You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@datafusion.apache.org by "Rachelint (via GitHub)" <gi...@apache.org> on 2026/09/28 18:55:20 UTC
[PR] feat: prototype frozen final aggregation buckets [datafusion]
Rachelint opened a new pull request, #25846:
URL: https://github.com/apache/datafusion/pull/25846
## Which issue does this PR close?
Related to #25724. This is a stacked PoC for benchmark validation and does not close an issue. Until #25724 is merged, this PR also contains its six commits; the new experiment is the top commit `14346eba7`.
## Rationale for this change
The final hash aggregation in #25724 moves its accumulated groups to buckets once the threshold is reached. Retaining those groups could reduce bucket input and the merge after input ends when later rows repeat the same keys. This draft tests that idea with an opt-in implementation and reproducible benchmarks. It is not proposed for merge yet.
## What changes are included in this PR?
- Add the off-by-default `hash_aggregate_bucket_adaptive` option. After a 65,536-row sample, probing is bypassed when fewer than 25% of rows hit the frozen keys.
- Retain the original final hash table while routing misses to buckets; merge its state into buckets at EOF so keys remain correct after bypass. Fall back to the original path for unsupported key types or memory pressure.
- Add counters and timing metrics, correctness tests, six synthetic SQL benchmarks, and [results with reproduction commands](https://github.com/Rachelint/arrow-datafusion/blob/codex/frozen-final-aggregation-poc/benchmarks/sql_benchmarks/aggregate_frozen/RESULTS.md).
The local optimized synthetic benchmark found this PoC **7%β21% slower than the original bucket path in all six queries**. It uses a second encoded-key index because `GroupValues` has no lookup-only API. This draft exists so the same branch can be measured on the benchmark host with real ClickBench data.
### Benchmark host commands
```bash
cargo build -p datafusion-benchmarks --release --bin benchmark_runner
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_THRESHOLD=0 target/release/benchmark_runner aggregate_frozen --iterations 10 --partitions 12
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_THRESHOLD=4000 target/release/benchmark_runner aggregate_frozen --iterations 10 --partitions 12
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_THRESHOLD=4000 DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_ADAPTIVE=true target/release/benchmark_runner aggregate_frozen --iterations 10 --partitions 12
```
Real ClickBench Q11/Q12/Q14 and Q18/Q33/Q34, using the host's `hits.parquet` and Jay's threshold of 262,144, are still needed for the decision. This local machine did not have the dataset.
## What is the testing strategy for this PR?
- `cargo fmt --all`: passed.
- `cargo clippy --all-targets --all-features -- -D warnings`: passed.
- Focused final aggregation Rust tests and `aggregate_bucketed` / `information_schema` SQL logic tests: passed.
- `git diff --check`: passed.
- The full extended workspace suite was stopped after the initial submodule issue was fixed, as the immediate objective is benchmark validation. `./dev/rust_lint.sh` could not start because local `/usr/bin/python3` lacks PyYAML and `uv` is unavailable. These gates remain open before any merge.
## Are there any user-facing changes?
Only the off-by-default execution option and additional physical plan metrics. Existing behavior remains the default.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] feat: prototype frozen final aggregation buckets [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25846:
URL: https://github.com/apache/datafusion/pull/25846#issuecomment-5876758059
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/25846#issuecomment-5876484027)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing codex/frozen-final-aggregation-poc (14346eba70d74cc5b90b5f7871a5bf3fedb78205) to fb42cf5 (merge-base) [diff](https://github.com/apache/datafusion/compare/fb42cf5911faf1c05f6ccd5bde3d9e2300d35b76..14346eba70d74cc5b90b5f7871a5bf3fedb78205)
<details><summary>Run configuration</summary>
```yaml
run benchmark clickbench_partitioned
changed:
env:
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_THRESHOLD: "262144"
```
</details>
<details><summary>CPU Details (lscpu)</summary>
```
Architecture: aarch64
CPU op-mode(s): 64-bit
Byte Order: Little Endian
CPU(s): 16
On-line CPU(s) list: 0-15
Vendor ID: ARM
Model name: Neoverse-V2
Model: 1
Thread(s) per core: 1
Core(s) per cluster: 16
Socket(s): -
Cluster(s): 1
Stepping: r0p1
BogoMIPS: 2000.00
Flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache: 1 MiB (16 instances)
L1i cache: 1 MiB (16 instances)
L2 cache: 32 MiB (16 instances)
L3 cache: 80 MiB (1 instance)
NUMA node(s): 1
NUMA node0 CPU(s): 0-15
Vulnerability Gather data sampling: Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Mmio stale data: Not affected
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Not affected
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; __user pointer sanitization
Vulnerability Spectre v2: Mitigation; CSV2, BHB
Vulnerability Srbds: Not affected
Vulnerability Tsa: Not affected
Vulnerability Tsx async abort: Not affected
Vulnerability Vmscape: Not affected
```
</details>
<details><summary>Details</summary>
<p>
```
Comparing HEAD and codex_frozen-final-aggregation-poc
--------------------
Benchmark clickbench_partitioned.json
--------------------
βββββββββββββ³βββββββββββββ³βββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β codex_frozen-final-aggregation-poc β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 0 β 1.58 ms β 1.38 ms β +1.15x faster β
β QQuery 1 β 12.39 ms β 12.62 ms β no change β
β QQuery 2 β 37.66 ms β 38.11 ms β no change β
β QQuery 3 β 33.15 ms β 33.73 ms β no change β
β QQuery 4 β 277.81 ms β 201.88 ms β +1.38x faster β
β QQuery 5 β 299.00 ms β 299.52 ms β no change β
β QQuery 6 β 1.50 ms β 1.37 ms β +1.09x faster β
β QQuery 7 β 14.03 ms β 13.90 ms β no change β
β QQuery 8 β 401.46 ms β 307.82 ms β +1.30x faster β
β QQuery 9 β 559.06 ms β 533.58 ms β no change β
β QQuery 10 β 70.17 ms β 71.99 ms β no change β
β QQuery 11 β 80.05 ms β 83.57 ms β no change β
β QQuery 12 β 307.21 ms β 309.55 ms β no change β
β QQuery 13 β 402.26 ms β 393.72 ms β no change β
β QQuery 14 β 324.77 ms β 326.70 ms β no change β
β QQuery 15 β 350.62 ms β 239.37 ms β +1.46x faster β
β QQuery 16 β 714.29 ms β 712.11 ms β no change β
β QQuery 17 β 733.67 ms β 641.19 ms β +1.14x faster β
β QQuery 18 β 1560.41 ms β 1319.91 ms β +1.18x faster β
β QQuery 19 β 30.70 ms β 29.63 ms β no change β
β QQuery 20 β 541.88 ms β 545.57 ms β no change β
β QQuery 21 β 541.13 ms β 538.27 ms β no change β
β QQuery 22 β 1084.40 ms β 1051.65 ms β no change β
β QQuery 23 β 3289.48 ms β 3435.76 ms β no change β
β QQuery 24 β 44.50 ms β 45.08 ms β no change β
β QQuery 25 β 114.18 ms β 111.78 ms β no change β
β QQuery 26 β 45.51 ms β 44.10 ms β no change β
β QQuery 27 β 560.68 ms β 529.03 ms β +1.06x faster β
β QQuery 28 β 3138.15 ms β 3149.42 ms β no change β
β QQuery 29 β 43.48 ms β 44.03 ms β no change β
β QQuery 30 β 369.62 ms β 318.21 ms β +1.16x faster β
β QQuery 31 β 323.43 ms β 289.95 ms β +1.12x faster β
β QQuery 32 β 1360.36 ms β 805.99 ms β +1.69x faster β
β QQuery 33 β 1864.62 ms β 1612.60 ms β +1.16x faster β
β QQuery 34 β 1976.46 ms β 1702.69 ms β +1.16x faster β
β QQuery 35 β 357.94 ms β 269.65 ms β +1.33x faster β
β QQuery 36 β 76.56 ms β 74.66 ms β no change β
β QQuery 37 β 38.15 ms β 38.15 ms β no change β
β QQuery 38 β 45.00 ms β 43.95 ms β no change β
β QQuery 39 β 173.68 ms β 163.53 ms β +1.06x faster β
β QQuery 40 β 15.72 ms β 16.23 ms β no change β
β QQuery 41 β 15.31 ms β 15.00 ms β no change β
β QQuery 42 β 14.75 ms β 14.55 ms β no change β
βββββββββββββ΄βββββββββββββ΄βββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 22246.79ms β
β Total Time (codex_frozen-final-aggregation-poc) β 20431.49ms β
β Average Time (HEAD) β 517.37ms β
β Average Time (codex_frozen-final-aggregation-poc) β 475.15ms β
β Queries Faster β 15 β
β Queries Slower β 0 β
β Queries with No Change β 28 β
β Queries with Failure β 0 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββ
Distribution per query (min / mean Β±stddev / max):
Comparing HEAD and codex_frozen-final-aggregation-poc
--------------------
Benchmark clickbench_partitioned.json
--------------------
βββββββββββββ³βββββββββββββββββββββββββββββββββββββββββ³βββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β codex_frozen-final-aggregation-poc β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 0 β 1.58 / 4.77 Β±6.07 / 16.90 ms β 1.38 / 4.50 Β±5.97 / 16.45 ms β +1.06x faster β
β QQuery 1 β 12.39 / 12.81 Β±0.32 / 13.23 ms β 12.62 / 12.79 Β±0.14 / 12.98 ms β no change β
β QQuery 2 β 37.66 / 38.08 Β±0.32 / 38.44 ms β 38.11 / 38.43 Β±0.23 / 38.70 ms β no change β
β QQuery 3 β 33.15 / 34.08 Β±0.77 / 35.39 ms β 33.73 / 34.96 Β±0.85 / 36.11 ms β no change β
β QQuery 4 β 277.81 / 286.82 Β±5.56 / 294.20 ms β 201.88 / 203.51 Β±1.79 / 206.59 ms β +1.41x faster β
β QQuery 5 β 299.00 / 309.12 Β±6.42 / 317.63 ms β 299.52 / 310.57 Β±8.09 / 321.91 ms β no change β
β QQuery 6 β 1.50 / 1.90 Β±0.39 / 2.59 ms β 1.37 / 1.60 Β±0.29 / 2.17 ms β +1.18x faster β
β QQuery 7 β 14.03 / 14.20 Β±0.14 / 14.41 ms β 13.90 / 14.15 Β±0.23 / 14.58 ms β no change β
β QQuery 8 β 401.46 / 408.37 Β±5.01 / 416.73 ms β 307.82 / 313.66 Β±3.42 / 318.15 ms β +1.30x faster β
β QQuery 9 β 559.06 / 570.24 Β±9.08 / 581.26 ms β 533.58 / 545.77 Β±8.90 / 557.09 ms β no change β
β QQuery 10 β 70.17 / 72.60 Β±3.93 / 80.41 ms β 71.99 / 80.88 Β±16.24 / 113.34 ms β 1.11x slower β
β QQuery 11 β 80.05 / 80.80 Β±0.49 / 81.31 ms β 83.57 / 87.90 Β±7.42 / 102.71 ms β 1.09x slower β
β QQuery 12 β 307.21 / 310.36 Β±2.46 / 314.19 ms β 309.55 / 317.78 Β±6.84 / 327.54 ms β no change β
β QQuery 13 β 402.26 / 413.41 Β±11.78 / 432.65 ms β 393.72 / 404.29 Β±16.65 / 437.40 ms β no change β
β QQuery 14 β 324.77 / 329.18 Β±4.18 / 335.55 ms β 326.70 / 345.75 Β±18.60 / 381.42 ms β 1.05x slower β
β QQuery 15 β 350.62 / 358.76 Β±5.05 / 364.15 ms β 239.37 / 246.37 Β±11.03 / 268.32 ms β +1.46x faster β
β QQuery 16 β 714.29 / 737.69 Β±21.06 / 769.48 ms β 712.11 / 757.13 Β±46.79 / 845.74 ms β no change β
β QQuery 17 β 733.67 / 748.46 Β±11.48 / 765.42 ms β 641.19 / 672.86 Β±18.34 / 693.89 ms β +1.11x faster β
β QQuery 18 β 1560.41 / 1590.04 Β±35.59 / 1659.93 ms β 1319.91 / 1352.65 Β±25.96 / 1388.63 ms β +1.18x faster β
β QQuery 19 β 30.70 / 47.30 Β±23.63 / 91.57 ms β 29.63 / 52.28 Β±38.62 / 128.98 ms β 1.11x slower β
β QQuery 20 β 541.88 / 554.07 Β±11.34 / 572.86 ms β 545.57 / 561.60 Β±8.99 / 573.00 ms β no change β
β QQuery 21 β 541.13 / 545.27 Β±3.31 / 549.14 ms β 538.27 / 555.44 Β±9.05 / 564.21 ms β no change β
β QQuery 22 β 1084.40 / 1116.59 Β±39.03 / 1191.43 ms β 1051.65 / 1114.99 Β±69.95 / 1239.40 ms β no change β
β QQuery 23 β 3289.48 / 3560.04 Β±151.84 / 3721.13 ms β 3435.76 / 3776.10 Β±304.35 / 4255.22 ms β 1.06x slower β
β QQuery 24 β 44.50 / 92.36 Β±58.53 / 177.72 ms β 45.08 / 54.98 Β±11.00 / 76.15 ms β +1.68x faster β
β QQuery 25 β 114.18 / 207.64 Β±183.94 / 575.51 ms β 111.78 / 115.39 Β±3.90 / 122.77 ms β +1.80x faster β
β QQuery 26 β 45.51 / 46.01 Β±0.46 / 46.57 ms β 44.10 / 56.89 Β±22.84 / 102.51 ms β 1.24x slower β
β QQuery 27 β 560.68 / 573.51 Β±10.69 / 590.19 ms β 529.03 / 543.26 Β±12.26 / 561.44 ms β +1.06x faster β
β QQuery 28 β 3138.15 / 3214.94 Β±71.92 / 3329.54 ms β 3149.42 / 3209.37 Β±55.89 / 3296.41 ms β no change β
β QQuery 29 β 43.48 / 44.82 Β±1.07 / 46.21 ms β 44.03 / 49.54 Β±9.89 / 69.31 ms β 1.11x slower β
β QQuery 30 β 369.62 / 396.46 Β±14.29 / 411.80 ms β 318.21 / 339.45 Β±12.16 / 351.13 ms β +1.17x faster β
β QQuery 31 β 323.43 / 342.97 Β±10.13 / 351.54 ms β 289.95 / 299.11 Β±5.63 / 306.06 ms β +1.15x faster β
β QQuery 32 β 1360.36 / 1560.03 Β±212.49 / 1969.87 ms β 805.99 / 914.58 Β±165.27 / 1242.01 ms β +1.71x faster β
β QQuery 33 β 1864.62 / 1964.40 Β±64.12 / 2062.33 ms β 1612.60 / 1902.57 Β±159.07 / 2040.19 ms β no change β
β QQuery 34 β 1976.46 / 2047.26 Β±92.68 / 2215.66 ms β 1702.69 / 1801.68 Β±83.28 / 1935.31 ms β +1.14x faster β
β QQuery 35 β 357.94 / 423.86 Β±107.15 / 636.58 ms β 269.65 / 299.85 Β±30.77 / 353.39 ms β +1.41x faster β
β QQuery 36 β 76.56 / 79.60 Β±2.35 / 82.60 ms β 74.66 / 79.28 Β±3.75 / 83.84 ms β no change β
β QQuery 37 β 38.15 / 42.17 Β±3.16 / 46.94 ms β 38.15 / 41.54 Β±4.82 / 50.96 ms β no change β
β QQuery 38 β 45.00 / 51.76 Β±7.92 / 62.23 ms β 43.95 / 49.01 Β±4.79 / 57.92 ms β +1.06x faster β
β QQuery 39 β 173.68 / 183.44 Β±8.46 / 196.22 ms β 163.53 / 176.46 Β±7.45 / 184.36 ms β no change β
β QQuery 40 β 15.72 / 20.00 Β±7.15 / 34.23 ms β 16.23 / 17.42 Β±1.67 / 20.74 ms β +1.15x faster β
β QQuery 41 β 15.31 / 15.59 Β±0.17 / 15.80 ms β 15.00 / 15.39 Β±0.22 / 15.60 ms β no change β
β QQuery 42 β 14.75 / 18.11 Β±3.66 / 22.85 ms β 14.55 / 16.18 Β±2.27 / 20.60 ms β +1.12x faster β
βββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 23469.86ms β
β Total Time (codex_frozen-final-aggregation-poc) β 21787.95ms β
β Average Time (HEAD) β 545.81ms β
β Average Time (codex_frozen-final-aggregation-poc) β 506.70ms β
β Queries Faster β 18 β
β Queries Slower β 7 β
β Queries with No Change β 18 β
β Queries with Failure β 0 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**clickbench_partitioned β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 120.0s |
| Peak memory | 17.7 GiB |
| Avg memory | 5.8 GiB |
| CPU user | 1139.5s |
| CPU sys | 123.3s |
| Peak spill | 0 B |
**clickbench_partitioned β branch**
| Metric | Value |
|--------|-------|
| Wall time | 110.0s |
| Peak memory | 18.0 GiB |
| Avg memory | 5.5 GiB |
| CPU user | 1029.0s |
| CPU sys | 112.0s |
| Peak spill | 0 B |
</details>
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] feat: prototype frozen final aggregation buckets [datafusion]
Posted by "Rachelint (via GitHub)" <gi...@apache.org>.
