You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@datafusion.apache.org by "Dandandan (via GitHub)" <gi...@apache.org> on 2026/04/16 15:47:25 UTC
[PR] perf(repartition): batch reservation + sends via per-partition local buffers [datafusion]
Dandandan opened a new pull request, #21677:
URL: https://github.com/apache/datafusion/pull/21677
## Which issue does this PR close?
<!-- Related to improving CPU utilization in parallel query execution (e.g. ClickBench). -->
## Rationale for this change
In `RepartitionExec::pull_from_input`, each input task partitions an input batch and sends the resulting sub-batches to output channels. On every sub-batch, the input task:
1. Acquires the output partition's memory reservation `Mutex` and calls `try_grow`
2. Sends the batch through the distributor channel (which internally takes `channel.state.lock()` and potentially the gate mutex)
When hash-partitioning an input batch of size `S` into `N` output partitions, each output gets a sub-batch of size `~S/N`. With many output partitions (e.g. 16 cores โ 16 output partitions), these sub-batches are very small, so the per-batch synchronization cost dominates.
## What changes are included in this PR?
Each input task now accumulates partitioned sub-batches in per-output-partition local buffers (`Vec<PartitionBuffer>`) and only flushes when a buffer reaches `FLUSH_THRESHOLD_BYTES` (256 KiB). Each flush performs a single `try_grow` call for the total buffered size instead of once per sub-batch. If `try_grow` fails, all buffered batches for that partition are spilled together.
End-of-input triggers a final flush of any remaining buffered data per partition.
## Are these changes tested?
Covered by the existing repartition test suite (41 tests pass), including the spilling, delayed-stream, dropped-output-stream, and ordering-preservation tests.
## Are there any user-facing changes?
No. The change only affects batching granularity inside `RepartitionExec`; memory semantics, spilling behavior, and output ordering are preserved.
๐ค Generated with [Claude Code](https://claude.com/claude-code)
--
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(repartition): batch reservation + sends via per-partition local buffers [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21677:
URL: https://github.com/apache/datafusion/pull/21677#issuecomment-4261561301
๐ค Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21677#issuecomment-4261425493)
**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 repartition-local-buffers
--------------------
Benchmark tpcds_sf1.json
--------------------
โโโโโโโโโโโโโณโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโโโโ
โ Query โ HEAD โ repartition-local-buffers โ Change โ
โกโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฉ
