Pdlp fix batch cuda graph #68
Merged
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
Kh4ster
requested review from
aliceb-nv,
chris-maes,
hlinsen and
kaatish
and removed request for
hlinsen and
kaatish
Kh4ster
added
bug
aliceb-nv
approved these changes
Jun 3, 2025
Contributor
Author
|
/merge |
Collaborator
|
@Kh4ster Just wanted to confirm if this needs to go to 25.05 |
Contributor
Author
|
@rgsl888prabhu if we can have it in 25.05 yes that's better, that avoid having a hard to read / understand and impossible to fix error when running in batch mode |
Collaborator
|
I think there are few other bug fixes that needs to go in, may be we can add this as well. |
Contributor
I agree |
Contributor
Author
|
@rgsl888prabhu so we can merge this right? |
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Outdated suggestions cannot be applied.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 aims at fixing the invalid operation while there is a graph capture we sometimes see when using batch solve.
Solution is to use a regular instead of a non-blocking stream to make sure that if any operation (like a cudaFree from Thrust) is being launched on the default stream, it will wait for all other operations on other stream to finish first, preventing any cudaMalloc/Free while another stream might be doing a CUDA Graph capture.