You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@datafusion.apache.org by "rluvaton (via GitHub)" <gi...@apache.org> on 2026/10/22 10:10:17 UTC
[PR] perf: try bucket sort [datafusion]
rluvaton opened a new pull request, #26587:
URL: https://github.com/apache/datafusion/pull/26587
(no comment)
--
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] perf: try bucket sort [datafusion]
Posted by "rluvaton (via GitHub)" <gi...@apache.org>.
rluvaton commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6339561598
run benchmark external_aggr
--
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] perf: try bucket sort [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6338475960
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/26587#issuecomment-6338228159)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing sort-bucket (6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0) to e12901d (merge-base) [diff](https://github.com/apache/datafusion/compare/e12901dadfd462231a093627449d3a7b51d45962..6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpcds
env:
prefer_hash_join: "false"
```
</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 sort-bucket
--------------------
Benchmark tpcds_sf1.json
--------------------
βββββββββββββ³βββββββββββββ³ββββββββββββββ³ββββββββββββββββ
β Query β HEAD β sort-bucket β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 5.23 ms β 5.30 ms β no change β
β QQuery 2 β 80.47 ms β 79.74 ms β no change β
β QQuery 3 β 28.46 ms β 28.67 ms β no change β
β QQuery 4 β 440.05 ms β 458.13 ms β no change β
β QQuery 5 β 50.48 ms β 49.43 ms β no change β
β QQuery 6 β 36.60 ms β 34.69 ms β +1.06x faster β
β QQuery 7 β 74.04 ms β 71.82 ms β no change β
β QQuery 8 β 36.43 ms β 36.07 ms β no change β
β QQuery 9 β 53.76 ms β 52.35 ms β no change β
β QQuery 10 β 61.87 ms β 60.03 ms β no change β
β QQuery 11 β 296.37 ms β 279.51 ms β +1.06x faster β
β QQuery 12 β 28.89 ms β 28.13 ms β no change β
β QQuery 13 β 114.96 ms β 112.00 ms β no change β
β QQuery 14 β 396.76 ms β 378.13 ms β no change β
β QQuery 15 β 60.59 ms β 55.24 ms β +1.10x faster β
β QQuery 16 β 6.41 ms β 6.32 ms β no change β
β QQuery 17 β 80.87 ms β 79.37 ms β no change β
β QQuery 18 β 108.13 ms β 105.48 ms β no change β
β QQuery 19 β 41.24 ms β 41.25 ms β no change β
β QQuery 20 β 35.93 ms β 35.49 ms β no change β
β QQuery 21 β 17.47 ms β 16.97 ms β no change β
β QQuery 22 β 65.40 ms β 63.99 ms β no change β
β QQuery 23 β 314.40 ms β 303.60 ms β no change β
β QQuery 24 β 201.32 ms β 198.38 ms β no change β
β QQuery 25 β 107.67 ms β 106.99 ms β no change β
β QQuery 26 β 46.01 ms β 45.52 ms β no change β
β QQuery 27 β 5.93 ms β 5.93 ms β no change β
β QQuery 28 β 55.60 ms β 55.54 ms β no change β
β QQuery 29 β 95.55 ms β 94.04 ms β no change β
β QQuery 30 β 32.47 ms β 31.96 ms β no change β
β QQuery 31 β 107.57 ms β 108.74 ms β no change β
β QQuery 32 β 19.91 ms β 20.59 ms β no change β
β QQuery 33 β 36.52 ms β 36.56 ms β no change β
β QQuery 34 β 9.96 ms β 10.35 ms β no change β
β QQuery 35 β 71.31 ms β 71.14 ms β no change β
β QQuery 36 β 6.04 ms β 6.12 ms β no change β
β QQuery 37 β 7.06 ms β 6.95 ms β no change β
β QQuery 38 β 60.64 ms β 61.15 ms β no change β
β QQuery 39 β 72.26 ms β 72.01 ms β no change β
β QQuery 40 β 23.36 ms β 23.57 ms β no change β
β QQuery 41 β 12.31 ms β 12.55 ms β no change β
β QQuery 42 β 23.75 ms β 23.82 ms β no change β
β QQuery 43 β 4.70 ms β 4.76 ms β no change β
β QQuery 44 β 8.90 ms β 8.69 ms β no change β
β QQuery 45 β 39.10 ms β 36.18 ms β +1.08x faster β
β QQuery 46 β 11.79 ms β 11.85 ms β no change β
β QQuery 47 β 234.88 ms β 228.10 ms β no change β
β QQuery 48 β 91.52 ms β 89.53 ms β no change β
β QQuery 49 β 70.28 ms β 68.16 ms β no change β
β QQuery 50 β 58.24 ms β 56.87 ms β no change β
β QQuery 51 β 91.30 ms β 118.62 ms β 1.30x slower β
β QQuery 52 β 24.38 ms β 24.41 ms β no change β
β QQuery 53 β 29.36 ms β 29.28 ms β no change β
β QQuery 54 β 53.63 ms β 53.40 ms β no change β
β QQuery 55 β 24.03 ms β 24.11 ms β no change β
β QQuery 56 β 38.24 ms β 37.90 ms β no change β
β QQuery 57 β 164.43 ms β 164.56 ms β no change β
β QQuery 58 β 106.10 ms β 105.30 ms β no change β
β QQuery 59 β 113.81 ms β 112.88 ms β no change β
β QQuery 60 β 37.17 ms β 38.32 ms β no change β
β QQuery 61 β 10.59 ms β 10.50 ms β no change β
β QQuery 62 β 44.09 ms β 43.25 ms β no change β
β QQuery 63 β 29.39 ms β 29.26 ms β no change β
β QQuery 64 β 357.89 ms β 351.28 ms β no change β
β QQuery 65 β 127.53 ms β 126.55 ms β no change β
β QQuery 66 β 76.48 ms β 75.16 ms β no change β
β QQuery 67 β 258.86 ms β 291.47 ms β 1.13x slower β
β QQuery 68 β 11.99 ms β 11.85 ms β no change β
β QQuery 69 β 56.07 ms β 56.21 ms β no change β
β QQuery 70 β 103.08 ms β 102.60 ms β no change β
β QQuery 71 β 34.89 ms β 35.03 ms β no change β
β QQuery 72 β 1447.17 ms β 1627.64 ms β 1.12x slower β
β QQuery 73 β 10.25 ms β 10.14 ms β no change β
β QQuery 74 β 167.94 ms β 165.50 ms β no change β
β QQuery 75 β 139.63 ms β 138.15 ms β no change β
β QQuery 76 β 34.57 ms β 34.09 ms β no change β
β QQuery 77 β 59.79 ms β 59.45 ms β no change β
β QQuery 78 β 164.05 ms β 160.89 ms β no change β
β QQuery 79 β 66.25 ms β 65.99 ms β no change β
β QQuery 80 β 94.99 ms β 93.68 ms β no change β
β QQuery 81 β 25.55 ms β 25.19 ms β no change β
β QQuery 82 β 16.47 ms β 16.20 ms β no change β
β QQuery 83 β 32.78 ms β 32.20 ms β no change β
β QQuery 84 β 29.65 ms β 29.95 ms β no change β