โ QQuery 1 โ 6.60 / 7.02 ยฑ0.72 / 8.45 ms โ 6.60 / 6.99 ยฑ0.73 / 8.45 ms โ no change โ
โ QQuery 2 โ 148.29 / 149.99 ยฑ1.15 / 151.26 ms โ 145.60 / 147.16 ยฑ0.99 / 148.42 ms โ no change โ
โ QQuery 3 โ 116.13 / 116.76 ยฑ0.76 / 118.22 ms โ 114.97 / 115.88 ยฑ0.80 / 117.31 ms โ no change โ
โ QQuery 4 โ 1386.41 / 1403.32 ยฑ17.24 / 1434.87 ms โ 1319.64 / 1387.12 ยฑ54.80 / 1447.15 ms โ no change โ
โ QQuery 5 โ 176.14 / 178.84 ยฑ2.83 / 182.36 ms โ 175.42 / 177.01 ยฑ1.44 / 178.75 ms โ no change โ
โ QQuery 6 โ 851.87 / 875.69 ยฑ17.57 / 897.56 ms โ 862.40 / 886.30 ยฑ29.15 / 941.89 ms โ no change โ
โ QQuery 7 โ 350.10 / 352.72 ยฑ2.70 / 357.49 ms โ 349.17 / 354.70 ยฑ3.74 / 359.75 ms โ no change โ
โ QQuery 8 โ 118.83 / 120.21 ยฑ0.76 / 120.86 ms โ 117.92 / 119.92 ยฑ1.05 / 120.85 ms โ no change โ
โ QQuery 9 โ 103.93 / 107.95 ยฑ2.01 / 109.28 ms โ 103.17 / 107.30 ยฑ3.38 / 111.21 ms โ no change โ
โ QQuery 10 โ 108.02 / 109.44 ยฑ0.84 / 110.54 ms โ 108.87 / 110.98 ยฑ1.65 / 113.96 ms โ no change โ
โ QQuery 11 โ 1004.59 / 1029.29 ยฑ18.95 / 1056.38 ms โ 959.64 / 1003.33 ยฑ49.26 / 1067.00 ms โ no change โ
โ QQuery 12 โ 45.71 / 47.56 ยฑ1.27 / 49.49 ms โ 46.77 / 48.24 ยฑ1.67 / 51.30 ms โ no change โ
โ QQuery 13 โ 416.52 / 419.39 ยฑ2.57 / 423.74 ms โ 406.43 / 418.03 ยฑ6.23 / 423.72 ms โ no change โ
โ QQuery 14 โ 1021.44 / 1031.97 ยฑ8.49 / 1043.68 ms โ 1009.81 / 1014.01 ยฑ5.60 / 1024.95 ms โ no change โ
โ QQuery 15 โ 15.53 / 16.23 ยฑ0.87 / 17.95 ms โ 16.62 / 18.44 ยฑ1.27 / 19.99 ms โ 1.14x slower โ
โ QQuery 16 โ 7.12 / 7.86 ยฑ0.73 / 8.76 ms โ 7.53 / 7.66 ยฑ0.16 / 7.98 ms โ no change โ
โ QQuery 17 โ 227.75 / 231.70 ยฑ2.92 / 236.81 ms โ 227.62 / 233.44 ยฑ4.49 / 239.85 ms โ no change โ
โ QQuery 18 โ 126.55 / 130.43 ยฑ2.51 / 133.10 ms โ 129.26 / 131.84 ยฑ2.23 / 135.29 ms โ no change โ
โ QQuery 19 โ 164.42 / 165.89 ยฑ1.23 / 167.39 ms โ 156.63 / 158.82 ยฑ1.34 / 160.44 ms โ no change โ
โ QQuery 20 โ 14.38 / 14.56 ยฑ0.11 / 14.69 ms โ 14.24 / 14.74 ยฑ0.28 / 15.09 ms โ no change โ
โ QQuery 21 โ 20.26 / 20.89 ยฑ0.70 / 22.21 ms โ 19.31 / 20.43 ยฑ0.83 / 21.88 ms โ no change โ
โ QQuery 22 โ 491.64 / 499.04 ยฑ7.14 / 510.56 ms โ 482.79 / 486.79 ยฑ4.39 / 495.09 ms โ no change โ
โ QQuery 23 โ 864.06 / 894.22 ยฑ16.34 / 909.61 ms โ 897.53 / 924.19 ยฑ21.89 / 947.21 ms โ no change โ
โ QQuery 24 โ 385.48 / 388.90 ยฑ2.03 / 391.86 ms โ 390.44 / 399.33 ยฑ6.29 / 405.89 ms โ no change โ
โ QQuery 25 โ 342.36 / 346.47 ยฑ4.30 / 352.58 ms โ 351.19 / 354.99 ยฑ2.12 / 357.49 ms โ no change โ
โ QQuery 26 โ 82.92 / 84.83 ยฑ1.65 / 87.04 ms โ 81.93 / 84.21 ยฑ1.52 / 86.53 ms โ no change โ
โ QQuery 27 โ 7.13 / 7.91 ยฑ0.88 / 9.53 ms โ 7.03 / 7.75 ยฑ0.62 / 8.60 ms โ no change โ
โ QQuery 28 โ 151.23 / 153.20 ยฑ1.42 / 155.06 ms โ 148.95 / 151.07 ยฑ2.39 / 155.15 ms โ no change โ
