[C++] Invoke storage allocation for CUDA Graph explicitly #3042
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
This PR adds a function that invokes the storage allocation function generated by CUDA Graph rewrite. With this function, we now manually trigger the storage allocation at initialization time. The reason we need this is because that the storage allocation may contain CUDA IPC memory alloc that has to run through a Disco session. So when a function that needs CUDA graph storage allocation runs first outside a Disco session, there might be error caused if we did not initialize the allocation in advance.
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Outdated suggestions cannot be applied.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 adds a function that invokes the storage allocation function generated by CUDA Graph rewrite. With this function, we now manually trigger the storage allocation at initialization time.
The reason we need this is because that the storage allocation may contain CUDA IPC memory alloc that has to run through a Disco session. So when a function that needs CUDA graph storage allocation runs first outside a Disco session, there might be error caused if we did not initialize the allocation in advance.