WIP speed up hashes (Primitive) by Rich-T-kid · Pull Request #26143 · apache/datafusion

Rich-T-kid · 2026-10-08T19:55:33Z

Which issue does this PR close?

Closes #.

Rationale for this change

What changes are included in this PR?

What is the testing strategy for this PR?

Are there any user-facing changes?

Rich-T-kid · 2026-10-08T19:57:12Z

show benchmark queue

adriangbot · 2026-10-08T19:57:13Z

Hi @Rich-T-kid, you asked to view the benchmark queue (#26143 (comment)).

Comment	Repo	PR	User	Benchmarks	Status
#6066055027	apache/arrow-rs	#11267	alamb	["arrow_writer"]	running
#6066413225	apache/arrow-rs	#11390	alamb	["row_selection_cursor"]	running
#6066541454	apache/arrow-rs	#11462	Rich-T-kid	["arrow_reader"]	running
#6066543381	apache/arrow-rs	#11463	Rich-T-kid	["arrow_reader"]	running
#6066544737	apache/arrow-rs	#11464	Rich-T-kid	["arrow_reader"]	running
#6066958247	apache/arrow-rs	#11449	adriangb	["arrow_writer"] — shard 2/4	running
#6066958247	apache/arrow-rs	#11449	adriangb	["arrow_writer"] — shard 3/4	running
#6066958247	apache/arrow-rs	#11449	adriangb	["arrow_writer"] — shard 4/4	running
#6066984849	apache/arrow-rs	#11267	adriangb	["arrow_writer"] — shard 1/4	running
#6066984849	apache/arrow-rs	#11267	adriangb	["arrow_writer"] — shard 2/4	running
#6066984849	apache/arrow-rs	#11267	adriangb	["arrow_writer"] — shard 3/4	pending
#6066984849	apache/arrow-rs	#11267	adriangb	["arrow_writer"] — shard 4/4	pending
#6067332837	apache/arrow-rs	#11464	Rich-T-kid	["arrow_reader_string_encoding"]	running
#6067335802	apache/arrow-rs	#11462	Rich-T-kid	["arrow_reader_string_encoding"]	running

File an issue against this benchmark runner

Rich-T-kid · 2026-10-08T19:57:31Z

run benchmarks

Rich-T-kid · 2026-10-08T19:57:39Z

run benchmark with_hashes

Rich-T-kid · 2026-10-08T20:00:47Z

run benchmark clickbench

adriangbot · 2026-10-08T20:09:12Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6067946271-3208-7lkxz 6.12.94+ #1 SMP Fri Aug 21 08:00:16 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 rich-T-kid/faster-hash-writes (5f5951c) to 409e3b8 (merge-base) diff

Run configuration

run benchmark clickbench_partitioned

Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-10-08T20:12:36Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6067948163-3211-x5smc 6.12.94+ #1 SMP Fri Aug 21 08:00:16 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 rich-T-kid/faster-hash-writes (5f5951c) to 409e3b8 (merge-base) diff

Run configuration

run benchmark with_hashes

Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-10-08T20:13:09Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6067946271-3209-j2b5t 6.12.94+ #1 SMP Fri Aug 21 08:00:16 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 rich-T-kid/faster-hash-writes (5f5951c) to 409e3b8 (merge-base) diff