β QQuery 85 β 99.36 ms β 97.38 ms β no change β
β QQuery 86 β 25.67 ms β 25.45 ms β no change β
β QQuery 87 β 61.04 ms β 61.64 ms β no change β
β QQuery 88 β 58.36 ms β 58.35 ms β no change β
β QQuery 89 β 35.47 ms β 34.79 ms β no change β
β QQuery 90 β 16.59 ms β 16.37 ms β no change β
β QQuery 91 β 41.68 ms β 40.42 ms β no change β
β QQuery 92 β 29.65 ms β 29.20 ms β no change β
β QQuery 93 β 48.48 ms β 47.91 ms β no change β
β QQuery 94 β 38.13 ms β 38.09 ms β no change β
β QQuery 95 β 80.52 ms β 78.60 ms β no change β
β QQuery 96 β 23.79 ms β 23.73 ms β no change β
β QQuery 97 β 50.09 ms β 50.57 ms β no change β
β QQuery 98 β 41.89 ms β 42.91 ms β no change β
β QQuery 99 β 64.83 ms β 63.90 ms β no change β
βββββββββββββ΄βββββββββββββ΄ββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 8747.47ms β
β Total Time (sort-bucket) β 8888.09ms β
β Average Time (HEAD) β 88.36ms β
β Average Time (sort-bucket) β 89.78ms β
β Queries Faster β 4 β
β Queries Slower β 3 β
β Queries with No Change β 92 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββ΄ββββββββββββ
Distribution per query (min / mean Β±stddev / max):
Comparing HEAD and sort-bucket
--------------------
Benchmark tpcds_sf1.json
--------------------
βββββββββββββ³ββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β sort-bucket β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 5.23 / 5.85 Β±1.07 / 7.99 ms β 5.30 / 5.94 Β±1.03 / 7.99 ms β no change β
β QQuery 2 β 80.47 / 80.75 Β±0.23 / 81.14 ms β 79.74 / 79.99 Β±0.26 / 80.48 ms β no change β
β QQuery 3 β 28.46 / 28.84 Β±0.27 / 29.19 ms β 28.67 / 28.90 Β±0.13 / 29.08 ms β no change β
β QQuery 4 β 440.05 / 471.15 Β±25.21 / 503.15 ms β 458.13 / 462.79 Β±4.38 / 469.13 ms β no change β
β QQuery 5 β 50.48 / 52.75 Β±2.78 / 58.16 ms β 49.43 / 50.99 Β±2.60 / 56.19 ms β no change β
β QQuery 6 β 36.60 / 37.03 Β±0.37 / 37.63 ms β 34.69 / 35.14 Β±0.38 / 35.82 ms β +1.05x faster β
β QQuery 7 β 74.04 / 74.65 Β±0.64 / 75.77 ms β 71.82 / 72.27 Β±0.45 / 73.10 ms β no change β
β QQuery 8 β 36.43 / 36.89 Β±0.46 / 37.71 ms β 36.07 / 36.24 Β±0.13 / 36.40 ms β no change β
β QQuery 9 β 53.76 / 57.15 Β±4.39 / 65.67 ms β 52.35 / 53.41 Β±0.72 / 54.57 ms β +1.07x faster β
β QQuery 10 β 61.87 / 62.24 Β±0.38 / 62.89 ms β 60.03 / 60.44 Β±0.38 / 60.96 ms β no change β
β QQuery 11 β 296.37 / 306.17 Β±11.57 / 327.76 ms β 279.51 / 290.36 Β±7.47 / 301.34 ms β +1.05x faster β
β QQuery 12 β 28.89 / 29.24 Β±0.39 / 29.98 ms β 28.13 / 29.17 Β±0.78 / 30.11 ms β no change β
β QQuery 13 β 114.96 / 117.04 Β±1.20 / 118.45 ms β 112.00 / 112.87 Β±0.58 / 113.70 ms β no change β
β QQuery 14 β 396.76 / 399.84 Β±2.81 / 403.61 ms β 378.13 / 384.94 Β±4.39 / 389.93 ms β no change β
β QQuery 15 β 60.59 / 64.55 Β±2.76 / 69.20 ms β 55.24 / 55.80 Β±0.80 / 57.37 ms β +1.16x faster β
β QQuery 16 β 6.41 / 6.70 Β±0.23 / 7.05 ms β 6.32 / 6.45 Β±0.13 / 6.69 ms β no change β
β QQuery 17 β 80.87 / 82.53 Β±1.28 / 84.64 ms β 79.37 / 80.39 Β±1.52 / 83.37 ms β no change β
β QQuery 18 β 108.13 / 111.21 Β±2.30 / 114.04 ms β 105.48 / 106.32 Β±1.06 / 108.37 ms β no change β
β QQuery 19 β 41.24 / 41.78 Β±0.43 / 42.37 ms β 41.25 / 41.64 Β±0.36 / 42.30 ms β no change β
β QQuery 20 β 35.93 / 36.13 Β±0.17 / 36.42 ms β 35.49 / 36.16 Β±0.68 / 37.36 ms β no change β
β QQuery 21 β 17.47 / 17.53 Β±0.07 / 17.65 ms β 16.97 / 17.18 Β±0.16 / 17.38 ms β no change β
β QQuery 22 β 65.40 / 65.72 Β±0.31 / 66.30 ms β 63.99 / 65.57 Β±0.87 / 66.59 ms β no change β
β QQuery 23 β 314.40 / 316.17 Β±1.47 / 318.70 ms β 303.60 / 309.99 Β±3.37 / 313.64 ms β no change β
β QQuery 24 β 201.32 / 204.40 Β±5.18 / 214.71 ms β 198.38 / 202.57 Β±4.17 / 209.96 ms β no change β
β QQuery 25 β 107.67 / 110.06 Β±1.75 / 112.70 ms β 106.99 / 108.67 Β±1.02 / 109.73 ms β no change β
β QQuery 26 β 46.01 / 46.29 Β±0.23 / 46.67 ms β 45.52 / 46.24 Β±0.64 / 47.26 ms β no change β
β QQuery 27 β 5.93 / 6.07 Β±0.15 / 6.35 ms β 5.93 / 6.11 Β±0.18 / 6.38 ms β no change β
β QQuery 28 β 55.60 / 59.81 Β±3.28 / 63.24 ms β 55.54 / 58.34 Β±1.85 / 60.31 ms β no change β
β QQuery 29 β 95.55 / 98.52 Β±2.86 / 103.77 ms β 94.04 / 96.59 Β±2.13 / 100.36 ms β no change β
β QQuery 30 β 32.47 / 32.86 Β±0.31 / 33.32 ms β 31.96 / 32.53 Β±0.38 / 33.02 ms β no change β
β QQuery 31 β 107.57 / 110.24 Β±3.29 / 116.58 ms β 108.74 / 109.28 Β±0.40 / 109.85 ms β no change β
β QQuery 32 β 19.91 / 20.16 Β±0.17 / 20.40 ms β 20.59 / 21.93 Β±1.90 / 25.69 ms β 1.09x slower β
β QQuery 33 β 36.52 / 36.93 Β±0.44 / 37.64 ms β 36.56 / 37.37 Β±0.89 / 39.09 ms β no change β
β QQuery 34 β 9.96 / 10.17 Β±0.15 / 10.41 ms β 10.35 / 10.73 Β±0.37 / 11.42 ms β 1.05x slower β
β QQuery 35 β 71.31 / 73.93 Β±1.77 / 75.99 ms β 71.14 / 72.82 Β±1.29 / 74.83 ms β no change β
β QQuery 36 β 6.04 / 6.14 Β±0.11 / 6.35 ms β 6.12 / 6.29 Β±0.14 / 6.54 ms β no change β
β QQuery 37 β 7.06 / 7.21 Β±0.11 / 7.33 ms β 6.95 / 7.04 Β±0.06 / 7.14 ms β no change β
β QQuery 38 β 60.64 / 62.97 Β±1.44 / 64.68 ms β 61.15 / 62.55 Β±0.99 / 63.82 ms β no change β
β QQuery 39 β 72.26 / 73.03 Β±0.57 / 73.70 ms β 72.01 / 72.71 Β±0.50 / 73.44 ms β no change β
β QQuery 40 β 23.36 / 23.65 Β±0.22 / 23.85 ms β 23.57 / 23.84 Β±0.17 / 24.07 ms β no change β
β QQuery 41 β 12.31 / 12.46 Β±0.16 / 12.76 ms β 12.55 / 13.00 Β±0.67 / 14.32 ms β no change β
β QQuery 42 β 23.75 / 24.14 Β±0.45 / 24.97 ms β 23.82 / 24.27 Β±0.71 / 25.69 ms β no change β
β QQuery 43 β 4.70 / 4.92 Β±0.20 / 5.29 ms β 4.76 / 4.92 Β±0.22 / 5.35 ms β no change β
