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 2026/04/20 12:57:10 UTC
[PR] perf(repartition): drop redundant reservation Mutex and halve per-batch Instant::now() calls [datafusion]
Dandandan opened a new pull request, #21745:
URL: https://github.com/apache/datafusion/pull/21745
## Which issue does this PR close?
- Closes #.
## Rationale for this change
`RepartitionExec`'s per-batch hot path has two small but measurable sources of overhead that are easy to remove without any semantic change:
1. The `SharedMemoryReservation` alias is `Arc<Mutex<MemoryReservation>>`, but every method on `MemoryReservation` (`try_grow`, `shrink`, `try_resize`, ...) already takes `&self` and mutates an `AtomicUsize` internally. The outer `Mutex` adds a lock acquire per batch on both the producer (`pull_from_input`'s `try_grow`) and the consumer (`PerPartitionStream`'s `shrink`) for no correctness benefit.
2. The per-sub-batch `send_time[partition].timer()` idiom calls `Instant::now()` twice per sub-batch (once on `timer()`, once on `done()`/drop). With a 16-output hash repartition that is 32 `Instant::now()` calls per input batch just for metrics.
A new criterion microbench (`datafusion/physical-plan/benches/repartition.rs`) was added to make both the current cost and the effect of the fix visible in isolation.
## What changes are included in this PR?
**`repartition/mod.rs`**
- Type the three `SharedMemoryReservation` fields (`OutputChannel`, `PartitionChannels`, `PerPartitionStream`) as `Arc<MemoryReservation>` and drop the three `.lock()` callsites. A doc comment on `OutputChannel` explains why the outer lock is unnecessary. The `SharedMemoryReservation` alias in `common.rs` is left alone — it is still used by `symmetric_hash_join`.
- Replace the per-sub-batch `ScopedTimerGuard` in the `pull_from_input` inner loop with a single advancing `Instant`. One `Instant::now()` per sub-batch, preserving the existing per-partition `send_time` metric.
**`benches/repartition.rs`** (new)
Criterion microbench that drives 1 M rows through `RepartitionExec` in three configurations: hash 16→16, round-robin 16→16, and hash 16→1 coalesce. Output partitions are drained concurrently so the measurement reflects real breaker behaviour (mpsc send, reservation lock, metric timers, `SpawnedTask` per input).
## Are these changes tested?
Yes — all 41 tests in `datafusion/physical-plan/src/repartition/tests` continue to pass, including the spill, memory-pool, order-preservation, and early-shutdown cases. Clippy and `cargo fmt` are clean.
Before / after on the new microbench (macOS ARM, 1 M rows):
| case | before | after | delta |
|---|---|---|---|
| `hash_16_to_16` | 1.62 ms | 1.55 ms | ~ −2.5% |
| `round_robin_16_to_16` | 303 µs | 268 µs | ~ −10% |
| `hash_16_to_1_coalesce` | 932 µs | 917 µs | ~ −5% |
The hash cases are dominated by `BatchPartitioner::partition_iter` (hashing + `take_arrays`) itself, so the plumbing wins show up most clearly on the round-robin variant where hashing is not the bottleneck.
## Are there any user-facing changes?
No. Metrics keep the same shape (per-partition `send_time`), no public API changes, no behavioural changes to spill / backpressure / ordering.
🤖 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): drop redundant reservation Mutex and halve per-batch Instant::now() calls [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21745:
URL: https://github.com/apache/datafusion/pull/21745#issuecomment-4281083558
🤖 Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21745#issuecomment-4280940728)
**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-perf-cheap-wins
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query ┃ HEAD ┃ repartition-perf-cheap-wins ┃ Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1 │ 6.59 / 7.15 ±0.81 / 8.74 ms │ 6.63 / 7.17 ±0.81 / 8.78 ms │ no change │
│ QQuery 2 │ 145.80 / 146.53 ±0.49 / 147.23 ms │ 146.97 / 147.33 ±0.37 / 147.95 ms │ no change │
│ QQuery 3 │ 113.67 / 114.24 ±0.47 / 115.11 ms │ 113.00 / 113.96 ±0.62 / 114.93 ms │ no change │
│ QQuery 4 │ 1313.91 / 1343.26 ±22.27 / 1377.07 ms │ 1327.57 / 1357.27 ±15.39 / 1369.42 ms │ no change │
│ QQuery 5 │ 173.10 / 175.65 ±1.68 / 177.98 ms │ 172.56 / 174.58 ±1.70 / 176.85 ms │ no change │
│ QQuery 6 │ 836.67 / 860.23 ±24.42 / 902.61 ms │ 854.60 / 865.47 ±7.76 / 873.83 ms │ no change │
│ QQuery 7 │ 341.92 / 343.32 ±1.23 / 345.07 ms │ 340.84 / 344.39 ±2.24 / 346.49 ms │ no change │
│ QQuery 8 │ 114.53 / 115.86 ±0.88 / 117.04 ms │ 117.49 / 118.55 ±0.87 / 119.41 ms │ no change │
│ QQuery 9 │ 100.69 / 102.50 ±2.57 / 107.55 ms │ 100.50 / 105.59 ±8.43 / 122.39 ms │ no change │
│ QQuery 10 │ 105.80 / 107.04 ±0.69 / 107.71 ms │ 108.35 / 109.18 ±0.57 / 109.95 ms │ no change │
│ QQuery 11 │ 957.74 / 967.41 ±5.71 / 973.51 ms │ 945.22 / 962.36 ±9.36 / 972.59 ms │ no change │
│ QQuery 12 │ 45.94 / 46.87 ±0.92 / 48.52 ms │ 44.41 / 46.03 ±1.11 / 47.24 ms │ no change │
│ QQuery 13 │ 399.75 / 402.46 ±2.00 / 404.89 ms │ 397.67 / 402.53 ±2.99 / 406.86 ms │ no change │
│ QQuery 14 │ 995.91 / 999.87 ±3.60 / 1005.75 ms │ 991.95 / 1000.45 ±6.26 / 1009.47 ms │ no change │
│ QQuery 15 │ 17.15 / 18.51 ±1.07 / 19.96 ms │ 15.37 / 15.82 ±0.34 / 16.21 ms │ +1.17x faster │
│ QQuery 16 │ 7.19 / 7.74 ±0.36 / 8.32 ms │ 7.34 / 8.22 ±1.42 / 11.04 ms │ 1.06x slower │
│ QQuery 17 │ 229.15 / 230.97 ±1.48 / 233.57 ms │ 226.76 / 229.47 ±1.75 / 231.35 ms │ no change │
│ QQuery 18 │ 127.30 / 128.56 ±0.70 / 129.27 ms │ 126.75 / 128.59 ±1.38 / 130.74 ms │ no change │
│ QQuery 19 │ 154.69 / 156.11 ±1.62 / 159.15 ms │ 154.23 / 156.32 ±1.11 / 157.42 ms │ no change │
│ QQuery 20 │ 13.40 / 14.03 ±0.45 / 14.79 ms │ 14.62 / 14.74 ±0.10 / 14.88 ms │ 1.05x slower │
│ QQuery 21 │ 19.85 / 20.55 ±0.76 / 21.94 ms │ 19.37 / 20.18 ±0.54 / 20.82 ms │ no change │
│ QQuery 22 │ 484.02 / 485.77 ±1.46 / 487.39 ms │ 481.25 / 484.04 ±2.84 / 489.50 ms │ no change │
│ QQuery 23 │ 865.42 / 881.42 ±9.53 / 893.21 ms │ 880.74 / 887.90 ±4.26 / 891.65 ms │ no change │