Run configuration

run benchmark tpcds

Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-10-08T20:19:56Z

🤖 Benchmark completed (GKE) | trigger

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

Comparing rich-T-kid/faster-hash-writes (5f5951c) to 409e3b8 (merge-base) diff

Run configuration

run benchmark tpcds

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 rich-T-kid_faster-hash-writes
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ rich-T-kid_faster-hash-writes ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │    5.82 ms │                       5.93 ms │    no change │
│ QQuery 2  │   82.01 ms │                      81.43 ms │    no change │
│ QQuery 3  │   29.87 ms │                      29.49 ms │    no change │
│ QQuery 4  │  497.42 ms │                     496.46 ms │    no change │
│ QQuery 5  │   52.04 ms │                      51.22 ms │    no change │
│ QQuery 6  │   36.94 ms │                      37.06 ms │    no change │
│ QQuery 7  │   73.64 ms │                      73.46 ms │    no change │
│ QQuery 8  │   37.39 ms │                      37.04 ms │    no change │
│ QQuery 9  │   53.61 ms │                      52.92 ms │    no change │
│ QQuery 10 │   61.89 ms │                      61.22 ms │    no change │
│ QQuery 11 │  314.64 ms │                     314.40 ms │    no change │
│ QQuery 12 │   29.37 ms │                      30.09 ms │    no change │
│ QQuery 13 │  115.99 ms │                     116.76 ms │    no change │
│ QQuery 14 │  397.07 ms │                     393.93 ms │    no change │
│ QQuery 15 │   59.97 ms │                      59.81 ms │    no change │
│ QQuery 16 │    6.84 ms │                       6.93 ms │    no change │
│ QQuery 17 │   81.68 ms │                      80.97 ms │    no change │
│ QQuery 18 │  109.91 ms │                     109.41 ms │    no change │
│ QQuery 19 │   42.83 ms │                      42.77 ms │    no change │
│ QQuery 20 │   36.99 ms │                      37.03 ms │    no change │
│ QQuery 21 │   17.55 ms │                      17.62 ms │    no change │
│ QQuery 22 │   67.91 ms │                      69.29 ms │    no change │
│ QQuery 23 │  330.37 ms │                     329.85 ms │    no change │
│ QQuery 24 │  208.45 ms │                     205.16 ms │    no change │
│ QQuery 25 │  112.24 ms │                     109.89 ms │    no change │
│ QQuery 26 │   47.17 ms │                      46.57 ms │    no change │
│ QQuery 27 │    6.37 ms │                       6.32 ms │    no change │
│ QQuery 28 │   58.52 ms │                      57.76 ms │    no change │
│ QQuery 29 │   98.00 ms │                      97.37 ms │    no change │
│ QQuery 30 │   34.82 ms │                      33.92 ms │    no change │
│ QQuery 31 │  113.58 ms │                     112.96 ms │    no change │
│ QQuery 32 │   21.28 ms │                      21.43 ms │    no change │
│ QQuery 33 │   38.00 ms │                      38.78 ms │    no change │
│ QQuery 34 │   10.62 ms │                      10.96 ms │    no change │
│ QQuery 35 │   76.92 ms │                      76.88 ms │    no change │
│ QQuery 36 │    6.13 ms │                       6.25 ms │    no change │
│ QQuery 37 │    7.39 ms │                       7.44 ms │    no change │
│ QQuery 38 │   64.28 ms │                      65.22 ms │    no change │
│ QQuery 39 │   76.12 ms │                      75.67 ms │    no change │
│ QQuery 40 │   24.61 ms │                      24.86 ms │    no change │
│ QQuery 41 │   12.10 ms │                      12.28 ms │    no change │
│ QQuery 42 │   24.58 ms │                      25.16 ms │    no change │
│ QQuery 43 │    5.22 ms │                       5.23 ms │    no change │
│ QQuery 44 │    9.32 ms │                       9.23 ms │    no change │
│ QQuery 45 │   41.45 ms │                      41.37 ms │    no change │
│ QQuery 46 │   12.25 ms │                      12.18 ms │    no change │
│ QQuery 47 │  247.50 ms │                     246.75 ms │    no change │
│ QQuery 48 │   94.18 ms │                      93.30 ms │    no change │
│ QQuery 49 │   71.59 ms │                      71.21 ms │    no change │
│ QQuery 50 │   59.13 ms │                      59.71 ms │    no change │
│ QQuery 51 │   95.38 ms │                      92.81 ms │    no change │
│ QQuery 52 │   24.87 ms │                      24.89 ms │    no change │
│ QQuery 53 │   30.04 ms │                      29.90 ms │    no change │
│ QQuery 54 │   55.32 ms │                      54.70 ms │    no change │
│ QQuery 55 │   24.45 ms │                      24.07 ms │    no change │
│ QQuery 56 │   39.93 ms │                      38.99 ms │    no change │
│ QQuery 57 │  172.03 ms │                     172.61 ms │    no change │
│ QQuery 58 │  110.39 ms │                     108.52 ms │    no change │
│ QQuery 59 │  117.34 ms │                     116.77 ms │    no change │
│ QQuery 60 │   39.52 ms │                      39.65 ms │    no change │
│ QQuery 61 │   11.90 ms │                      11.62 ms │    no change │
│ QQuery 62 │   43.87 ms │                      44.08 ms │    no change │
│ QQuery 63 │   30.12 ms │                      29.76 ms │    no change │
│ QQuery 64 │  378.23 ms │                     371.59 ms │    no change │
│ QQuery 65 │  137.90 ms │                     137.46 ms │    no change │
│ QQuery 66 │   78.73 ms │                      79.45 ms │    no change │
│ QQuery 67 │  269.88 ms │                     272.35 ms │    no change │
│ QQuery 68 │   12.60 ms │                      12.58 ms │    no change │
│ QQuery 69 │   57.66 ms │                      57.97 ms │    no change │
│ QQuery 70 │  105.90 ms │                     105.36 ms │    no change │
│ QQuery 71 │   35.92 ms │                      35.54 ms │    no change │
│ QQuery 72 │ 1626.67 ms │                    1758.04 ms │ 1.08x slower │
│ QQuery 73 │   10.63 ms │                      10.46 ms │    no change │
│ QQuery 74 │  177.15 ms │                     175.33 ms │    no change │
│ QQuery 75 │  144.70 ms │                     142.63 ms │    no change │
│ QQuery 76 │   34.84 ms │                      35.13 ms │    no change │
│ QQuery 77 │   61.51 ms │                      61.85 ms │    no change │
│ QQuery 78 │  172.48 ms │                     164.21 ms │    no change │
│ QQuery 79 │   68.85 ms │                      67.10 ms │    no change │
│ QQuery 80 │   99.25 ms │                      96.98 ms │    no change │
│ QQuery 81 │   27.00 ms │                      26.82 ms │    no change │
│ QQuery 82 │   17.38 ms │                      16.87 ms │    no change │
│ QQuery 83 │   33.96 ms │                      33.08 ms │    no change │
│ QQuery 84 │   30.38 ms │                      29.92 ms │    no change │
│ QQuery 85 │   99.70 ms │                      99.30 ms │    no change │
│ QQuery 86 │   26.68 ms │                      25.83 ms │    no change │
│ QQuery 87 │   64.69 ms │                      63.86 ms │    no change │
│ QQuery 88 │   60.21 ms │                      60.48 ms │    no change │
│ QQuery 89 │   35.86 ms │                      35.65 ms │    no change │
│ QQuery 90 │   16.99 ms │                      16.99 ms │    no change │
│ QQuery 91 │   42.14 ms │                      41.39 ms │    no change │
│ QQuery 92 │   30.47 ms │                      30.70 ms │    no change │
│ QQuery 93 │   50.23 ms │                      50.30 ms │    no change │
│ QQuery 94 │   39.55 ms │                      40.12 ms │    no change │
│ QQuery 95 │   83.13 ms │                      81.79 ms │    no change │
│ QQuery 96 │   24.15 ms │                      24.23 ms │    no change │
│ QQuery 97 │   51.50 ms │                      51.20 ms │    no change │
│ QQuery 98 │   44.10 ms │                      43.32 ms │    no change │
│ QQuery 99 │   65.39 ms │                      65.11 ms │    no change │
└───────────┴────────────┴───────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 9235.05ms │
│ Total Time (rich-T-kid_faster-hash-writes)   │ 9318.32ms │
│ Average Time (HEAD)                          │   93.28ms │
│ Average Time (rich-T-kid_faster-hash-writes) │   94.12ms │
│ Queries Faster                               │         0 │
│ Queries Slower                               │         1 │
│ Queries with No Change                       │        98 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