Rachelint commented on PR #25846:
URL: https://github.com/apache/datafusion/pull/25846#issuecomment-5876484027
run benchmark clickbench_partitioned
changed:
env:
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_THRESHOLD: "262144"
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] feat: prototype frozen final aggregation buckets [datafusion]
Posted by "codecov-commenter (via GitHub)" <gi...@apache.org>.
codecov-commenter commented on PR #25846:
URL: https://github.com/apache/datafusion/pull/25846#issuecomment-5877683253
## [Codecov](https://app.codecov.io/gh/apache/datafusion/pull/25846?dropdown=coverage&src=pr&el=h1&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) Report
:x: Patch coverage is `86.19303%` with `206 lines` in your changes missing coverage. Please review.
:white_check_mark: Project coverage is 82.56%. Comparing base ([`fb42cf5`](https://app.codecov.io/gh/apache/datafusion/commit/fb42cf5911faf1c05f6ccd5bde3d9e2300d35b76?dropdown=coverage&el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache)) to head ([`14346eb`](https://app.codecov.io/gh/apache/datafusion/commit/14346eba70d74cc5b90b5f7871a5bf3fedb78205?dropdown=coverage&el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache)).
:warning: Report is 38 commits behind head on main.
| [Files with missing lines](https://app.codecov.io/gh/apache/datafusion/pull/25846?dropdown=coverage&src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) | Patch % | Lines |
|---|---|---|
| [...fusion/physical-plan/src/aggregates/hash\_stream.rs](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Faggregates%2Fhash_stream.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9hZ2dyZWdhdGVzL2hhc2hfc3RyZWFtLnJz) | 82.60% | [68 Missing and 56 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [...sion/physical-plan/src/aggregates/final\_buckets.rs](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Faggregates%2Ffinal_buckets.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9hZ2dyZWdhdGVzL2ZpbmFsX2J1Y2tldHMucnM=) | 86.69% | [8 Missing and 29 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [...ysical-plan/src/aggregates/bucketed\_aggregation.rs](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Faggregates%2Fbucketed_aggregation.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9hZ2dyZWdhdGVzL2J1Y2tldGVkX2FnZ3JlZ2F0aW9uLnJz) | 85.84% | [9 Missing and 23 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [...src/aggregates/aggregate\_hash\_table/final\_table.rs](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Faggregates%2Faggregate_hash_table%2Ffinal_table.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9hZ2dyZWdhdGVzL2FnZ3JlZ2F0ZV9oYXNoX3RhYmxlL2ZpbmFsX3RhYmxlLnJz) | 90.62% | [3 Missing :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [...gregates/group\_values/multi\_group\_by/bytes\_view.rs](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Faggregates%2Fgroup_values%2Fmulti_group_by%2Fbytes_view.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9hZ2dyZWdhdGVzL2dyb3VwX3ZhbHVlcy9tdWx0aV9ncm91cF9ieS9ieXRlc192aWV3LnJz) | 97.34% | [1 Missing and 2 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [.../src/aggregates/group\_values/multi\_group\_by/mod.rs](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Faggregates%2Fgroup_values%2Fmulti_group_by%2Fmod.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9hZ2dyZWdhdGVzL2dyb3VwX3ZhbHVlcy9tdWx0aV9ncm91cF9ieS9tb2QucnM=) | 92.30% | [0 Missing and 3 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [...plan/src/aggregates/aggregate\_hash\_table/common.rs](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Faggregates%2Faggregate_hash_table%2Fcommon.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9hZ2dyZWdhdGVzL2FnZ3JlZ2F0ZV9oYXNoX3RhYmxlL2NvbW1vbi5ycw==) | 97.18% | [1 Missing and 1 partial :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [...n/physical-plan/src/aggregates/group\_values/mod.rs](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Faggregates%2Fgroup_values%2Fmod.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9hZ2dyZWdhdGVzL2dyb3VwX3ZhbHVlcy9tb2QucnM=) | 86.66% | [2 Missing :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25846?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
<details><summary>Additional details and impacted files</summary>
```diff
@@ Coverage Diff @@
## main #25846 +/- ##
==========================================
+ Coverage 82.49% 82.56% +0.07%
==========================================
Files 1140 1143 +3
Lines 439034 441804 +2770
Branches 439034 441804 +2770
==========================================
+ Hits 362165 364771 +2606
+ Misses 54962 54929 -33
- Partials 21907 22104 +197
```
</details>
[:umbrella: View full report in Codecov by Harness](https://app.codecov.io/gh/apache/datafusion/pull/25846?dropdown=coverage&src=pr&el=continue&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache).
:loudspeaker: Have feedback on the report? [Share it here](https://about.codecov.io/codecov-pr-comment-feedback/?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache).
<details><summary> :rocket: New features to boost your workflow: </summary>
- :snowflake: [Test Analytics](https://docs.codecov.com/docs/test-analytics): Detect flaky tests, report on failures, and find test suite problems.
- :package: [JS Bundle Analysis](https://docs.codecov.com/docs/javascript-bundle-analysis): Save yourself from yourself by tracking and limiting bundle sizes in JS merges.
</details>
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] feat: prototype frozen final aggregation buckets [datafusion]
Posted by "Rachelint (via GitHub)" <gi...@apache.org>.
