Updated pytorch and disabled sparse tests #168
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
Contributor
|
Please add more information to the PR description explaining why we are doing this. |
|
The change and PR description LGTM. Although, I'm not sure why the tests still showing pytorch2.5.1 even though we already changed the workflow. |
dakinggg
reviewed
Mar 11, 2025
dakinggg
approved these changes
Mar 12, 2025
rithwik-db
force-pushed
the
sparse-disable
branch
from
5388045 to
129335b
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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Outdated suggestions cannot be applied.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.
What does this PR do?
Updated PyTorch like in #167, but also disabling
sparsesupport (only supportinggroupedsupport) until we fix the issue withtriton. Also updatedpyclnversion that's breaking pre-commit checks.For full context:
triton 3.2.0introduced some change to how it handles dtype promotion when two binary operands have different dtypes, and as a result we're encountering an int16 overflow in thestkdependency of megablocks which results in an illegal memory access (IMA). A fix like https://github.com/stanford-futuredata/stk/pull/17/files needs to be applied to to stk instead of megablocks or triton to resolve this initial issue. Even if we don't hit an IMA, not having this casting will lead to different outputs intriton 3.1andtriton 3.2given the same inputs.