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 2026/07/05 21:20:10 UTC
[PR] perf(sort): passthrough last stream that left in `SortPreservingMergeStream` [datafusion]
rluvaton opened a new pull request, #23328:
URL: https://github.com/apache/datafusion/pull/23328
## Which issue does this PR close?
N/A
## Rationale for this change
No need to go through the entire Loser tree mechanism when only 1 stream left, so in that case, just passthrough last stream
## What changes are included in this PR?
added passthrough for last stream in `SortPreservingMergeStream`
## Are these changes tested?
Yes, basic
## Are there any user-facing changes?
No
----
FYI, this is one of many upcoming optimizations that I will do for Sort
--
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(sort): passthrough last stream that left in `SortPreservingMergeStream` [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #23328:
URL: https://github.com/apache/datafusion/pull/23328#issuecomment-4896080809
๐ค Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/23328#issuecomment-4896056824)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4896056824-870-dknxj 6.12.85+ #1 SMP Mon May 11 08:17:35 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 passthrough-last-stream (d7156e6ebc32377d63ebcb7609bba609a2962924) to 6a0e76e (merge-base) [diff](https://github.com/apache/datafusion/compare/6a0e76ed692afaafd616e0d98d51bef63c1ff02d..d7156e6ebc32377d63ebcb7609bba609a2962924) using: clickbench_partitioned
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(sort): passthrough last stream that left in `SortPreservingMergeStream` [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #23328:
URL: https://github.com/apache/datafusion/pull/23328#issuecomment-4896201555
๐ค Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/23328#issuecomment-4896056824)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
<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 passthrough-last-stream
--------------------
Benchmark tpch_sf1.json
--------------------
โโโโโโโโโโโโโณโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโโโโ
โ Query โ HEAD โ passthrough-last-stream โ Change โ
โกโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฉ
โ QQuery 1 โ 38.64 / 40.36 ยฑ2.93 / 46.22 ms โ 38.30 / 39.91 ยฑ1.70 / 42.00 ms โ no change โ
โ QQuery 2 โ 19.32 / 19.96 ยฑ0.56 / 20.88 ms โ 19.28 / 19.50 ยฑ0.17 / 19.72 ms โ no change โ
โ QQuery 3 โ 30.70 / 33.46 ยฑ2.25 / 36.94 ms โ 30.86 / 32.27 ยฑ1.29 / 33.96 ms โ no change โ
โ QQuery 4 โ 17.60 / 17.91 ยฑ0.23 / 18.17 ms โ 17.68 / 17.80 ยฑ0.11 / 18.00 ms โ no change โ
โ QQuery 5 โ 38.81 / 41.79 ยฑ1.79 / 43.83 ms โ 37.89 / 39.28 ยฑ1.47 / 41.09 ms โ +1.06x faster โ
โ QQuery 6 โ 16.31 / 17.30 ยฑ0.96 / 18.57 ms โ 16.34 / 16.70 ยฑ0.49 / 17.58 ms โ no change โ
โ QQuery 7 โ 45.02 / 47.78 ยฑ2.45 / 51.56 ms โ 43.66 / 46.26 ยฑ2.10 / 49.38 ms โ no change โ
โ QQuery 8 โ 43.59 / 44.52 ยฑ1.44 / 47.36 ms โ 43.29 / 43.59 ยฑ0.21 / 43.85 ms โ no change โ
โ QQuery 9 โ 50.39 / 50.84 ยฑ0.50 / 51.73 ms โ 48.97 / 50.52 ยฑ1.33 / 52.71 ms โ no change โ
โ QQuery 10 โ 42.89 / 43.58 ยฑ0.86 / 45.23 ms โ 42.36 / 42.48 ยฑ0.10 / 42.61 ms โ no change โ
โ QQuery 11 โ 13.72 / 14.25 ยฑ0.58 / 15.28 ms โ 13.79 / 14.11 ยฑ0.41 / 14.91 ms โ no change โ
โ QQuery 12 โ 24.22 / 24.63 ยฑ0.45 / 25.46 ms โ 24.22 / 24.52 ยฑ0.28 / 24.93 ms โ no change โ
โ QQuery 13 โ 32.55 / 34.50 ยฑ1.41 / 36.90 ms โ 34.31 / 35.19 ยฑ0.76 / 36.39 ms โ no change โ
โ QQuery 14 โ 24.09 / 24.17 ยฑ0.06 / 24.28 ms โ 23.76 / 23.95 ยฑ0.16 / 24.21 ms โ no change โ
โ QQuery 15 โ 31.33 / 31.98 ยฑ0.55 / 32.90 ms โ 31.29 / 32.17 ยฑ0.84 / 33.74 ms โ no change โ
โ QQuery 16 โ 14.07 / 14.27 ยฑ0.12 / 14.42 ms โ 14.02 / 14.43 ยฑ0.27 / 14.85 ms โ no change โ
โ QQuery 17 โ 74.23 / 74.74 ยฑ0.38 / 75.38 ms โ 73.37 / 74.89 ยฑ1.46 / 77.68 ms โ no change โ
