fix: Destory cuda graphs before setting weight streaming #3461
+38
−14
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
github-actions
bot
added
component: core
narendasan
reviewed
Apr 2, 2025
py/torch_tensorrt/dynamo/runtime/_CudaGraphsTorchTensorRTModule.py
Outdated
Show resolved
Hide resolved
narendasan
reviewed
Apr 4, 2025
| @@ -103,9 +103,13 @@ def validate_input_shapes(self, inputs: Sequence[torch.Tensor]) -> bool: | |||
|
|
|||
| return False | |||
|
|
|||
| def __del__(self) -> None: | |||
| def reset_captured_graph(self) -> None: | |||
|
@zewenli98 please cherry pick this to 2.7 |
narendasan
added
needs-release-cherrypick
and removed
cherry-pick
7 tasks
zewenli98
added a commit
that referenced
this pull request
Apr 5, 2025
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Outdated suggestions cannot be applied.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.
Description
When cuda graphs and weigh streaming are used together, cuda graphs is destroyed after setting the weight streaming.
Weight streaming recreates the context and load new module. Destroying cudagraphs with old reference caused application crash. Fix is to move cuda graphs reset before the weight streaming setting.
The timing of del is not entirely predictable in python. Moved cudagraph reset logic from del to dedicated reset_cudagraph method and it's called when exiting from CudaGraphsTorchTensorRTModule context block
Fixes #3460
Type of change
Please delete options that are not relevant and/or add your own.
Checklist: