[10125] [encode path] Minor optimizations to arrow-flight#10137
Open
Rich-T-kid wants to merge 6 commits into
Open
[10125] [encode path] Minor optimizations to arrow-flight#10137Rich-T-kid wants to merge 6 commits into
Rich-T-kid wants to merge 6 commits into
Conversation
github-actions
Bot
added
arrow
Rich-T-kid
commented
Jun 12, 2026
| } | ||
|
|
||
| /// Place the `FlightData` in the queue to send | ||
| #[inline] |
Contributor
Author
There was a problem hiding this comment.
The compiler very likely could have inlined this, but I think its work adding this explicitly.
Contributor
|
run benchmarks flight |
|
🤖 Arrow criterion benchmark running (GKE) | trigger CPU Details (lscpu)Comparing rich-T-kid/minor-arrow-flight-opt (d02e297) to 826b808 (merge-base) diff File an issue against this benchmark runner |
|
🤖 Arrow criterion benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagebase (merge-base)
branch
File an issue against this benchmark runner |
Contributor
Author
|
seems like its mostly noise |
Contributor
Author
its interesting that this seems to always regress |
Rich-T-kid
changed the title
Rich-T-kid
force-pushed
the
rich-T-kid/minor-arrow-flight-opt
branch
2 times, most recently
from
2c00600 to
337abd5
Compare
Rich-T-kid
commented
Jun 12, 2026
…mory to fill array data
Rich-T-kid
force-pushed
the
rich-T-kid/minor-arrow-flight-opt
branch
from
337abd5 to
094579b
Compare
Contributor
Author
|
@Jefffrey I meant to ping you on this PR . Sorry about that! |
Contributor
|
run benchmarks flight |
|
🤖 Arrow criterion benchmark running (GKE) | trigger CPU Details (lscpu)Comparing rich-T-kid/minor-arrow-flight-opt (505fb20) to 826b808 (merge-base) diff File an issue against this benchmark runner |
|
🤖 Arrow criterion benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagebase (merge-base)
branch
File an issue against this benchmark runner |
Contributor
Author
|
Nice, regressions are gone. should re-run when 54faeda gets merged. I expected a larger improvement for larger rows/columns batches. I'll profile & update the PR |
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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?
starting small 😄
Rationale for this change
The arrow-flight encode path was allocating intermediate Vecs to hold data that was immediately iterated and discarded. Replacing these with lazy iterators and inlining the one helper that existed only to loop removes allocations that served no purpose beyond bridging two adjacent lines of code.
What changes are included in this PR?
[commit #1]
Impl<Iterator>.[commit #2]
[commit #3]
CompressionContexttoIpcWriteContextand added anfbb: FlatBufferBuilder<'static>field to it[commit #4]
IpcWriteContextgains a scratch:Vec<u8>field. When set before a call toIpcDataGenerator::encode(), the existing allocation is reused instead of allocating a fresh buffer for each batch's arrow data body.FlightIpcEncodermaintains anArrowDataPool, a small pool of Arc<Mutex<Vec<Vec>>> buffers pre-sized to the gRPC message limit (2 MiB). Before eachencode()call, a buffer is acquired from the pool and placed inIpcWriteContext::scratch. After encoding, the buffer is wrapped inPooledBufand handed to Bytes::from_owner; when the Bytes is dropped (after the gRPC frame is sent), the buffer is automatically returned to the pool rather than freed.[commit #5+]
acquiremethod to also pre-allocate 2MB of space in the vectorAre these changes tested?
yes
Are there any user-facing changes?
no