โ QQuery 29 โ 288.81 / 291.18 ยฑ1.24 / 292.42 ms โ 280.46 / 284.73 ยฑ3.99 / 291.28 ms โ no change โ
โ QQuery 30 โ 42.72 / 45.31 ยฑ1.82 / 47.37 ms โ 45.21 / 46.73 ยฑ1.22 / 48.23 ms โ no change โ
โ QQuery 31 โ 171.21 / 176.84 ยฑ3.43 / 181.32 ms โ 177.01 / 180.27 ยฑ2.36 / 183.52 ms โ no change โ
โ QQuery 32 โ 57.86 / 59.04 ยฑ1.00 / 60.11 ms โ 57.71 / 67.86 ยฑ17.53 / 102.89 ms โ 1.15x slower โ
โ QQuery 33 โ 144.17 / 147.56 ยฑ2.43 / 150.32 ms โ 144.66 / 146.16 ยฑ1.29 / 147.74 ms โ no change โ
โ QQuery 34 โ 7.06 / 7.27 ยฑ0.20 / 7.64 ms โ 7.24 / 8.09 ยฑ0.88 / 9.39 ms โ 1.11x slower โ
โ QQuery 35 โ 110.15 / 113.08 ยฑ1.47 / 113.98 ms โ 110.05 / 112.58 ยฑ1.53 / 114.66 ms โ no change โ
โ QQuery 36 โ 6.92 / 7.03 ยฑ0.12 / 7.25 ms โ 6.79 / 6.98 ยฑ0.11 / 7.13 ms โ no change โ
โ QQuery 37 โ 8.56 / 9.76 ยฑ0.91 / 11.11 ms โ 8.49 / 9.05 ยฑ0.41 / 9.49 ms โ +1.08x faster โ
โ QQuery 38 โ 88.03 / 92.10 ยฑ4.31 / 100.38 ms โ 88.97 / 92.52 ยฑ4.46 / 101.03 ms โ no change โ
โ QQuery 39 โ 133.25 / 135.44 ยฑ1.73 / 137.77 ms โ 121.48 / 129.39 ยฑ5.74 / 136.33 ms โ no change โ
โ QQuery 40 โ 113.82 / 119.10 ยฑ7.54 / 133.92 ms โ 110.06 / 116.00 ยฑ7.61 / 130.98 ms โ no change โ
โ QQuery 41 โ 14.34 / 15.61 ยฑ1.26 / 17.57 ms โ 14.51 / 15.75 ยฑ0.99 / 16.93 ms โ no change โ
โ QQuery 42 โ 112.01 / 112.72 ยฑ0.41 / 113.15 ms โ 111.03 / 112.19 ยฑ0.80 / 113.44 ms โ no change โ
โ QQuery 43 โ 5.95 / 6.15 ยฑ0.28 / 6.70 ms โ 6.14 / 6.38 ยฑ0.23 / 6.78 ms โ no change โ
โ QQuery 44 โ 11.83 / 12.27 ยฑ0.60 / 13.47 ms โ 12.19 / 12.72 ยฑ0.51 / 13.62 ms โ no change โ
โ QQuery 45 โ 50.64 / 51.83 ยฑ0.62 / 52.43 ms โ 53.92 / 55.48 ยฑ1.34 / 57.95 ms โ 1.07x slower โ
โ QQuery 46 โ 8.31 / 8.56 ยฑ0.19 / 8.88 ms โ 9.26 / 9.55 ยฑ0.17 / 9.74 ms โ 1.12x slower โ
โ QQuery 47 โ 777.15 / 804.13 ยฑ18.61 / 822.99 ms โ 810.01 / 819.67 ยฑ6.14 / 828.01 ms โ no change โ
โ QQuery 48 โ 291.43 / 293.89 ยฑ2.02 / 295.94 ms โ 299.21 / 304.72 ยฑ4.48 / 311.50 ms โ no change โ
โ QQuery 49 โ 254.58 / 258.99 ยฑ2.89 / 263.31 ms โ 258.90 / 262.33 ยฑ3.74 / 269.33 ms โ no change โ
โ QQuery 50 โ 226.77 / 234.55 ยฑ8.96 / 251.63 ms โ 228.71 / 241.65 ยฑ8.08 / 253.64 ms โ no change โ
โ QQuery 51 โ 188.19 / 189.64 ยฑ1.26 / 191.85 ms โ 176.43 / 180.18 ยฑ3.13 / 184.73 ms โ no change โ
โ QQuery 52 โ 112.22 / 114.77 ยฑ1.97 / 117.69 ms โ 109.43 / 110.95 ยฑ1.22 / 112.74 ms โ no change โ
โ QQuery 53 โ 105.86 / 108.10 ยฑ2.59 / 112.62 ms โ 105.36 / 105.99 ยฑ0.47 / 106.72 ms โ no change โ
โ QQuery 54 โ 147.65 / 150.17 ยฑ2.82 / 155.52 ms โ 149.64 / 150.94 ยฑ0.84 / 152.27 ms โ no change โ
โ QQuery 55 โ 108.98 / 111.19 ยฑ1.98 / 113.66 ms โ 109.43 / 110.20 ยฑ0.62 / 111.31 ms โ no change โ
โ QQuery 56 โ 142.56 / 145.70 ยฑ2.30 / 149.23 ms โ 140.62 / 144.76 ยฑ2.50 / 148.10 ms โ no change โ
โ QQuery 57 โ 176.52 / 179.14 ยฑ1.72 / 181.29 ms โ 173.27 / 177.30 ยฑ2.64 / 181.37 ms โ no change โ