Rachelint commented on PR #25846:
URL: https://github.com/apache/datafusion/pull/25846#issuecomment-5876802086
run benchmark clickbench_partitioned
changed:
env:
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_THRESHOLD: "262144"
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] feat: prototype frozen final aggregation buckets [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25846:
URL: https://github.com/apache/datafusion/pull/25846#issuecomment-5876542556
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/25846#issuecomment-5876484027)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c5876484027-2894-kdktl 6.12.94+ #1 SMP Wed Aug 19 07:47:20 UTC 2026 aarch64 GNU/Linux`
<details><summary>CPU Details (lscpu)</summary>
```
Architecture: aarch64
CPU op-mode(s): 64-bit
Byte Order: Little Endian
CPU(s): 16
On-line CPU(s) list: 0-15
Vendor ID: ARM
Model name: Neoverse-V2
Model: 1
Thread(s) per core: 1
Core(s) per cluster: 16
Socket(s): -
Cluster(s): 1
Stepping: r0p1
BogoMIPS: 2000.00
Flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache: 1 MiB (16 instances)
L1i cache: 1 MiB (16 instances)
L2 cache: 32 MiB (16 instances)
L3 cache: 80 MiB (1 instance)
NUMA node(s): 1
NUMA node0 CPU(s): 0-15
Vulnerability Gather data sampling: Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Mmio stale data: Not affected
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Not affected
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; __user pointer sanitization
Vulnerability Spectre v2: Mitigation; CSV2, BHB
Vulnerability Srbds: Not affected
Vulnerability Tsa: Not affected
Vulnerability Tsx async abort: Not affected
Vulnerability Vmscape: Not affected
```
</details>
Comparing codex/frozen-final-aggregation-poc (14346eba70d74cc5b90b5f7871a5bf3fedb78205) to fb42cf5 (merge-base) [diff](https://github.com/apache/datafusion/compare/fb42cf5911faf1c05f6ccd5bde3d9e2300d35b76..14346eba70d74cc5b90b5f7871a5bf3fedb78205)
<details><summary>Run configuration</summary>
```yaml
run benchmark clickbench_partitioned
changed:
env:
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_THRESHOLD: "262144"
```
</details>
Results will be posted here when complete
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] feat: prototype frozen final aggregation buckets [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25846:
URL: https://github.com/apache/datafusion/pull/25846#issuecomment-5876967451
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/25846#issuecomment-5876802086)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing codex/frozen-final-aggregation-poc (14346eba70d74cc5b90b5f7871a5bf3fedb78205) to fb42cf5 (merge-base) [diff](https://github.com/apache/datafusion/compare/fb42cf5911faf1c05f6ccd5bde3d9e2300d35b76..14346eba70d74cc5b90b5f7871a5bf3fedb78205)
<details><summary>Run configuration</summary>
```yaml
run benchmark clickbench_partitioned
changed:
env:
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_THRESHOLD: "262144"
```
</details>
<details><summary>CPU Details (lscpu)</summary>
```
Architecture: aarch64
CPU op-mode(s): 64-bit
Byte Order: Little Endian
CPU(s): 16
On-line CPU(s) list: 0-15
Vendor ID: ARM
Model name: Neoverse-V2
Model: 1
Thread(s) per core: 1
Core(s) per cluster: 16
Socket(s): -
Cluster(s): 1
Stepping: r0p1
BogoMIPS: 2000.00
Flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache: 1 MiB (16 instances)
L1i cache: 1 MiB (16 instances)
L2 cache: 32 MiB (16 instances)
L3 cache: 80 MiB (1 instance)
NUMA node(s): 1
NUMA node0 CPU(s): 0-15
Vulnerability Gather data sampling: Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Mmio stale data: Not affected
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Not affected
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; __user pointer sanitization
Vulnerability Spectre v2: Mitigation; CSV2, BHB
Vulnerability Srbds: Not affected
Vulnerability Tsa: Not affected
Vulnerability Tsx async abort: Not affected
Vulnerability Vmscape: Not affected
```
</details>
<details><summary>Details</summary>
<p>
```
Comparing HEAD and codex_frozen-final-aggregation-poc
--------------------
Benchmark clickbench_partitioned.json
--------------------
βββββββββββββ³βββββββββββββ³βββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β codex_frozen-final-aggregation-poc β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 0 β 1.30 ms β 1.21 ms β +1.08x faster β
β QQuery 1 β 12.11 ms β 11.73 ms β no change β
β QQuery 2 β 37.02 ms β 36.99 ms β no change β
β QQuery 3 β 32.31 ms β 31.31 ms β no change β
β QQuery 4 β 259.10 ms β 187.54 ms β +1.38x faster β
β QQuery 5 β 288.55 ms β 282.14 ms β no change β
β QQuery 6 β 1.42 ms β 1.31 ms β +1.09x faster β
β QQuery 7 β 13.53 ms β 13.32 ms β no change β
β QQuery 8 β 357.78 ms β 295.64 ms β +1.21x faster β
β QQuery 9 β 519.51 ms β 487.36 ms β +1.07x faster β
β QQuery 10 β 65.43 ms β 66.68 ms β no change β
β QQuery 11 β 76.90 ms β 77.54 ms β no change β
β QQuery 12 β 283.85 ms β 272.10 ms β no change β
β QQuery 13 β 386.79 ms β 366.35 ms β +1.06x faster β
β QQuery 14 β 288.33 ms β 289.15 ms β no change β
β QQuery 15 β 309.18 ms β 221.45 ms β +1.40x faster β
β QQuery 16 β 668.06 ms β 656.81 ms β no change β
β QQuery 17 β 670.79 ms β 583.40 ms β +1.15x faster β
β QQuery 18 β 1354.53 ms β 1148.86 ms β +1.18x faster β
β QQuery 19 β 28.32 ms β 28.34 ms β no change β
β QQuery 20 β 528.14 ms β 522.17 ms β no change β
β QQuery 21 β 519.43 ms β 534.49 ms β no change β
β QQuery 22 β 1054.83 ms β 1126.59 ms β 1.07x slower β
β QQuery 23 β 3265.46 ms β 3394.65 ms β no change β
β QQuery 24 β 40.70 ms β 41.74 ms β no change β
β QQuery 25 β 105.98 ms β 105.03 ms β no change β
β QQuery 26 β 41.24 ms β 40.72 ms β no change β
β QQuery 27 β 514.06 ms β 513.00 ms β no change β
β QQuery 28 β 2869.20 ms β 3016.56 ms β 1.05x slower β
β QQuery 29 β 41.54 ms β 41.97 ms β no change β
β QQuery 30 β 304.32 ms β 292.35 ms β no change β
β QQuery 31 β 277.97 ms β 252.62 ms β +1.10x faster β
β QQuery 32 β 1117.09 ms β 751.19 ms β +1.49x faster β
β QQuery 33 β 1565.69 ms β 1543.74 ms β no change β
β QQuery 34 β 1576.25 ms β 1527.98 ms β no change β
β QQuery 35 β 289.24 ms β 238.27 ms β +1.21x faster β
β QQuery 36 β 69.12 ms β 68.44 ms β no change β
β QQuery 37 β 35.44 ms β 36.17 ms β no change β
β QQuery 38 β 40.59 ms β 42.92 ms β 1.06x slower β
β QQuery 39 β 153.00 ms β 143.74 ms β +1.06x faster β
β QQuery 40 β 14.20 ms β 14.96 ms β 1.05x slower β
β QQuery 41 β 13.42 ms β 14.28 ms β 1.06x slower β
β QQuery 42 β 12.96 ms β 14.06 ms β 1.09x slower β
βββββββββββββ΄βββββββββββββ΄βββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 20104.66ms β
β Total Time (codex_frozen-final-aggregation-poc) β 19336.85ms β
β Average Time (HEAD) β 467.55ms β
β Average Time (codex_frozen-final-aggregation-poc) β 449.69ms β
β Queries Faster β 13 β
β Queries Slower β 6 β
β Queries with No Change β 24 β
β Queries with Failure β 0 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββ
Distribution per query (min / mean Β±stddev / max):
Comparing HEAD and codex_frozen-final-aggregation-poc
--------------------
Benchmark clickbench_partitioned.json
--------------------
βββββββββββββ³βββββββββββββββββββββββββββββββββββββββββ³βββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β codex_frozen-final-aggregation-poc β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 0 β 1.30 / 4.28 Β±5.73 / 15.74 ms β 1.21 / 4.08 Β±5.55 / 15.18 ms β no change β
