Skip to content

Benchmark arrow-rs #12008 on dictionary group-bys - #27077

Closed
Rich-T-kid wants to merge 9 commits into
apache:mainfrom
Rich-T-kid:codex/bench-arrow-12008-current
Closed

Benchmark arrow-rs #12008 on dictionary group-bys#27077
Rich-T-kid wants to merge 9 commits into
apache:mainfrom
Rich-T-kid:codex/bench-arrow-12008-current

Conversation

@Rich-T-kid

@Rich-T-kid Rich-T-kid commented Nov 4, 2026

Copy link
Copy Markdown
Contributor

Which issue does this PR close?

Benchmark-only branch; do not merge.

Rationale for this change

Measure the current head of apache/arrow-rs#12008 on the current head of DataFusion #26816 with Parquet RLE-to-dictionary reads enabled.

What changes are included in this PR?

Two dependency-only commits provide the A/B refs:

The shared lockfile update to tonic and tonic-prost 0.14.9 is required by both Arrow revisions.

What is the testing strategy for this PR?

Run partitioned ClickBench with dictionary decoding enabled on both refs and compare Q16, Q17, Q33, and Q34.

Are there any user-facing changes?

No. This PR only supplies benchmark refs.

Rich-T-kid and others added 9 commits October 28, 2026 00:46
@github-actions github-actions Bot added common Related to common crate physical-plan Changes to the physical-plan crate labels Nov 4, 2026
@Rich-T-kid

Rich-T-kid commented Nov 4, 2026

Copy link
Copy Markdown
Contributor Author

run benchmark clickbench_partitioned

env:
DATAFUSION_RUNTIME_MEMORY_LIMIT: "32G"
DATAFUSION_EXECUTION_PARQUET_ENABLE_RLE_TO_DICTIONARY: "true"
baseline:
ref: "6d790c5f7441b0f7b786dadc0535140f41a5fa42"
changed:
ref: "e15a3816bdf45033ec1e3edfc8d05981641ad663"

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6622738060-4434-psw46 6.12.105+ #1 SMP Sat Sep 19 07:55:25 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
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

Comparing e15a381 (e15a381) to 6d790c5 diff

Run configuration 
run benchmark clickbench_partitioned
env:
  DATAFUSION_EXECUTION_PARQUET_ENABLE_RLE_TO_DICTIONARY: "true"
  DATAFUSION_RUNTIME_MEMORY_LIMIT: "32G"
baseline:
  ref: "6d790c5f7441b0f7b786dadc0535140f41a5fa42"
changed:
  ref: "e15a3816bdf45033ec1e3edfc8d05981641ad663"

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing e15a381 (e15a381) to 6d790c5 diff