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

Comparing HEAD and rich-T-kid_faster-hash-writes
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                   HEAD ┃         rich-T-kid_faster-hash-writes ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │            5.82 / 6.38 ±0.93 / 8.24 ms │           5.93 / 6.52 ±0.88 / 8.28 ms │     no change │
│ QQuery 2  │         82.01 / 82.37 ±0.24 / 82.72 ms │        81.43 / 81.88 ±0.30 / 82.24 ms │     no change │
│ QQuery 3  │         29.87 / 30.06 ±0.19 / 30.40 ms │        29.49 / 29.65 ±0.09 / 29.76 ms │     no change │
│ QQuery 4  │      497.42 / 503.33 ±3.88 / 508.51 ms │     496.46 / 503.43 ±9.58 / 522.21 ms │     no change │
│ QQuery 5  │         52.04 / 52.47 ±0.50 / 53.31 ms │        51.22 / 52.01 ±0.65 / 53.00 ms │     no change │
│ QQuery 6  │         36.94 / 37.44 ±0.45 / 38.24 ms │        37.06 / 37.56 ±0.49 / 38.46 ms │     no change │
│ QQuery 7  │         73.64 / 74.15 ±0.39 / 74.76 ms │        73.46 / 73.74 ±0.35 / 74.31 ms │     no change │
│ QQuery 8  │         37.39 / 38.41 ±1.84 / 42.08 ms │        37.04 / 38.12 ±1.68 / 41.45 ms │     no change │
│ QQuery 9  │         53.61 / 55.23 ±0.82 / 55.93 ms │        52.92 / 55.55 ±1.96 / 57.87 ms │     no change │
│ QQuery 10 │         61.89 / 62.32 ±0.44 / 63.09 ms │        61.22 / 61.71 ±0.44 / 62.30 ms │     no change │
│ QQuery 11 │      314.64 / 319.94 ±3.63 / 325.82 ms │     314.40 / 318.11 ±2.60 / 321.41 ms │     no change │
│ QQuery 12 │         29.37 / 30.68 ±0.92 / 32.24 ms │        30.09 / 30.42 ±0.35 / 31.07 ms │     no change │
│ QQuery 13 │      115.99 / 117.03 ±1.03 / 118.68 ms │     116.76 / 117.65 ±1.25 / 119.98 ms │     no change │
│ QQuery 14 │      397.07 / 400.60 ±2.96 / 405.52 ms │     393.93 / 401.67 ±4.98 / 406.74 ms │     no change │
│ QQuery 15 │         59.97 / 61.14 ±0.70 / 61.83 ms │        59.81 / 60.35 ±0.46 / 61.07 ms │     no change │
│ QQuery 16 │            6.84 / 6.94 ±0.14 / 7.21 ms │           6.93 / 7.05 ±0.11 / 7.21 ms │     no change │
│ QQuery 17 │         81.68 / 83.54 ±2.36 / 88.18 ms │        80.97 / 81.64 ±0.51 / 82.49 ms │     no change │
│ QQuery 18 │      109.91 / 110.75 ±0.49 / 111.32 ms │     109.41 / 110.54 ±0.95 / 111.88 ms │     no change │
│ QQuery 19 │         42.83 / 43.34 ±0.41 / 43.91 ms │        42.77 / 43.25 ±0.39 / 43.85 ms │     no change │
│ QQuery 20 │         36.99 / 37.78 ±0.43 / 38.22 ms │        37.03 / 37.94 ±0.66 / 38.70 ms │     no change │
│ QQuery 21 │         17.55 / 19.00 ±2.10 / 23.16 ms │        17.62 / 18.30 ±1.12 / 20.52 ms │     no change │
│ QQuery 22 │         67.91 / 69.38 ±1.44 / 72.10 ms │        69.29 / 69.48 ±0.20 / 69.84 ms │     no change │
│ QQuery 23 │      330.37 / 334.95 ±3.38 / 340.60 ms │     329.85 / 333.24 ±2.08 / 335.38 ms │     no change │
│ QQuery 24 │      208.45 / 210.87 ±1.54 / 212.35 ms │     205.16 / 210.82 ±6.89 / 224.30 ms │     no change │
│ QQuery 25 │      112.24 / 113.45 ±1.33 / 115.26 ms │     109.89 / 112.34 ±2.17 / 116.00 ms │     no change │
│ QQuery 26 │         47.17 / 47.89 ±0.89 / 49.63 ms │        46.57 / 47.40 ±0.47 / 47.98 ms │     no change │
│ QQuery 27 │            6.37 / 6.52 ±0.17 / 6.84 ms │           6.32 / 6.63 ±0.25 / 7.06 ms │     no change │
│ QQuery 28 │         58.52 / 62.43 ±3.77 / 69.52 ms │        57.76 / 63.18 ±4.75 / 72.07 ms │     no change │
│ QQuery 29 │        98.00 / 99.64 ±1.40 / 102.03 ms │       97.37 / 99.74 ±2.44 / 103.97 ms │     no change │
│ QQuery 30 │         34.82 / 35.17 ±0.23 / 35.38 ms │        33.92 / 34.28 ±0.22 / 34.53 ms │     no change │
│ QQuery 31 │      113.58 / 114.00 ±0.33 / 114.44 ms │     112.96 / 114.28 ±1.80 / 117.84 ms │     no change │
│ QQuery 32 │         21.28 / 21.55 ±0.20 / 21.80 ms │        21.43 / 21.62 ±0.13 / 21.79 ms │     no change │
│ QQuery 33 │         38.00 / 38.49 ±0.40 / 39.13 ms │        38.78 / 39.14 ±0.19 / 39.35 ms │     no change │