β QQuery 1 β 12.11 / 12.36 Β±0.20 / 12.71 ms β 11.73 / 12.05 Β±0.22 / 12.41 ms β no change β
β QQuery 2 β 37.02 / 37.46 Β±0.28 / 37.80 ms β 36.99 / 37.20 Β±0.20 / 37.52 ms β no change β
β QQuery 3 β 32.31 / 32.93 Β±0.77 / 34.42 ms β 31.31 / 31.62 Β±0.24 / 31.89 ms β no change β
β QQuery 4 β 259.10 / 262.67 Β±3.21 / 268.16 ms β 187.54 / 192.41 Β±3.21 / 196.20 ms β +1.37x faster β
β QQuery 5 β 288.55 / 292.56 Β±2.58 / 296.00 ms β 282.14 / 290.58 Β±8.91 / 305.53 ms β no change β
β QQuery 6 β 1.42 / 4.59 Β±5.97 / 16.52 ms β 1.31 / 1.46 Β±0.23 / 1.92 ms β +3.15x faster β
β QQuery 7 β 13.53 / 13.78 Β±0.14 / 13.94 ms β 13.32 / 13.35 Β±0.06 / 13.46 ms β no change β
β QQuery 8 β 357.78 / 364.04 Β±3.41 / 367.86 ms β 295.64 / 307.44 Β±12.67 / 327.58 ms β +1.18x faster β
β QQuery 9 β 519.51 / 525.43 Β±6.82 / 536.50 ms β 487.36 / 493.77 Β±4.17 / 498.00 ms β +1.06x faster β
β QQuery 10 β 65.43 / 66.34 Β±0.64 / 67.12 ms β 66.68 / 67.93 Β±1.44 / 70.64 ms β no change β
β QQuery 11 β 76.90 / 82.59 Β±9.00 / 100.50 ms β 77.54 / 79.34 Β±1.90 / 82.89 ms β no change β
β QQuery 12 β 283.85 / 289.09 Β±4.63 / 295.81 ms β 272.10 / 280.58 Β±7.32 / 293.39 ms β no change β
β QQuery 13 β 386.79 / 394.91 Β±6.08 / 403.90 ms β 366.35 / 378.21 Β±15.26 / 407.26 ms β no change β
β QQuery 14 β 288.33 / 297.66 Β±5.45 / 303.67 ms β 289.15 / 306.20 Β±17.65 / 339.15 ms β no change β
β QQuery 15 β 309.18 / 320.34 Β±8.41 / 332.19 ms β 221.45 / 230.29 Β±10.11 / 249.93 ms β +1.39x faster β
β QQuery 16 β 668.06 / 676.45 Β±8.00 / 690.53 ms β 656.81 / 675.92 Β±11.03 / 689.50 ms β no change β
β QQuery 17 β 670.79 / 681.82 Β±8.12 / 694.48 ms β 583.40 / 596.62 Β±12.08 / 614.17 ms β +1.14x faster β
β QQuery 18 β 1354.53 / 1395.89 Β±26.40 / 1433.96 ms β 1148.86 / 1226.26 Β±43.66 / 1271.48 ms β +1.14x faster β
β QQuery 19 β 28.32 / 29.79 Β±2.72 / 35.22 ms β 28.34 / 41.93 Β±16.41 / 65.89 ms β 1.41x slower β
β QQuery 20 β 528.14 / 535.75 Β±4.94 / 543.38 ms β 522.17 / 533.36 Β±9.32 / 547.60 ms β no change β
β QQuery 21 β 519.43 / 532.92 Β±8.04 / 542.20 ms β 534.49 / 561.41 Β±16.56 / 585.22 ms β 1.05x slower β
β QQuery 22 β 1054.83 / 1075.73 Β±24.43 / 1119.86 ms β 1126.59 / 1145.82 Β±15.29 / 1163.65 ms β 1.07x slower β
β QQuery 23 β 3265.46 / 3385.78 Β±85.83 / 3487.71 ms β 3394.65 / 3538.11 Β±120.61 / 3729.57 ms β no change β
β QQuery 24 β 40.70 / 120.24 Β±128.93 / 373.95 ms β 41.74 / 104.01 Β±54.27 / 166.18 ms β +1.16x faster β
β QQuery 25 β 105.98 / 155.45 Β±95.42 / 346.25 ms β 105.03 / 175.14 Β±119.89 / 412.71 ms β 1.13x slower β
β QQuery 26 β 41.24 / 42.20 Β±0.52 / 42.73 ms β 40.72 / 41.23 Β±0.62 / 42.41 ms β no change β
β QQuery 27 β 514.06 / 520.12 Β±5.28 / 528.88 ms β 513.00 / 529.39 Β±9.37 / 540.28 ms β no change β
β QQuery 28 β 2869.20 / 2895.49 Β±16.79 / 2922.20 ms β 3016.56 / 3056.48 Β±29.41 / 3107.49 ms β 1.06x slower β
β QQuery 29 β 41.54 / 45.29 Β±6.58 / 58.39 ms β 41.97 / 50.27 Β±15.84 / 81.95 ms β 1.11x slower β
β QQuery 30 β 304.32 / 311.89 Β±5.37 / 318.50 ms β 292.35 / 312.37 Β±16.04 / 338.77 ms β no change β
β QQuery 31 β 277.97 / 300.02 Β±23.39 / 344.61 ms β 252.62 / 260.86 Β±7.23 / 274.29 ms β +1.15x faster β
β QQuery 32 β 1117.09 / 1219.80 Β±70.14 / 1297.64 ms β 751.19 / 787.54 Β±52.62 / 891.82 ms β +1.55x faster β
β QQuery 33 β 1565.69 / 1672.27 Β±166.73 / 2004.27 ms β 1543.74 / 1632.01 Β±96.35 / 1753.61 ms β no change β
β QQuery 34 β 1576.25 / 1607.85 Β±31.61 / 1668.12 ms β 1527.98 / 1566.28 Β±36.21 / 1630.11 ms β no change β
β QQuery 35 β 289.24 / 408.11 Β±211.83 / 831.46 ms β 238.27 / 277.49 Β±43.58 / 341.96 ms β +1.47x faster β
β QQuery 36 β 69.12 / 76.12 Β±4.03 / 80.15 ms β 68.44 / 72.68 Β±3.58 / 79.03 ms β no change β
β QQuery 37 β 35.44 / 36.49 Β±0.92 / 38.12 ms β 36.17 / 39.44 Β±2.02 / 42.42 ms β 1.08x slower β
β QQuery 38 β 40.59 / 44.85 Β±4.85 / 54.19 ms β 42.92 / 47.78 Β±3.99 / 53.70 ms β 1.07x slower β
β QQuery 39 β 153.00 / 160.46 Β±9.08 / 178.19 ms β 143.74 / 153.90 Β±6.63 / 161.62 ms β no change β
β QQuery 40 β 14.20 / 14.72 Β±0.46 / 15.56 ms β 14.96 / 16.48 Β±1.88 / 20.07 ms β 1.12x slower β
β QQuery 41 β 13.42 / 13.69 Β±0.22 / 14.07 ms β 14.28 / 18.32 Β±7.33 / 32.97 ms β 1.34x slower β
β QQuery 42 β 12.96 / 13.21 Β±0.23 / 13.61 ms β 14.06 / 16.86 Β±3.92 / 24.18 ms β 1.28x slower β
βββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 20973.40ms β
β Total Time (codex_frozen-final-aggregation-poc) β 20204.49ms β
β Average Time (HEAD) β 487.75ms β
β Average Time (codex_frozen-final-aggregation-poc) β 469.87ms β
β Queries Faster β 11 β
β Queries Slower β 11 β
β Queries with No Change β 21 β
β Queries with Failure β 0 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**clickbench_partitioned β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 110.0s |
| Peak memory | 17.3 GiB |
| Avg memory | 5.5 GiB |
| CPU user | 1024.2s |
| CPU sys | 102.1s |
| Peak spill | 0 B |
**clickbench_partitioned β branch**
| Metric | Value |
|--------|-------|
| Wall time | 105.0s |
| Peak memory | 18.1 GiB |
| Avg memory | 5.2 GiB |
| CPU user | 961.9s |
| CPU sys | 101.5s |
| Peak spill | 0 B |
</details>
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] feat: prototype frozen final aggregation buckets [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25846:
URL: https://github.com/apache/datafusion/pull/25846#issuecomment-5881566719
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/25846#issuecomment-5881538944)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c5881538944-2900-58c7x 6.12.94+ #1 SMP Wed Aug 19 07:47:20 UTC 2026 aarch64 GNU/Linux`
<details><summary>CPU Details (lscpu)</summary>
```
Architecture: aarch64
CPU op-mode(s): 64-bit
Byte Order: Little Endian
CPU(s): 16
On-line CPU(s) list: 0-15
Vendor ID: ARM
Model name: Neoverse-V2
Model: 1
Thread(s) per core: 1
Core(s) per cluster: 16
Socket(s): -
Cluster(s): 1
Stepping: r0p1
BogoMIPS: 2000.00
Flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache: 1 MiB (16 instances)
L1i cache: 1 MiB (16 instances)
L2 cache: 32 MiB (16 instances)
L3 cache: 80 MiB (1 instance)
NUMA node(s): 1
NUMA node0 CPU(s): 0-15
Vulnerability Gather data sampling: Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Mmio stale data: Not affected
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Not affected
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; __user pointer sanitization
Vulnerability Spectre v2: Mitigation; CSV2, BHB
Vulnerability Srbds: Not affected
Vulnerability Tsa: Not affected
Vulnerability Tsx async abort: Not affected
Vulnerability Vmscape: Not affected
```
</details>
Comparing 14346eba70d74cc5b90b5f7871a5bf3fedb78205 (14346eba70d74cc5b90b5f7871a5bf3fedb78205) to f9df89e2b1077deaae9be106682c9f1e4d0d61d2 [diff](https://github.com/apache/datafusion/compare/f9df89e2b1077deaae9be106682c9f1e4d0d61d2..14346eba70d74cc5b90b5f7871a5bf3fedb78205)
<details><summary>Run configuration</summary>
```yaml
run benchmark clickbench_partitioned
env:
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_THRESHOLD: "262144"
baseline:
ref: "f9df89e2b1077deaae9be106682c9f1e4d0d61d2"
changed:
ref: "14346eba70d74cc5b90b5f7871a5bf3fedb78205"
env:
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_ADAPTIVE: "true"
```
</details>
Results will be posted here when complete
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] feat: prototype frozen final aggregation buckets [datafusion]
Posted by "Rachelint (via GitHub)" <gi...@apache.org>.