โ QQuery 18 โ 58.86 / 60.98 ยฑ1.20 / 62.27 ms โ 57.31 / 59.72 ยฑ1.42 / 61.34 ms โ no change โ
โ QQuery 19 โ 33.37 / 34.08 ยฑ1.05 / 36.16 ms โ 33.14 / 33.80 ยฑ0.61 / 34.83 ms โ no change โ
โ QQuery 20 โ 32.40 / 32.91 ยฑ0.56 / 33.78 ms โ 31.63 / 32.18 ยฑ0.50 / 33.01 ms โ no change โ
โ QQuery 21 โ 55.93 / 58.17 ยฑ1.48 / 59.77 ms โ 55.62 / 57.02 ยฑ1.26 / 59.30 ms โ no change โ
โ QQuery 22 โ 14.22 / 14.42 ยฑ0.16 / 14.65 ms โ 14.23 / 14.52 ยฑ0.29 / 15.04 ms โ no change โ
โโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโโ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโ
โ Benchmark Summary โ โ
โกโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฉ
โ Total Time (HEAD) โ 776.61ms โ
โ Total Time (passthrough-last-stream) โ 764.79ms โ
โ Average Time (HEAD) โ 35.30ms โ
โ Average Time (passthrough-last-stream) โ 34.76ms โ
โ Queries Faster โ 1 โ
โ Queries Slower โ 0 โ
โ Queries with No Change โ 21 โ
โ Queries with Failure โ 0 โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**tpch โ base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 5.0s |
| Peak memory | 1.2 GiB |
| Avg memory | 522.1 MiB |
| CPU user | 22.5s |
| CPU sys | 1.7s |
| Peak spill | 0 B |
**tpch โ branch**
| Metric | Value |
|--------|-------|
| Wall time | 5.0s |
| Peak memory | 1.2 GiB |
| Avg memory | 513.8 MiB |
| CPU user | 22.0s |
| CPU sys | 1.6s |
| 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(sort): passthrough last stream that left in `SortPreservingMergeStream` [datafusion]
Posted by "rluvaton (via GitHub)" <gi...@apache.org>.
rluvaton commented on PR #23328:
URL: https://github.com/apache/datafusion/pull/23328#issuecomment-4896056824
run benchmarks
--
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(sort): passthrough last stream that left in `SortPreservingMergeStream` [datafusion]
Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] closed pull request #23328: perf(sort): passthrough last stream that left in `SortPreservingMergeStream`
URL: https://github.com/apache/datafusion/pull/23328
--
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(sort): passthrough last stream that left in `SortPreservingMergeStream` [datafusion]
Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #23328:
URL: https://github.com/apache/datafusion/pull/23328#issuecomment-5577813287
Thank you for your contribution. Unfortunately, this pull request is stale because it has been open 60 days with no activity. Please remove the stale label or comment or this will be closed in 7 days.
--
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(sort): passthrough last stream that left in `SortPreservingMergeStream` [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #23328:
URL: https://github.com/apache/datafusion/pull/23328#issuecomment-4896083617
๐ค Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/23328#issuecomment-4896056824)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4896056824-871-npgfc 6.12.85+ #1 SMP Mon May 11 08:17:35 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 passthrough-last-stream (d7156e6ebc32377d63ebcb7609bba609a2962924) to 6a0e76e (merge-base) [diff](https://github.com/apache/datafusion/compare/6a0e76ed692afaafd616e0d98d51bef63c1ff02d..d7156e6ebc32377d63ebcb7609bba609a2962924) using: tpcds
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(sort): passthrough last stream that left in `SortPreservingMergeStream` [datafusion]
Posted by "rluvaton (via GitHub)" <gi...@apache.org>.
rluvaton commented on PR #23328:
URL: https://github.com/apache/datafusion/pull/23328#issuecomment-4902711352
@alamb the current failing tests are from a fuzz tests you added that verify that we output the same size for all the batches except 1, however this prevent one of the optimization here that just passthrough the stream and avoid the copy or slicing, WDYT?
