perf: Implement groups accumulator count distinct primitive types#21561
Open
coderfender wants to merge 20 commits intoapache:mainfrom
Open
perf: Implement groups accumulator count distinct primitive types#21561coderfender wants to merge 20 commits intoapache:mainfrom
coderfender wants to merge 20 commits intoapache:mainfrom
Conversation
Contributor
Author
|
Requesting to run benchmarks TPCH , TPCDS , clickbench etc to see how groups accumulator impl performs in count(distinct) use case |
Dandandan
reviewed
Apr 12, 2026
Contributor
|
run benchmark tpch tpcds clickbench_partitiomed clickbemch_extended |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (f0b2a4a) to 16e578d (merge-base) diff using: tpch File an issue against this benchmark runner |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (f0b2a4a) to 16e578d (merge-base) diff using: tpcds File an issue against this benchmark runner |
|
🤖 Criterion benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (f0b2a4a) to 16e578d (merge-base) diff File an issue against this benchmark runner |
|
🤖 Criterion benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (f0b2a4a) to 16e578d (merge-base) diff File an issue against this benchmark runner |
|
Benchmark for this request failed. Last 20 lines of output: Click to expandFile an issue against this benchmark runner |
1 similar comment
|
Benchmark for this request failed. Last 20 lines of output: Click to expandFile an issue against this benchmark runner |
coderfender
force-pushed
the
implement_groups_accumulator_count_distinct_primitive_types
branch
from
5692893 to
ee0c865
Compare
|
🤖 Benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagetpcds — base (merge-base)
tpcds — branch
File an issue against this benchmark runner |
Contributor
Author
|
@Dandandan could you retrigger the benchmarks again please? |
Contributor
|
run benchmark tpch tpcds clickbench_partitioned clickbench_extended |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (ee0c865) to 16e578d (merge-base) diff using: tpch File an issue against this benchmark runner |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (ee0c865) to 16e578d (merge-base) diff using: clickbench_extended File an issue against this benchmark runner |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (ee0c865) to 16e578d (merge-base) diff using: clickbench_partitioned File an issue against this benchmark runner |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (ee0c865) to 16e578d (merge-base) diff using: tpcds 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 |
|
🤖 Benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usageclickbench_extended — base (merge-base)
clickbench_extended — branch
File an issue against this benchmark runner |
coderfender
force-pushed
the
implement_groups_accumulator_count_distinct_primitive_types
branch
from
5b7f593 to
067a6b6
Compare
Contributor
Author
|
Comparison with main |
Contributor
Author
|
@Dandandan true! I also benchmarked through |
coderfender
force-pushed
the
implement_groups_accumulator_count_distinct_primitive_types
branch
from
2aa6549 to
cbfee35
Compare
coderfender
force-pushed
the
implement_groups_accumulator_count_distinct_primitive_types
branch
from
cbfee35 to
c2bc9d3
Compare
Contributor
Author
|
@Dandandan , could you re run the benchmarks please ? I do see a consistent speedup with Q9 and seems like this PR is ready for review |
Contributor
|
run benchmark clickbench_partitioned |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (47110a8) to 882be98 (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 |
Dandandan
approved these changes
Apr 17, 2026
Contributor
Author
|
Thank you for the approval @Dandandan . but it doesn't seem like the benchmarks ran correctly. Could you please trigger it again and perhaps even toch tpcds and other relevant queries please? |
Dandandan
reviewed
Apr 17, 2026
| EmitTo::First(n) => n, | ||
| }; | ||
|
|
||
| let mut group_values: Vec<Vec<T::Native>> = vec![Vec::new(); num_emitted]; |
Contributor
There was a problem hiding this comment.
Something like this (Claude-generated) probably is faster?
// Prefix-sum counts[..num_emitted] into offsets
let mut offsets = Vec::with_capacity(num_emitted + 1);
offsets.push(0i32);
let mut total = 0i32;
for &c in &self.counts[..num_emitted] {
total += c as i32;
offsets.push(total);
}
let mut all_values = vec![T::Native::default(); total as usize];
let mut cursors: Vec<i32> = offsets[..num_emitted].to_vec();
if matches!(emit_to, EmitTo::All) {
for (group_idx, value) in self.seen.drain() {
let pos = cursors[group_idx] as usize;
all_values[pos] = value;
cursors[group_idx] += 1;
}
self.counts.clear();
} else {
let mut remaining = HashSet::default();
for (group_idx, value) in self.seen.drain() {
if group_idx < num_emitted {
let pos = cursors[group_idx] as usize;
all_values[pos] = value;
cursors[group_idx] += 1;
} else {
remaining.insert((group_idx - num_emitted, value));
}
}
self.seen = remaining;
let _ = emit_to.take_needed(&mut self.counts);
Contributor
Author
There was a problem hiding this comment.
Ah yes . I tried having one vector but that my implementation more complex than I thought I could maintain . But let me give this a try and get back to you to see if that helps with benchmarks
Dandandan
reviewed
Apr 17, 2026
| let list_array = values[0].as_list::<i32>(); | ||
|
|
||
| for (row_idx, &group_idx) in group_indices.iter().enumerate() { | ||
| let inner = list_array.value(row_idx); |
Contributor
There was a problem hiding this comment.
Something like this should be slightly faster:
let inner_values = inner.values();
let offsets = list_array.offsets();
for (row_idx, &group_idx) in group_indices.iter().enumerate() {
let start = offsets[row_idx] as usize;
let end = offsets[row_idx + 1] as usize;
for &value in &inner_values[start..end] {
if self.seen.insert((group_idx, value)) {
self.counts[group_idx] += 1;
}
}
}
Dandandan
reviewed
Apr 17, 2026
| offsets.push(all_values.len() as i32); | ||
| } | ||
|
|
||
| let values_array = Arc::new(PrimitiveArray::<T>::from_iter_values(all_values)); |
Contributor
There was a problem hiding this comment.
Suggested change
| let values_array = Arc::new(PrimitiveArray::<T>::from_iter_values(all_values)); | |
| let values_array = Arc::new(PrimitiveArray::<T>::new(ScalarBuffer::from(all_values), None)); |
Contributor
There was a problem hiding this comment.
This is faster as well @coderfender (avoids a copy).
Contributor
|
run benchmarks |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (066504d) to 5a427cb (merge-base) diff using: tpch File an issue against this benchmark runner |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (066504d) to 5a427cb (merge-base) diff using: tpcds File an issue against this benchmark runner |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (066504d) to 5a427cb (merge-base) diff using: clickbench_partitioned File an issue against this benchmark runner |
Contributor
Author
|
@Dandandan , pushed both optimizations. |
|
🤖 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 |
|
🤖 Benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagetpcds — base (merge-base)
tpcds — branch
File an issue against this benchmark runner |
Contributor
|
run benchmarks |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (617e790) to 5a427cb (merge-base) diff using: tpcds File an issue against this benchmark runner |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (617e790) to 5a427cb (merge-base) diff using: clickbench_partitioned File an issue against this benchmark runner |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing implement_groups_accumulator_count_distinct_primitive_types (617e790) to 5a427cb (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 |
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?
Evaluate perf with group accumulators for count distinct
Related : #21575 benchmark PR
Rationale for this change
What changes are included in this PR?
Are these changes tested?
Are there any user-facing changes?