This repository has been archived by the owner on Mar 21, 2024. It is now read-only.
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Outdated suggestions cannot be applied.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.
I was trying to get cuDF to use per-thread default stream. It mostly worked, but I was still seeing some stuff happening on the legacy default stream. Finally traced to the
default_stream()
function here. Making the following change got rid of the default stream activities.This seems like the right thing to do. If someone is trying to use per-thread default stream through compiler flags, it seems wrong for thrust to use the legacy default stream behind the user's back. There might be additional race conditions the user needs to take care of, but at least the thrust code would be doing what the user tells it to do.
See rapidsai/rmm#354 and rapidsai/cudf#4995.