β QQuery 44 β 8.90 / 10.65 Β±2.82 / 16.26 ms β 8.69 / 8.84 Β±0.12 / 9.03 ms β +1.20x faster β
β QQuery 45 β 39.10 / 39.80 Β±0.42 / 40.42 ms β 36.18 / 37.14 Β±0.96 / 38.94 ms β +1.07x faster β
β QQuery 46 β 11.79 / 12.61 Β±0.48 / 13.20 ms β 11.85 / 13.05 Β±2.09 / 17.22 ms β no change β
β QQuery 47 β 234.88 / 240.05 Β±5.80 / 251.35 ms β 228.10 / 237.47 Β±6.14 / 245.56 ms β no change β
β QQuery 48 β 91.52 / 91.71 Β±0.16 / 91.97 ms β 89.53 / 89.94 Β±0.32 / 90.43 ms β no change β
β QQuery 49 β 70.28 / 72.83 Β±2.81 / 78.04 ms β 68.16 / 69.04 Β±0.67 / 70.22 ms β +1.05x faster β
β QQuery 50 β 58.24 / 58.79 Β±0.41 / 59.47 ms β 56.87 / 57.69 Β±0.63 / 58.65 ms β no change β
β QQuery 51 β 91.30 / 93.10 Β±2.07 / 97.03 ms β 118.62 / 120.20 Β±0.86 / 121.13 ms β 1.29x slower β
β QQuery 52 β 24.38 / 24.71 Β±0.43 / 25.51 ms β 24.41 / 25.80 Β±2.21 / 30.17 ms β no change β
β QQuery 53 β 29.36 / 29.70 Β±0.30 / 30.20 ms β 29.28 / 29.62 Β±0.46 / 30.52 ms β no change β
β QQuery 54 β 53.63 / 54.28 Β±0.76 / 55.70 ms β 53.40 / 54.10 Β±0.36 / 54.33 ms β no change β
β QQuery 55 β 24.03 / 24.20 Β±0.16 / 24.51 ms β 24.11 / 24.73 Β±0.66 / 25.90 ms β no change β
β QQuery 56 β 38.24 / 38.46 Β±0.17 / 38.69 ms β 37.90 / 38.37 Β±0.31 / 38.65 ms β no change β
β QQuery 57 β 164.43 / 168.51 Β±2.43 / 171.44 ms β 164.56 / 168.30 Β±2.67 / 172.79 ms β no change β
β QQuery 58 β 106.10 / 108.17 Β±2.62 / 113.32 ms β 105.30 / 106.50 Β±1.06 / 107.97 ms β no change β
β QQuery 59 β 113.81 / 114.35 Β±0.47 / 115.14 ms β 112.88 / 113.98 Β±0.70 / 115.07 ms β no change β
β QQuery 60 β 37.17 / 38.66 Β±1.84 / 41.77 ms β 38.32 / 38.54 Β±0.20 / 38.84 ms β no change β
β QQuery 61 β 10.59 / 10.71 Β±0.10 / 10.89 ms β 10.50 / 10.59 Β±0.08 / 10.72 ms β no change β
β QQuery 62 β 44.09 / 44.42 Β±0.35 / 44.94 ms β 43.25 / 44.51 Β±2.09 / 48.68 ms β no change β
β QQuery 63 β 29.39 / 30.53 Β±0.79 / 31.67 ms β 29.26 / 29.58 Β±0.47 / 30.51 ms β no change β
β QQuery 64 β 357.89 / 361.53 Β±2.07 / 364.00 ms β 351.28 / 356.38 Β±2.77 / 358.84 ms β no change β
β QQuery 65 β 127.53 / 132.72 Β±3.03 / 136.15 ms β 126.55 / 129.23 Β±1.62 / 131.41 ms β no change β
β QQuery 66 β 76.48 / 78.28 Β±1.91 / 80.81 ms β 75.16 / 76.09 Β±0.49 / 76.58 ms β no change β
β QQuery 67 β 258.86 / 263.41 Β±4.18 / 269.27 ms β 291.47 / 301.36 Β±6.85 / 312.13 ms β 1.14x slower β
β QQuery 68 β 11.99 / 12.17 Β±0.20 / 12.54 ms β 11.85 / 11.98 Β±0.10 / 12.16 ms β no change β
β QQuery 69 β 56.07 / 57.24 Β±0.73 / 58.21 ms β 56.21 / 58.34 Β±3.49 / 65.29 ms β no change β
β QQuery 70 β 103.08 / 105.89 Β±2.92 / 111.48 ms β 102.60 / 105.83 Β±3.36 / 111.78 ms β no change β
β QQuery 71 β 34.89 / 35.25 Β±0.40 / 35.92 ms β 35.03 / 35.50 Β±0.36 / 36.06 ms β no change β
β QQuery 72 β 1447.17 / 1569.82 Β±66.26 / 1621.99 ms β 1627.64 / 1685.08 Β±47.44 / 1755.76 ms β 1.07x slower β
β QQuery 73 β 10.25 / 10.72 Β±0.46 / 11.51 ms β 10.14 / 10.43 Β±0.34 / 11.09 ms β no change β
β QQuery 74 β 167.94 / 170.55 Β±3.19 / 176.61 ms β 165.50 / 168.47 Β±2.78 / 172.81 ms β no change β
β QQuery 75 β 139.63 / 143.53 Β±3.25 / 149.20 ms β 138.15 / 139.62 Β±1.00 / 141.28 ms β no change β
β QQuery 76 β 34.57 / 34.90 Β±0.41 / 35.67 ms β 34.09 / 34.38 Β±0.27 / 34.75 ms β no change β
β QQuery 77 β 59.79 / 60.31 Β±0.50 / 61.22 ms β 59.45 / 62.13 Β±2.87 / 65.84 ms β no change β
β QQuery 78 β 164.05 / 166.92 Β±1.52 / 168.47 ms β 160.89 / 163.10 Β±1.81 / 165.90 ms β no change β
β QQuery 79 β 66.25 / 67.35 Β±1.39 / 70.09 ms β 65.99 / 67.14 Β±1.50 / 70.06 ms β no change β
β QQuery 80 β 94.99 / 98.78 Β±6.33 / 111.37 ms β 93.68 / 94.28 Β±0.58 / 95.36 ms β no change β
β QQuery 81 β 25.55 / 25.82 Β±0.16 / 26.04 ms β 25.19 / 25.39 Β±0.21 / 25.80 ms β no change β
β QQuery 82 β 16.47 / 17.16 Β±0.93 / 18.99 ms β 16.20 / 19.18 Β±4.39 / 27.83 ms β 1.12x slower β
β QQuery 83 β 32.78 / 33.23 Β±0.30 / 33.60 ms β 32.20 / 33.61 Β±2.28 / 38.15 ms β no change β
β QQuery 84 β 29.65 / 30.18 Β±0.87 / 31.91 ms β 29.95 / 30.52 Β±0.54 / 31.43 ms β no change β
β QQuery 85 β 99.36 / 102.46 Β±3.05 / 108.02 ms β 97.38 / 98.60 Β±0.79 / 99.44 ms β no change β
β QQuery 86 β 25.67 / 26.37 Β±0.66 / 27.59 ms β 25.45 / 25.62 Β±0.16 / 25.91 ms β no change β
β QQuery 87 β 61.04 / 62.79 Β±1.09 / 64.21 ms β 61.64 / 63.94 Β±2.29 / 67.81 ms β no change β
β QQuery 88 β 58.36 / 60.71 Β±3.67 / 67.93 ms β 58.35 / 59.02 Β±0.50 / 59.57 ms β no change β
β QQuery 89 β 35.47 / 35.83 Β±0.36 / 36.50 ms β 34.79 / 35.33 Β±0.63 / 36.54 ms β no change β
β QQuery 90 β 16.59 / 16.68 Β±0.06 / 16.76 ms β 16.37 / 16.55 Β±0.15 / 16.79 ms β no change β
β QQuery 91 β 41.68 / 42.43 Β±0.59 / 43.18 ms β 40.42 / 40.99 Β±0.35 / 41.40 ms β no change β
β QQuery 92 β 29.65 / 30.21 Β±0.59 / 31.02 ms β 29.20 / 30.63 Β±1.17 / 32.40 ms β no change β
β QQuery 93 β 48.48 / 49.40 Β±0.67 / 50.12 ms β 47.91 / 49.56 Β±1.20 / 51.63 ms β no change β
β QQuery 94 β 38.13 / 40.22 Β±2.65 / 45.42 ms β 38.09 / 39.00 Β±0.86 / 40.64 ms β no change β
β QQuery 95 β 80.52 / 81.51 Β±0.96 / 82.74 ms β 78.60 / 80.09 Β±0.88 / 81.05 ms β no change β
β QQuery 96 β 23.79 / 24.14 Β±0.22 / 24.39 ms β 23.73 / 24.82 Β±1.83 / 28.48 ms β no change β
β QQuery 97 β 50.09 / 51.10 Β±1.03 / 53.08 ms β 50.57 / 52.65 Β±1.65 / 55.54 ms β no change β
β QQuery 98 β 41.89 / 43.80 Β±1.86 / 47.36 ms β 42.91 / 43.57 Β±0.35 / 43.89 ms β no change β
β QQuery 99 β 64.83 / 65.46 Β±0.48 / 65.95 ms β 63.90 / 64.52 Β±0.45 / 65.19 ms β no change β
βββββββββββββ΄ββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 9040.97ms β
β Total Time (sort-bucket) β 9091.66ms β