- https://github.com/apache/datafusion/pull/1678
Same idea as
- #23345
--
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(sort): passthrough last stream that left in `SortPreservingMergeStream` [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #23328:
URL: https://github.com/apache/datafusion/pull/23328#issuecomment-4896214423
๐ค Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/23328#issuecomment-4896056824)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
<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 passthrough-last-stream
--------------------
Benchmark tpcds_sf1.json
--------------------
โโโโโโโโโโโโโณโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโโโโ
โ Query โ HEAD โ passthrough-last-stream โ Change โ
โกโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฉ
โ QQuery 1 โ 5.77 / 6.23 ยฑ0.84 / 7.92 ms โ 5.58 / 6.05 ยฑ0.85 / 7.75 ms โ no change โ
โ QQuery 2 โ 82.01 / 82.41 ยฑ0.24 / 82.67 ms โ 80.10 / 80.42 ยฑ0.24 / 80.81 ms โ no change โ
โ QQuery 3 โ 29.65 / 29.89 ยฑ0.24 / 30.31 ms โ 29.18 / 29.43 ยฑ0.18 / 29.63 ms โ no change โ
โ QQuery 4 โ 487.60 / 490.81 ยฑ2.27 / 493.88 ms โ 478.77 / 484.42 ยฑ5.74 / 494.38 ms โ no change โ
โ QQuery 5 โ 52.50 / 53.24 ยฑ0.54 / 54.17 ms โ 51.75 / 55.25 ยฑ4.97 / 65.02 ms โ no change โ
โ QQuery 6 โ 36.74 / 37.08 ยฑ0.39 / 37.78 ms โ 36.44 / 37.04 ยฑ0.41 / 37.50 ms โ no change โ
โ QQuery 7 โ 94.96 / 95.97 ยฑ0.57 / 96.45 ms โ 94.72 / 94.92 ยฑ0.14 / 95.06 ms โ no change โ
โ QQuery 8 โ 37.37 / 39.00 ยฑ2.27 / 43.40 ms โ 36.60 / 38.84 ยฑ3.26 / 45.32 ms โ no change โ
โ QQuery 9 โ 51.86 / 54.85 ยฑ2.19 / 58.09 ms โ 53.18 / 54.72 ยฑ1.61 / 57.77 ms โ no change โ
โ QQuery 10 โ 63.92 / 64.18 ยฑ0.31 / 64.75 ms โ 63.27 / 63.72 ยฑ0.40 / 64.30 ms โ no change โ
โ QQuery 11 โ 306.33 / 312.69 ยฑ6.46 / 324.57 ms โ 297.26 / 302.73 ยฑ3.94 / 307.64 ms โ no change โ
โ QQuery 12 โ 28.89 / 29.25 ยฑ0.20 / 29.50 ms โ 28.34 / 28.95 ยฑ0.49 / 29.79 ms โ no change โ
โ QQuery 13 โ 119.69 / 120.71 ยฑ1.35 / 123.29 ms โ 117.74 / 118.51 ยฑ0.54 / 119.41 ms โ no change โ
โ QQuery 14 โ 418.48 / 421.40 ยฑ1.56 / 422.97 ms โ 411.13 / 417.78 ยฑ3.53 / 420.79 ms โ no change โ
โ QQuery 15 โ 57.61 / 58.64 ยฑ1.10 / 60.33 ms โ 57.07 / 58.03 ยฑ0.74 / 59.33 ms โ no change โ
โ QQuery 16 โ 6.90 / 7.02 ยฑ0.11 / 7.23 ms โ 6.67 / 6.85 ยฑ0.16 / 7.13 ms โ no change โ
โ QQuery 17 โ 81.05 / 82.53 ยฑ1.82 / 86.05 ms โ 80.48 / 81.80 ยฑ1.54 / 84.80 ms โ no change โ
โ QQuery 18 โ 124.04 / 125.73 ยฑ1.10 / 127.13 ms โ 124.74 / 125.92 ยฑ1.09 / 127.77 ms โ no change โ
โ QQuery 19 โ 41.94 / 42.18 ยฑ0.23 / 42.56 ms โ 41.74 / 42.13 ยฑ0.31 / 42.49 ms โ no change โ
โ QQuery 20 โ 36.42 / 37.20 ยฑ0.97 / 39.10 ms โ 35.30 / 35.84 ยฑ0.47 / 36.59 ms โ no change โ
โ QQuery 21 โ 17.45 / 18.38 ยฑ1.17 / 20.68 ms โ 17.48 / 17.64 ยฑ0.18 / 17.96 ms โ no change โ
โ QQuery 22 โ 62.01 / 64.28 ยฑ1.86 / 67.02 ms โ 62.41 / 63.21 ยฑ0.60 / 63.75 ms โ no change โ
โ QQuery 23 โ 347.27 / 351.33 ยฑ3.62 / 357.06 ms โ 345.49 / 350.03 ยฑ4.67 / 357.78 ms โ no change โ
โ QQuery 24 โ 227.15 / 230.61 ยฑ4.71 / 239.53 ms โ 226.11 / 229.31 ยฑ3.21 / 235.37 ms โ no change โ
