Reduce reallocation costs in high-cardinality GROUP BY#21385
Closed
Dandandan wants to merge 1 commit intoapache:mainfrom
Closed
Reduce reallocation costs in high-cardinality GROUP BY#21385Dandandan wants to merge 1 commit intoapache:mainfrom
Dandandan wants to merge 1 commit intoapache:mainfrom
Conversation
- Combine AvgGroupsAccumulator's separate `counts` and `sums` Vecs into a single `Vec<AvgState>`, halving reallocations and improving cache locality for high-cardinality GROUP BY queries - Add `reserve(rows.len())` in PrimitiveGroupValueBuilder::vectorized_append to avoid repeated small reallocations during push loops - Add `reserve(rows.len())` in ByteViewGroupValueBuilder::vectorized_append_inner for the same reason Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Contributor
Author
|
run benchmarks |
github-actions
Bot
added
functions
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing reduce-reallocation-costs (45c2e44) to c17c87c (merge-base) diff using: clickbench_partitioned File an issue against this benchmark runner |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing reduce-reallocation-costs (45c2e44) to c17c87c (merge-base) diff using: tpcds File an issue against this benchmark runner |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing reduce-reallocation-costs (45c2e44) to c17c87c (merge-base) diff using: tpch File an issue against this benchmark runner |
|
🤖 Benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagetpcds — base (merge-base)
tpcds — branch
File an issue against this benchmark runner |
|
🤖 Benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usageclickbench_partitioned — base (merge-base)
clickbench_partitioned — branch
File an issue against this benchmark runner |
Contributor
Author
|
run benchmark clickbench_partitioned |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing reduce-reallocation-costs (45c2e44) to c17c87c (merge-base) diff using: clickbench_partitioned File an issue against this benchmark runner |
|
🤖 Benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usageclickbench_partitioned — base (merge-base)
clickbench_partitioned — branch
File an issue against this benchmark runner |
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
2 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?
Performance optimization - no specific issue.
Rationale for this change
Profiling high-cardinality GROUP BY queries (e.g.
GROUP BY WatchID, ClientIPon ClickBench) shows ~40% of CPU time inmemmove/reallocfrom vector growing in accumulators and group value builders. This PR reduces that overhead.What changes are included in this PR?
AvgGroupsAccumulator: combine
countsandsumsinto singleVec<AvgState>— The AVG accumulator previously maintained two separateVecs (counts: Vec<u64>andsums: Vec<T::Native>) that both resized tototal_num_groupson every batch. By combining them into a singleVec<AvgState>, we halve the number of reallocations and improve cache locality (count and sum for the same group are now adjacent in memory).PrimitiveGroupValueBuilder: add
reserve(rows.len())before push loops —vectorized_appendwas pushing elements one by one without pre-reserving capacity, causing repeated small reallocations when capacity was exceeded mid-loop.ByteViewGroupValueBuilder: add
reserve(rows.len())before push loops — Same issue as above for the byte view variant.Are these changes tested?
Covered by existing aggregate sqllogic tests (all pass).
Are there any user-facing changes?
No. Internal optimization only.