โ QQuery 58 โ 293.54 / 300.07 ยฑ6.79 / 309.17 ms โ 289.99 / 302.80 ยฑ12.44 / 323.94 ms โ no change โ
โ QQuery 59 โ 199.84 / 202.90 ยฑ2.63 / 207.23 ms โ 202.62 / 204.66 ยฑ1.36 / 206.24 ms โ no change โ
โ QQuery 60 โ 145.20 / 148.00 ยฑ1.89 / 150.37 ms โ 142.59 / 146.24 ยฑ2.97 / 151.66 ms โ no change โ
โ QQuery 61 โ 12.77 / 13.00 ยฑ0.25 / 13.46 ms โ 12.93 / 13.40 ยฑ0.51 / 14.24 ms โ no change โ
โ QQuery 62 โ 897.56 / 921.32 ยฑ14.98 / 938.14 ms โ 910.94 / 954.43 ยฑ33.47 / 1014.02 ms โ no change โ
โ QQuery 63 โ 105.85 / 108.32 ยฑ2.43 / 112.41 ms โ 104.41 / 107.10 ยฑ1.90 / 109.93 ms โ no change โ
โ QQuery 64 โ 700.92 / 711.06 ยฑ7.09 / 719.08 ms โ 696.48 / 714.51 ยฑ10.44 / 725.61 ms โ no change โ
โ QQuery 65 โ 259.01 / 268.16 ยฑ5.93 / 275.28 ms โ 253.90 / 257.28 ยฑ2.18 / 260.54 ms โ no change โ
โ QQuery 66 โ 249.48 / 266.26 ยฑ10.99 / 279.31 ms โ 245.63 / 272.37 ยฑ14.73 / 287.05 ms โ no change โ
โ QQuery 67 โ 313.44 / 332.59 ยฑ10.98 / 345.15 ms โ 323.40 / 330.13 ยฑ5.72 / 338.75 ms โ no change โ
โ QQuery 68 โ 9.36 / 10.54 ยฑ1.08 / 12.53 ms โ 9.32 / 11.11 ยฑ1.95 / 14.90 ms โ 1.05x slower โ
โ QQuery 69 โ 103.21 / 105.63 ยฑ1.64 / 108.13 ms โ 106.07 / 107.02 ยฑ0.92 / 108.46 ms โ no change โ
โ QQuery 70 โ 350.00 / 363.41 ยฑ13.65 / 389.16 ms โ 340.16 / 356.82 ยฑ17.54 / 389.64 ms โ no change โ
โ QQuery 71 โ 141.67 / 144.69 ยฑ1.75 / 146.38 ms โ 134.62 / 138.65 ยฑ3.54 / 144.28 ms โ no change โ
โ QQuery 72 โ 637.85 / 651.44 ยฑ9.20 / 662.21 ms โ 611.86 / 627.77 ยฑ12.53 / 646.80 ms โ no change โ
โ QQuery 73 โ 7.54 / 9.60 ยฑ1.58 / 11.67 ms โ 6.65 / 7.51 ยฑ0.69 / 8.39 ms โ +1.28x faster โ
โ QQuery 74 โ 628.77 / 657.96 ยฑ22.58 / 681.49 ms โ 595.40 / 614.01 ยฑ17.38 / 641.08 ms โ +1.07x faster โ
โ QQuery 75 โ 282.11 / 283.20 ยฑ0.82 / 284.25 ms โ 275.74 / 282.49 ยฑ3.95 / 287.32 ms โ no change โ
โ QQuery 76 โ 136.30 / 138.12 ยฑ1.48 / 140.29 ms โ 135.66 / 136.69 ยฑ0.96 / 138.00 ms โ no change โ
โ QQuery 77 โ 191.38 / 194.86 ยฑ2.54 / 198.92 ms โ 191.55 / 192.84 ยฑ1.45 / 195.07 ms โ no change โ
โ QQuery 78 โ 342.37 / 349.58 ยฑ4.37 / 354.07 ms โ 342.33 / 348.09 ยฑ3.96 / 354.44 ms โ no change โ
โ QQuery 79 โ 235.68 / 239.87 ยฑ4.70 / 249.03 ms โ 250.73 / 252.31 ยฑ1.06 / 254.04 ms โ 1.05x slower โ
โ QQuery 80 โ 324.08 / 327.59 ยฑ2.90 / 332.64 ms โ 327.37 / 330.01 ยฑ2.53 / 334.06 ms โ no change โ
โ QQuery 81 โ 26.24 / 27.42 ยฑ0.93 / 28.52 ms โ 26.09 / 27.14 ยฑ0.65 / 28.09 ms โ no change โ
โ QQuery 82 โ 204.41 / 208.53 ยฑ3.65 / 215.22 ms โ 206.40 / 210.35 ยฑ4.02 / 216.51 ms โ no change โ
โ QQuery 83 โ 39.08 / 40.31 ยฑ1.29 / 42.75 ms โ 40.91 / 42.61 ยฑ1.34 / 44.01 ms โ 1.06x slower โ
โ QQuery 84 โ 48.91 / 49.92 ยฑ1.09 / 51.92 ms โ 48.90 / 50.26 ยฑ0.75 / 51.18 ms โ no change โ
โ QQuery 85 โ 148.07 / 151.07 ยฑ2.25 / 154.47 ms โ 154.60 / 156.76 ยฑ1.54 / 159.31 ms โ no change โ