β Average Time (HEAD) β 91.32ms β
β Average Time (sort-bucket) β 91.83ms β
β Queries Faster β 7 β
β Queries Slower β 6 β
β Queries with No Change β 86 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββ΄ββββββββββββ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**tpcds β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 50.0s |
| Peak memory | 2.2 GiB |
| Avg memory | 1.4 GiB |
| CPU user | 188.8s |
| CPU sys | 5.7s |
| Peak spill | 0 B |
**tpcds β branch**
| Metric | Value |
|--------|-------|
| Wall time | 50.0s |
| Peak memory | 2.4 GiB |
| Avg memory | 1.6 GiB |
| CPU user | 194.4s |
| CPU sys | 5.6s |
| 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] perf: try bucket sort [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6338466816
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/26587#issuecomment-6338228159)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing sort-bucket (6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0) to e12901d (merge-base) [diff](https://github.com/apache/datafusion/compare/e12901dadfd462231a093627449d3a7b51d45962..6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpch10
env:
prefer_hash_join: "false"
```
</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 sort-bucket
--------------------
Benchmark tpch_sf10.json
--------------------
βββββββββββββ³ββββββββββββ³ββββββββββββββ³ββββββββββββ
β Query β HEAD β sort-bucket β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 330.89 ms β 327.63 ms β no change β
β QQuery 2 β 90.44 ms β 88.49 ms β no change β
β QQuery 3 β 216.01 ms β 215.02 ms β no change β
β QQuery 4 β 113.30 ms β 113.15 ms β no change β
β QQuery 5 β 343.97 ms β 344.15 ms β no change β
β QQuery 6 β 99.84 ms β 100.71 ms β no change β
β QQuery 7 β 451.96 ms β 444.10 ms β no change β
β QQuery 8 β 354.08 ms β 354.93 ms β no change β
β QQuery 9 β 509.09 ms β 511.98 ms β no change β
β QQuery 10 β 285.47 ms β 281.40 ms β no change β
β QQuery 11 β 61.77 ms β 61.71 ms β no change β
β QQuery 12 β 117.71 ms β 117.09 ms β no change β
β QQuery 13 β 294.59 ms β 299.82 ms β no change β
β QQuery 14 β 168.69 ms β 170.39 ms β no change β
β QQuery 15 β 298.00 ms β 299.91 ms β no change β
β QQuery 16 β 63.15 ms β 66.01 ms β no change β
β QQuery 17 β 578.33 ms β 574.42 ms β no change β
β QQuery 18 β 719.77 ms β 702.98 ms β no change β
β QQuery 19 β 233.06 ms β 229.91 ms β no change β
β QQuery 20 β 265.78 ms β 265.56 ms β no change β
β QQuery 21 β 656.19 ms β 645.64 ms β no change β
β QQuery 22 β 58.90 ms β 59.69 ms β no change β
βββββββββββββ΄ββββββββββββ΄ββββββββββββββ΄ββββββββββββ
ββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 6310.97ms β
β Total Time (sort-bucket) β 6274.68ms β
β Average Time (HEAD) β 286.86ms β
β Average Time (sort-bucket) β 285.21ms β
β Queries Faster β 0 β
β Queries Slower β 0 β
β Queries with No Change β 22 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββ΄ββββββββββββ
Distribution per query (min / mean Β±stddev / max):
Comparing HEAD and sort-bucket
--------------------
Benchmark tpch_sf10.json
--------------------
βββββββββββββ³βββββββββββββββββββββββββββββββββββββ³βββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β sort-bucket β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 330.89 / 332.05 Β±0.85 / 333.51 ms β 327.63 / 329.66 Β±1.18 / 331.12 ms β no change β
β QQuery 2 β 90.44 / 91.68 Β±0.90 / 92.78 ms β 88.49 / 90.97 Β±2.54 / 94.24 ms β no change β
β QQuery 3 β 216.01 / 220.32 Β±2.85 / 224.72 ms β 215.02 / 217.86 Β±1.79 / 220.38 ms β no change β
β QQuery 4 β 113.30 / 115.06 Β±1.73 / 117.95 ms β 113.15 / 114.49 Β±1.30 / 116.78 ms β no change β
β QQuery 5 β 343.97 / 354.78 Β±6.17 / 361.30 ms β 344.15 / 351.37 Β±6.81 / 362.71 ms β no change β
β QQuery 6 β 99.84 / 100.45 Β±0.63 / 101.56 ms β 100.71 / 102.98 Β±3.27 / 109.42 ms β no change β
β QQuery 7 β 451.96 / 457.39 Β±3.30 / 462.01 ms β 444.10 / 457.85 Β±8.34 / 469.53 ms β no change β
β QQuery 8 β 354.08 / 359.44 Β±3.22 / 363.81 ms β 354.93 / 363.06 Β±5.90 / 372.46 ms β no change β
β QQuery 9 β 509.09 / 515.75 Β±4.38 / 520.80 ms β 511.98 / 519.18 Β±8.93 / 535.38 ms β no change β
β QQuery 10 β 285.47 / 299.40 Β±10.71 / 310.66 ms β 281.40 / 290.33 Β±7.34 / 303.21 ms β no change β
β QQuery 11 β 61.77 / 63.01 Β±0.82 / 64.02 ms β 61.71 / 63.63 Β±1.50 / 65.63 ms β no change β
β QQuery 12 β 117.71 / 119.85 Β±1.66 / 122.45 ms β 117.09 / 124.07 Β±9.42 / 142.35 ms β no change β
β QQuery 13 β 294.59 / 307.96 Β±12.92 / 329.70 ms β 299.82 / 306.48 Β±4.36 / 312.37 ms β no change β
β QQuery 14 β 168.69 / 174.68 Β±5.44 / 182.06 ms β 170.39 / 173.50 Β±4.07 / 181.54 ms β no change β
β QQuery 15 β 298.00 / 300.95 Β±3.10 / 305.30 ms β 299.91 / 301.94 Β±2.20 / 305.83 ms β no change β
β QQuery 16 β 63.15 / 65.67 Β±1.64 / 68.25 ms β 66.01 / 66.81 Β±1.02 / 68.69 ms β no change β
β QQuery 17 β 578.33 / 587.66 Β±6.28 / 596.74 ms β 574.42 / 584.32 Β±8.41 / 598.79 ms β no change β
β QQuery 18 β 719.77 / 738.42 Β±21.28 / 774.63 ms β 702.98 / 725.00 Β±22.41 / 763.92 ms β no change β
β QQuery 19 β 233.06 / 247.20 Β±14.00 / 272.36 ms β 229.91 / 243.12 Β±15.02 / 265.98 ms β no change β
β QQuery 20 β 265.78 / 284.09 Β±14.81 / 310.07 ms β 265.56 / 275.31 Β±6.74 / 285.82 ms β no change β
β QQuery 21 β 656.19 / 661.19 Β±3.13 / 664.79 ms β 645.64 / 655.46 Β±8.46 / 670.07 ms β no change β
β QQuery 22 β 58.90 / 65.34 Β±5.44 / 71.84 ms β 59.69 / 61.65 Β±1.83 / 64.30 ms β +1.06x faster β
