-
Notifications
You must be signed in to change notification settings - Fork 41
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Move trace_link.py from facebookresearch/param to mlcommons/chakra #36
Closed
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
The `handle_kineto_segmentation` function is intended to support kineto traces cross multiple iterations by splitting a trace into several segments according to the provided annotations. Unfortunately, this function is not operating as expected, leading to errors. It is advisable to remove it.
The multi-iteration support feature for PyTorch execution traces is designed to facilitate the handling of traces over multiple iterations. Unfortunately, this feature is not functioning as expected and is leading to errors. It is advisable to remove it.
This commit introduces support for inter-thread dependencies within the Chakra framework. By examining Kineto traces via chrome://tracing, one can observe multiple CPU threads and their implicit dependencies. This update explicitly encodes these dependencies in the output trace, enabling accurate handling by subsequent tools.
This commit adds stream ID encoding to GPU operators. This ensures that all operators within the same stream are executed in the correct order, supporting intra-stream dependencies.
Introduced exclusive duration calculation for Kineto operators in the TraceLinker class. This update differentiates between inclusive and exclusive durations, providing a clearer distinction in the profiling data. Exclusive durations are now calculated to identify the actual time spent in individual operations, excluding overlaps with child operators.
MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅ |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
Summary
Move trace_link.py from facebookresearch/param to mlcommons/chakra based on the discussion between Meta and NVIDIA.
Test Plan