โ QQuery 86 โ 39.43 / 40.39 ยฑ1.28 / 42.89 ms โ 40.74 / 42.14 ยฑ0.81 / 42.82 ms โ no change โ
โ QQuery 87 โ 87.95 / 91.26 ยฑ2.84 / 95.62 ms โ 91.87 / 95.31 ยฑ3.34 / 101.50 ms โ no change โ
โ QQuery 88 โ 101.47 / 102.93 ยฑ0.81 / 103.83 ms โ 103.45 / 105.45 ยฑ1.39 / 107.11 ms โ no change โ
โ QQuery 89 โ 120.35 / 122.48 ยฑ2.12 / 126.04 ms โ 121.82 / 124.74 ยฑ1.75 / 126.87 ms โ no change โ
โ QQuery 90 โ 23.84 / 24.67 ยฑ0.92 / 26.47 ms โ 25.13 / 25.88 ยฑ0.65 / 26.93 ms โ no change โ
โ QQuery 91 โ 61.96 / 63.65 ยฑ1.02 / 64.67 ms โ 66.28 / 67.70 ยฑ0.96 / 68.66 ms โ 1.06x slower โ
โ QQuery 92 โ 59.21 / 60.53 ยฑ1.13 / 62.30 ms โ 59.97 / 60.37 ยฑ0.33 / 60.90 ms โ no change โ
โ QQuery 93 โ 193.36 / 194.97 ยฑ1.85 / 198.11 ms โ 206.09 / 209.08 ยฑ2.65 / 213.92 ms โ 1.07x slower โ
โ QQuery 94 โ 63.45 / 63.97 ยฑ0.72 / 65.37 ms โ 64.25 / 65.16 ยฑ0.79 / 66.29 ms โ no change โ
โ QQuery 95 โ 130.90 / 131.83 ยฑ1.05 / 133.52 ms โ 137.88 / 139.19 ยฑ0.85 / 140.29 ms โ 1.06x slower โ
โ QQuery 96 โ 74.75 / 76.51 ยฑ1.17 / 78.24 ms โ 76.88 / 77.63 ยฑ0.42 / 78.18 ms โ no change โ
โ QQuery 97 โ 130.75 / 133.82 ยฑ2.30 / 137.46 ms โ 128.54 / 132.99 ยฑ2.66 / 135.90 ms โ no change โ
โ QQuery 98 โ 155.20 / 157.96 ยฑ2.25 / 161.25 ms โ 155.98 / 159.87 ยฑ2.17 / 162.21 ms โ no change โ
โ QQuery 99 โ 10812.01 / 10860.24 ยฑ30.26 / 10899.73 ms โ 10843.05 / 10872.31 ยฑ25.18 / 10906.98 ms โ no change โ
โโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโโ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโ
โ Benchmark Summary โ โ
โกโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฉ
โ Total Time (HEAD) โ 32264.06ms โ
โ Total Time (repartition-local-buffers) โ 32284.96ms โ
โ Average Time (HEAD) โ 325.90ms โ
โ Average Time (repartition-local-buffers) โ 326.11ms โ
โ Queries Faster โ 3 โ
โ Queries Slower โ 11 โ
โ Queries with No Change โ 85 โ
โ Queries with Failure โ 0 โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**tpcds โ base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 161.7s |
| Peak memory | 5.7 GiB |
| Avg memory | 4.4 GiB |
| CPU user | 266.9s |
| CPU sys | 17.9s |
| Peak spill | 0 B |
**tpcds โ branch**
| Metric | Value |
|--------|-------|
| Wall time | 161.7s |
| Peak memory | 5.2 GiB |
| Avg memory | 4.3 GiB |
| CPU user | 263.7s |
| CPU sys | 17.4s |
| 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(repartition): batch reservation + sends via per-partition local buffers [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21677:
URL: https://github.com/apache/datafusion/pull/21677#issuecomment-4261443448