Run configuration 
run benchmark clickbench_partitioned
env:
  DATAFUSION_EXECUTION_PARQUET_ENABLE_RLE_TO_DICTIONARY: "true"
  DATAFUSION_RUNTIME_MEMORY_LIMIT: "32G"
baseline:
  ref: "6d790c5f7441b0f7b786dadc0535140f41a5fa42"
changed:
  ref: "e15a3816bdf45033ec1e3edfc8d05981641ad663"
CPU Details (lscpu) 
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 HEAD and codex_bench-arrow-12008-current
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ codex_bench-arrow-12008-current ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │    1.17 ms │                         1.16 ms │     no change │
│ QQuery 1  │   11.42 ms │                        11.22 ms │     no change │
│ QQuery 2  │   35.51 ms │                        34.71 ms │     no change │
│ QQuery 3  │   31.69 ms │                        31.27 ms │     no change │
│ QQuery 4  │  241.14 ms │                       265.24 ms │  1.10x slower │
│ QQuery 5  │ 1079.97 ms │                       888.24 ms │ +1.22x faster │
│ QQuery 6  │    1.17 ms │                         1.18 ms │     no change │
│ QQuery 7  │   12.42 ms │                        12.49 ms │     no change │
│ QQuery 8  │  335.76 ms │                       356.05 ms │  1.06x slower │
│ QQuery 9  │  532.03 ms │                       508.76 ms │     no change │
│ QQuery 10 │  203.16 ms │                       203.18 ms │     no change │
│ QQuery 11 │  216.59 ms │                       213.07 ms │     no change │
│ QQuery 12 │ 1120.09 ms │                       865.28 ms │ +1.29x faster │
│ QQuery 13 │ 1625.21 ms │                      1321.95 ms │ +1.23x faster │
│ QQuery 14 │ 1209.91 ms │                       963.13 ms │ +1.26x faster │
│ QQuery 15 │  307.09 ms │                       322.07 ms │     no change │
│ QQuery 16 │ 2308.67 ms │                      1404.65 ms │ +1.64x faster │
│ QQuery 17 │ 2297.98 ms │                      1456.22 ms │ +1.58x faster │
│ QQuery 18 │ 3852.86 ms │                      2213.09 ms │ +1.74x faster │
│ QQuery 19 │   28.12 ms │                        28.15 ms │     no change │
│ QQuery 20 │ 1070.52 ms │                      1071.84 ms │     no change │
│ QQuery 21 │ 1410.44 ms │                      1393.51 ms │     no change │
│ QQuery 22 │ 3301.14 ms │                      3278.77 ms │     no change │
│ QQuery 23 │ 9407.57 ms │                      9387.49 ms │     no change │
│ QQuery 24 │   99.12 ms │                        96.30 ms │     no change │
│ QQuery 25 │  362.63 ms │                       359.35 ms │     no change │
│ QQuery 26 │   99.52 ms │                        97.05 ms │     no change │
│ QQuery 27 │ 1168.27 ms │                      1159.77 ms │     no change │
│ QQuery 28 │ 2679.33 ms │                      2635.88 ms │     no change │
│ QQuery 29 │   38.33 ms │                        38.14 ms │     no change │
│ QQuery 30 │  558.29 ms │                       563.16 ms │     no change │
│ QQuery 31 │  552.86 ms │                       553.92 ms │     no change │
│ QQuery 32 │ 1100.39 ms │                      1137.18 ms │     no change │
│ QQuery 33 │ 7959.94 ms │                      4073.52 ms │ +1.95x faster │
│ QQuery 34 │ 7794.07 ms │                      4041.45 ms │ +1.93x faster │
│ QQuery 35 │  304.52 ms │                       319.17 ms │     no change │
│ QQuery 36 │  130.81 ms │                       141.17 ms │  1.08x slower │
│ QQuery 37 │   98.51 ms │                        99.46 ms │     no change │
│ QQuery 38 │   79.78 ms │                        80.51 ms │     no change │
│ QQuery 39 │  296.73 ms │                       302.40 ms │     no change │
│ QQuery 40 │   14.30 ms │                        15.61 ms │  1.09x slower │
│ QQuery 41 │   13.48 ms │                        14.99 ms │  1.11x slower │
│ QQuery 42 │   12.96 ms │                        14.34 ms │  1.11x slower │
└───────────┴────────────┴─────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                              ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                              │ 54005.45ms │
│ Total Time (codex_bench-arrow-12008-current)   │ 41976.11ms │
│ Average Time (HEAD)                            │  1255.94ms │
│ Average Time (codex_bench-arrow-12008-current) │   976.19ms │
│ Queries Faster                                 │          9 │
│ Queries Slower                                 │          6 │
│ Queries with No Change                         │         28 │
│ Queries with Failure                           │          0 │
└────────────────────────────────────────────────┴────────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and codex_bench-arrow-12008-current
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                   HEAD ┃          codex_bench-arrow-12008-current ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │            1.17 / 1.96 ±1.29 / 4.53 ms │              1.16 / 1.79 ±1.15 / 4.08 ms │ +1.10x faster │
│ QQuery 1  │         11.42 / 11.67 ±0.14 / 11.80 ms │           11.22 / 11.65 ±0.63 / 12.90 ms │     no change │
│ QQuery 2  │         35.51 / 36.62 ±0.84 / 38.01 ms │           34.71 / 34.91 ±0.20 / 35.24 ms │     no change │
│ QQuery 3  │         31.69 / 32.70 ±0.88 / 34.03 ms │           31.27 / 31.76 ±0.39 / 32.28 ms │     no change │
