the same OpenMP implementation as `mkl`, which is `llvm-openmp`. In generic variant, the upstream can use either `llvm-openmp` or `libgomp`, but `llvm-openmp` takes precedence if it is installed. In CUDA-enabled builds, `llvm-openmp` was installed as an indirect dependency, and therefore the package used it rather than `libgomp`. For consistency and to avoid need to patch the build system, just depend on `llvm-openmp` for all Linux targets.
…nda-forge-pinning 2025.03.12.11.34.20
Contributor
|
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 ( I do have some suggestions for making it better though... For recipe/meta.yaml:
This message was generated by GitHub Actions workflow run https://github.com/conda-forge/conda-forge-webservices/actions/runs/13941794401. Examine the logs at this URL for more detail. |
Contributor
Author
|
Confirmed that it's looking for the library via |
mgorny
requested review from
Tobias-Fischer,
beckermr,
benjaminrwilson,
hmaarrfk and
sodre
as code owners
Contributor
Author
|
I think this should be good enough. To avoid further changes to the build system, I've went ahead and removed |
mgorny
changed the title
isuruf
reviewed
Mar 17, 2025
recipe/patches/0008-Fix-CUPTI-lookup-to-include-target-directory.patch
Outdated
Show resolved
Hide resolved
Contributor
This is effectively a waste of our time. Other people's tech debt should not be your burden @mgorny Feel free to just "stop" on this front. |
Contributor
Author
|
Not a big deal — it's just a matter of copying the build |
h-vetinari
reviewed
Mar 18, 2025
h-vetinari
reviewed
Mar 19, 2025
Contributor
|
To expedite the process, I am locally building 'linux_aarch64_c_compiler_version13' and will share the build log once it's complete. |
Member
That's kind of you, but IMO not necessary. There's no big congestion on the server at the moment, so we can just let this build normally. |
Agreed. I was concerned that the build was being pushed back due to other PR. Uploading the log, just in case: |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
7 participants
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.
Checklist
0(if the version changed)conda-smithy(Use the phrase@conda-forge-admin, please rerenderin a comment in this PR for automated rerendering)This is a fork of #341, on top of which I attempted enabling Kineto with CUPTI. Unfortunately, I couldn't get it to work — possibly it needs more CUDA lookup fixes from 2.5.x. But that's too much for my feverish brain right now. Still, opening a draft in case anyone wanted to try some more.