๐ค Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21677#issuecomment-4261425493)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4261425493-1376-8j85d 6.12.55+ #1 SMP Sun Feb 1 08:59:41 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 repartition-local-buffers (f2270ad3c4f04dd917b5019d5274b499c18d4e18) to 3b5008a (merge-base) [diff](https://github.com/apache/datafusion/compare/3b5008a47c8a8854bd0fe6e9eed1236dce59e216..f2270ad3c4f04dd917b5019d5274b499c18d4e18) 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(repartition): batch reservation + sends via per-partition local buffers [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21677:
URL: https://github.com/apache/datafusion/pull/21677#issuecomment-4261445563
๐ค Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21677#issuecomment-4261425493)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4261425493-1378-bnxwp 6.12.55+ #1 SMP Sun Feb 1 08:59:41 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 repartition-local-buffers (f2270ad3c4f04dd917b5019d5274b499c18d4e18) to 3b5008a (merge-base) [diff](https://github.com/apache/datafusion/compare/3b5008a47c8a8854bd0fe6e9eed1236dce59e216..f2270ad3c4f04dd917b5019d5274b499c18d4e18) 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(repartition): batch reservation + sends via per-partition local buffers [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21677:
URL: https://github.com/apache/datafusion/pull/21677#issuecomment-4261562385
๐ค Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21677#issuecomment-4261425493)
**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 repartition-local-buffers
--------------------
Benchmark clickbench_partitioned.json
--------------------
โโโโโโโโโโโโโณโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโโโ
โ Query โ HEAD โ repartition-local-buffers โ Change โ
โกโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฉ
โ QQuery 0 โ 1.24 / 4.58 ยฑ6.51 / 17.59 ms โ 1.22 / 4.50 ยฑ6.42 / 17.33 ms โ no change โ
โ QQuery 1 โ 14.59 / 15.43 ยฑ0.51 / 16.14 ms โ 14.58 / 14.94 ยฑ0.19 / 15.11 ms โ no change โ
โ QQuery 2 โ 44.56 / 44.98 ยฑ0.27 / 45.33 ms โ 44.83 / 45.77 ยฑ1.08 / 47.77 ms โ no change โ
โ QQuery 3 โ 40.52 / 41.49 ยฑ1.19 / 43.79 ms โ 40.06 / 41.10 ยฑ0.89 / 42.67 ms โ no change โ
โ QQuery 4 โ 296.27 / 300.34 ยฑ2.70 / 303.39 ms โ 288.32 / 295.48 ยฑ4.55 / 300.98 ms โ no change โ
โ QQuery 5 โ 349.97 / 356.52 ยฑ4.84 / 363.82 ms โ 350.46 / 360.58 ยฑ7.58 / 369.48 ms โ no change โ
โ QQuery 6 โ 5.05 / 6.00 ยฑ1.08 / 8.10 ms โ 5.16 / 6.78 ยฑ1.13 / 8.40 ms โ 1.13x slower โ
โ QQuery 7 โ 17.31 / 17.65 ยฑ0.32 / 18.24 ms โ 17.75 / 18.81 ยฑ1.02 / 20.73 ms โ 1.07x slower โ
โ QQuery 8 โ 439.08 / 443.10 ยฑ3.30 / 447.21 ms โ 429.23 / 442.88 ยฑ7.85 / 449.46 ms โ no change โ
โ QQuery 9 โ 681.98 / 685.11 ยฑ4.23 / 693.48 ms โ 659.42 / 688.39 ยฑ18.79 / 717.77 ms โ no change โ
โ QQuery 10 โ 94.63 / 96.67 ยฑ2.79 / 102.16 ms โ 94.37 / 99.50 ยฑ5.49 / 109.86 ms โ no change โ