│ QQuery 4  │      241.14 / 250.96 ±9.93 / 267.91 ms │        265.24 / 270.55 ±3.65 / 275.23 ms │  1.08x slower │
│ QQuery 5  │  1079.97 / 1129.67 ±25.23 / 1150.41 ms │       888.24 / 903.47 ±11.84 / 919.87 ms │ +1.25x faster │
│ QQuery 6  │            1.17 / 1.33 ±0.25 / 1.82 ms │              1.18 / 1.33 ±0.24 / 1.82 ms │     no change │
│ QQuery 7  │         12.42 / 12.67 ±0.13 / 12.75 ms │           12.49 / 12.87 ±0.24 / 13.24 ms │     no change │
│ QQuery 8  │     335.76 / 357.96 ±13.61 / 374.60 ms │        356.05 / 365.57 ±7.45 / 378.71 ms │     no change │
│ QQuery 9  │      532.03 / 545.00 ±6.89 / 551.02 ms │       508.76 / 532.59 ±18.64 / 554.97 ms │     no change │
│ QQuery 10 │      203.16 / 206.00 ±1.75 / 208.61 ms │        203.18 / 203.69 ±0.46 / 204.29 ms │     no change │
│ QQuery 11 │      216.59 / 219.82 ±2.68 / 224.39 ms │        213.07 / 215.75 ±2.97 / 221.35 ms │     no change │
│ QQuery 12 │  1120.09 / 1147.11 ±17.98 / 1172.01 ms │       865.28 / 901.02 ±20.72 / 921.44 ms │ +1.27x faster │
│ QQuery 13 │  1625.21 / 1714.37 ±83.20 / 1838.07 ms │    1321.95 / 1362.21 ±35.45 / 1423.98 ms │ +1.26x faster │
│ QQuery 14 │  1209.91 / 1273.84 ±38.49 / 1313.77 ms │      963.13 / 985.75 ±23.54 / 1017.01 ms │ +1.29x faster │
│ QQuery 15 │     307.09 / 330.70 ±16.32 / 353.01 ms │        322.07 / 333.99 ±8.85 / 349.07 ms │     no change │
│ QQuery 16 │  2308.67 / 2375.47 ±35.00 / 2407.52 ms │    1404.65 / 1461.88 ±44.03 / 1536.13 ms │ +1.62x faster │
│ QQuery 17 │  2297.98 / 2396.72 ±88.07 / 2510.76 ms │    1456.22 / 1500.98 ±40.59 / 1548.32 ms │ +1.60x faster │
│ QQuery 18 │  3852.86 / 3973.61 ±77.94 / 4051.60 ms │    2213.09 / 2307.44 ±75.54 / 2416.31 ms │ +1.72x faster │
│ QQuery 19 │        28.12 / 37.44 ±13.10 / 63.37 ms │           28.15 / 33.60 ±8.42 / 50.04 ms │ +1.11x faster │
│ QQuery 20 │  1070.52 / 1089.57 ±12.27 / 1107.24 ms │    1071.84 / 1091.36 ±18.30 / 1123.16 ms │     no change │
│ QQuery 21 │   1410.44 / 1424.40 ±9.24 / 1437.09 ms │    1393.51 / 1421.27 ±17.43 / 1439.22 ms │     no change │
│ QQuery 22 │  3301.14 / 3342.28 ±38.65 / 3407.72 ms │    3278.77 / 3313.07 ±32.32 / 3362.82 ms │     no change │
│ QQuery 23 │  9407.57 / 9569.75 ±86.68 / 9647.29 ms │    9387.49 / 9561.33 ±95.16 / 9679.65 ms │     no change │
│ QQuery 24 │       99.12 / 109.25 ±9.81 / 125.97 ms │        96.30 / 119.01 ±28.55 / 170.31 ms │  1.09x slower │
│ QQuery 25 │      362.63 / 365.35 ±4.35 / 374.03 ms │       359.35 / 369.92 ±18.59 / 407.01 ms │     no change │
│ QQuery 26 │       99.52 / 101.06 ±1.63 / 103.85 ms │           97.05 / 98.14 ±0.86 / 99.66 ms │     no change │
│ QQuery 27 │  1168.27 / 1189.92 ±17.09 / 1216.10 ms │    1159.77 / 1180.17 ±18.11 / 1205.50 ms │     no change │
│ QQuery 28 │  2679.33 / 2712.37 ±26.28 / 2743.36 ms │    2635.88 / 2655.91 ±11.93 / 2670.21 ms │     no change │
│ QQuery 29 │         38.33 / 49.25 ±9.25 / 62.17 ms │          38.14 / 52.68 ±15.43 / 80.74 ms │  1.07x slower │
│ QQuery 30 │     558.29 / 573.31 ±10.09 / 588.85 ms │        563.16 / 571.75 ±7.93 / 586.61 ms │     no change │
│ QQuery 31 │      552.86 / 567.02 ±8.36 / 578.83 ms │       553.92 / 571.14 ±14.50 / 595.99 ms │     no change │
│ QQuery 32 │  1100.39 / 1135.42 ±31.77 / 1186.10 ms │    1137.18 / 1212.14 ±61.84 / 1277.77 ms │  1.07x slower │
│ QQuery 33 │ 7959.94 / 8152.64 ±191.02 / 8479.41 ms │ 4073.52 / 9072.24 ±6792.21 / 21624.53 ms │  1.11x slower │
│ QQuery 34 │ 7794.07 / 8085.33 ±230.12 / 8491.55 ms │    4041.45 / 4149.95 ±82.12 / 4294.46 ms │ +1.95x faster │
│ QQuery 35 │     304.52 / 328.01 ±14.59 / 344.43 ms │       319.17 / 343.04 ±19.44 / 363.34 ms │     no change │
│ QQuery 36 │     130.81 / 152.08 ±13.87 / 165.50 ms │       141.17 / 152.77 ±13.32 / 177.22 ms │     no change │
│ QQuery 37 │       98.51 / 102.41 ±5.95 / 114.22 ms │         99.46 / 103.27 ±4.33 / 111.39 ms │     no change │
│ QQuery 38 │         79.78 / 84.58 ±4.07 / 89.76 ms │           80.51 / 83.98 ±3.23 / 89.78 ms │     no change │
│ QQuery 39 │     296.73 / 322.63 ±18.53 / 341.09 ms │       302.40 / 323.22 ±16.05 / 351.42 ms │     no change │
│ QQuery 40 │         14.30 / 17.82 ±5.20 / 28.17 ms │           15.61 / 18.71 ±3.43 / 23.55 ms │     no change │
│ QQuery 41 │         13.48 / 13.99 ±0.47 / 14.62 ms │           14.99 / 15.49 ±0.39 / 16.08 ms │  1.11x slower │
│ QQuery 42 │         12.96 / 14.33 ±1.91 / 18.11 ms │           14.34 / 14.51 ±0.17 / 14.77 ms │     no change │
└───────────┴────────────────────────────────────────┴──────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                              ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                              │ 55558.43ms │
│ Total Time (codex_bench-arrow-12008-current)   │ 47967.91ms │
│ Average Time (HEAD)                            │  1292.06ms │
│ Average Time (codex_bench-arrow-12008-current) │  1115.53ms │
│ Queries Faster                                 │         10 │
│ Queries Slower                                 │          6 │
│ Queries with No Change                         │         27 │
│ Queries with Failure                           │          0 │
└────────────────────────────────────────────────┴────────────┘

