Add triton dependency, readd cudss and cusparselt, mention dev speedup tricks in the README #309
Conversation
This file contains 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
mgorny
requested review from
Tobias-Fischer,
beckermr,
benjaminrwilson,
hmaarrfk,
jeongseok-meta and
sodre
as code owners
|
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/12601832807. Examine the logs at this URL for more detail. |
5 tasks
hmaarrfk
reviewed
Dec 27, 2024
|
Started only a single job for now (x64+CUDA+generic). If this one passes we can run the rest (modulo other changes from review). |
1 task
|
So the CI here for x64+CUDA+generic passed. I think you could push whatever new commits you have @mgorny and we can do a full CI run. |
|
@h-vetinari, I think the latest build of libabseil is now broken on OSX. See e.g. https://dev.azure.com/conda-forge/feedstock-builds/_build/results?buildId=1127302&view=logs&j=1dfee543-4d72-5193-7012-7358e3fc4acd&t=344bb75a-1253-525e-defa-7c6fedd7579f&l=2594 |
|
Hmm, or maybe protobuf needs to be rebuilt. Though it's still weird that ABI changed on rebuild. |
It's something to do with clang 17 to clang 18. I've looked at the breaking changes at the time, and nothing stood out in particular. I guess abseil does unusual things, relatively speaking... In any case, I'm hoping that once we've rebuilt the core protobuf stack (conda-forge/libprotobuf-feedstock#244 + maybe one for grpc), the issue will resolve itself... (edit: or maybe not: arrow also seems to fail) |
|
Could you trigger the Linux CI builds, please? |
h-vetinari
force-pushed
the
triton-dep
branch
from
ccbaebb to
4de17af
Compare
Was waiting for the |
…onda-forge-pinning 2024.12.30.17.35.18
|
I'm leaving the linux CI running for now; to fix OSX, we should pick up the changes from #310 here. |
|
Hmm, looks like two jobs failed with "This request was automatically failed because there were no enabled runners online to process". The other four passed. |
|
i clicked rerun failed jobs. that is my contribution ^_^ |
|
Oh, right, we don't build Triton for aarch64 (yet). Let's see if that is possible… |
|
Ok, so the current status is:
|
|
Okay, I'm going to wait for new triton builds to propagate, then merge main and #310 here to trigger new CI. |
|
@h-vetinari, I think we can limit CI to the aarch64 + CUDA job. There were no other changes, and everything else passed the previous time. |
|
Sigh, I hope that the failing build was a runner GPU problem and not something with the package. |
|
It failed at ~55% of the test through, with: detailed logs follow: |
|
Could you rerun failing builds, please? I think it was a fluke — there's almost no changes (I've only merged |
|
There's a limitation in the UI that I can only restart all failed jobs (includes cancelled jobs), or one. So I prioritised the aarch GPU build first, but now restarted the previous failure too. |
|
Thanks. So all passed, except for the 3 cancelled jobs. Not sure if you want to rerun them, or just assume they didn't change since the last run. |
h-vetinari
approved these changes
Jan 5, 2025
|
Thank you! |
|
Failure of Perhaps also due to resource issues? |
|
Hrm, no clue. Could it have OOM-ed perhaps? |
|
Do you want me to try addressing it separately, or wait and see if we can get kineto passing first, and perhaps the same solution will work for both cases? |
|
You can keep working on it of course. In the meantime I'll see if I can nurse the CI on main to success by restarting (but the queue seems long at the moment) |
5 tasks
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.
Checklist
0(if the version changed)conda-smithy(Use the phrase@conda-forge-admin, please rerenderin a comment in this PR for automated rerendering)