βββββββββββββ΄βββββββββββββββββββββββββββββββββββββ΄βββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 6462.35ms β
β Total Time (sort-bucket) β 6419.05ms β
β Average Time (HEAD) β 293.74ms β
β Average Time (sort-bucket) β 291.77ms β
β Queries Faster β 1 β
β Queries Slower β 0 β
β Queries with No Change β 21 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββ΄ββββββββββββ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**tpch10 β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 35.0s |
| Peak memory | 5.2 GiB |
| Avg memory | 1.6 GiB |
| CPU user | 328.2s |
| CPU sys | 19.0s |
| Peak spill | 0 B |
**tpch10 β branch**
| Metric | Value |
|--------|-------|
| Wall time | 35.0s |
| Peak memory | 5.2 GiB |
| Avg memory | 1.6 GiB |
| CPU user | 326.7s |
| CPU sys | 19.4s |
| 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] perf: try bucket sort [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6338460531
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/26587#issuecomment-6338228159)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing sort-bucket (6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0) to e12901d (merge-base) [diff](https://github.com/apache/datafusion/compare/e12901dadfd462231a093627449d3a7b51d45962..6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0)
<details><summary>Run configuration</summary>
```yaml
run benchmark sort_tpch
env:
prefer_hash_join: "false"
```
</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 sort-bucket
--------------------
Benchmark sort_tpch1.json
--------------------
βββββββββ³ββββββββββββ³ββββββββββββββ³ββββββββββββββββ
β Query β HEAD β sort-bucket β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Q1 β 113.02 ms β 124.13 ms β 1.10x slower β
β Q2 β 100.20 ms β 89.93 ms β +1.11x faster β
β Q3 β 657.58 ms β 660.68 ms β no change β
β Q4 β 162.80 ms β 133.46 ms β +1.22x faster β
β Q5 β 235.50 ms β 221.85 ms β +1.06x faster β
β Q6 β 248.87 ms β 238.01 ms β no change β
β Q7 β 423.51 ms β 446.88 ms β 1.06x slower β
β Q8 β 299.85 ms β 729.86 ms β 2.43x slower β
β Q9 β 319.20 ms β 761.06 ms β 2.38x slower β
β Q10 β 491.89 ms β 954.17 ms β 1.94x slower β
β Q11 β 248.80 ms β 241.57 ms β no change β
βββββββββ΄ββββββββββββ΄ββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 3301.23ms β
β Total Time (sort-bucket) β 4601.61ms β
β Average Time (HEAD) β 300.11ms β
β Average Time (sort-bucket) β 418.33ms β
β Queries Faster β 3 β
β Queries Slower β 5 β
β Queries with No Change β 3 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββ΄ββββββββββββ
Distribution per query (min / mean Β±stddev / max):
Comparing HEAD and sort-bucket
--------------------
Benchmark sort_tpch1.json
--------------------
βββββββββ³βββββββββββββββββββββββββββββββββββββ³ββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β sort-bucket β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Q1 β 113.02 / 114.21 Β±1.20 / 116.40 ms β 124.13 / 125.18 Β±1.08 / 127.26 ms β 1.10x slower β
β Q2 β 100.20 / 100.78 Β±0.86 / 102.48 ms β 89.93 / 90.91 Β±1.42 / 93.73 ms β +1.11x faster β
β Q3 β 657.58 / 662.82 Β±2.96 / 666.44 ms β 660.68 / 665.61 Β±3.80 / 670.03 ms β no change β
β Q4 β 162.80 / 163.58 Β±0.73 / 164.63 ms β 133.46 / 135.85 Β±3.19 / 142.15 ms β +1.20x faster β
β Q5 β 235.50 / 237.74 Β±2.12 / 241.67 ms β 221.85 / 223.44 Β±1.54 / 225.77 ms β +1.06x faster β
β Q6 β 248.87 / 251.77 Β±1.48 / 252.91 ms β 238.01 / 239.37 Β±0.90 / 240.26 ms β no change β
β Q7 β 423.51 / 425.37 Β±1.71 / 428.52 ms β 446.88 / 450.02 Β±3.96 / 457.50 ms β 1.06x slower β
β Q8 β 299.85 / 306.24 Β±6.93 / 317.44 ms β 729.86 / 748.24 Β±13.03 / 764.93 ms β 2.44x slower β
β Q9 β 319.20 / 329.03 Β±11.55 / 351.52 ms β 761.06 / 787.47 Β±19.57 / 807.46 ms β 2.39x slower β
β Q10 β 491.89 / 506.15 Β±10.55 / 522.59 ms β 954.17 / 979.45 Β±21.28 / 1004.87 ms β 1.94x slower β
β Q11 β 248.80 / 259.26 Β±14.18 / 287.12 ms β 241.57 / 254.60 Β±12.17 / 270.45 ms β no change β
βββββββββ΄βββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 3356.96ms β
β Total Time (sort-bucket) β 4700.13ms β
β Average Time (HEAD) β 305.18ms β
β Average Time (sort-bucket) β 427.28ms β
β Queries Faster β 3 β
β Queries Slower β 5 β
β Queries with No Change β 3 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββ΄ββββββββββββ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**sort_tpch β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 20.0s |
| Peak memory | 2.4 GiB |
| Avg memory | 1.1 GiB |
| CPU user | 64.1s |
| CPU sys | 3.4s |
| Peak spill | 0 B |
**sort_tpch β branch**
| Metric | Value |
|--------|-------|
| Wall time | 25.0s |
| Peak memory | 1.8 GiB |
| Avg memory | 997.1 MiB |
| CPU user | 70.7s |
| CPU sys | 4.0s |
| 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] perf: try bucket sort [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6338283402
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/26587#issuecomment-6338228159)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c6338228159-3754-skztd 6.12.94+ #1 SMP Fri Aug 21 08:00:16 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 sort-bucket (6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0) to e12901d (merge-base) [diff](https://github.com/apache/datafusion/compare/e12901dadfd462231a093627449d3a7b51d45962..6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpcds
env:
prefer_hash_join: "false"
```
</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] perf: try bucket sort [datafusion]
Posted by "rluvaton (via GitHub)" <gi...@apache.org>.