│ QQuery 24 │ 381.59 / 385.80 ±3.29 / 391.26 ms │ 383.46 / 386.82 ±3.75 / 393.85 ms │ no change │
│ QQuery 25 │ 340.65 / 342.50 ±1.47 / 344.28 ms │ 340.47 / 342.72 ±1.82 / 345.93 ms │ no change │
│ QQuery 26 │ 80.86 / 85.49 ±3.27 / 90.33 ms │ 82.82 / 84.46 ±0.94 / 85.73 ms │ no change │
│ QQuery 27 │ 6.71 / 7.28 ±0.51 / 8.23 ms │ 6.93 / 7.49 ±0.49 / 8.26 ms │ no change │
│ QQuery 28 │ 149.35 / 150.92 ±1.10 / 152.67 ms │ 148.33 / 149.77 ±1.25 / 151.49 ms │ no change │
│ QQuery 29 │ 281.14 / 283.35 ±1.68 / 286.15 ms │ 283.46 / 284.48 ±1.01 / 286.27 ms │ no change │
│ QQuery 30 │ 42.09 / 44.88 ±1.91 / 47.63 ms │ 43.02 / 44.69 ±1.03 / 45.85 ms │ no change │
│ QQuery 31 │ 170.27 / 171.55 ±0.66 / 172.02 ms │ 168.25 / 171.55 ±2.15 / 174.02 ms │ no change │
│ QQuery 32 │ 13.57 / 14.07 ±0.34 / 14.56 ms │ 14.02 / 15.06 ±1.30 / 17.62 ms │ 1.07x slower │
│ QQuery 33 │ 141.68 / 143.49 ±1.06 / 144.50 ms │ 140.96 / 142.48 ±0.98 / 143.99 ms │ no change │
│ QQuery 34 │ 6.99 / 7.66 ±0.64 / 8.86 ms │ 6.99 / 7.19 ±0.19 / 7.46 ms │ +1.06x faster │
│ QQuery 35 │ 107.14 / 108.23 ±1.32 / 110.66 ms │ 106.60 / 109.33 ±1.70 / 111.24 ms │ no change │
│ QQuery 36 │ 6.42 / 7.07 ±0.62 / 8.24 ms │ 6.70 / 6.91 ±0.14 / 7.12 ms │ no change │
│ QQuery 37 │ 8.93 / 9.22 ±0.15 / 9.34 ms │ 8.39 / 10.04 ±1.15 / 11.13 ms │ 1.09x slower │
│ QQuery 38 │ 83.47 / 87.01 ±3.12 / 91.35 ms │ 83.81 / 87.95 ±4.27 / 96.18 ms │ no change │
│ QQuery 39 │ 123.01 / 126.36 ±3.37 / 132.62 ms │ 126.04 / 128.83 ±1.49 / 130.16 ms │ no change │
│ QQuery 40 │ 111.95 / 115.34 ±4.35 / 123.77 ms │ 111.98 / 116.40 ±5.48 / 126.91 ms │ no change │
│ QQuery 41 │ 14.32 / 15.08 ±1.08 / 17.21 ms │ 14.34 / 15.11 ±0.53 / 15.86 ms │ no change │
│ QQuery 42 │ 108.59 / 109.74 ±0.86 / 111.22 ms │ 109.21 / 109.82 ±0.51 / 110.45 ms │ no change │
│ QQuery 43 │ 5.96 / 6.36 ±0.61 / 7.55 ms │ 6.00 / 6.16 ±0.19 / 6.53 ms │ no change │
│ QQuery 44 │ 11.53 / 11.92 ±0.27 / 12.23 ms │ 11.83 / 12.13 ±0.29 / 12.64 ms │ no change │
│ QQuery 45 │ 50.90 / 52.35 ±1.15 / 54.31 ms │ 49.98 / 51.39 ±1.20 / 53.34 ms │ no change │
│ QQuery 46 │ 8.34 / 8.67 ±0.18 / 8.89 ms │ 8.68 / 8.81 ±0.10 / 8.97 ms │ no change │
│ QQuery 47 │ 720.08 / 723.74 ±3.25 / 728.37 ms │ 718.50 / 729.33 ±8.94 / 745.12 ms │ no change │
│ QQuery 48 │ 286.30 / 290.29 ±2.67 / 293.08 ms │ 286.70 / 291.60 ±5.10 / 300.33 ms │ no change │
│ QQuery 49 │ 249.89 / 252.82 ±1.88 / 255.03 ms │ 248.62 / 253.87 ±2.92 / 257.18 ms │ no change │
│ QQuery 50 │ 219.37 / 221.84 ±1.42 / 223.65 ms │ 218.64 / 223.43 ±5.49 / 233.72 ms │ no change │
│ QQuery 51 │ 181.21 / 182.03 ±0.67 / 182.97 ms │ 177.33 / 180.96 ±2.24 / 184.33 ms │ no change │
│ QQuery 52 │ 108.16 / 110.17 ±1.23 / 111.99 ms │ 108.28 / 109.09 ±0.56 / 109.90 ms │ no change │