│ QQuery 34 │         10.62 / 10.87 ±0.23 / 11.20 ms │        10.96 / 11.13 ±0.15 / 11.32 ms │     no change │
│ QQuery 35 │         76.92 / 78.81 ±2.10 / 82.70 ms │        76.88 / 78.17 ±1.44 / 80.79 ms │     no change │
│ QQuery 36 │            6.13 / 6.25 ±0.10 / 6.42 ms │           6.25 / 6.79 ±0.87 / 8.51 ms │  1.09x slower │
│ QQuery 37 │            7.39 / 7.52 ±0.10 / 7.69 ms │           7.44 / 7.70 ±0.16 / 7.88 ms │     no change │
│ QQuery 38 │         64.28 / 65.26 ±0.61 / 66.02 ms │        65.22 / 65.71 ±0.35 / 66.21 ms │     no change │
│ QQuery 39 │         76.12 / 78.07 ±2.52 / 82.91 ms │        75.67 / 77.06 ±1.13 / 79.12 ms │     no change │
│ QQuery 40 │         24.61 / 25.48 ±0.50 / 26.14 ms │        24.86 / 25.87 ±0.64 / 26.54 ms │     no change │
│ QQuery 41 │         12.10 / 12.46 ±0.18 / 12.61 ms │        12.28 / 12.41 ±0.08 / 12.50 ms │     no change │
│ QQuery 42 │         24.58 / 25.21 ±0.72 / 26.42 ms │        25.16 / 25.86 ±0.54 / 26.68 ms │     no change │
│ QQuery 43 │            5.22 / 5.32 ±0.10 / 5.50 ms │           5.23 / 5.27 ±0.05 / 5.36 ms │     no change │
│ QQuery 44 │            9.32 / 9.49 ±0.14 / 9.66 ms │           9.23 / 9.39 ±0.08 / 9.46 ms │     no change │
│ QQuery 45 │         41.45 / 42.21 ±0.68 / 43.21 ms │        41.37 / 42.23 ±0.66 / 43.03 ms │     no change │
│ QQuery 46 │         12.25 / 12.53 ±0.20 / 12.76 ms │        12.18 / 12.62 ±0.34 / 13.20 ms │     no change │
│ QQuery 47 │      247.50 / 250.78 ±4.55 / 259.72 ms │     246.75 / 250.93 ±2.94 / 254.13 ms │     no change │
│ QQuery 48 │         94.18 / 95.00 ±0.64 / 95.89 ms │        93.30 / 94.68 ±1.71 / 97.96 ms │     no change │
│ QQuery 49 │         71.59 / 72.00 ±0.43 / 72.81 ms │        71.21 / 72.12 ±0.56 / 72.72 ms │     no change │
│ QQuery 50 │         59.13 / 60.56 ±0.89 / 61.68 ms │        59.71 / 62.03 ±3.84 / 69.63 ms │     no change │
│ QQuery 51 │         95.38 / 95.94 ±0.37 / 96.44 ms │       92.81 / 96.46 ±3.73 / 103.27 ms │     no change │
│ QQuery 52 │         24.87 / 25.10 ±0.22 / 25.45 ms │        24.89 / 25.31 ±0.48 / 26.24 ms │     no change │
│ QQuery 53 │         30.04 / 30.27 ±0.20 / 30.57 ms │        29.90 / 30.08 ±0.25 / 30.56 ms │     no change │
│ QQuery 54 │         55.32 / 57.95 ±4.45 / 66.82 ms │        54.70 / 55.71 ±0.71 / 56.79 ms │     no change │
│ QQuery 55 │         24.45 / 25.25 ±0.75 / 26.53 ms │        24.07 / 25.16 ±0.69 / 26.15 ms │     no change │
│ QQuery 56 │         39.93 / 40.03 ±0.11 / 40.23 ms │        38.99 / 39.65 ±0.57 / 40.54 ms │     no change │
│ QQuery 57 │      172.03 / 176.65 ±4.95 / 185.78 ms │     172.61 / 176.48 ±3.05 / 181.96 ms │     no change │
│ QQuery 58 │      110.39 / 111.31 ±0.75 / 112.57 ms │     108.52 / 110.02 ±1.42 / 112.70 ms │     no change │
│ QQuery 59 │      117.34 / 119.85 ±2.45 / 123.95 ms │     116.77 / 118.07 ±0.77 / 119.15 ms │     no change │
│ QQuery 60 │         39.52 / 40.68 ±1.06 / 42.61 ms │        39.65 / 40.19 ±0.47 / 40.84 ms │     no change │
│ QQuery 61 │         11.90 / 12.00 ±0.09 / 12.12 ms │        11.62 / 11.81 ±0.15 / 12.08 ms │     no change │
│ QQuery 62 │         43.87 / 44.66 ±0.70 / 45.98 ms │        44.08 / 44.51 ±0.38 / 45.15 ms │     no change │
│ QQuery 63 │         30.12 / 30.76 ±0.86 / 32.39 ms │        29.76 / 31.42 ±2.37 / 36.02 ms │     no change │
│ QQuery 64 │      378.23 / 382.85 ±3.73 / 389.22 ms │     371.59 / 374.83 ±2.45 / 378.03 ms │     no change │
│ QQuery 65 │      137.90 / 143.04 ±7.10 / 156.87 ms │     137.46 / 139.61 ±1.59 / 142.39 ms │     no change │
│ QQuery 66 │         78.73 / 80.97 ±3.19 / 87.22 ms │        79.45 / 82.07 ±4.51 / 91.09 ms │     no change │
│ QQuery 67 │      269.88 / 276.91 ±7.01 / 287.90 ms │     272.35 / 279.28 ±6.44 / 287.65 ms │     no change │
│ QQuery 68 │         12.60 / 12.69 ±0.07 / 12.78 ms │        12.58 / 12.71 ±0.14 / 12.94 ms │     no change │