rluvaton commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6338224774
run benchmark sort
env:
BENCH_FILTER: sort_axis/sort.*.1M.High.cardinality.Unsorted..100cols
--
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] perf: try bucket sort [datafusion]
Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6347017293
<!-- 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 v55.1.0 (current)
Built [ 39.872s] (current)
Parsing datafusion v55.1.0 (current)
Parsed [ 0.025s] (current)
Building datafusion v55.1.0 (baseline)
Built [ 39.177s] (baseline)
Parsing datafusion v55.1.0 (baseline)
Parsed [ 0.025s] (baseline)
Checking datafusion v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 0.408s] 229 checks: 229 pass, 31 skip
Summary no semver update required
Finished [ 81.338s] datafusion
Building datafusion-common v55.1.0 (current)
Built [ 22.969s] (current)
Parsing datafusion-common v55.1.0 (current)
Parsed [ 0.051s] (current)
Building datafusion-common v55.1.0 (baseline)
Built [ 22.689s] (baseline)
Parsing datafusion-common v55.1.0 (baseline)
Parsed [ 0.054s] (baseline)
Checking datafusion-common v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 0.506s] 229 checks: 228 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/v0.51.0/src/lints/constructible_struct_adds_field.ron
Failed in:
field ExecutionOptions.sort_bucket_sort_enabled in /home/runner/work/datafusion/datafusion/datafusion/common/src/config.rs:897
field ExecutionOptions.sort_bucket_count in /home/runner/work/datafusion/datafusion/datafusion/common/src/config.rs:897
field ExecutionOptions.sort_bucket_target_bytes in /home/runner/work/datafusion/datafusion/datafusion/common/src/config.rs:897
field ExecutionOptions.sort_bucket_sample_rows in /home/runner/work/datafusion/datafusion/datafusion/common/src/config.rs:897
field ExecutionOptions.sort_bucket_mode in /home/runner/work/datafusion/datafusion/datafusion/common/src/config.rs:897
Summary semver requires new major version: 1 major and 0 minor checks failed
Finished [ 47.249s] datafusion-common
Building datafusion-physical-plan v55.1.0 (current)
Built [ 26.611s] (current)
Parsing datafusion-physical-plan v55.1.0 (current)
Parsed [ 0.172s] (current)
Building datafusion-physical-plan v55.1.0 (baseline)
Built [ 26.449s] (baseline)
Parsing datafusion-physical-plan v55.1.0 (baseline)
Parsed [ 0.159s] (baseline)
Checking datafusion-physical-plan v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 0.508s] 229 checks: 229 pass, 31 skip
Summary no semver update required
Finished [ 55.126s] datafusion-physical-plan
```
</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] perf: try bucket sort [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6338284020
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/26587#issuecomment-6338228159)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c6338228159-3751-g6sg6 6.12.94+ #1 SMP Fri Aug 21 08:00:16 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 sort-bucket (6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0) to e12901d (merge-base) [diff](https://github.com/apache/datafusion/compare/e12901dadfd462231a093627449d3a7b51d45962..6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0)
<details><summary>Run configuration</summary>
```yaml
run benchmark sort_tpch10
env:
prefer_hash_join: "false"
```
</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] perf: try bucket sort [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6338283894
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/26587#issuecomment-6338228159)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c6338228159-3752-58dg4 6.12.94+ #1 SMP Fri Aug 21 08:00:16 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 sort-bucket (6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0) to e12901d (merge-base) [diff](https://github.com/apache/datafusion/compare/e12901dadfd462231a093627449d3a7b51d45962..6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpch
env:
prefer_hash_join: "false"
```
</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] perf: try bucket sort [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6338669152
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/26587#issuecomment-6338228159)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing sort-bucket (6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0) to e12901d (merge-base) [diff](https://github.com/apache/datafusion/compare/e12901dadfd462231a093627449d3a7b51d45962..6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0)
<details><summary>Run configuration</summary>
```yaml
run benchmark sort_tpch10
env:
prefer_hash_join: "false"
```
</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 sort-bucket
--------------------
Benchmark sort_tpch10.json
--------------------
βββββββββ³βββββββββββββ³ββββββββββββββ³ββββββββββββββββ
β Query β HEAD β sort-bucket β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Q1 β 1430.65 ms β 2820.00 ms β 1.97x slower β
β Q2 β 1306.66 ms β 856.90 ms β +1.52x faster β
β Q3 β 8603.50 ms β 11947.53 ms β 1.39x slower β
β Q4 β 2241.84 ms β 1333.59 ms β +1.68x faster β
β Q5 β 2394.53 ms β 3357.19 ms β 1.40x slower β
β Q6 β 2552.77 ms β 3523.90 ms β 1.38x slower β
β Q7 β 5085.74 ms β 5601.80 ms β 1.10x slower β
β Q8 β 3547.52 ms β 21598.67 ms β 6.09x slower β
β Q9 β 3917.73 ms β 22336.40 ms β 5.70x slower β
β Q10 β 6089.13 ms β 25417.61 ms β 4.17x slower β
β Q11 β 3480.29 ms β 2321.78 ms β +1.50x faster β
βββββββββ΄βββββββββββββ΄ββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββ³ββββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 40650.37ms β
β Total Time (sort-bucket) β 101115.38ms β
β Average Time (HEAD) β 3695.49ms β
β Average Time (sort-bucket) β 9192.31ms β
β Queries Faster β 3 β
β Queries Slower β 8 β
β Queries with No Change β 0 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββ΄ββββββββββββββ
Distribution per query (min / mean Β±stddev / max):
Comparing HEAD and sort-bucket
--------------------
Benchmark sort_tpch10.json
--------------------
βββββββββ³βββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β sort-bucket β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Q1 β 1430.65 / 1435.88 Β±4.04 / 1442.09 ms β 2820.00 / 2877.28 Β±48.96 / 2954.12 ms β 2.00x slower β
β Q2 β 1306.66 / 1343.85 Β±29.42 / 1375.65 ms β 856.90 / 883.30 Β±14.85 / 901.96 ms β +1.52x faster β
β Q3 β 8603.50 / 8673.01 Β±55.22 / 8734.38 ms β 11947.53 / 12201.59 Β±174.62 / 12387.73 ms β 1.41x slower β
β Q4 β 2241.84 / 2268.49 Β±32.66 / 2332.29 ms β 1333.59 / 1392.70 Β±53.83 / 1493.74 ms β +1.63x faster β
β Q5 β 2394.53 / 2407.15 Β±10.33 / 2422.10 ms β 3357.19 / 3391.39 Β±17.94 / 3405.43 ms β 1.41x slower β
β Q6 β 2552.77 / 2577.90 Β±17.25 / 2599.04 ms β 3523.90 / 3552.53 Β±20.47 / 3582.31 ms β 1.38x slower β
β Q7 β 5085.74 / 5107.02 Β±16.46 / 5134.07 ms β 5601.80 / 5649.46 Β±30.13 / 5693.15 ms β 1.11x slower β
β Q8 β 3547.52 / 3656.85 Β±71.66 / 3759.53 ms β 21598.67 / 22051.78 Β±254.52 / 22273.80 ms β 6.03x slower β
β Q9 β 3917.73 / 3959.05 Β±42.84 / 4022.14 ms β 22336.40 / 22592.02 Β±257.09 / 23010.50 ms β 5.71x slower β
β Q10 β 6089.13 / 6436.57 Β±195.28 / 6673.16 ms β 25417.61 / 25661.11 Β±234.05 / 26037.50 ms β 3.99x slower β
β Q11 β 3480.29 / 3505.22 Β±21.71 / 3540.14 ms β 2321.78 / 2730.82 Β±469.44 / 3639.91 ms β +1.28x faster β
βββββββββ΄βββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββ³ββββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 41370.98ms β
β Total Time (sort-bucket) β 102983.97ms β
β Average Time (HEAD) β 3761.00ms β
β Average Time (sort-bucket) β 9362.18ms β
β Queries Faster β 3 β
β Queries Slower β 8 β
β Queries with No Change β 0 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββ΄ββββββββββββββ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**sort_tpch10 β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 210.0s |
| Peak memory | 19.0 GiB |
| Avg memory | 7.2 GiB |
| CPU user | 862.8s |
| CPU sys | 76.3s |
| Peak spill | 0 B |
**sort_tpch10 β branch**
| Metric | Value |
|--------|-------|
| Wall time | 520.1s |
| Peak memory | 14.8 GiB |
| Avg memory | 5.1 GiB |
| CPU user | 1017.9s |
| CPU sys | 82.1s |
| 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] perf: try bucket sort [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6339622629
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/26587#issuecomment-6339561598)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c6339561598-3755-nr27m 6.12.94+ #1 SMP Fri Aug 21 08:00:16 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 sort-bucket (6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0) to e12901d (merge-base) [diff](https://github.com/apache/datafusion/compare/e12901dadfd462231a093627449d3a7b51d45962..6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0)
<details><summary>Run configuration</summary>
```yaml
run benchmark external_aggr
```
</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] perf: try bucket sort [datafusion]
Posted by "rluvaton (via GitHub)" <gi...@apache.org>.