│ QQuery 53 │ 102.48 / 103.50 ±0.79 / 104.48 ms │ 102.79 / 103.94 ±0.75 / 105.17 ms │ no change │
│ QQuery 54 │ 145.99 / 148.45 ±1.59 / 150.28 ms │ 145.40 / 146.79 ±1.48 / 149.54 ms │ no change │
│ QQuery 55 │ 107.36 / 108.48 ±1.11 / 110.31 ms │ 106.52 / 107.56 ±0.61 / 108.31 ms │ no change │
│ QQuery 56 │ 139.42 / 141.96 ±1.74 / 144.21 ms │ 141.42 / 142.29 ±0.81 / 143.51 ms │ no change │
│ QQuery 57 │ 173.73 / 175.45 ±1.10 / 177.18 ms │ 174.50 / 177.63 ±1.69 / 179.34 ms │ no change │
│ QQuery 58 │ 274.77 / 276.42 ±1.66 / 279.04 ms │ 272.91 / 273.85 ±0.95 / 275.64 ms │ no change │
│ QQuery 59 │ 193.80 / 196.90 ±2.03 / 199.87 ms │ 195.07 / 196.57 ±0.87 / 197.45 ms │ no change │
│ QQuery 60 │ 143.56 / 144.46 ±0.66 / 145.54 ms │ 143.95 / 146.22 ±1.69 / 149.11 ms │ no change │
│ QQuery 61 │ 13.20 / 13.37 ±0.10 / 13.51 ms │ 13.45 / 13.62 ±0.19 / 13.92 ms │ no change │
│ QQuery 62 │ 896.71 / 921.77 ±44.03 / 1009.58 ms │ 903.09 / 916.20 ±11.78 / 937.20 ms │ no change │
│ QQuery 63 │ 105.06 / 105.82 ±0.50 / 106.55 ms │ 102.42 / 106.97 ±2.36 / 109.14 ms │ no change │
│ QQuery 64 │ 685.75 / 695.09 ±5.52 / 701.37 ms │ 680.95 / 688.08 ±5.16 / 694.29 ms │ no change │
│ QQuery 65 │ 253.86 / 257.83 ±4.15 / 265.60 ms │ 253.66 / 257.24 ±2.22 / 259.56 ms │ no change │
│ QQuery 66 │ 235.19 / 255.45 ±13.22 / 275.63 ms │ 234.22 / 250.82 ±11.55 / 267.51 ms │ no change │
│ QQuery 67 │ 315.22 / 323.48 ±6.38 / 334.39 ms │ 312.30 / 319.42 ±5.46 / 327.71 ms │ no change │
│ QQuery 68 │ 8.58 / 9.95 ±1.09 / 11.32 ms │ 9.27 / 10.76 ±1.10 / 12.19 ms │ 1.08x slower │
│ QQuery 69 │ 101.64 / 104.84 ±2.51 / 107.03 ms │ 102.44 / 105.60 ±1.75 / 107.31 ms │ no change │
│ QQuery 70 │ 340.72 / 358.44 ±14.57 / 379.67 ms │ 335.32 / 347.01 ±8.36 / 361.40 ms │ no change │
│ QQuery 71 │ 137.63 / 138.79 ±1.20 / 141.04 ms │ 135.03 / 137.21 ±1.55 / 139.60 ms │ no change │
│ QQuery 72 │ 608.60 / 624.81 ±13.19 / 640.74 ms │ 613.74 / 627.24 ±8.53 / 638.43 ms │ no change │
│ QQuery 73 │ 6.57 / 7.93 ±1.31 / 10.33 ms │ 6.64 / 7.73 ±0.75 / 8.64 ms │ no change │
│ QQuery 74 │ 599.25 / 602.36 ±2.67 / 605.70 ms │ 601.09 / 610.41 ±8.58 / 620.80 ms │ no change │
│ QQuery 75 │ 274.30 / 276.79 ±2.28 / 280.61 ms │ 277.25 / 279.06 ±1.31 / 280.54 ms │ no change │
│ QQuery 76 │ 131.54 / 135.18 ±2.26 / 137.93 ms │ 131.96 / 134.36 ±1.95 / 136.80 ms │ no change │
│ QQuery 77 │ 188.44 / 190.66 ±1.71 / 193.31 ms │ 190.12 / 191.11 ±0.74 / 192.00 ms │ no change │
│ QQuery 78 │ 337.55 / 342.46 ±3.35 / 347.73 ms │ 341.43 / 346.89 ±4.48 / 352.44 ms │ no change │
│ QQuery 79 │ 232.33 / 234.98 ±2.89 / 240.41 ms │ 231.55 / 237.30 ±3.50 / 240.33 ms │ no change │
│ QQuery 80 │ 322.50 / 325.64 ±2.12 / 328.04 ms │ 320.70 / 325.25 ±2.59 / 327.97 ms │ no change │
│ QQuery 81 │ 26.39 / 27.48 ±1.08 / 29.27 ms │ 26.82 / 28.18 ±1.05 / 29.39 ms │ no change │