│ QQuery 69 │         57.66 / 57.98 ±0.27 / 58.33 ms │        57.97 / 58.32 ±0.27 / 58.76 ms │     no change │
│ QQuery 70 │      105.90 / 109.74 ±3.97 / 115.01 ms │     105.36 / 108.88 ±3.53 / 114.97 ms │     no change │
│ QQuery 71 │         35.92 / 36.67 ±0.54 / 37.45 ms │        35.54 / 36.09 ±0.33 / 36.46 ms │     no change │
│ QQuery 72 │ 1626.67 / 1738.54 ±111.17 / 1919.50 ms │ 1758.04 / 1815.54 ±55.16 / 1906.50 ms │     no change │
│ QQuery 73 │         10.63 / 11.40 ±0.75 / 12.66 ms │        10.46 / 11.02 ±0.34 / 11.41 ms │     no change │
│ QQuery 74 │      177.15 / 183.07 ±3.67 / 186.60 ms │     175.33 / 177.38 ±1.48 / 179.56 ms │     no change │
│ QQuery 75 │      144.70 / 145.86 ±1.14 / 147.88 ms │     142.63 / 145.38 ±2.96 / 150.81 ms │     no change │
│ QQuery 76 │         34.84 / 36.81 ±3.14 / 43.06 ms │        35.13 / 36.05 ±1.09 / 38.07 ms │     no change │
│ QQuery 77 │         61.51 / 62.34 ±0.56 / 63.28 ms │        61.85 / 62.10 ±0.35 / 62.78 ms │     no change │
│ QQuery 78 │      172.48 / 173.44 ±0.97 / 174.83 ms │     164.21 / 170.01 ±3.66 / 174.81 ms │     no change │
│ QQuery 79 │         68.85 / 69.83 ±1.09 / 71.91 ms │        67.10 / 67.72 ±0.49 / 68.53 ms │     no change │
│ QQuery 80 │       99.25 / 102.61 ±5.53 / 113.62 ms │       96.98 / 99.34 ±2.29 / 103.62 ms │     no change │
│ QQuery 81 │         27.00 / 27.82 ±0.85 / 29.32 ms │        26.82 / 27.12 ±0.30 / 27.69 ms │     no change │
│ QQuery 82 │         17.38 / 17.65 ±0.26 / 17.99 ms │        16.87 / 17.20 ±0.24 / 17.51 ms │     no change │
│ QQuery 83 │         33.96 / 34.58 ±0.38 / 35.17 ms │        33.08 / 33.79 ±0.45 / 34.49 ms │     no change │
│ QQuery 84 │         30.38 / 31.41 ±1.19 / 33.67 ms │        29.92 / 30.15 ±0.24 / 30.48 ms │     no change │
│ QQuery 85 │       99.70 / 102.13 ±3.18 / 108.42 ms │      99.30 / 103.92 ±4.12 / 111.40 ms │     no change │
│ QQuery 86 │         26.68 / 27.68 ±1.22 / 30.05 ms │        25.83 / 26.18 ±0.42 / 26.97 ms │ +1.06x faster │
│ QQuery 87 │         64.69 / 65.51 ±0.72 / 66.77 ms │        63.86 / 64.48 ±0.59 / 65.40 ms │     no change │
│ QQuery 88 │         60.21 / 60.69 ±0.47 / 61.60 ms │        60.48 / 63.25 ±4.97 / 73.19 ms │     no change │
│ QQuery 89 │         35.86 / 36.43 ±0.44 / 37.11 ms │        35.65 / 36.73 ±1.09 / 38.74 ms │     no change │
│ QQuery 90 │         16.99 / 17.09 ±0.08 / 17.20 ms │        16.99 / 17.22 ±0.13 / 17.36 ms │     no change │
│ QQuery 91 │         42.14 / 43.45 ±1.47 / 46.24 ms │        41.39 / 42.16 ±0.51 / 42.71 ms │     no change │
│ QQuery 92 │         30.47 / 31.02 ±0.41 / 31.51 ms │        30.70 / 31.18 ±0.37 / 31.82 ms │     no change │
│ QQuery 93 │         50.23 / 51.38 ±0.87 / 52.54 ms │        50.30 / 52.33 ±2.05 / 55.92 ms │     no change │
│ QQuery 94 │         39.55 / 41.33 ±2.93 / 47.19 ms │        40.12 / 40.76 ±0.93 / 42.62 ms │     no change │
│ QQuery 95 │         83.13 / 84.44 ±0.77 / 85.51 ms │        81.79 / 82.59 ±0.47 / 83.12 ms │     no change │
│ QQuery 96 │         24.15 / 24.35 ±0.10 / 24.44 ms │        24.23 / 24.50 ±0.17 / 24.67 ms │     no change │
│ QQuery 97 │         51.50 / 52.07 ±0.38 / 52.51 ms │        51.20 / 54.73 ±2.47 / 58.90 ms │  1.05x slower │
│ QQuery 98 │         44.10 / 47.40 ±3.69 / 54.56 ms │        43.32 / 44.75 ±1.00 / 46.40 ms │ +1.06x faster │
│ QQuery 99 │         65.39 / 66.05 ±0.56 / 66.86 ms │        65.11 / 65.32 ±0.15 / 65.45 ms │     no change │
└───────────┴────────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 9482.71ms │
│ Total Time (rich-T-kid_faster-hash-writes)   │ 9518.80ms │
│ Average Time (HEAD)                          │   95.78ms │
│ Average Time (rich-T-kid_faster-hash-writes) │   96.15ms │
│ Queries Faster                               │         2 │
│ Queries Slower                               │         2 │
│ Queries with No Change                       │        95 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Resource Usage