โ QQuery 25 โ 111.48 / 112.47 ยฑ0.88 / 113.91 ms โ 110.27 / 111.36 ยฑ0.68 / 112.28 ms โ no change โ
โ QQuery 26 โ 58.31 / 60.60 ยฑ2.57 / 65.44 ms โ 58.03 / 59.99 ยฑ2.18 / 64.06 ms โ no change โ
โ QQuery 27 โ 6.63 / 6.77 ยฑ0.12 / 7.00 ms โ 6.50 / 6.63 ยฑ0.15 / 6.92 ms โ no change โ
โ QQuery 28 โ 57.39 / 60.86 ยฑ1.76 / 62.12 ms โ 60.42 / 61.66 ยฑ1.06 / 63.57 ms โ no change โ
โ QQuery 29 โ 97.87 / 101.39 ยฑ3.74 / 107.73 ms โ 96.78 / 98.27 ยฑ1.06 / 99.48 ms โ no change โ
โ QQuery 30 โ 32.90 / 33.82 ยฑ0.84 / 35.21 ms โ 33.22 / 34.63 ยฑ2.16 / 38.92 ms โ no change โ
โ QQuery 31 โ 112.95 / 113.38 ยฑ0.45 / 114.22 ms โ 111.76 / 112.63 ยฑ0.86 / 114.25 ms โ no change โ
โ QQuery 32 โ 20.67 / 21.14 ยฑ0.32 / 21.62 ms โ 20.45 / 21.03 ยฑ0.29 / 21.22 ms โ no change โ
โ QQuery 33 โ 38.89 / 39.24 ยฑ0.39 / 39.96 ms โ 38.11 / 39.62 ยฑ2.06 / 43.70 ms โ no change โ
โ QQuery 34 โ 10.08 / 10.61 ยฑ0.50 / 11.38 ms โ 10.13 / 10.97 ยฑ1.00 / 12.83 ms โ no change โ
โ QQuery 35 โ 74.37 / 74.83 ยฑ0.39 / 75.50 ms โ 73.00 / 74.79 ยฑ1.22 / 76.61 ms โ no change โ
โ QQuery 36 โ 5.95 / 6.12 ยฑ0.18 / 6.47 ms โ 5.88 / 6.05 ยฑ0.18 / 6.39 ms โ no change โ
โ QQuery 37 โ 7.07 / 7.17 ยฑ0.06 / 7.25 ms โ 7.03 / 7.15 ยฑ0.08 / 7.28 ms โ no change โ
โ QQuery 38 โ 63.51 / 63.86 ยฑ0.44 / 64.72 ms โ 62.76 / 62.99 ยฑ0.21 / 63.29 ms โ no change โ
โ QQuery 39 โ 92.95 / 95.46 ยฑ2.51 / 99.15 ms โ 89.57 / 93.06 ยฑ3.46 / 99.41 ms โ no change โ
โ QQuery 40 โ 24.59 / 25.00 ยฑ0.61 / 26.19 ms โ 23.68 / 24.11 ยฑ0.29 / 24.57 ms โ no change โ
โ QQuery 41 โ 11.78 / 12.01 ยฑ0.22 / 12.41 ms โ 11.59 / 11.75 ยฑ0.20 / 12.14 ms โ no change โ
โ QQuery 42 โ 24.63 / 25.29 ยฑ0.58 / 26.24 ms โ 23.72 / 24.06 ยฑ0.23 / 24.39 ms โ no change โ
โ QQuery 43 โ 5.19 / 5.28 ยฑ0.15 / 5.58 ms โ 5.04 / 5.16 ยฑ0.15 / 5.46 ms โ no change โ
โ QQuery 44 โ 9.73 / 9.85 ยฑ0.11 / 9.97 ms โ 9.61 / 9.73 ยฑ0.14 / 10.00 ms โ no change โ
โ QQuery 45 โ 39.06 / 39.93 ยฑ0.91 / 41.43 ms โ 37.86 / 38.76 ยฑ0.61 / 39.61 ms โ no change โ
โ QQuery 46 โ 12.31 / 12.78 ยฑ0.29 / 13.05 ms โ 11.75 / 11.96 ยฑ0.19 / 12.30 ms โ +1.07x faster โ
โ QQuery 47 โ 231.59 / 235.46 ยฑ5.10 / 245.02 ms โ 224.36 / 228.48 ยฑ4.91 / 237.98 ms โ no change โ
โ QQuery 48 โ 98.31 / 98.84 ยฑ0.34 / 99.22 ms โ 96.44 / 97.31 ยฑ0.74 / 98.60 ms โ no change โ
โ QQuery 49 โ 78.10 / 78.56 ยฑ0.59 / 79.71 ms โ 76.31 / 77.48 ยฑ0.70 / 78.04 ms โ no change โ
โ QQuery 50 โ 59.85 / 61.44 ยฑ2.21 / 65.79 ms โ 58.81 / 59.78 ยฑ0.65 / 60.57 ms โ no change โ
โ QQuery 51 โ 93.12 / 94.80 ยฑ1.00 / 95.97 ms โ 91.64 / 93.27 ยฑ1.54 / 96.01 ms โ no change โ
โ QQuery 52 โ 24.27 / 26.44 ยฑ3.22 / 32.80 ms โ 23.90 / 24.43 ยฑ0.33 / 24.77 ms โ +1.08x faster โ
โ QQuery 53 โ 30.23 / 31.21 ยฑ1.08 / 33.31 ms โ 29.50 / 29.69 ยฑ0.12 / 29.84 ms โ no change โ
โ QQuery 54 โ 56.53 / 56.95 ยฑ0.35 / 57.48 ms โ 55.66 / 58.36 ยฑ3.15 / 64.51 ms โ no change โ
โ QQuery 55 โ 23.80 / 24.45 ยฑ0.39 / 24.79 ms โ 23.71 / 24.13 ยฑ0.39 / 24.85 ms โ no change โ
โ QQuery 56 โ 40.26 / 40.90 ยฑ0.44 / 41.39 ms โ 40.12 / 40.54 ยฑ0.37 / 41.11 ms โ no change โ
โ QQuery 57 โ 179.39 / 181.19 ยฑ3.09 / 187.33 ms โ 175.41 / 178.34 ยฑ2.06 / 180.94 ms โ no change โ