│ QQuery 82 │ 40.73 / 41.91 ±1.19 / 43.68 ms │ 40.63 / 41.70 ±0.78 / 42.51 ms │ no change │
│ QQuery 83 │ 38.53 / 39.75 ±0.91 / 41.32 ms │ 39.99 / 41.14 ±0.85 / 42.37 ms │ no change │
│ QQuery 84 │ 48.67 / 50.21 ±1.19 / 51.84 ms │ 48.14 / 48.89 ±0.48 / 49.63 ms │ no change │
│ QQuery 85 │ 148.21 / 149.59 ±1.40 / 151.98 ms │ 150.34 / 151.15 ±0.67 / 152.21 ms │ no change │
│ QQuery 86 │ 38.75 / 39.84 ±0.65 / 40.51 ms │ 39.02 / 40.03 ±0.63 / 40.73 ms │ no change │
│ QQuery 87 │ 84.49 / 87.50 ±2.75 / 92.24 ms │ 85.07 / 89.74 ±3.96 / 94.59 ms │ no change │
│ QQuery 88 │ 99.22 / 100.38 ±0.86 / 101.81 ms │ 100.44 / 106.22 ±9.65 / 125.37 ms │ 1.06x slower │
│ QQuery 89 │ 120.42 / 121.19 ±0.55 / 121.85 ms │ 119.44 / 120.69 ±0.80 / 121.83 ms │ no change │
│ QQuery 90 │ 23.16 / 24.39 ±0.75 / 25.24 ms │ 23.50 / 24.27 ±0.69 / 25.54 ms │ no change │
│ QQuery 91 │ 64.80 / 65.90 ±0.77 / 67.13 ms │ 63.75 / 64.78 ±0.61 / 65.39 ms │ no change │
│ QQuery 92 │ 57.98 / 58.55 ±0.29 / 58.76 ms │ 57.85 / 59.37 ±1.01 / 60.87 ms │ no change │
│ QQuery 93 │ 186.36 / 188.54 ±1.76 / 191.39 ms │ 186.51 / 188.42 ±2.08 / 192.17 ms │ no change │
│ QQuery 94 │ 62.61 / 63.59 ±0.84 / 65.14 ms │ 62.08 / 63.34 ±1.12 / 65.24 ms │ no change │
│ QQuery 95 │ 129.64 / 131.75 ±1.11 / 132.90 ms │ 129.56 / 131.48 ±1.81 / 134.54 ms │ no change │
│ QQuery 96 │ 72.07 / 73.53 ±1.09 / 74.67 ms │ 71.56 / 73.95 ±1.22 / 74.92 ms │ no change │
│ QQuery 97 │ 126.35 / 128.22 ±1.14 / 129.84 ms │ 127.15 / 128.70 ±1.14 / 130.22 ms │ no change │
│ QQuery 98 │ 156.02 / 156.72 ±0.66 / 157.70 ms │ 153.87 / 154.24 ±0.33 / 154.71 ms │ no change │
│ QQuery 99 │ 10740.26 / 10767.86 ±19.95 / 10787.36 ms │ 10705.19 / 10740.70 ±31.02 / 10785.30 ms │ no change │
└───────────┴──────────────────────────────────────────┴──────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary ┃ ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD) │ 31289.46ms │
│ Total Time (repartition-perf-cheap-wins) │ 31292.12ms │
│ Average Time (HEAD) │ 316.06ms │
│ Average Time (repartition-perf-cheap-wins) │ 316.08ms │
│ Queries Faster │ 2 │
│ Queries Slower │ 6 │
│ Queries with No Change │ 91 │
│ Queries with Failure │ 0 │
└────────────────────────────────────────────┴────────────┘
```
</p>
</details>
<details><summary>Resource Usage</summary>
**tpcds — base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 160.0s |
| Peak memory | 5.0 GiB |
| Avg memory | 4.3 GiB |
| CPU user | 259.8s |
| CPU sys | 16.1s |
| Peak spill | 0 B |
**tpcds — branch**
| Metric | Value |
|--------|-------|
| Wall time | 160.0s |
| Peak memory | 5.4 GiB |
| Avg memory | 4.4 GiB |
| CPU user | 259.8s |
| CPU sys | 16.3s |
| 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): drop redundant reservation Mutex and halve per-batch Instant::now() calls [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21745:
