Ballista: Avoid sleeping between polling for tasks #698
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
Dandandan
changed the title
Dandandan
changed the title
H0TB0X420
pushed a commit
to H0TB0X420/datafusion
that referenced
this pull request
Oct 7, 2025
Closes apache#672 rustls Closes #682 syn Closes apache#653 parking_lot closes apache#648 object_store Closes apache#625 h2 Closes apache#623 tokio Closes apache#608 mio Closes apache#597 pyo3 Closes apache#642 pyo3-build-config Closes apache#627 prost Closes apache#634 prost-types Closes apache#637 async-trait
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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Outdated suggestions cannot be applied.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.
Which issue does this PR close?
Closes #697
Rationale for this change
Currently, when polling for work, the executor always waits 250 ms in between polling for tasks.
This means if the tasks themselves are small, it can be unproductive for a large percent of the time.
We can track whether we received tasks in the last iterations to avoid sleeping.
Also a slightly lower sleep time (e.g. 100ms) can help reduce the initial delay.
What changes are included in this PR?
Keep track of whether we received some tasks last time. If it's the case we don't sleep, otherwise we do.
Also reduce hard-coded 250 to 100ms to reduce initial delay a bit.
For a single executor on query 12 / SF=1, the time to execute goes from >5.5s to ~2.5s
Are there any user-facing changes?
FYI @andygrove