โ QQuery 58 โ 118.16 / 119.18 ยฑ0.67 / 119.93 ms โ 114.15 / 116.08 ยฑ1.68 / 118.89 ms โ no change โ
โ QQuery 59 โ 120.67 / 121.07 ยฑ0.36 / 121.66 ms โ 117.74 / 118.76 ยฑ0.80 / 120.16 ms โ no change โ
โ QQuery 60 โ 41.30 / 42.48 ยฑ1.39 / 45.18 ms โ 40.05 / 40.53 ยฑ0.44 / 41.24 ms โ no change โ
โ QQuery 61 โ 12.97 / 13.40 ยฑ0.64 / 14.67 ms โ 12.90 / 13.03 ยฑ0.18 / 13.38 ms โ no change โ
โ QQuery 62 โ 47.39 / 47.60 ยฑ0.28 / 48.14 ms โ 46.29 / 47.88 ยฑ2.25 / 52.33 ms โ no change โ
โ QQuery 63 โ 30.25 / 30.83 ยฑ0.42 / 31.53 ms โ 29.67 / 30.23 ยฑ0.37 / 30.83 ms โ no change โ
โ QQuery 64 โ 412.31 / 421.30 ยฑ7.43 / 434.33 ms โ 408.66 / 413.42 ยฑ4.12 / 419.03 ms โ no change โ
โ QQuery 65 โ 148.44 / 150.19 ยฑ1.23 / 151.65 ms โ 143.87 / 147.51 ยฑ2.30 / 150.96 ms โ no change โ
โ QQuery 66 โ 79.81 / 80.19 ยฑ0.45 / 81.05 ms โ 79.56 / 79.76 ยฑ0.16 / 79.97 ms โ no change โ
โ QQuery 67 โ 240.44 / 247.90 ยฑ8.64 / 264.84 ms โ 238.53 / 241.33 ยฑ2.73 / 245.00 ms โ no change โ
โ QQuery 68 โ 12.14 / 16.13 ยฑ6.62 / 29.26 ms โ 11.80 / 12.03 ยฑ0.20 / 12.40 ms โ +1.34x faster โ
โ QQuery 69 โ 58.21 / 59.03 ยฑ0.57 / 59.96 ms โ 57.50 / 58.10 ยฑ0.35 / 58.45 ms โ no change โ
โ QQuery 70 โ 105.60 / 108.98 ยฑ4.22 / 117.26 ms โ 105.21 / 110.12 ยฑ7.15 / 124.18 ms โ no change โ
โ QQuery 71 โ 35.83 / 39.34 ยฑ6.39 / 52.12 ms โ 35.56 / 36.49 ยฑ1.18 / 38.76 ms โ +1.08x faster โ
โ QQuery 72 โ 2090.55 / 2177.11 ยฑ72.94 / 2310.24 ms โ 1992.06 / 2081.20 ยฑ53.68 / 2147.84 ms โ no change โ
โ QQuery 73 โ 9.87 / 14.06 ยฑ7.99 / 30.04 ms โ 9.83 / 10.22 ยฑ0.51 / 11.21 ms โ +1.38x faster โ
โ QQuery 74 โ 171.86 / 175.28 ยฑ3.92 / 181.77 ms โ 166.75 / 171.12 ยฑ4.48 / 178.97 ms โ no change โ
โ QQuery 75 โ 150.92 / 155.81 ยฑ5.93 / 167.17 ms โ 149.08 / 151.60 ยฑ2.32 / 155.14 ms โ no change โ
โ QQuery 76 โ 35.88 / 36.50 ยฑ0.47 / 37.30 ms โ 35.46 / 35.72 ยฑ0.14 / 35.85 ms โ no change โ
โ QQuery 77 โ 62.26 / 65.91 ยฑ4.59 / 74.85 ms โ 62.91 / 65.09 ยฑ2.14 / 69.09 ms โ no change โ
โ QQuery 78 โ 197.88 / 202.38 ยฑ5.55 / 211.47 ms โ 195.18 / 198.65 ยฑ2.02 / 201.49 ms โ no change โ
โ QQuery 79 โ 67.26 / 67.79 ยฑ0.30 / 68.18 ms โ 67.30 / 68.05 ยฑ0.85 / 69.66 ms โ no change โ
โ QQuery 80 โ 99.75 / 102.54 ยฑ4.32 / 111.12 ms โ 98.51 / 100.36 ยฑ1.74 / 103.50 ms โ no change โ
โ QQuery 81 โ 26.53 / 28.00 ยฑ2.17 / 32.28 ms โ 25.95 / 26.93 ยฑ1.29 / 29.45 ms โ no change โ
โ QQuery 82 โ 16.85 / 17.43 ยฑ0.54 / 18.12 ms โ 16.68 / 17.01 ยฑ0.22 / 17.25 ms โ no change โ
โ QQuery 83 โ 40.75 / 41.05 ยฑ0.43 / 41.90 ms โ 40.38 / 40.95 ยฑ0.85 / 42.65 ms โ no change โ
โ QQuery 84 โ 30.40 / 30.77 ยฑ0.33 / 31.37 ms โ 30.45 / 30.74 ยฑ0.40 / 31.54 ms โ no change โ
โ QQuery 85 โ 107.82 / 111.37 ยฑ4.68 / 120.09 ms โ 106.65 / 107.42 ยฑ0.66 / 108.53 ms โ no change โ
โ QQuery 86 โ 25.23 / 25.76 ยฑ0.38 / 26.30 ms โ 25.09 / 25.28 ยฑ0.14 / 25.49 ms โ no change โ
โ QQuery 87 โ 63.21 / 64.24 ยฑ0.79 / 65.32 ms โ 62.51 / 63.90 ยฑ0.94 / 64.98 ms โ no change โ
โ QQuery 88 โ 63.67 / 64.33 ยฑ0.35 / 64.70 ms โ 63.36 / 63.84 ยฑ0.34 / 64.42 ms โ no change โ
โ QQuery 89 โ 36.47 / 39.26 ยฑ4.18 / 47.52 ms โ 36.30 / 38.08 ยฑ3.07 / 44.21 ms โ no change โ
