Merge OpenAI Triton commit f436c9e
#3124
Merged
+606
−495
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
It was reported that triton compilation times have heavily increased lately. The cause of this is that we very often create the associated LL to check properties of a given Layout. We do this thousands of times, and this gets very expensive. In this PR, we implement a thread-safe cache for LinearLayouts. We clear this cache after we are done with the TTGIR -> LLVM conversion. In the future, we will make `DistributedEncoding` inherit from `LinearLayoutEncoding`, which will mean that `DistributedEncoding`s will always have access to their associated LinearLayout. Even in this scenario I still think that caching will be good, as there is no real 1-to-1 correspondence between `DistributedEncoding`s and `LinearLayout`s due to broadcasting, where we tile a layout along the tensor or we make it smaller. As such, I think this cache may be also useful in the future.
<!--- The core Triton is a small number of people, and we receive many PRs (thank you!). To help us review your code more quickly, **if you are a new contributor (less than 3 PRs merged) we ask that you complete the following tasks and include the filled-out checklist in your PR description.** Complete the following tasks before sending your PR, and replace `[ ]` with `[x]` to indicate you have done them. --> Currently, torch is required for importing triton and performing autotuning. This seems like a relatively heavy runtime dependency in the context of the cpu backend, as numpy can easily be used instead. Opening here as suggested in triton-lang/triton-cpu#205 to minimize future merge conflicts. Ideally there would be a test for this, but with the cpu backend out-of-tree this seems hard to test. See also triton-lang/triton-cpu#204, triton-lang/triton-cpu#205. # New contributor declaration - [x] I am not making a trivial change, such as fixing a typo in a comment. - [x] I have written a PR description following these [rules](https://cbea.ms/git-commit/#why-not-how). - [x] I have run `pre-commit run --from-ref origin/main --to-ref HEAD`. - Select one of the following. - [ ] I have added tests. - `/test` for `lit` tests - `/unittest` for C++ tests - `/python/test` for end-to-end tests - [x] This PR does not need a test because not (currently) easy to test and basic functionality should be covered by existing tests. - Select one of the following. - [x] I have not added any `lit` tests. - [ ] The `lit` tests I have added follow these [best practices](https://mlir.llvm.org/getting_started/TestingGuide/#filecheck-best-practices), including the "tests should be minimal" section. (Usually running Python code and using the instructions it generates is not minimal.)
This reverts commit c81ed27.
Unifying lowering path of `local_alloc` with `local_store` for the case shared mem layout has `leadingOffset`.
pbchekin
approved these changes
Jan 9, 2025
whitneywhtsang
force-pushed
the
whitneywhtsang/merge
branch
from
81721af to
6d600e2
Compare
whitneywhtsang
force-pushed
the
whitneywhtsang/merge
branch
from
6d600e2 to
a15b458
Compare
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.
This PR change the Triton base from 51dddd3 to f436c9e (Jan 8).
Pass rate: 99.86%
Please do not squash and merge this PR.