tpcds — base (merge-base)

Metric	Value
Wall time	50.0s
Peak memory	2.2 GiB
Avg memory	1.5 GiB
CPU user	200.2s
CPU sys	6.2s
Peak spill	0 B

tpcds — branch

Metric	Value
Wall time	50.0s
Peak memory	2.2 GiB
Avg memory	1.5 GiB
CPU user	204.7s
CPU sys	6.2s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-10-08T20:21:08Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6067998159-3212-2d5bx 6.12.94+ #1 SMP Fri Aug 21 08:00:16 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 rich-T-kid/faster-hash-writes (5f5951c) to 409e3b8 (merge-base) diff

Run configuration

run benchmark clickbench

Results will be posted here when complete

File an issue against this benchmark runner

codecov-commenter · 2026-10-08T20:21:37Z

Codecov Report

❌ Patch coverage is 97.87234% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 82.75%. Comparing base (409e3b8) to head (926cd62).
⚠️ Report is 2 commits behind head on main.

Files with missing lines	Patch %	Lines
datafusion/common/src/hash_utils.rs	97.87%	1 Missing and 2 partials ⚠️

Additional details and impacted files

@@           Coverage Diff            @@
##             main   #26143    +/-   ##
========================================
  Coverage   82.75%   82.75%            
========================================
  Files        1147     1147            
  Lines      449886   450007   +121     
  Branches   449886   450007   +121     
========================================
+ Hits       372303   372419   +116     
- Misses      54911    54914     +3     
- Partials    22672    22674     +2

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

adriangbot · 2026-10-08T20:21:59Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6067946271-3210-fmqbw 6.12.94+ #1 SMP Fri Aug 21 08:00:16 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 rich-T-kid/faster-hash-writes (5f5951c) to 409e3b8 (merge-base) diff

Run configuration

run benchmark tpch

Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-10-08T20:22:53Z

🤖 Benchmark completed (GKE) | trigger

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

Comparing rich-T-kid/faster-hash-writes (5f5951c) to 409e3b8 (merge-base) diff

Run configuration

run benchmark clickbench_partitioned

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 rich-T-kid_faster-hash-writes
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ rich-T-kid_faster-hash-writes ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 0  │    1.22 ms │                       1.21 ms │    no change │
│ QQuery 1  │   11.92 ms │                      11.80 ms │    no change │
│ QQuery 2  │   37.35 ms │                      36.59 ms │    no change │
│ QQuery 3  │   31.66 ms │                      31.62 ms │    no change │
│ QQuery 4  │  235.99 ms │                     234.81 ms │    no change │
│ QQuery 5  │  273.36 ms │                     270.72 ms │    no change │
│ QQuery 6  │    1.28 ms │                       1.26 ms │    no change │
│ QQuery 7  │   12.89 ms │                      12.79 ms │    no change │
│ QQuery 8  │  328.71 ms │                     324.54 ms │    no change │
│ QQuery 9  │  457.70 ms │                     469.62 ms │    no change │
│ QQuery 10 │   63.91 ms │                      63.79 ms │    no change │
│ QQuery 11 │   74.68 ms │                      74.56 ms │    no change │
│ QQuery 12 │  262.82 ms │                     259.41 ms │    no change │
│ QQuery 13 │  359.91 ms │                     366.71 ms │    no change │
│ QQuery 14 │  278.57 ms │                     276.87 ms │    no change │
│ QQuery 15 │  281.60 ms │                     289.98 ms │    no change │
│ QQuery 16 │  615.75 ms │                     621.33 ms │    no change │
│ QQuery 17 │  614.93 ms │                     630.15 ms │    no change │
│ QQuery 18 │ 1250.90 ms │                    1278.33 ms │    no change │
│ QQuery 19 │   27.96 ms │                      28.52 ms │    no change │
│ QQuery 20 │  517.07 ms │                     518.83 ms │    no change │
│ QQuery 21 │  508.00 ms │                     519.58 ms │    no change │
│ QQuery 22 │  987.96 ms │                    1035.12 ms │    no change │
│ QQuery 23 │ 3135.19 ms │                    3263.22 ms │    no change │
│ QQuery 24 │   41.15 ms │                      41.18 ms │    no change │
│ QQuery 25 │  105.23 ms │                     105.46 ms │    no change │
│ QQuery 26 │   41.85 ms │                      41.31 ms │    no change │
│ QQuery 27 │  513.40 ms │                     512.14 ms │    no change │
│ QQuery 28 │ 2907.43 ms │                    2896.68 ms │    no change │
│ QQuery 29 │   40.01 ms │                      39.11 ms │    no change │
│ QQuery 30 │  304.35 ms │                     301.77 ms │    no change │
│ QQuery 31 │  286.92 ms │                     279.50 ms │    no change │
│ QQuery 32 │ 1068.43 ms │                    1033.69 ms │    no change │
│ QQuery 33 │ 1540.79 ms │                    1560.26 ms │    no change │
│ QQuery 34 │ 1563.59 ms │                    1602.24 ms │    no change │
│ QQuery 35 │  300.44 ms │                     295.53 ms │    no change │
│ QQuery 36 │   66.41 ms │                      72.67 ms │ 1.09x slower │
│ QQuery 37 │   35.22 ms │                      36.78 ms │    no change │
│ QQuery 38 │   41.56 ms │                      44.33 ms │ 1.07x slower │
│ QQuery 39 │  138.52 ms │                     148.37 ms │ 1.07x slower │
│ QQuery 40 │   14.02 ms │                      13.69 ms │    no change │
│ QQuery 41 │   13.61 ms │                      13.23 ms │    no change │
│ QQuery 42 │   13.00 ms │                      12.64 ms │    no change │
└───────────┴────────────┴───────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 19407.26ms │
│ Total Time (rich-T-kid_faster-hash-writes)   │ 19671.95ms │
│ Average Time (HEAD)                          │   451.33ms │
│ Average Time (rich-T-kid_faster-hash-writes) │   457.49ms │
│ Queries Faster                               │          0 │
│ Queries Slower                               │          3 │
│ Queries with No Change                       │         40 │
│ Queries with Failure                         │          0 │
└──────────────────────────────────────────────┴────────────┘