โ QQuery 90 โ 17.63 / 17.79 ยฑ0.17 / 18.01 ms โ 17.30 / 17.97 ยฑ0.96 / 19.87 ms โ no change โ
โ QQuery 91 โ 47.15 / 47.42 ยฑ0.19 / 47.63 ms โ 46.71 / 47.24 ยฑ0.42 / 47.79 ms โ no change โ
โ QQuery 92 โ 29.29 / 30.17 ยฑ0.82 / 31.52 ms โ 29.13 / 29.82 ยฑ0.53 / 30.44 ms โ no change โ
โ QQuery 93 โ 50.64 / 51.39 ยฑ0.47 / 52.09 ms โ 50.02 / 50.60 ยฑ0.45 / 51.38 ms โ no change โ
โ QQuery 94 โ 38.77 / 41.23 ยฑ3.92 / 49.03 ms โ 37.90 / 39.95 ยฑ2.51 / 44.76 ms โ no change โ
โ QQuery 95 โ 82.05 / 84.06 ยฑ1.98 / 87.80 ms โ 82.12 / 83.61 ยฑ1.10 / 85.38 ms โ no change โ
โ QQuery 96 โ 24.31 / 24.62 ยฑ0.31 / 25.15 ms โ 24.10 / 24.48 ยฑ0.37 / 25.17 ms โ no change โ
โ QQuery 97 โ 46.63 / 47.04 ยฑ0.24 / 47.27 ms โ 46.32 / 46.50 ยฑ0.15 / 46.77 ms โ no change โ
โ QQuery 98 โ 42.27 / 43.23 ยฑ0.85 / 44.52 ms โ 41.95 / 44.11 ยฑ2.98 / 50.01 ms โ no change โ
โ QQuery 99 โ 70.29 / 70.66 ยฑ0.33 / 71.16 ms โ 70.13 / 70.65 ยฑ0.83 / 72.29 ms โ no change โ
โโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโโ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโ
โ Benchmark Summary โ โ
โกโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฉ
โ Total Time (HEAD) โ 10098.46ms โ
โ Total Time (passthrough-last-stream) โ 9878.11ms โ
โ Average Time (HEAD) โ 102.00ms โ
โ Average Time (passthrough-last-stream) โ 99.78ms โ
โ Queries Faster โ 5 โ
โ Queries Slower โ 0 โ
โ Queries with No Change โ 94 โ
โ Queries with Failure โ 0 โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**tpcds โ base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 55.0s |
| Peak memory | 2.0 GiB |
| Avg memory | 1.3 GiB |
| CPU user | 227.5s |
| CPU sys | 6.2s |
| Peak spill | 0 B |
**tpcds โ branch**
| Metric | Value |
|--------|-------|
| Wall time | 50.0s |
| Peak memory | 2.1 GiB |
| Avg memory | 1.5 GiB |
| CPU user | 220.9s |
| CPU sys | 6.0s |
| 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(sort): passthrough last stream that left in `SortPreservingMergeStream` [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #23328:
URL: https://github.com/apache/datafusion/pull/23328#issuecomment-4896249743
๐ค Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/23328#issuecomment-4896056824)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
<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 passthrough-last-stream
--------------------
Benchmark clickbench_partitioned.json
--------------------
โโโโโโโโโโโโโณโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโโโโ
โ Query โ HEAD โ passthrough-last-stream โ Change โ
โกโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฉ
โ QQuery 0 โ 1.24 / 3.98 ยฑ5.40 / 14.79 ms โ 1.23 / 3.97 ยฑ5.40 / 14.77 ms โ no change โ
โ QQuery 1 โ 12.53 / 12.91 ยฑ0.23 / 13.16 ms โ 12.46 / 12.69 ยฑ0.13 / 12.82 ms โ no change โ
โ QQuery 2 โ 36.74 / 37.10 ยฑ0.35 / 37.65 ms โ 35.30 / 35.50 ยฑ0.26 / 36.02 ms โ no change โ
โ QQuery 3 โ 30.64 / 31.74 ยฑ1.34 / 34.37 ms โ 30.14 / 30.38 ยฑ0.32 / 30.98 ms โ no change โ
โ QQuery 4 โ 220.89 / 225.12 ยฑ2.65 / 229.20 ms โ 221.16 / 227.13 ยฑ4.08 / 232.89 ms โ no change โ