rluvaton commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6338228159
run benchmarks sort_tpch sort_tpch10 tpch tpch10 tpcds
```
env:
prefer_hash_join: false
```
--
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] perf: try bucket sort [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6339753975
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/26587#issuecomment-6339561598)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing sort-bucket (6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0) to e12901d (merge-base) [diff](https://github.com/apache/datafusion/compare/e12901dadfd462231a093627449d3a7b51d45962..6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0)
<details><summary>Run configuration</summary>
```yaml
run benchmark external_aggr
```
</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 sort-bucket
--------------------
Benchmark external_aggr.json
--------------------
ββββββββββββββββ³ββββββββββββ³ββββββββββββββ³ββββββββββββ
β Query β HEAD β sort-bucket β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Q1(64.0 MB) β 54.69 ms β 56.10 ms β no change β
β Q1(32.0 MB) β 49.66 ms β 49.90 ms β no change β
β Q1(16.0 MB) β 49.75 ms β 49.89 ms β no change β
β Q2(512.0 MB) β 246.80 ms β 256.34 ms β no change β
β Q2(256.0 MB) β 233.32 ms β 232.68 ms β no change β
β Q2(128.0 MB) β 208.16 ms β 207.25 ms β no change β
β Q2(64.0 MB) β 204.81 ms β 203.52 ms β no change β
β Q2(32.0 MB) β 264.07 ms β 261.44 ms β no change β
ββββββββββββββββ΄ββββββββββββ΄ββββββββββββββ΄ββββββββββββ
ββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 1311.26ms β
β Total Time (sort-bucket) β 1317.11ms β
β Average Time (HEAD) β 163.91ms β
β Average Time (sort-bucket) β 164.64ms β
β Queries Faster β 0 β
β Queries Slower β 0 β
β Queries with No Change β 8 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββ΄ββββββββββββ
Distribution per query (min / mean Β±stddev / max):
Comparing HEAD and sort-bucket
--------------------
Benchmark external_aggr.json
--------------------
ββββββββββββββββ³βββββββββββββββββββββββββββββββββββββ³ββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β sort-bucket β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Q1(64.0 MB) β 54.69 / 58.55 Β±4.01 / 66.18 ms β 56.10 / 60.81 Β±4.11 / 66.89 ms β no change β
β Q1(32.0 MB) β 49.66 / 51.33 Β±1.34 / 53.46 ms β 49.90 / 51.52 Β±0.87 / 52.46 ms β no change β
β Q1(16.0 MB) β 49.75 / 52.62 Β±3.05 / 58.43 ms β 49.89 / 50.99 Β±0.84 / 52.44 ms β no change β
β Q2(512.0 MB) β 246.80 / 257.70 Β±6.84 / 265.85 ms β 256.34 / 259.90 Β±3.24 / 265.17 ms β no change β
β Q2(256.0 MB) β 233.32 / 260.73 Β±22.01 / 285.95 ms β 232.68 / 240.74 Β±8.62 / 256.29 ms β +1.08x faster β
β Q2(128.0 MB) β 208.16 / 217.67 Β±8.63 / 230.79 ms β 207.25 / 210.87 Β±5.92 / 222.66 ms β no change β
β Q2(64.0 MB) β 204.81 / 211.69 Β±11.14 / 233.88 ms β 203.52 / 205.49 Β±1.06 / 206.72 ms β no change β
β Q2(32.0 MB) β 264.07 / 269.51 Β±3.63 / 274.03 ms β 261.44 / 266.94 Β±3.44 / 271.78 ms β no change β
ββββββββββββββββ΄βββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 1379.79ms β
β Total Time (sort-bucket) β 1347.25ms β
β Average Time (HEAD) β 172.47ms β
β Average Time (sort-bucket) β 168.41ms β
β Queries Faster β 1 β
β Queries Slower β 0 β
β Queries with No Change β 7 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββ΄ββββββββββββ
```
</p>
</details>
<details><summary>Memory Pool Peaks</summary>
Peak `MemoryPool` reservation per query β what DataFusion's accounting believes it reserved. Recorded only for benchmarks that write a results JSON (the `dfbench` suites, and the suites `bench.sh` runs through the Criterion SQL harness since apache/datafusion#25644), and only under a memory limit: `DATAFUSION_RUNTIME_MEMORY_LIMIT`, or one the suite sets itself.
Base: `e12901d (merge-base)` | Changed: `sort-bucket`
**`external_aggr`**
| Query | Base | Changed | Change |
| --- | --- | --- | --- |
| 1(64.0 MB) | 31.9 MiB | 28.8 MiB | -9.8% |
| 1(32.0 MB) | 14.3 MiB | 13.9 MiB | -2.6% |
| 1(16.0 MB) | 8.9 MiB | 8.9 MiB | -0.0% |
| 2(512.0 MB) | 137.8 MiB | 139.4 MiB | +1.2% |
| 2(256.0 MB) | 98.4 MiB | 98.1 MiB | -0.3% |
| 2(128.0 MB) | 49.5 MiB | 49.4 MiB | -0.3% |
| 2(64.0 MB) | 36.6 MiB | 36.6 MiB | -0.0% |
| 2(32.0 MB) | 17.8 MiB | 17.7 MiB | -0.5% |
**Pool accounting vs. process RSS**
Max pool peak is the largest reservation any single query in the run reached; peak RSS covers the whole invocation, including data loading and allocator retention, and the two high-water marks need not coincide in time. The gap is therefore an upper bound on what the pool did not account for, not a measurement of it.
| Benchmark | Side | Max pool peak | Peak RSS | Gap | RSS / pool |
| --- | --- | --- | --- | --- | --- |
| external_aggr | base (`e12901d (merge-base)`) | 137.8 MiB | 434.3 MiB | 296.5 MiB | 3.2Γ |
| external_aggr | changed (`sort-bucket`) | 139.4 MiB | 449.3 MiB | 309.8 MiB | 3.2Γ |
</details>
<details><summary>Resource Usage</summary>
**external_aggr β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 105.0s |
| Peak memory | 434.3 MiB |
| Avg memory | 24.9 MiB |
| CPU user | 21.3s |
| CPU sys | 3.7s |
| Peak spill | 93.1 MiB |
**external_aggr β branch**
| Metric | Value |
|--------|-------|
| Wall time | 115.0s |
| Peak memory | 449.3 MiB |
| Avg memory | 22.2 MiB |
| CPU user | 18.0s |
| CPU sys | 3.2s |
| Peak spill | 93.1 MiB |
</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] perf: try bucket sort [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6340142322
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/26587#issuecomment-6338224774)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing sort-bucket (6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0) to e12901d (merge-base) [diff](https://github.com/apache/datafusion/compare/e12901dadfd462231a093627449d3a7b51d45962..6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0)
<details><summary>Run configuration</summary>
```yaml
run benchmark sort
env:
BENCH_FILTER: "sort_axis/sort.*.1M.High.cardinality.Unsorted..100cols"
```
</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>
```
group HEAD sort-bucket
----- ---- -----------
sort_axis/sort i64 1M High cardinality Unsorted +100cols 1.00 2.3Β±0.04s ? ?/sec 2.30 5.4Β±0.12s ? ?/sec
sort_axis/sort mixed tuple 1M High cardinality Unsorted +100cols 1.00 2.4Β±0.05s ? ?/sec 2.41 5.8Β±0.17s ? ?/sec
sort_axis/sort utf8 view 1M High cardinality Unsorted +100cols 1.00 2.3Β±0.03s ? ?/sec 2.52 5.9Β±0.23s ? ?/sec
```
</p>
</details>
<details><summary>Resource Usage</summary>
**sort β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 2530.6s |
| Peak memory | 2.4 GiB |
| Avg memory | 1.4 GiB |
| CPU user | 2259.4s |
| CPU sys | 107.2s |
| Peak spill | 0 B |
**sort β branch**
| Metric | Value |
|--------|-------|
| Wall time | 3620.8s |
| Peak memory | 1.2 GiB |
| Avg memory | 1015.7 MiB |
| CPU user | 3153.1s |
| CPU sys | 8.4s |
| 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] perf: try bucket sort [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6338283910
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/26587#issuecomment-6338228159)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c6338228159-3753-4br79 6.12.94+ #1 SMP Fri Aug 21 08:00:16 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 sort-bucket (6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0) to e12901d (merge-base) [diff](https://github.com/apache/datafusion/compare/e12901dadfd462231a093627449d3a7b51d45962..6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpch10
env:
prefer_hash_join: "false"
```