URL: https://github.com/apache/datafusion/pull/21745#issuecomment-4281089982
🤖 Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21745#issuecomment-4280940728)
**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-perf-cheap-wins
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query ┃ HEAD ┃ repartition-perf-cheap-wins ┃ Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0 │ 1.18 / 4.49 ±6.45 / 17.38 ms │ 1.20 / 4.51 ±6.45 / 17.41 ms │ no change │
│ QQuery 1 │ 14.99 / 15.44 ±0.33 / 15.98 ms │ 14.59 / 14.96 ±0.23 / 15.22 ms │ no change │
│ QQuery 2 │ 44.49 / 44.92 ±0.35 / 45.38 ms │ 46.03 / 46.27 ±0.24 / 46.68 ms │ no change │
│ QQuery 3 │ 39.73 / 40.98 ±1.02 / 42.69 ms │ 40.07 / 40.85 ±0.61 / 41.86 ms │ no change │
│ QQuery 4 │ 286.79 / 294.95 ±6.24 / 305.05 ms │ 292.65 / 301.57 ±7.39 / 313.22 ms │ no change │
│ QQuery 5 │ 337.45 / 342.48 ±4.04 / 348.84 ms │ 349.37 / 353.33 ±2.48 / 356.74 ms │ no change │
│ QQuery 6 │ 5.25 / 6.31 ±0.54 / 6.71 ms │ 5.31 / 6.36 ±1.28 / 8.86 ms │ no change │
│ QQuery 7 │ 17.83 / 18.19 ±0.21 / 18.50 ms │ 17.26 / 17.42 ±0.14 / 17.59 ms │ no change │
│ QQuery 8 │ 411.24 / 424.01 ±6.90 / 432.11 ms │ 429.76 / 439.87 ±13.73 / 465.72 ms │ no change │
│ QQuery 9 │ 655.11 / 665.80 ±5.69 / 670.39 ms │ 675.32 / 690.30 ±12.47 / 709.14 ms │ no change │
│ QQuery 10 │ 91.31 / 92.88 ±2.28 / 97.40 ms │ 94.13 / 96.18 ±3.08 / 102.27 ms │ no change │
│ QQuery 11 │ 103.72 / 105.82 ±1.67 / 107.95 ms │ 105.60 / 107.23 ±1.07 / 108.43 ms │ no change │
│ QQuery 12 │ 337.47 / 344.30 ±6.02 / 355.15 ms │ 342.76 / 351.74 ±11.41 / 373.82 ms │ no change │
│ QQuery 13 │ 452.26 / 468.64 ±8.31 / 475.26 ms │ 464.36 / 473.84 ±9.37 / 490.81 ms │ no change │
│ QQuery 14 │ 344.37 / 352.37 ±4.50 / 357.65 ms │ 352.66 / 357.31 ±4.24 / 365.32 ms │ no change │
│ QQuery 15 │ 370.28 / 383.01 ±13.72 / 409.56 ms │ 367.18 / 382.05 ±11.06 / 394.03 ms │ no change │
│ QQuery 16 │ 715.03 / 722.59 ±5.24 / 730.50 ms │ 714.75 / 727.74 ±13.11 / 752.16 ms │ no change │
│ QQuery 17 │ 708.67 / 714.61 ±4.39 / 722.01 ms │ 721.38 / 727.77 ±5.18 / 734.40 ms │ no change │
│ QQuery 18 │ 1457.18 / 1486.15 ±32.19 / 1542.50 ms │ 1422.70 / 1500.49 ±40.85 / 1541.81 ms │ no change │
│ QQuery 19 │ 36.94 / 59.75 ±42.51 / 144.69 ms │ 36.44 / 38.72 ±2.53 / 43.02 ms │ +1.54x faster │
│ QQuery 20 │ 715.18 / 739.05 ±21.04 / 769.22 ms │ 724.54 / 742.75 ±17.11 / 767.13 ms │ no change │
│ QQuery 21 │ 759.40 / 761.30 ±1.22 / 762.70 ms │ 770.92 / 781.12 ±10.62 / 800.91 ms │ no change │
│ QQuery 22 │ 1142.10 / 1147.97 ±7.56 / 1162.79 ms │ 1150.71 / 1156.72 ±4.49 / 1163.01 ms │ no change │
│ QQuery 23 │ 3110.42 / 3141.21 ±23.29 / 3171.86 ms │ 3144.47 / 3161.68 ±12.99 / 3177.26 ms │ no change │