โ QQuery 5 โ 276.47 / 278.52 ยฑ2.99 / 284.46 ms โ 275.93 / 277.83 ยฑ1.65 / 280.70 ms โ no change โ
โ QQuery 6 โ 1.30 / 1.46 ยฑ0.24 / 1.94 ms โ 1.27 / 1.41 ยฑ0.22 / 1.84 ms โ no change โ
โ QQuery 7 โ 13.97 / 14.33 ยฑ0.34 / 14.90 ms โ 13.62 / 14.13 ยฑ0.31 / 14.45 ms โ no change โ
โ QQuery 8 โ 323.98 / 336.18 ยฑ16.50 / 368.85 ms โ 321.06 / 325.37 ยฑ4.28 / 333.41 ms โ no change โ
โ QQuery 9 โ 494.47 / 521.56 ยฑ19.74 / 542.40 ms โ 459.87 / 477.26 ยฑ28.03 / 532.57 ms โ +1.09x faster โ
โ QQuery 10 โ 71.90 / 72.53 ยฑ0.44 / 73.01 ms โ 70.59 / 71.85 ยฑ1.29 / 74.26 ms โ no change โ
โ QQuery 11 โ 82.80 / 84.07 ยฑ0.86 / 85.49 ms โ 81.53 / 82.71 ยฑ1.73 / 86.10 ms โ no change โ
โ QQuery 12 โ 272.72 / 276.46 ยฑ4.03 / 283.38 ms โ 267.63 / 286.23 ยฑ17.08 / 311.32 ms โ no change โ
โ QQuery 13 โ 371.44 / 376.59 ยฑ2.76 / 379.22 ms โ 409.98 / 424.60 ยฑ13.45 / 447.20 ms โ 1.13x slower โ
โ QQuery 14 โ 287.12 / 291.06 ยฑ3.28 / 295.09 ms โ 286.15 / 301.22 ยฑ14.16 / 327.74 ms โ no change โ
โ QQuery 15 โ 274.06 / 279.41 ยฑ5.64 / 289.41 ms โ 269.90 / 283.96 ยฑ9.60 / 299.41 ms โ no change โ
โ QQuery 16 โ 618.69 / 629.81 ยฑ7.66 / 640.46 ms โ 609.26 / 621.65 ยฑ9.06 / 632.70 ms โ no change โ
โ QQuery 17 โ 625.46 / 638.15 ยฑ7.26 / 646.57 ms โ 615.90 / 660.14 ยฑ52.03 / 757.58 ms โ no change โ
โ QQuery 18 โ 1266.62 / 1363.60 ยฑ106.12 / 1518.84 ms โ 1277.23 / 1308.56 ยฑ45.81 / 1399.44 ms โ no change โ
โ QQuery 19 โ 28.87 / 35.04 ยฑ11.81 / 58.65 ms โ 27.48 / 35.61 ยฑ16.03 / 67.67 ms โ no change โ
โ QQuery 20 โ 520.77 / 529.54 ยฑ7.51 / 542.33 ms โ 517.35 / 522.49 ยฑ4.92 / 530.73 ms โ no change โ
โ QQuery 21 โ 516.06 / 521.79 ยฑ4.25 / 528.50 ms โ 512.71 / 518.20 ยฑ5.49 / 525.82 ms โ no change โ
โ QQuery 22 โ 995.28 / 1024.92 ยฑ25.58 / 1056.18 ms โ 981.25 / 992.80 ยฑ6.75 / 1001.57 ms โ no change โ
โ QQuery 23 โ 3114.47 / 3138.87 ยฑ27.82 / 3189.13 ms โ 3133.73 / 3226.42 ยฑ55.58 / 3290.41 ms โ no change โ
โ QQuery 24 โ 41.44 / 42.41 ยฑ0.73 / 43.62 ms โ 42.10 / 47.24 ยฑ8.85 / 64.90 ms โ 1.11x slower โ
โ QQuery 25 โ 112.45 / 114.91 ยฑ1.84 / 118.15 ms โ 113.76 / 114.63 ยฑ0.84 / 116.22 ms โ no change โ
โ QQuery 26 โ 42.24 / 43.68 ยฑ1.77 / 47.12 ms โ 41.97 / 42.28 ยฑ0.29 / 42.80 ms โ no change โ
โ QQuery 27 โ 681.52 / 703.81 ยฑ12.18 / 716.72 ms โ 670.38 / 676.21 ยฑ6.43 / 688.48 ms โ no change โ
โ QQuery 28 โ 3052.80 / 3090.05 ยฑ39.03 / 3141.74 ms โ 3025.28 / 3089.54 ยฑ62.09 / 3203.71 ms โ no change โ
โ QQuery 29 โ 40.91 / 49.09 ยฑ15.56 / 80.21 ms โ 40.55 / 44.14 ยฑ6.58 / 57.30 ms โ +1.11x faster โ
โ QQuery 30 โ 297.86 / 305.71 ยฑ4.76 / 312.50 ms โ 301.65 / 305.71 ยฑ3.79 / 312.80 ms โ no change โ
โ QQuery 31 โ 290.15 / 303.55 ยฑ20.05 / 343.16 ms โ 278.28 / 292.89 ยฑ10.18 / 307.05 ms โ no change โ
โ QQuery 32 โ 927.53 / 964.78 ยฑ30.72 / 1005.05 ms โ 929.25 / 969.80 ยฑ40.72 / 1047.56 ms โ no change โ
โ QQuery 33 โ 1477.42 / 1564.61 ยฑ72.68 / 1656.80 ms โ 1496.36 / 1568.97 ยฑ84.66 / 1734.65 ms โ no change โ
โ QQuery 34 โ 1467.56 / 1554.49 ยฑ81.72 / 1693.14 ms โ 1509.73 / 1526.34 ยฑ10.08 / 1537.24 ms โ no change โ
โ QQuery 35 โ 280.89 / 308.21 ยฑ28.28 / 358.10 ms โ 295.81 / 358.90 ยฑ54.88 / 457.38 ms โ 1.16x slower โ