Memory Pool Peaks

Peak MemoryPool reservation per query — what DataFusion's accounting believes it reserved. Recorded only for benchmarks that write a results JSON (the dfbench suites, and the suites bench.sh runs through the Criterion SQL harness since #25644), and only under a memory limit: DATAFUSION_RUNTIME_MEMORY_LIMIT, or one the suite sets itself.

Base: 6d790c5f7441b0f7b786dadc0535140f41a5fa42 | Changed: e15a3816bdf45033ec1e3edfc8d05981641ad663

clickbench_partitioned

Query Base Changed Change
Query 0 0 B 0 B 0.0%
Query 1 104 B 104 B +0.0%
Query 2 936 B 936 B +0.0%
Query 3 312 B 312 B +0.0%
Query 4 810.2 MiB 831.6 MiB +2.6%
Query 5 1.9 GiB 1.9 GiB -1.0%
Query 6 0 B 0 B 0.0%
Query 7 50.3 MiB 60.3 MiB +19.7%
Query 8 945.1 MiB 974.1 MiB +3.1%
Query 9 635.9 MiB 593.9 MiB -6.6%
Query 10 73.3 MiB 73.7 MiB +0.5%
Query 11 78.1 MiB 77.0 MiB -1.4%
Query 12 1.9 GiB 1.9 GiB -1.1%
Query 13 2.4 GiB 1.7 GiB -29.3%
Query 14 1.9 GiB 1.9 GiB -2.0%
Query 15 1.2 GiB 1.2 GiB -2.4%
Query 16 3.6 GiB 3.4 GiB -4.9%
Query 17 3.5 GiB 3.7 GiB +7.2%
Query 18 7.7 GiB 7.2 GiB -6.4%
Query 19 0 B 0 B 0.0%
Query 20 104 B 104 B +0.0%
Query 21 1.5 MiB 1.5 MiB -1.4%
Query 22 2.4 MiB 2.5 MiB +4.1%
Query 23 13.0 MiB 12.9 MiB -0.3%
Query 24 6.7 MiB 6.3 MiB -6.2%
Query 25 8.1 MiB 7.6 MiB -6.2%
Query 26 7.8 MiB 7.3 MiB -6.4%
Query 27 2.5 MiB 2.6 MiB +5.9%
Query 28 2.0 GiB 1.9 GiB -5.3%
Query 29 624 B 624 B +0.0%
Query 30 642.7 MiB 668.1 MiB +3.9%
Query 31 1.5 GiB 1.5 GiB +3.1%
Query 32 7.1 GiB 7.2 GiB +0.2%
Query 33 12.7 GiB 15.3 GiB +20.5%
Query 34 12.7 GiB 11.9 GiB -6.7%
Query 35 659.6 MiB 660.0 MiB +0.1%
Query 36 474.1 MiB 505.5 MiB +6.6%
Query 37 8.1 MiB 8.1 MiB +0.0%
Query 38 5.6 MiB 5.3 MiB -5.4%
Query 39 909.9 MiB 843.8 MiB -7.3%
Query 40 2.5 MiB 2.3 MiB -6.8%
Query 41 3.3 MiB 3.3 MiB +0.0%
Query 42 1.7 MiB 1.9 MiB +12.3%

Pool accounting vs. process RSS

Max pool peak is the largest reservation any single query in the run reached; peak RSS covers the whole invocation, including data loading and allocator retention, and the two high-water marks need not coincide in time. The gap is therefore an upper bound on what the pool did not account for, not a measurement of it.