│ QQuery 24 │ 102.56 / 105.27 ±2.25 / 108.00 ms │ 103.42 / 106.32 ±2.66 / 110.48 ms │ no change │
│ QQuery 25 │ 138.47 / 142.08 ±2.34 / 144.74 ms │ 139.07 / 142.67 ±2.39 / 145.73 ms │ no change │
│ QQuery 26 │ 102.36 / 105.26 ±1.91 / 108.13 ms │ 103.28 / 104.26 ±1.35 / 106.91 ms │ no change │
│ QQuery 27 │ 850.00 / 859.44 ±13.44 / 885.44 ms │ 847.76 / 851.82 ±2.73 / 855.71 ms │ no change │
│ QQuery 28 │ 3277.10 / 3324.21 ±32.42 / 3372.54 ms │ 3259.23 / 3292.64 ±18.71 / 3317.31 ms │ no change │
│ QQuery 29 │ 52.21 / 56.17 ±3.05 / 61.05 ms │ 51.61 / 58.10 ±4.94 / 65.69 ms │ no change │
│ QQuery 30 │ 370.81 / 378.51 ±4.23 / 383.42 ms │ 362.19 / 371.18 ±5.30 / 376.75 ms │ no change │
│ QQuery 31 │ 376.27 / 392.93 ±10.15 / 408.05 ms │ 374.13 / 395.06 ±13.06 / 409.49 ms │ no change │
│ QQuery 32 │ 1052.66 / 1081.27 ±25.57 / 1128.72 ms │ 1068.87 / 1082.77 ±19.04 / 1120.10 ms │ no change │
│ QQuery 33 │ 1482.17 / 1495.78 ±12.87 / 1516.91 ms │ 1495.06 / 1519.30 ±15.65 / 1544.22 ms │ no change │
│ QQuery 34 │ 1498.99 / 1526.03 ±19.33 / 1553.96 ms │ 1491.78 / 1523.03 ±20.00 / 1550.83 ms │ no change │
│ QQuery 35 │ 396.93 / 411.79 ±10.52 / 429.33 ms │ 400.12 / 405.28 ±3.60 / 410.90 ms │ no change │
│ QQuery 36 │ 121.64 / 123.55 ±1.60 / 126.10 ms │ 111.79 / 120.48 ±5.01 / 125.74 ms │ no change │
│ QQuery 37 │ 48.38 / 49.09 ±0.59 / 49.99 ms │ 49.85 / 51.00 ±1.01 / 52.57 ms │ no change │
│ QQuery 38 │ 74.85 / 77.12 ±1.63 / 79.26 ms │ 76.85 / 78.61 ±1.90 / 82.13 ms │ no change │
│ QQuery 39 │ 210.79 / 221.44 ±6.66 / 231.69 ms │ 208.67 / 219.01 ±5.22 / 222.51 ms │ no change │
│ QQuery 40 │ 24.52 / 26.36 ±1.95 / 30.14 ms │ 25.18 / 27.23 ±1.62 / 29.69 ms │ no change │
│ QQuery 41 │ 21.06 / 21.83 ±0.66 / 22.90 ms │ 20.92 / 22.65 ±1.56 / 25.01 ms │ no change │
│ QQuery 42 │ 20.20 / 20.89 ±0.74 / 22.27 ms │ 19.86 / 20.50 ±0.39 / 21.08 ms │ no change │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary ┃ ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD) │ 22796.23ms │
│ Total Time (repartition-perf-cheap-wins) │ 22912.72ms │
│ Average Time (HEAD) │ 530.14ms │
│ Average Time (repartition-perf-cheap-wins) │ 532.85ms │
│ Queries Faster │ 1 │
│ Queries Slower │ 0 │
│ Queries with No Change │ 42 │
│ Queries with Failure │ 0 │
└────────────────────────────────────────────┴────────────┘
```
</p>
</details>
<details><summary>Resource Usage</summary>
**clickbench_partitioned — base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 120.0s |
| Peak memory | 38.7 GiB |
| Avg memory | 27.0 GiB |
| CPU user | 1082.6s |
| CPU sys | 88.4s |
| Peak spill | 0 B |
**clickbench_partitioned — branch**
| Metric | Value |
|--------|-------|
| Wall time | 120.0s |
| Peak memory | 36.2 GiB |
| Avg memory | 26.5 GiB |
| CPU user | 1087.7s |
| CPU sys | 89.3s |