โ QQuery 11 โ 105.79 / 109.97 ยฑ2.80 / 113.86 ms โ 108.40 / 111.46 ยฑ2.87 / 115.26 ms โ no change โ
โ QQuery 12 โ 351.76 / 361.26 ยฑ8.57 / 376.16 ms โ 350.55 / 360.76 ยฑ10.52 / 379.62 ms โ no change โ
โ QQuery 13 โ 474.93 / 478.11 ยฑ1.80 / 479.97 ms โ 476.59 / 486.51 ยฑ8.08 / 497.81 ms โ no change โ
โ QQuery 14 โ 344.28 / 351.16 ยฑ3.78 / 355.15 ms โ 349.09 / 354.30 ยฑ3.98 / 360.95 ms โ no change โ
โ QQuery 15 โ 361.56 / 377.33 ยฑ12.04 / 396.27 ms โ 359.97 / 385.76 ยฑ14.84 / 401.63 ms โ no change โ
โ QQuery 16 โ 721.96 / 731.05 ยฑ7.07 / 739.86 ms โ 723.93 / 738.57 ยฑ19.63 / 777.49 ms โ no change โ
โ QQuery 17 โ 719.36 / 725.06 ยฑ5.01 / 731.90 ms โ 725.96 / 731.66 ยฑ4.35 / 738.99 ms โ no change โ
โ QQuery 18 โ 1468.44 / 1510.21 ยฑ46.67 / 1600.01 ms โ 1439.12 / 1527.79 ยฑ57.17 / 1603.09 ms โ no change โ
โ QQuery 19 โ 35.95 / 38.30 ยฑ2.72 / 42.89 ms โ 37.33 / 45.03 ยฑ13.96 / 72.91 ms โ 1.18x slower โ
โ QQuery 20 โ 720.03 / 737.18 ยฑ18.88 / 760.74 ms โ 723.13 / 742.16 ยฑ19.95 / 776.75 ms โ no change โ
โ QQuery 21 โ 762.64 / 769.66 ยฑ5.62 / 779.83 ms โ 775.43 / 790.33 ยฑ15.60 / 819.15 ms โ no change โ
โ QQuery 22 โ 1138.60 / 1143.87 ยฑ5.29 / 1151.57 ms โ 1151.93 / 1157.50 ยฑ5.84 / 1167.64 ms โ no change โ
โ QQuery 23 โ 3124.67 / 3141.19 ยฑ9.00 / 3151.39 ms โ 3157.05 / 3174.81 ยฑ15.52 / 3194.38 ms โ no change โ
โ QQuery 24 โ 101.41 / 106.55 ยฑ3.50 / 111.37 ms โ 106.75 / 112.29 ยฑ3.36 / 116.82 ms โ 1.05x slower โ
โ QQuery 25 โ 139.75 / 142.21 ยฑ1.53 / 144.11 ms โ 140.98 / 142.99 ยฑ1.96 / 146.28 ms โ no change โ
โ QQuery 26 โ 101.89 / 103.70 ยฑ1.54 / 105.92 ms โ 104.36 / 106.93 ยฑ1.92 / 109.07 ms โ no change โ
โ QQuery 27 โ 861.42 / 864.46 ยฑ2.05 / 866.87 ms โ 858.49 / 867.52 ยฑ5.57 / 873.63 ms โ no change โ
โ QQuery 28 โ 3240.13 / 3276.67 ยฑ20.85 / 3304.10 ms โ 3276.58 / 3290.85 ยฑ11.22 / 3309.26 ms โ no change โ
โ QQuery 29 โ 49.69 / 56.75 ยฑ6.48 / 68.28 ms โ 51.32 / 56.37 ยฑ2.84 / 58.99 ms โ no change โ
โ QQuery 30 โ 362.28 / 378.87 ยฑ9.93 / 387.21 ms โ 365.09 / 373.19 ยฑ6.80 / 382.23 ms โ no change โ
โ QQuery 31 โ 361.32 / 382.67 ยฑ13.69 / 399.29 ms โ 354.62 / 377.14 ยฑ24.46 / 423.81 ms โ no change โ
โ QQuery 32 โ 1212.45 / 1261.47 ยฑ39.79 / 1333.25 ms โ 1222.86 / 1283.09 ยฑ40.81 / 1325.01 ms โ no change โ
โ QQuery 33 โ 1569.06 / 1600.85 ยฑ26.23 / 1639.04 ms โ 1487.39 / 1564.94 ยฑ43.49 / 1621.41 ms โ no change โ
โ QQuery 34 โ 1524.14 / 1535.54 ยฑ14.16 / 1563.33 ms โ 1509.99 / 1531.31 ยฑ16.28 / 1556.46 ms โ no change โ
โ QQuery 35 โ 401.93 / 412.90 ยฑ13.65 / 439.39 ms โ 397.64 / 406.84 ยฑ13.04 / 432.61 ms โ no change โ
โ QQuery 36 โ 119.73 / 123.17 ยฑ3.83 / 130.42 ms โ 117.17 / 123.79 ยฑ4.38 / 129.63 ms โ no change โ
โ QQuery 37 โ 48.78 / 50.32 ยฑ1.50 / 53.20 ms โ 48.62 / 50.66 ยฑ2.18 / 54.82 ms โ no change โ