Benchmark Side Max pool peak Peak RSS Gap RSS / pool
clickbench_partitioned base (6d790c5f7441b0f7b786dadc0535140f41a5fa42) 12.7 GiB 15.9 GiB 3.2 GiB 1.3×
clickbench_partitioned changed (e15a3816bdf45033ec1e3edfc8d05981641ad663) 15.3 GiB 14.7 GiB −649.0 MiB 1.0×
Resource Usage

clickbench_partitioned — base (merge-base)

Metric Value
Wall time 280.1s
Peak memory 15.9 GiB
Avg memory 6.1 GiB
CPU user 2189.3s
CPU sys 188.4s
Peak spill 0 B

clickbench_partitioned — branch

Metric Value
Wall time 245.1s
Peak memory 14.7 GiB
Avg memory 5.7 GiB
CPU user 2068.3s
CPU sys 179.2s
Peak spill 9.6 GiB

File an issue against this benchmark runner

@Rich-T-kid

Rich-T-kid commented Nov 4, 2026

Copy link
Copy Markdown
Contributor Author

run benchmark clickbench_partitioned

env:
DATAFUSION_RUNTIME_MEMORY_LIMIT: "32G"
DATAFUSION_EXECUTION_PARQUET_ENABLE_RLE_TO_DICTIONARY: "true"
baseline:
ref: "6d790c5f7441b0f7b786dadc0535140f41a5fa42"
changed:
ref: "e15a3816bdf45033ec1e3edfc8d05981641ad663"

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6623915133-4435-5nxq8 6.12.105+ #1 SMP Sat Sep 19 07:55:25 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
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

Comparing e15a381 (e15a381) to 6d790c5 diff

Run configuration 
run benchmark clickbench_partitioned
env:
  DATAFUSION_EXECUTION_PARQUET_ENABLE_RLE_TO_DICTIONARY: "true"
  DATAFUSION_RUNTIME_MEMORY_LIMIT: "32G"
baseline:
  ref: "6d790c5f7441b0f7b786dadc0535140f41a5fa42"
changed:
  ref: "e15a3816bdf45033ec1e3edfc8d05981641ad663"

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing e15a381 (e15a381) to 6d790c5 diff