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

Comparing HEAD and rich-T-kid_faster-hash-writes
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃         rich-T-kid_faster-hash-writes ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │          1.22 / 3.97 ±5.44 / 14.86 ms │          1.21 / 3.95 ±5.39 / 14.74 ms │     no change │
│ QQuery 1  │        11.92 / 12.02 ±0.11 / 12.19 ms │        11.80 / 11.84 ±0.05 / 11.93 ms │     no change │
│ QQuery 2  │        37.35 / 37.82 ±0.29 / 38.13 ms │        36.59 / 36.87 ±0.20 / 37.11 ms │     no change │
│ QQuery 3  │        31.66 / 32.64 ±0.96 / 34.26 ms │        31.62 / 32.13 ±0.81 / 33.74 ms │     no change │
│ QQuery 4  │     235.99 / 239.25 ±2.12 / 242.22 ms │     234.81 / 237.61 ±2.88 / 243.15 ms │     no change │
│ QQuery 5  │     273.36 / 277.84 ±3.10 / 281.74 ms │     270.72 / 273.43 ±2.21 / 277.09 ms │     no change │
│ QQuery 6  │           1.28 / 1.44 ±0.23 / 1.89 ms │           1.26 / 1.51 ±0.26 / 1.87 ms │  1.05x slower │
│ QQuery 7  │        12.89 / 13.00 ±0.13 / 13.23 ms │        12.79 / 13.02 ±0.15 / 13.22 ms │     no change │
│ QQuery 8  │     328.71 / 335.43 ±5.19 / 342.77 ms │     324.54 / 327.87 ±3.93 / 335.36 ms │     no change │
│ QQuery 9  │    457.70 / 473.96 ±11.64 / 492.03 ms │     469.62 / 479.74 ±8.04 / 493.70 ms │     no change │
│ QQuery 10 │        63.91 / 66.34 ±2.44 / 69.73 ms │        63.79 / 64.55 ±0.66 / 65.58 ms │     no change │
│ QQuery 11 │        74.68 / 76.12 ±1.61 / 79.25 ms │        74.56 / 75.96 ±1.55 / 78.82 ms │     no change │
│ QQuery 12 │     262.82 / 268.12 ±4.30 / 274.77 ms │     259.41 / 263.70 ±3.28 / 268.96 ms │     no change │
│ QQuery 13 │    359.91 / 371.82 ±14.24 / 395.19 ms │    366.71 / 388.24 ±14.35 / 403.68 ms │     no change │
│ QQuery 14 │     278.57 / 281.11 ±2.25 / 283.78 ms │    276.87 / 295.54 ±24.86 / 344.78 ms │  1.05x slower │
│ QQuery 15 │     281.60 / 290.63 ±8.36 / 305.69 ms │     289.98 / 293.97 ±3.30 / 298.71 ms │     no change │
│ QQuery 16 │     615.75 / 622.96 ±7.87 / 637.42 ms │    621.33 / 635.51 ±14.01 / 662.14 ms │     no change │
│ QQuery 17 │     614.93 / 628.78 ±9.14 / 642.20 ms │     630.15 / 641.88 ±8.67 / 656.17 ms │     no change │
│ QQuery 18 │ 1250.90 / 1277.77 ±19.30 / 1300.82 ms │ 1278.33 / 1304.69 ±20.09 / 1333.41 ms │     no change │
│ QQuery 19 │        27.96 / 28.72 ±0.82 / 30.14 ms │        28.52 / 29.13 ±0.39 / 29.76 ms │     no change │
│ QQuery 20 │     517.07 / 528.54 ±9.58 / 543.43 ms │     518.83 / 527.21 ±8.98 / 544.01 ms │     no change │
│ QQuery 21 │    508.00 / 524.51 ±10.65 / 539.28 ms │    519.58 / 531.74 ±10.85 / 547.36 ms │     no change │
│ QQuery 22 │    987.96 / 995.97 ±5.38 / 1004.69 ms │ 1035.12 / 1058.52 ±18.29 / 1080.01 ms │  1.06x slower │
│ QQuery 23 │ 3135.19 / 3198.98 ±46.48 / 3255.54 ms │ 3263.22 / 3327.20 ±70.97 / 3459.29 ms │     no change │
│ QQuery 24 │      41.15 / 99.60 ±82.35 / 252.44 ms │    41.18 / 166.60 ±207.04 / 574.37 ms │  1.67x slower │
│ QQuery 25 │   105.23 / 165.74 ±117.50 / 400.70 ms │     105.46 / 107.40 ±1.13 / 108.55 ms │ +1.54x faster │
│ QQuery 26 │        41.85 / 43.02 ±1.76 / 46.50 ms │        41.31 / 41.77 ±0.27 / 42.05 ms │     no change │
│ QQuery 27 │     513.40 / 522.73 ±6.69 / 530.72 ms │    512.14 / 525.85 ±12.19 / 542.80 ms │     no change │
│ QQuery 28 │ 2907.43 / 2925.86 ±16.59 / 2945.07 ms │ 2896.68 / 2921.06 ±20.73 / 2959.47 ms │     no change │
│ QQuery 29 │       40.01 / 58.46 ±21.46 / 98.80 ms │      39.11 / 59.70 ±23.23 / 101.98 ms │     no change │
│ QQuery 30 │    304.35 / 316.82 ±16.29 / 349.02 ms │     301.77 / 313.31 ±9.82 / 329.05 ms │     no change │
│ QQuery 31 │     286.92 / 294.50 ±7.12 / 305.28 ms │    279.50 / 295.36 ±13.63 / 313.64 ms │     no change │
│ QQuery 32 │ 1068.43 / 1143.38 ±58.90 / 1234.97 ms │ 1033.69 / 1082.18 ±46.18 / 1156.46 ms │ +1.06x faster │
│ QQuery 33 │ 1540.79 / 1578.36 ±24.87 / 1603.92 ms │ 1560.26 / 1575.54 ±11.67 / 1592.17 ms │     no change │
│ QQuery 34 │ 1563.59 / 1633.74 ±44.49 / 1697.52 ms │ 1602.24 / 1640.81 ±31.08 / 1679.36 ms │     no change │
│ QQuery 35 │    300.44 / 335.61 ±55.02 / 445.33 ms │    295.53 / 348.21 ±68.60 / 478.58 ms │     no change │
│ QQuery 36 │        66.41 / 68.89 ±2.36 / 72.93 ms │        72.67 / 81.72 ±7.34 / 91.72 ms │  1.19x slower │
│ QQuery 37 │       35.22 / 48.01 ±18.98 / 85.63 ms │       36.78 / 46.03 ±10.95 / 66.63 ms │     no change │
│ QQuery 38 │        41.56 / 46.44 ±4.27 / 52.88 ms │        44.33 / 45.64 ±1.46 / 48.10 ms │     no change │
│ QQuery 39 │    138.52 / 158.29 ±14.62 / 182.27 ms │    148.37 / 161.01 ±11.38 / 180.21 ms │     no change │
│ QQuery 40 │        14.02 / 14.85 ±0.52 / 15.56 ms │        13.69 / 14.08 ±0.41 / 14.83 ms │ +1.05x faster │
│ QQuery 41 │        13.61 / 14.82 ±2.09 / 19.01 ms │        13.23 / 15.59 ±4.32 / 24.21 ms │  1.05x slower │
│ QQuery 42 │        13.00 / 15.04 ±3.68 / 22.40 ms │        12.64 / 12.91 ±0.19 / 13.22 ms │ +1.17x faster │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 20071.89ms │
│ Total Time (rich-T-kid_faster-hash-writes)   │ 20310.57ms │
│ Average Time (HEAD)                          │   466.79ms │
│ Average Time (rich-T-kid_faster-hash-writes) │   472.34ms │
│ Queries Faster                               │          4 │
│ Queries Slower                               │          6 │
│ Queries with No Change                       │         33 │
│ Queries with Failure                         │          0 │
└──────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric	Value
Wall time	105.0s
Peak memory	18.5 GiB
Avg memory	6.0 GiB
CPU user	993.4s
CPU sys	95.5s
Peak spill	0 B