</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] perf: try bucket sort [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6338449364
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/26587#issuecomment-6338228159)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing sort-bucket (6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0) to e12901d (merge-base) [diff](https://github.com/apache/datafusion/compare/e12901dadfd462231a093627449d3a7b51d45962..6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpch
env:
prefer_hash_join: "false"
```
</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 sort-bucket
--------------------
Benchmark tpch_sf1.json
--------------------
βββββββββββββ³βββββββββββ³ββββββββββββββ³ββββββββββββ
β Query β HEAD β sort-bucket β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 40.90 ms β 41.47 ms β no change β
β QQuery 2 β 18.70 ms β 18.87 ms β no change β
β QQuery 3 β 29.12 ms β 29.54 ms β no change β
β QQuery 4 β 17.44 ms β 18.15 ms β no change β
β QQuery 5 β 35.85 ms β 37.13 ms β no change β
β QQuery 6 β 14.01 ms β 14.14 ms β no change β
β QQuery 7 β 43.53 ms β 43.93 ms β no change β
β QQuery 8 β 42.12 ms β 42.28 ms β no change β
β QQuery 9 β 51.74 ms β 51.10 ms β no change β
β QQuery 10 β 41.68 ms β 42.12 ms β no change β
β QQuery 11 β 14.21 ms β 14.38 ms β no change β
β QQuery 12 β 18.65 ms β 18.85 ms β no change β
β QQuery 13 β 42.30 ms β 42.27 ms β no change β
β QQuery 14 β 25.50 ms β 25.66 ms β no change β
β QQuery 15 β 32.64 ms β 32.14 ms β no change β
β QQuery 16 β 13.95 ms β 14.24 ms β no change β
β QQuery 17 β 79.33 ms β 79.90 ms β no change β
β QQuery 18 β 46.54 ms β 46.86 ms β no change β
β QQuery 19 β 33.04 ms β 32.97 ms β no change β
β QQuery 20 β 33.26 ms β 33.68 ms β no change β
β QQuery 21 β 59.93 ms β 58.49 ms β no change β
β QQuery 22 β 14.63 ms β 14.59 ms β no change β
βββββββββββββ΄βββββββββββ΄ββββββββββββββ΄ββββββββββββ
ββββββββββββββββββββββββββββββ³βββββββββββ
β Benchmark Summary β β
β‘ββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 749.09ms β
β Total Time (sort-bucket) β 752.74ms β
β Average Time (HEAD) β 34.05ms β
β Average Time (sort-bucket) β 34.22ms β
β Queries Faster β 0 β
β Queries Slower β 0 β
β Queries with No Change β 22 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββ΄βββββββββββ
Distribution per query (min / mean Β±stddev / max):
Comparing HEAD and sort-bucket
--------------------
Benchmark tpch_sf1.json
--------------------
βββββββββββββ³βββββββββββββββββββββββββββββββββ³βββββββββββββββββββββββββββββββββ³ββββββββββββ
β Query β HEAD β sort-bucket β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 40.90 / 41.73 Β±0.95 / 43.57 ms β 41.47 / 43.33 Β±2.19 / 47.13 ms β no change β
β QQuery 2 β 18.70 / 18.99 Β±0.27 / 19.48 ms β 18.87 / 19.25 Β±0.23 / 19.46 ms β no change β
β QQuery 3 β 29.12 / 29.81 Β±1.07 / 31.94 ms β 29.54 / 29.76 Β±0.24 / 30.18 ms β no change β
β QQuery 4 β 17.44 / 17.72 Β±0.18 / 17.87 ms β 18.15 / 18.53 Β±0.40 / 19.21 ms β no change β
β QQuery 5 β 35.85 / 36.76 Β±0.81 / 38.04 ms β 37.13 / 37.72 Β±0.39 / 38.30 ms β no change β
β QQuery 6 β 14.01 / 14.57 Β±0.54 / 15.47 ms β 14.14 / 14.28 Β±0.12 / 14.45 ms β no change β
β QQuery 7 β 43.53 / 45.86 Β±2.21 / 49.60 ms β 43.93 / 46.57 Β±4.06 / 54.61 ms β no change β
β QQuery 8 β 42.12 / 42.57 Β±0.64 / 43.83 ms β 42.28 / 42.82 Β±0.50 / 43.59 ms β no change β
β QQuery 9 β 51.74 / 52.51 Β±0.53 / 53.28 ms β 51.10 / 53.50 Β±1.82 / 55.32 ms β no change β
β QQuery 10 β 41.68 / 43.14 Β±1.42 / 45.33 ms β 42.12 / 42.53 Β±0.28 / 42.92 ms β no change β
β QQuery 11 β 14.21 / 14.71 Β±0.52 / 15.65 ms β 14.38 / 14.59 Β±0.14 / 14.78 ms β no change β
β QQuery 12 β 18.65 / 19.05 Β±0.25 / 19.35 ms β 18.85 / 19.46 Β±0.55 / 20.46 ms β no change β
β QQuery 13 β 42.30 / 43.53 Β±1.64 / 46.74 ms β 42.27 / 44.09 Β±3.11 / 50.29 ms β no change β
β QQuery 14 β 25.50 / 26.53 Β±1.57 / 29.65 ms β 25.66 / 26.00 Β±0.29 / 26.45 ms β no change β
β QQuery 15 β 32.64 / 33.52 Β±0.77 / 34.90 ms β 32.14 / 33.47 Β±1.61 / 36.61 ms β no change β
β QQuery 16 β 13.95 / 14.33 Β±0.20 / 14.54 ms β 14.24 / 14.52 Β±0.18 / 14.75 ms β no change β
β QQuery 17 β 79.33 / 81.03 Β±1.23 / 82.68 ms β 79.90 / 80.78 Β±0.62 / 81.78 ms β no change β
β QQuery 18 β 46.54 / 48.03 Β±0.97 / 49.55 ms β 46.86 / 47.88 Β±1.07 / 49.92 ms β no change β
β QQuery 19 β 33.04 / 33.74 Β±0.93 / 35.56 ms β 32.97 / 33.10 Β±0.11 / 33.30 ms β no change β
β QQuery 20 β 33.26 / 33.70 Β±0.33 / 34.20 ms β 33.68 / 33.78 Β±0.10 / 33.90 ms β no change β
β QQuery 21 β 59.93 / 61.26 Β±0.72 / 61.93 ms β 58.49 / 59.72 Β±0.84 / 60.94 ms β no change β
β QQuery 22 β 14.63 / 14.77 Β±0.10 / 14.94 ms β 14.59 / 14.74 Β±0.13 / 14.93 ms β no change β
βββββββββββββ΄βββββββββββββββββββββββββββββββββ΄βββββββββββββββββββββββββββββββββ΄ββββββββββββ
ββββββββββββββββββββββββββββββ³βββββββββββ
β Benchmark Summary β β
β‘ββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 767.87ms β
β Total Time (sort-bucket) β 770.43ms β
β Average Time (HEAD) β 34.90ms β
β Average Time (sort-bucket) β 35.02ms β
β Queries Faster β 0 β
β Queries Slower β 0 β
β Queries with No Change β 22 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββ΄βββββββββββ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**tpch β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 5.0s |
| Peak memory | 1.1 GiB |
| Avg memory | 476.7 MiB |
| CPU user | 21.1s |
| CPU sys | 1.9s |
| Peak spill | 0 B |
**tpch β branch**
| Metric | Value |
|--------|-------|
| Wall time | 5.0s |
| Peak memory | 1.1 GiB |
| Avg memory | 465.1 MiB |
| CPU user | 21.3s |
| CPU sys | 1.9s |
| 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] perf: try bucket sort [datafusion]
Posted by "rluvaton (via GitHub)" <gi...@apache.org>.
rluvaton commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6343194691
bad perf
--
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] perf: try bucket sort [datafusion]
Posted by "rluvaton (via GitHub)" <gi...@apache.org>.
rluvaton closed pull request #26587: perf: try bucket sort
URL: https://github.com/apache/datafusion/pull/26587
--
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] perf: try bucket sort [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6338283719
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/26587#issuecomment-6338228159)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c6338228159-3750-4jwj2 6.12.94+ #1 SMP Fri Aug 21 08:00:16 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 sort-bucket (6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0) to e12901d (merge-base) [diff](https://github.com/apache/datafusion/compare/e12901dadfd462231a093627449d3a7b51d45962..6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0)
<details><summary>Run configuration</summary>
```yaml
run benchmark sort_tpch
env:
prefer_hash_join: "false"
```
</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] perf: try bucket sort [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #26587:
URL: https://github.com/apache/datafusion/pull/26587#issuecomment-6338284039
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/26587#issuecomment-6338224774)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c6338224774-3749-bt4kb 6.12.94+ #1 SMP Fri Aug 21 08:00:16 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 sort-bucket (6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0) to e12901d (merge-base) [diff](https://github.com/apache/datafusion/compare/e12901dadfd462231a093627449d3a7b51d45962..6fdf9a9157d19f07a813a98cdf98edc4a00e7ca0)
<details><summary>Run configuration</summary>
```yaml
run benchmark sort
env:
BENCH_FILTER: "sort_axis/sort.*.1M.High.cardinality.Unsorted..100cols"
```
</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