Rachelint commented on PR #25846:
URL: https://github.com/apache/datafusion/pull/25846#issuecomment-5881538944
run benchmark clickbench_partitioned
env:
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_THRESHOLD: "262144"
baseline:
ref: f9df89e2b1077deaae9be106682c9f1e4d0d61d2
changed:
ref: 14346eba70d74cc5b90b5f7871a5bf3fedb78205
env:
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_ADAPTIVE: "true"
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] feat: prototype frozen final aggregation buckets [datafusion]
Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #25846:
URL: https://github.com/apache/datafusion/pull/25846#issuecomment-5877289719
<!-- semver-check-comment -->
Thank you for opening this pull request!
Reviewer note: [cargo-semver-checks](https://github.com/obi1kenobi/cargo-semver-checks) reported the current version number is not SemVer-compatible with the changes in this pull request (compared against the base branch).
<details>
<summary>Details</summary>
```
Cloning apache/main
Building datafusion-common v55.1.0 (current)
Built [ 21.875s] (current)
Parsing datafusion-common v55.1.0 (current)
Parsed [ 0.043s] (current)
Building datafusion-common v55.1.0 (baseline)
Built [ 21.924s] (baseline)
Parsing datafusion-common v55.1.0 (baseline)
Parsed [ 0.046s] (baseline)
Checking datafusion-common v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 0.544s] 223 checks: 222 pass, 1 fail, 0 warn, 31 skip
--- failure constructible_struct_adds_field: struct exhaustively constructible through public API adds field ---
Description:
A pub struct that could be exhaustively constructed with a literal using only public API has a new pub field, breaking existing exhaustive literals.
ref: https://doc.rust-lang.org/reference/expressions/struct-expr.html
impl: https://github.com/obi1kenobi/cargo-semver-checks/tree/v0.50.0/src/lints/constructible_struct_adds_field.ron
Failed in:
field ExecutionOptions.hash_aggregate_bucket_threshold in /home/runner/work/datafusion/datafusion/datafusion/common/src/config.rs:896
field ExecutionOptions.hash_aggregate_bucket_adaptive in /home/runner/work/datafusion/datafusion/datafusion/common/src/config.rs:896
Summary semver requires new major version: 1 major and 0 minor checks failed
Finished [ 46.346s] datafusion-common
Building datafusion-physical-plan v55.1.0 (current)
Built [ 26.001s] (current)
Parsing datafusion-physical-plan v55.1.0 (current)
Parsed [ 0.121s] (current)
Building datafusion-physical-plan v55.1.0 (baseline)
Built [ 25.405s] (baseline)
Parsing datafusion-physical-plan v55.1.0 (baseline)
Parsed [ 0.123s] (baseline)
Checking datafusion-physical-plan v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 0.467s] 223 checks: 223 pass, 31 skip
Summary no semver update required
Finished [ 53.381s] datafusion-physical-plan
Building datafusion-sqllogictest v55.1.0 (current)
Built [ 64.861s] (current)
Parsing datafusion-sqllogictest v55.1.0 (current)
Parsed [ 0.015s] (current)
Building datafusion-sqllogictest v55.1.0 (baseline)
Built [ 64.899s] (baseline)
Parsing datafusion-sqllogictest v55.1.0 (baseline)
Parsed [ 0.018s] (baseline)
Checking datafusion-sqllogictest v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 0.067s] 223 checks: 223 pass, 31 skip
Summary no semver update required
Finished [ 132.169s] datafusion-sqllogictest
```
</details>
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] feat: prototype frozen final aggregation buckets [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25846:
URL: https://github.com/apache/datafusion/pull/25846#issuecomment-5881705556
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/25846#issuecomment-5881538944)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing 14346eba70d74cc5b90b5f7871a5bf3fedb78205 (14346eba70d74cc5b90b5f7871a5bf3fedb78205) to f9df89e2b1077deaae9be106682c9f1e4d0d61d2 [diff](https://github.com/apache/datafusion/compare/f9df89e2b1077deaae9be106682c9f1e4d0d61d2..14346eba70d74cc5b90b5f7871a5bf3fedb78205)
<details><summary>Run configuration</summary>
```yaml
run benchmark clickbench_partitioned
env:
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_THRESHOLD: "262144"
baseline:
ref: "f9df89e2b1077deaae9be106682c9f1e4d0d61d2"
changed:
ref: "14346eba70d74cc5b90b5f7871a5bf3fedb78205"
env:
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_ADAPTIVE: "true"
```
</details>
<details><summary>CPU Details (lscpu)</summary>
```
Architecture: aarch64
CPU op-mode(s): 64-bit
Byte Order: Little Endian
CPU(s): 16
On-line CPU(s) list: 0-15
Vendor ID: ARM
Model name: Neoverse-V2
Model: 1
Thread(s) per core: 1
Core(s) per cluster: 16
Socket(s): -
Cluster(s): 1
Stepping: r0p1
BogoMIPS: 2000.00
Flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache: 1 MiB (16 instances)
L1i cache: 1 MiB (16 instances)
L2 cache: 32 MiB (16 instances)
L3 cache: 80 MiB (1 instance)
NUMA node(s): 1
NUMA node0 CPU(s): 0-15
Vulnerability Gather data sampling: Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Mmio stale data: Not affected
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Not affected
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; __user pointer sanitization
Vulnerability Spectre v2: Mitigation; CSV2, BHB
Vulnerability Srbds: Not affected
Vulnerability Tsa: Not affected
Vulnerability Tsx async abort: Not affected
Vulnerability Vmscape: Not affected
```
</details>
<details><summary>Details</summary>
<p>
```
Comparing HEAD and codex_frozen-final-aggregation-poc
--------------------
Benchmark clickbench_partitioned.json
--------------------
βββββββββββββ³βββββββββββββ³βββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β codex_frozen-final-aggregation-poc β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 0 β 1.29 ms β 1.25 ms β no change β
β QQuery 1 β 12.01 ms β 11.91 ms β no change β
β QQuery 2 β 37.23 ms β 36.95 ms β no change β
β QQuery 3 β 31.76 ms β 31.28 ms β no change β
β QQuery 4 β 190.96 ms β 197.12 ms β no change β
β QQuery 5 β 292.57 ms β 298.06 ms β no change β
β QQuery 6 β 1.32 ms β 1.28 ms β no change β
β QQuery 7 β 13.58 ms β 13.22 ms β no change β