| 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): drop redundant reservation Mutex and halve per-batch Instant::now() calls [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21745:
URL: https://github.com/apache/datafusion/pull/21745#issuecomment-4280960569
🤖 Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21745#issuecomment-4280940728)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4280940728-1599-wg5qx 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-perf-cheap-wins (b8fba5627c2dc2b43810071690f67a2e6f33da91) to 43d32a8 (merge-base) [diff](https://github.com/apache/datafusion/compare/43d32a8c14b12f388e04719170a7704f84b973b6..b8fba5627c2dc2b43810071690f67a2e6f33da91) 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): drop redundant reservation Mutex and halve per-batch Instant::now() calls [datafusion]
Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #21745:
URL: https://github.com/apache/datafusion/pull/21745#issuecomment-4280940728
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(repartition): drop redundant reservation Mutex and halve per-batch Instant::now() calls [datafusion]
Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan closed pull request #21745: perf(repartition): drop redundant reservation Mutex and halve per-batch Instant::now() calls
URL: https://github.com/apache/datafusion/pull/21745
--
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): drop redundant reservation Mutex and halve per-batch Instant::now() calls [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21745:
URL: https://github.com/apache/datafusion/pull/21745#issuecomment-4280961191
🤖 Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21745#issuecomment-4280940728)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4280940728-1598-27pk2 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-perf-cheap-wins (b8fba5627c2dc2b43810071690f67a2e6f33da91) to 43d32a8 (merge-base) [diff](https://github.com/apache/datafusion/compare/43d32a8c14b12f388e04719170a7704f84b973b6..b8fba5627c2dc2b43810071690f67a2e6f33da91) 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): drop redundant reservation Mutex and halve per-batch Instant::now() calls [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21745:
URL: https://github.com/apache/datafusion/pull/21745#issuecomment-4280960819
🤖 Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21745#issuecomment-4280940728)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4280940728-1600-7dthm 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-perf-cheap-wins (b8fba5627c2dc2b43810071690f67a2e6f33da91) to 43d32a8 (merge-base) [diff](https://github.com/apache/datafusion/compare/43d32a8c14b12f388e04719170a7704f84b973b6..b8fba5627c2dc2b43810071690f67a2e6f33da91) 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