Run configuration 
run benchmark clickbench_partitioned
env:
  DATAFUSION_EXECUTION_PARQUET_ENABLE_RLE_TO_DICTIONARY: "true"
  DATAFUSION_RUNTIME_MEMORY_LIMIT: "32G"
baseline:
  ref: "6d790c5f7441b0f7b786dadc0535140f41a5fa42"
changed:
  ref: "e15a3816bdf45033ec1e3edfc8d05981641ad663"
CPU Details (lscpu) 
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 HEAD and codex_bench-arrow-12008-current
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ codex_bench-arrow-12008-current ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │    1.21 ms │                         1.24 ms │     no change │
│ QQuery 1  │   11.53 ms │                        11.54 ms │     no change │
│ QQuery 2  │   35.05 ms │                        35.67 ms │     no change │
│ QQuery 3  │   31.48 ms │                        31.87 ms │     no change │
│ QQuery 4  │  264.65 ms │                       263.04 ms │     no change │
│ QQuery 5  │ 1148.81 ms │                       860.84 ms │ +1.33x faster │
│ QQuery 6  │    1.30 ms │                         1.24 ms │     no change │
│ QQuery 7  │   13.03 ms │                        12.98 ms │     no change │
│ QQuery 8  │  366.98 ms │                       365.21 ms │     no change │
│ QQuery 9  │  517.31 ms │                       526.60 ms │     no change │
│ QQuery 10 │  204.99 ms │                       205.10 ms │     no change │
│ QQuery 11 │  215.84 ms │                       215.34 ms │     no change │
│ QQuery 12 │ 1164.53 ms │                       902.27 ms │ +1.29x faster │
│ QQuery 13 │ 1713.99 ms │                      1376.52 ms │ +1.25x faster │
│ QQuery 14 │ 1311.69 ms │                       989.62 ms │ +1.33x faster │
│ QQuery 15 │  330.59 ms │                       321.08 ms │     no change │
│ QQuery 16 │ 2352.52 ms │                      1464.47 ms │ +1.61x faster │
│ QQuery 17 │ 2438.17 ms │                      1432.44 ms │ +1.70x faster │
│ QQuery 18 │ 3821.58 ms │                      2212.64 ms │ +1.73x faster │
│ QQuery 19 │   28.46 ms │                        28.65 ms │     no change │
│ QQuery 20 │ 1109.11 ms │                      1090.17 ms │     no change │
│ QQuery 21 │ 1423.34 ms │                      1416.33 ms │     no change │
│ QQuery 22 │ 3406.54 ms │                      3344.35 ms │     no change │
│ QQuery 23 │ 9521.92 ms │                      9591.87 ms │     no change │
│ QQuery 24 │   98.40 ms │                       100.87 ms │     no change │
│ QQuery 25 │  361.51 ms │                       359.02 ms │     no change │
│ QQuery 26 │   99.72 ms │                        99.46 ms │     no change │
│ QQuery 27 │ 1220.91 ms │                      1203.09 ms │     no change │
│ QQuery 28 │ 2805.58 ms │                      2665.95 ms │     no change │
│ QQuery 29 │   39.26 ms │                        38.76 ms │     no change │
│ QQuery 30 │  584.08 ms │                       569.33 ms │     no change │
│ QQuery 31 │  570.65 ms │                       569.23 ms │     no change │
│ QQuery 32 │ 1126.25 ms │                      1092.32 ms │     no change │
│ QQuery 33 │ 7893.80 ms │                      4356.48 ms │ +1.81x faster │
│ QQuery 34 │ 8315.91 ms │                      4342.55 ms │ +1.91x faster │
│ QQuery 35 │  336.29 ms │                       334.95 ms │     no change │
│ QQuery 36 │  156.07 ms │                       144.02 ms │ +1.08x faster │
│ QQuery 37 │   99.04 ms │                        99.25 ms │     no change │
│ QQuery 38 │   84.67 ms │                        84.69 ms │     no change │
│ QQuery 39 │  334.70 ms │                       310.78 ms │ +1.08x faster │
│ QQuery 40 │   15.86 ms │                        14.27 ms │ +1.11x faster │
│ QQuery 41 │   14.54 ms │                        14.52 ms │     no change │
│ QQuery 42 │   14.03 ms │                        14.12 ms │     no change │
└───────────┴────────────┴─────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                              ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                              │ 55605.93ms │
│ Total Time (codex_bench-arrow-12008-current)   │ 43114.74ms │
│ Average Time (HEAD)                            │  1293.16ms │
│ Average Time (codex_bench-arrow-12008-current) │  1002.67ms │
│ Queries Faster                                 │         12 │
│ Queries Slower                                 │          0 │
│ Queries with No Change                         │         31 │
│ Queries with Failure                           │          0 │
└────────────────────────────────────────────────┴────────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and codex_bench-arrow-12008-current
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                   HEAD ┃        codex_bench-arrow-12008-current ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │            1.21 / 1.94 ±1.25 / 4.42 ms │            1.24 / 1.91 ±1.19 / 4.29 ms │     no change │
│ QQuery 1  │         11.53 / 11.66 ±0.08 / 11.77 ms │         11.54 / 11.68 ±0.13 / 11.90 ms │     no change │
│ QQuery 2  │         35.05 / 35.43 ±0.28 / 35.83 ms │         35.67 / 35.81 ±0.21 / 36.22 ms │     no change │