β QQuery 8 β 299.24 ms β 306.37 ms β no change β
β QQuery 9 β 506.42 ms β 472.13 ms β +1.07x faster β
β QQuery 10 β 70.27 ms β 80.26 ms β 1.14x slower β
β QQuery 11 β 82.95 ms β 89.70 ms β 1.08x slower β
β QQuery 12 β 286.74 ms β 290.10 ms β no change β
β QQuery 13 β 399.56 ms β 491.42 ms β 1.23x slower β
β QQuery 14 β 308.08 ms β 303.53 ms β no change β
β QQuery 15 β 234.78 ms β 232.97 ms β no change β
β QQuery 16 β 699.72 ms β 650.68 ms β +1.08x faster β
β QQuery 17 β 633.48 ms β 598.48 ms β +1.06x faster β
β QQuery 18 β 1245.88 ms β 1224.17 ms β no change β
β QQuery 19 β 30.17 ms β 28.11 ms β +1.07x faster β
β QQuery 20 β 540.14 ms β 524.01 ms β no change β
β QQuery 21 β 526.26 ms β 529.17 ms β no change β
β QQuery 22 β 1083.78 ms β 1034.20 ms β no change β
β QQuery 23 β 3211.07 ms β 3222.59 ms β no change β
β QQuery 24 β 42.66 ms β 41.07 ms β no change β
β QQuery 25 β 107.39 ms β 105.34 ms β no change β
β QQuery 26 β 40.85 ms β 40.67 ms β no change β
β QQuery 27 β 516.52 ms β 521.83 ms β no change β
β QQuery 28 β 2939.57 ms β 2991.33 ms β no change β
β QQuery 29 β 42.03 ms β 42.20 ms β no change β
β QQuery 30 β 295.04 ms β 312.12 ms β 1.06x slower β
β QQuery 31 β 256.19 ms β 270.22 ms β 1.05x slower β
β QQuery 32 β 777.89 ms β 764.73 ms β no change β
β QQuery 33 β 1597.34 ms β 1584.53 ms β no change β
β QQuery 34 β 1554.07 ms β 1511.35 ms β no change β
β QQuery 35 β 228.66 ms β 323.22 ms β 1.41x slower β
β QQuery 36 β 65.34 ms β 66.13 ms β no change β
β QQuery 37 β 36.30 ms β 35.77 ms β no change β
β QQuery 38 β 42.92 ms β 41.85 ms β no change β
β QQuery 39 β 151.97 ms β 143.36 ms β +1.06x faster β
β QQuery 40 β 14.84 ms β 14.63 ms β no change β
β QQuery 41 β 14.08 ms β 13.94 ms β no change β
β QQuery 42 β 13.36 ms β 13.60 ms β no change β
βββββββββββββ΄βββββββββββββ΄βββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 19476.27ms β
β Total Time (codex_frozen-final-aggregation-poc) β 19506.83ms β
β Average Time (HEAD) β 452.94ms β
β Average Time (codex_frozen-final-aggregation-poc) β 453.65ms β
β Queries Faster β 5 β
β Queries Slower β 6 β
β Queries with No Change β 32 β
β Queries with Failure β 0 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββ
Distribution per query (min / mean Β±stddev / max):
Comparing HEAD and codex_frozen-final-aggregation-poc
--------------------
Benchmark clickbench_partitioned.json
--------------------
βββββββββββββ³βββββββββββββββββββββββββββββββββββββββββ³βββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β codex_frozen-final-aggregation-poc β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 0 β 1.29 / 4.23 Β±5.71 / 15.64 ms β 1.25 / 4.11 Β±5.58 / 15.26 ms β no change β
β QQuery 1 β 12.01 / 12.33 Β±0.28 / 12.78 ms β 11.91 / 12.12 Β±0.11 / 12.23 ms β no change β
β QQuery 2 β 37.23 / 37.82 Β±0.50 / 38.62 ms β 36.95 / 37.07 Β±0.09 / 37.16 ms β no change β
β QQuery 3 β 31.76 / 32.51 Β±0.55 / 33.38 ms β 31.28 / 32.08 Β±0.53 / 32.89 ms β no change β
β QQuery 4 β 190.96 / 194.02 Β±3.23 / 199.46 ms β 197.12 / 199.73 Β±1.76 / 202.12 ms β no change β
β QQuery 5 β 292.57 / 301.02 Β±13.01 / 326.93 ms β 298.06 / 306.14 Β±6.80 / 314.69 ms β no change β
β QQuery 6 β 1.32 / 1.46 Β±0.23 / 1.92 ms β 1.28 / 1.43 Β±0.24 / 1.90 ms β no change β
β QQuery 7 β 13.58 / 13.65 Β±0.07 / 13.76 ms β 13.22 / 13.40 Β±0.10 / 13.50 ms β no change β
β QQuery 8 β 299.24 / 305.62 Β±5.43 / 314.41 ms β 306.37 / 313.38 Β±4.14 / 317.72 ms β no change β
β QQuery 9 β 506.42 / 513.00 Β±8.53 / 529.59 ms β 472.13 / 493.96 Β±15.06 / 513.18 ms β no change β
β QQuery 10 β 70.27 / 71.65 Β±0.84 / 72.89 ms β 80.26 / 81.52 Β±1.89 / 85.28 ms β 1.14x slower β
β QQuery 11 β 82.95 / 87.95 Β±5.21 / 94.88 ms β 89.70 / 98.54 Β±14.12 / 126.68 ms β 1.12x slower β
β QQuery 12 β 286.74 / 302.78 Β±9.83 / 315.96 ms β 290.10 / 298.94 Β±6.64 / 305.49 ms β no change β
β QQuery 13 β 399.56 / 413.06 Β±9.80 / 429.19 ms β 491.42 / 497.28 Β±3.15 / 500.43 ms β 1.20x slower β
β QQuery 14 β 308.08 / 330.68 Β±14.03 / 350.19 ms β 303.53 / 308.55 Β±3.79 / 313.71 ms β +1.07x faster β
β QQuery 15 β 234.78 / 249.40 Β±19.73 / 287.23 ms β 232.97 / 235.22 Β±2.04 / 238.87 ms β +1.06x faster β
β QQuery 16 β 699.72 / 729.34 Β±30.96 / 786.64 ms β 650.68 / 664.97 Β±9.06 / 677.00 ms β +1.10x faster β
β QQuery 17 β 633.48 / 645.92 Β±9.70 / 657.25 ms β 598.48 / 620.79 Β±21.28 / 658.40 ms β no change β
β QQuery 18 β 1245.88 / 1301.28 Β±42.61 / 1346.64 ms β 1224.17 / 1278.20 Β±45.39 / 1353.18 ms β no change β
β QQuery 19 β 30.17 / 44.03 Β±18.38 / 76.93 ms β 28.11 / 44.79 Β±32.82 / 110.43 ms β no change β
β QQuery 20 β 540.14 / 580.03 Β±57.12 / 692.44 ms β 524.01 / 546.93 Β±29.47 / 603.68 ms β +1.06x faster β
β QQuery 21 β 526.26 / 546.92 Β±14.88 / 571.00 ms β 529.17 / 536.72 Β±5.84 / 547.13 ms β no change β
β QQuery 22 β 1083.78 / 1124.26 Β±43.87 / 1201.87 ms β 1034.20 / 1070.92 Β±32.46 / 1119.71 ms β no change β
β QQuery 23 β 3211.07 / 3409.23 Β±139.97 / 3641.13 ms β 3222.59 / 3526.74 Β±233.58 / 3822.52 ms β no change β
β QQuery 24 β 42.66 / 129.96 Β±130.03 / 381.08 ms β 41.07 / 98.15 Β±88.14 / 269.58 ms β +1.32x faster β
β QQuery 25 β 107.39 / 109.49 Β±2.75 / 114.84 ms β 105.34 / 119.88 Β±26.66 / 173.16 ms β 1.09x slower β
β QQuery 26 β 40.85 / 42.36 Β±0.95 / 43.44 ms β 40.67 / 42.23 Β±0.84 / 43.16 ms β no change β
β QQuery 27 β 516.52 / 529.72 Β±11.03 / 545.13 ms β 521.83 / 533.98 Β±11.44 / 553.89 ms β no change β
β QQuery 28 β 2939.57 / 2990.41 Β±41.65 / 3058.92 ms β 2991.33 / 3056.21 Β±42.80 / 3108.24 ms β no change β
β QQuery 29 β 42.03 / 59.93 Β±26.29 / 111.06 ms β 42.20 / 69.94 Β±29.11 / 113.26 ms β 1.17x slower β
β QQuery 30 β 295.04 / 313.24 Β±16.04 / 337.79 ms β 312.12 / 321.76 Β±8.07 / 331.75 ms β no change β
β QQuery 31 β 256.19 / 265.73 Β±5.67 / 273.82 ms β 270.22 / 279.05 Β±8.57 / 294.15 ms β 1.05x slower β
β QQuery 32 β 777.89 / 793.33 Β±16.78 / 825.30 ms β 764.73 / 847.81 Β±107.07 / 1056.57 ms β 1.07x slower β
β QQuery 33 β 1597.34 / 1653.29 Β±77.97 / 1807.76 ms β 1584.53 / 1625.99 Β±35.63 / 1671.49 ms β no change β
β QQuery 34 β 1554.07 / 1664.90 Β±149.95 / 1958.94 ms β 1511.35 / 1673.58 Β±133.54 / 1894.95 ms β no change β
