[Dynamo] Introduce hook receiving set of traced files #153622
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
[ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/153622
Note: Links to docs will display an error until the docs builds have been completed. ❌ 2 New Failures, 3 Unrelated FailuresAs of commit dc5a663 with merge base 9c11493 ( NEW FAILURES - The following jobs have failed:
BROKEN TRUNK - The following jobs failed but were present on the merge base:👉 Rebase onto the `viable/strict` branch to avoid these failures
UNSTABLE - The following job is marked as unstable, possibly due to flakiness on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This PR:
* Expands `Hooks` with a new, optional `frame_traced_fn` field. It should be a callable receiving the set of traced filenames
* Maintains a set of `traced_files` in the `TracingContext` of an `OutputGraph`
* Whenever an `inline_call()` is encountered, the corresponding filename is added to this set
* `OutputGraph`'s `co_fields["co_filename"]` is added to the set before the hook is called
I believe use of this hook should enable the source code hashing that vLLM does in a better way than monkey-patching `inline_call()`.
cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames
[ghstack-poisoned]
jansel
requested changes
May 15, 2025
| @@ -865,6 +865,8 @@ def __init__(self, fake_mode): | |||
| # see note: [Returning Fake Tensors on First AOT Autograd Call] | |||
| self.fakify_first_call = False | |||
| self.hop_dispatch_set_cache = HopDispatchSetCache() | |||
| # set of filenames for inlined functions | |||
| self.traced_files: set[str] = set() | |||
| # they are benign and do not generate any new graphs. | ||
| hooks.guard_export_fn(output.guards) | ||
| if hooks.frame_traced_fn is not None: | ||
| output.tracing_context.traced_files.add(output.co_fields["co_filename"]) |
There was a problem hiding this comment.
Is the filename enough? One file can have many things in it. Maybe a function or code object would enable more use cases.
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.
Stack from ghstack (oldest at bottom):
This PR:
Hookswith a new, optionalframe_traced_fnfield. It should be a callable receiving the set of traced filenamestraced_filesin theTracingContextof anOutputGraphinline_call()is encountered, the corresponding filename is added to this setOutputGraph'sco_fields["co_filename"]is added to the set before the hook is calledI believe use of this hook should enable the source code hashing that vLLM does in a better way than monkey-patching
inline_call().cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @chenyang78 @kadeng @chauhang @amjames