│ QQuery 3  │         31.48 / 32.53 ±0.88 / 34.11 ms │         31.87 / 32.45 ±0.69 / 33.70 ms │     no change │
│ QQuery 4  │      264.65 / 267.91 ±3.12 / 272.23 ms │      263.04 / 266.34 ±2.60 / 270.04 ms │     no change │
│ QQuery 5  │  1148.81 / 1161.31 ±11.83 / 1177.66 ms │     860.84 / 902.91 ±27.09 / 940.81 ms │ +1.29x faster │
│ QQuery 6  │            1.30 / 1.45 ±0.23 / 1.90 ms │            1.24 / 1.41 ±0.25 / 1.90 ms │     no change │
│ QQuery 7  │         13.03 / 13.20 ±0.10 / 13.31 ms │         12.98 / 13.15 ±0.08 / 13.20 ms │     no change │
│ QQuery 8  │      366.98 / 373.59 ±4.04 / 378.87 ms │     365.21 / 377.79 ±12.34 / 399.49 ms │     no change │
│ QQuery 9  │      517.31 / 526.96 ±8.03 / 538.17 ms │      526.60 / 535.83 ±8.91 / 546.93 ms │     no change │
│ QQuery 10 │      204.99 / 208.57 ±2.81 / 213.46 ms │      205.10 / 206.29 ±0.71 / 207.28 ms │     no change │
│ QQuery 11 │      215.84 / 217.02 ±0.74 / 218.05 ms │      215.34 / 216.92 ±1.07 / 218.31 ms │     no change │
│ QQuery 12 │  1164.53 / 1199.96 ±19.00 / 1220.10 ms │     902.27 / 925.23 ±20.62 / 958.43 ms │ +1.30x faster │
│ QQuery 13 │  1713.99 / 1763.83 ±48.01 / 1845.08 ms │  1376.52 / 1450.94 ±53.88 / 1501.35 ms │ +1.22x faster │
│ QQuery 14 │  1311.69 / 1331.13 ±10.09 / 1338.86 ms │   989.62 / 1024.51 ±30.73 / 1068.11 ms │ +1.30x faster │
│ QQuery 15 │      330.59 / 339.96 ±7.33 / 350.96 ms │      321.08 / 331.23 ±8.60 / 345.94 ms │     no change │
│ QQuery 16 │  2352.52 / 2412.21 ±43.54 / 2482.79 ms │  1464.47 / 1488.68 ±18.58 / 1519.65 ms │ +1.62x faster │
│ QQuery 17 │  2438.17 / 2470.03 ±25.78 / 2500.17 ms │  1432.44 / 1508.52 ±69.34 / 1635.70 ms │ +1.64x faster │
│ QQuery 18 │ 3821.58 / 4134.48 ±183.78 / 4333.08 ms │ 2212.64 / 2337.59 ±114.57 / 2527.00 ms │ +1.77x faster │
│ QQuery 19 │        28.46 / 37.57 ±12.10 / 59.09 ms │         28.65 / 29.13 ±0.40 / 29.81 ms │ +1.29x faster │
│ QQuery 20 │   1109.11 / 1120.58 ±6.43 / 1127.97 ms │  1090.17 / 1112.87 ±20.22 / 1143.95 ms │     no change │
│ QQuery 21 │  1423.34 / 1441.22 ±11.91 / 1459.46 ms │  1416.33 / 1438.10 ±13.46 / 1457.92 ms │     no change │
│ QQuery 22 │  3406.54 / 3422.79 ±18.41 / 3447.82 ms │  3344.35 / 3406.18 ±56.83 / 3496.02 ms │     no change │
│ QQuery 23 │ 9521.92 / 9721.49 ±125.15 / 9887.97 ms │ 9591.87 / 9755.79 ±112.19 / 9926.82 ms │     no change │
│ QQuery 24 │      98.40 / 111.59 ±21.66 / 154.77 ms │     100.87 / 134.37 ±41.42 / 212.09 ms │  1.20x slower │
│ QQuery 25 │     361.51 / 375.79 ±23.74 / 423.20 ms │      359.02 / 361.57 ±2.17 / 365.06 ms │     no change │
│ QQuery 26 │       99.72 / 100.87 ±1.27 / 103.05 ms │       99.46 / 100.85 ±2.25 / 105.33 ms │     no change │
│ QQuery 27 │  1220.91 / 1241.83 ±16.17 / 1270.82 ms │   1203.09 / 1215.35 ±8.38 / 1229.00 ms │     no change │
│ QQuery 28 │  2805.58 / 2821.17 ±11.57 / 2837.89 ms │  2665.95 / 2698.20 ±29.78 / 2735.84 ms │     no change │
│ QQuery 29 │         39.26 / 43.92 ±7.77 / 59.43 ms │         38.76 / 41.93 ±5.36 / 52.64 ms │     no change │
│ QQuery 30 │     584.08 / 599.52 ±22.98 / 644.95 ms │      569.33 / 580.49 ±8.13 / 592.25 ms │     no change │
│ QQuery 31 │     570.65 / 587.39 ±23.43 / 633.15 ms │      569.23 / 573.82 ±3.69 / 579.30 ms │     no change │
│ QQuery 32 │ 1126.25 / 1371.62 ±291.80 / 1881.33 ms │  1092.32 / 1213.77 ±77.40 / 1314.60 ms │ +1.13x faster │
│ QQuery 33 │  7893.80 / 8046.80 ±86.97 / 8160.64 ms │ 4356.48 / 4455.49 ±131.86 / 4701.72 ms │ +1.81x faster │
│ QQuery 34 │ 8315.91 / 8532.97 ±196.52 / 8834.13 ms │ 4342.55 / 4515.90 ±109.84 / 4653.54 ms │ +1.89x faster │
│ QQuery 35 │      336.29 / 346.68 ±8.40 / 358.67 ms │     334.95 / 362.45 ±22.29 / 402.25 ms │     no change │
│ QQuery 36 │      156.07 / 163.45 ±5.67 / 173.27 ms │     144.02 / 166.80 ±23.77 / 209.60 ms │     no change │
│ QQuery 37 │       99.04 / 106.69 ±5.36 / 114.09 ms │       99.25 / 102.67 ±3.62 / 108.48 ms │     no change │
│ QQuery 38 │         84.67 / 87.90 ±2.54 / 90.26 ms │         84.69 / 88.45 ±2.78 / 92.16 ms │     no change │
│ QQuery 39 │      334.70 / 339.09 ±3.35 / 343.79 ms │     310.78 / 328.54 ±11.23 / 342.20 ms │     no change │
│ QQuery 40 │        15.86 / 21.65 ±11.41 / 44.46 ms │         14.27 / 15.09 ±0.57 / 16.06 ms │ +1.43x faster │
│ QQuery 41 │         14.54 / 14.87 ±0.30 / 15.38 ms │         14.52 / 17.02 ±4.61 / 26.24 ms │  1.14x slower │
│ QQuery 42 │         14.03 / 14.24 ±0.17 / 14.47 ms │         14.12 / 14.51 ±0.55 / 15.59 ms │     no change │
└───────────┴────────────────────────────────────────┴────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                              ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                              │ 57174.87ms │
│ Total Time (codex_bench-arrow-12008-current)   │ 44398.53ms │
│ Average Time (HEAD)                            │  1329.65ms │
│ Average Time (codex_bench-arrow-12008-current) │  1032.52ms │
│ Queries Faster                                 │         12 │
│ Queries Slower                                 │          2 │
│ Queries with No Change                         │         29 │
│ Queries with Failure                           │          0 │
└────────────────────────────────────────────────┴────────────┘