β QQuery 35 β 228.66 / 331.76 Β±196.77 / 725.23 ms β 323.22 / 338.63 Β±26.00 / 390.45 ms β no change β
β QQuery 36 β 65.34 / 82.94 Β±31.49 / 145.82 ms β 66.13 / 70.79 Β±3.74 / 75.22 ms β +1.17x faster β
β QQuery 37 β 36.30 / 37.54 Β±0.93 / 38.79 ms β 35.77 / 39.54 Β±3.31 / 45.10 ms β 1.05x slower β
β QQuery 38 β 42.92 / 47.61 Β±4.63 / 56.22 ms β 41.85 / 44.84 Β±1.79 / 47.00 ms β +1.06x faster β
β QQuery 39 β 151.97 / 157.98 Β±8.20 / 174.23 ms β 143.36 / 159.78 Β±11.87 / 176.64 ms β no change β
β QQuery 40 β 14.84 / 15.41 Β±0.44 / 16.19 ms β 14.63 / 15.04 Β±0.38 / 15.73 ms β no change β
β QQuery 41 β 14.08 / 15.40 Β±2.46 / 20.33 ms β 13.94 / 14.32 Β±0.33 / 14.86 ms β +1.08x faster β
β QQuery 42 β 13.36 / 15.01 Β±2.40 / 19.79 ms β 13.60 / 19.58 Β±11.69 / 42.96 ms β 1.30x slower β
βββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
βββββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 20508.21ms β
β Total Time (codex_frozen-final-aggregation-poc) β 20594.61ms β
β Average Time (HEAD) β 476.94ms β
β Average Time (codex_frozen-final-aggregation-poc) β 478.94ms β
β Queries Faster β 8 β
β Queries Slower β 9 β
β Queries with No Change β 26 β
β Queries with Failure β 0 β
βββββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**clickbench_partitioned β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 105.0s |
| Peak memory | 18.1 GiB |
| Avg memory | 5.5 GiB |
| CPU user | 971.1s |
| CPU sys | 100.6s |
| Peak spill | 0 B |
**clickbench_partitioned β branch**
| Metric | Value |
|--------|-------|
| Wall time | 105.0s |
| Peak memory | 18.0 GiB |
| Avg memory | 5.6 GiB |
| CPU user | 984.7s |
| CPU sys | 98.8s |
| Peak spill | 0 B |
</details>
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] feat: prototype frozen final aggregation buckets [datafusion]
Posted by "Rachelint (via GitHub)" <gi...@apache.org>.
Rachelint closed pull request #25846: feat: prototype frozen final aggregation buckets
URL: https://github.com/apache/datafusion/pull/25846
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] feat: prototype frozen final aggregation buckets [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25846:
URL: https://github.com/apache/datafusion/pull/25846#issuecomment-5876843520
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/25846#issuecomment-5876802086)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c5876802086-2895-r72k7 6.12.94+ #1 SMP Wed Aug 19 07:47:20 UTC 2026 aarch64 GNU/Linux`
<details><summary>CPU Details (lscpu)</summary>
```
Architecture: aarch64
CPU op-mode(s): 64-bit
Byte Order: Little Endian
CPU(s): 16
On-line CPU(s) list: 0-15
Vendor ID: ARM
Model name: Neoverse-V2
Model: 1
Thread(s) per core: 1
Core(s) per cluster: 16
Socket(s): -
Cluster(s): 1
Stepping: r0p1
BogoMIPS: 2000.00
Flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache: 1 MiB (16 instances)
L1i cache: 1 MiB (16 instances)
L2 cache: 32 MiB (16 instances)
L3 cache: 80 MiB (1 instance)
NUMA node(s): 1
NUMA node0 CPU(s): 0-15
Vulnerability Gather data sampling: Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Mmio stale data: Not affected
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Not affected
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; __user pointer sanitization
Vulnerability Spectre v2: Mitigation; CSV2, BHB
Vulnerability Srbds: Not affected
Vulnerability Tsa: Not affected
Vulnerability Tsx async abort: Not affected
Vulnerability Vmscape: Not affected
```
</details>
Comparing codex/frozen-final-aggregation-poc (14346eba70d74cc5b90b5f7871a5bf3fedb78205) to fb42cf5 (merge-base) [diff](https://github.com/apache/datafusion/compare/fb42cf5911faf1c05f6ccd5bde3d9e2300d35b76..14346eba70d74cc5b90b5f7871a5bf3fedb78205)
<details><summary>Run configuration</summary>
```yaml
run benchmark clickbench_partitioned
changed:
env:
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_THRESHOLD: "262144"
```
</details>
Results will be posted here when complete
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] feat: prototype frozen final aggregation buckets [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25846:
URL: https://github.com/apache/datafusion/pull/25846#issuecomment-5876790717
Hi @Rachelint, your benchmark configuration could not be parsed (https://github.com/apache/datafusion/pull/25846#issuecomment-5876790170).
**Error:** `invalid configuration: unknown field `DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_THRESHOLD`, expected one of `env`, `baseline`, `changed`, `resources` at line 3 column 1`
Usage:
```
run benchmark <name> # run specific benchmark(s)
run benchmarks # run default suite
run benchmarks <name1> <name2> # run specific benchmarks
```
Any benchmark name is accepted: `bench.sh` suite names (e.g. `tpch`, `clickbench_partitioned`, `wide_schema`) and Criterion bench targets (e.g. `sql_planner`) are resolved automatically. A name that matches neither fails on the runner.
Per-side configuration (`run benchmark tpch` followed by):
```yaml
env:
# shared env is inherited by BOTH the build and the run, so build
# flags go here. Builds default to no debuginfo for speed; opt back
# in for hung-job gdb dumps and cap jobs to stay within memory:
CARGO_PROFILE_RELEASE_DEBUG: "1"
CARGO_BUILD_JOBS: "1"
baseline:
ref: v45.0.0
env:
# per-side env only reaches the benchmark run, not the build
DATAFUSION_RUNTIME_MEMORY_LIMIT: 1G
changed:
ref: v46.0.0
env:
DATAFUSION_RUNTIME_MEMORY_LIMIT: 2G
```
Pod size and CPU architecture (each key is optional and keeps the default when absent; requests and limits are set equal):
```yaml
resources:
cpu: "16"
memory: "128Gi"
arch: arm64 # arm64 or amd64
```
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] feat: prototype frozen final aggregation buckets [datafusion]
Posted by "Rachelint (via GitHub)" <gi...@apache.org>.
Rachelint commented on PR #25846:
URL: https://github.com/apache/datafusion/pull/25846#issuecomment-5876790170
run benchmark clickbench_partitioned
changed:
env:
DATAFUSION_EXECUTION_HASH_AGGREGATE_BUCKET_THRESHOLD: "262144"
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org