โ QQuery 36 โ 67.27 / 76.41 ยฑ9.07 / 90.00 ms โ 77.23 / 81.71 ยฑ5.38 / 92.11 ms โ 1.07x slower โ
โ QQuery 37 โ 35.84 / 37.49 ยฑ2.53 / 42.51 ms โ 38.45 / 42.77 ยฑ4.88 / 51.94 ms โ 1.14x slower โ
โ QQuery 38 โ 39.89 / 50.33 ยฑ8.30 / 60.89 ms โ 45.04 / 53.83 ยฑ9.79 / 72.22 ms โ 1.07x slower โ
โ QQuery 39 โ 132.58 / 146.74 ยฑ7.18 / 151.38 ms โ 155.64 / 175.75 ยฑ13.21 / 193.60 ms โ 1.20x slower โ
โ QQuery 40 โ 13.90 / 14.33 ยฑ0.45 / 15.19 ms โ 15.43 / 17.75 ยฑ2.26 / 21.18 ms โ 1.24x slower โ
โ QQuery 41 โ 13.69 / 14.13 ยฑ0.34 / 14.53 ms โ 14.79 / 14.96 ยฑ0.15 / 15.17 ms โ 1.06x slower โ
โ QQuery 42 โ 13.30 / 16.75 ยฑ4.62 / 25.22 ms โ 14.03 / 15.78 ยฑ2.99 / 21.74 ms โ +1.06x faster โ
โโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโโ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโ
โ Benchmark Summary โ โ
โกโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฉ
โ Total Time (HEAD) โ 20126.23ms โ
โ Total Time (passthrough-last-stream) โ 20181.29ms โ
โ Average Time (HEAD) โ 468.05ms โ
โ Average Time (passthrough-last-stream) โ 469.33ms โ
โ Queries Faster โ 3 โ
โ Queries Slower โ 9 โ
โ Queries with No Change โ 31 โ
โ Queries with Failure โ 0 โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**clickbench_partitioned โ base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 105.0s |
| Peak memory | 11.8 GiB |
| Avg memory | 4.6 GiB |
| CPU user | 1033.9s |
| CPU sys | 68.6s |
| Peak spill | 0 B |
**clickbench_partitioned โ branch**
| Metric | Value |
|--------|-------|
| Wall time | 105.0s |
| Peak memory | 11.4 GiB |
| Avg memory | 4.2 GiB |
| CPU user | 1034.6s |
| CPU sys | 71.2s |
| 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(sort): passthrough last stream that left in `SortPreservingMergeStream` [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #23328:
URL: https://github.com/apache/datafusion/pull/23328#issuecomment-4896083737
๐ค Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/23328#issuecomment-4896056824)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4896056824-872-9z9x9 6.12.85+ #1 SMP Mon May 11 08:17:35 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 passthrough-last-stream (d7156e6ebc32377d63ebcb7609bba609a2962924) to 6a0e76e (merge-base) [diff](https://github.com/apache/datafusion/compare/6a0e76ed692afaafd616e0d98d51bef63c1ff02d..d7156e6ebc32377d63ebcb7609bba609a2962924) using: tpch
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(sort): passthrough last stream that left in `SortPreservingMergeStream` [datafusion]
Posted by "alamb (via GitHub)" <gi...@apache.org>.
alamb commented on PR #23328:
URL: https://github.com/apache/datafusion/pull/23328#issuecomment-4910098110
> @alamb the current failing tests are from a fuzz tests you added that verify that we output the same size for all the batches except 1, however this prevent one of the optimization here that just passthrough the stream and avoid the copy or slicing, WDYT?
I think we should look as to why it was added the first time and make sure the intent is still satisfied
--
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