โ QQuery 38 โ 75.61 / 77.65 ยฑ1.36 / 79.35 ms โ 76.06 / 78.16 ยฑ1.99 / 81.32 ms โ no change โ
โ QQuery 39 โ 207.65 / 216.83 ยฑ8.12 / 230.79 ms โ 206.92 / 217.05 ยฑ6.52 / 224.95 ms โ no change โ
โ QQuery 40 โ 22.94 / 26.28 ยฑ2.20 / 29.32 ms โ 25.01 / 27.79 ยฑ2.09 / 30.48 ms โ 1.06x slower โ
โ QQuery 41 โ 20.93 / 21.95 ยฑ0.86 / 23.39 ms โ 20.85 / 22.24 ยฑ1.44 / 24.93 ms โ no change โ
โ QQuery 42 โ 20.06 / 20.76 ยฑ0.64 / 21.91 ms โ 20.01 / 20.64 ยฑ0.49 / 21.45 ms โ no change โ
โโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโ
โ Benchmark Summary โ โ
โกโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฉ
โ Total Time (HEAD) โ 23145.83ms โ
โ Total Time (repartition-local-buffers) โ 23279.15ms โ
โ Average Time (HEAD) โ 538.28ms โ
โ Average Time (repartition-local-buffers) โ 541.38ms โ
โ Queries Faster โ 0 โ
โ Queries Slower โ 5 โ
โ Queries with No Change โ 38 โ
โ Queries with Failure โ 0 โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**clickbench_partitioned โ base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 116.8s |
| Peak memory | 33.8 GiB |
| Avg memory | 25.9 GiB |
| CPU user | 1090.1s |
| CPU sys | 100.0s |
| Peak spill | 0 B |
**clickbench_partitioned โ branch**
| Metric | Value |
|--------|-------|
| Wall time | 117.3s |
| Peak memory | 35.2 GiB |
| Avg memory | 24.8 GiB |
| CPU user | 1087.5s |
| CPU sys | 104.7s |
| 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(repartition): batch reservation + sends via per-partition local buffers [datafusion]
Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan closed pull request #21677: perf(repartition): batch reservation + sends via per-partition local buffers
URL: https://github.com/apache/datafusion/pull/21677
--
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(repartition): batch reservation + sends via per-partition local buffers [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21677:
URL: https://github.com/apache/datafusion/pull/21677#issuecomment-4261443806
๐ค Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21677#issuecomment-4261425493)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4261425493-1377-vq4kk 6.12.55+ #1 SMP Sun Feb 1 08:59:41 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 repartition-local-buffers (f2270ad3c4f04dd917b5019d5274b499c18d4e18) to 3b5008a (merge-base) [diff](https://github.com/apache/datafusion/compare/3b5008a47c8a8854bd0fe6e9eed1236dce59e216..f2270ad3c4f04dd917b5019d5274b499c18d4e18) 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(repartition): batch reservation + sends via per-partition local buffers [datafusion]
Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #21677:
URL: https://github.com/apache/datafusion/pull/21677#issuecomment-4261425493
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