Memory Pool Peaks

Peak MemoryPool reservation per query — what DataFusion's accounting believes it reserved. Recorded only for benchmarks that write a results JSON (the dfbench suites, and the suites bench.sh runs through the Criterion SQL harness since #25644), and only under a memory limit: DATAFUSION_RUNTIME_MEMORY_LIMIT, or one the suite sets itself.

Base: 6d790c5f7441b0f7b786dadc0535140f41a5fa42 | Changed: e15a3816bdf45033ec1e3edfc8d05981641ad663

clickbench_partitioned

Query Base Changed Change
Query 0 0 B 0 B 0.0%
Query 1 104 B 104 B +0.0%
Query 2 936 B 936 B +0.0%
Query 3 312 B 312 B +0.0%
Query 4 829.3 MiB 828.5 MiB -0.1%
Query 5 1.8 GiB 1.9 GiB +4.4%
Query 6 0 B 0 B 0.0%
Query 7 50.3 MiB 60.1 MiB +19.6%
Query 8 921.2 MiB 942.3 MiB +2.3%
Query 9 552.0 MiB 552.0 MiB -0.0%
Query 10 73.6 MiB 73.7 MiB +0.1%
Query 11 77.0 MiB 78.3 MiB +1.7%
Query 12 1.9 GiB 1.9 GiB -0.7%
Query 13 2.3 GiB 2.5 GiB +6.0%
Query 14 2.0 GiB 2.2 GiB +12.8%
Query 15 1.2 GiB 1.2 GiB +0.7%
Query 16 3.5 GiB 3.5 GiB -0.6%
Query 17 3.5 GiB 3.5 GiB +0.8%
Query 18 7.6 GiB 7.0 GiB -8.5%
Query 19 0 B 0 B 0.0%
Query 20 104 B 104 B +0.0%
Query 21 1.4 MiB 1.5 MiB +12.3%
Query 22 2.5 MiB 2.9 MiB +12.9%
Query 23 12.2 MiB 13.1 MiB +7.0%
Query 24 6.8 MiB 6.2 MiB -7.8%
Query 25 6.9 MiB 7.6 MiB +9.5%
Query 26 8.0 MiB 7.0 MiB -11.9%
Query 27 2.5 MiB 2.7 MiB +7.4%
Query 28 2.1 GiB 2.1 GiB -0.3%
Query 29 624 B 624 B +0.0%
Query 30 648.1 MiB 654.9 MiB +1.1%
Query 31 1.5 GiB 1.5 GiB +0.0%
Query 32 7.1 GiB 7.2 GiB +0.3%
Query 33 14.2 GiB 16.9 GiB +18.4%
Query 34 12.3 GiB 18.2 GiB +47.6%
Query 35 664.8 MiB 655.1 MiB -1.5%
Query 36 474.1 MiB 474.1 MiB +0.0%
Query 37 8.1 MiB 8.1 MiB -0.0%
Query 38 4.8 MiB 5.1 MiB +6.3%
Query 39 844.2 MiB 844.2 MiB +0.0%
Query 40 2.3 MiB 2.5 MiB +9.6%
Query 41 3.3 MiB 3.3 MiB +0.0%
Query 42 1.6 MiB 1.8 MiB +15.4%

Pool accounting vs. process RSS

Max pool peak is the largest reservation any single query in the run reached; peak RSS covers the whole invocation, including data loading and allocator retention, and the two high-water marks need not coincide in time. The gap is therefore an upper bound on what the pool did not account for, not a measurement of it.

Benchmark Side Max pool peak Peak RSS Gap RSS / pool
clickbench_partitioned base (6d790c5f7441b0f7b786dadc0535140f41a5fa42) 14.2 GiB 14.7 GiB 438.7 MiB 1.0×
clickbench_partitioned changed (e15a3816bdf45033ec1e3edfc8d05981641ad663) 18.2 GiB 15.6 GiB −2.6 GiB 0.9×
Resource Usage

clickbench_partitioned — base (merge-base)

Metric Value
Wall time 290.1s
Peak memory 14.7 GiB
Avg memory 6.3 GiB
CPU user 2245.0s
CPU sys 200.2s
Peak spill 0 B

clickbench_partitioned — branch

Metric Value
Wall time 225.1s
Peak memory 15.6 GiB
Avg memory 5.9 GiB
CPU user 2096.4s
CPU sys 161.7s
Peak spill 400.3 MiB

File an issue against this benchmark runner

@Rich-T-kid Rich-T-kid mentioned this pull request Nov 4, 2026
Open
@Rich-T-kid Rich-T-kid closed this Nov 4, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

common Related to common crate physical-plan Changes to the physical-plan crate

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@Rich-T-kid @adriangbot