clickbench_partitioned — branch

Metric	Value
Wall time	105.0s
Peak memory	17.9 GiB
Avg memory	5.6 GiB
CPU user	992.9s
CPU sys	98.9s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-10-08T20:24:41Z

Benchmark for this request failed before finishing (Kubernetes reason: BackoffLimitExceeded).

Benchmarks requested: clickbench

Runner log (last 40 lines)

h2o_medium_window_sorted_parquet: Window Top-N over a declared-sorted h2o input, medium dataset (1e8 rows), source file format is parquet
h2o_big_window_sorted_parquet:    Window Top-N over a declared-sorted h2o input, large dataset (1e9 rows),  source file format is parquet
h2o_small_parquet:              h2oai benchmark with small dataset (1e7 rows) for groupby,  file format is parquet
h2o_medium_parquet:             h2oai benchmark with medium dataset (1e8 rows) for groupby, file format is parquet
h2o_big_parquet:                h2oai benchmark with large dataset (1e9 rows) for groupby,  file format is parquet
h2o_small_join_parquet:         h2oai benchmark with small dataset (1e7 rows) for join,  file format is parquet
h2o_medium_join_parquet:        h2oai benchmark with medium dataset (1e8 rows) for join, file format is parquet
h2o_big_join_parquet:           h2oai benchmark with large dataset (1e9 rows) for join,  file format is parquet
h2o_small_window_parquet:       Extended h2oai benchmark with small dataset (1e7 rows) for window,  file format is parquet
h2o_medium_window_parquet:      Extended h2oai benchmark with medium dataset (1e8 rows) for window, file format is parquet
h2o_big_window_parquet:         Extended h2oai benchmark with large dataset (1e9 rows) for window,  file format is parquet

# Join Order Benchmark (IMDB)
imdb:                   Join Order Benchmark (JOB) using the IMDB dataset converted to parquet

# Micro-Benchmarks (specific operators and features)
cancellation:           How long cancelling a query takes
asof_join:              ASOF join workloads varying size, ordering, grouping, match direction, and payload width
nlj:                    Benchmark for simple nested loop joins, testing various join scenarios
hj:                     Benchmark for simple hash joins, testing various join scenarios
smj:                    Benchmark for simple sort merge joins, testing various join scenarios
dict:                   Benchmark for dictionary-encoded group-by scenarios
array_agg_distinct:     1000K-group, two-row-per-group array_agg(DISTINCT) benchmark
compile_profile:        Compile and execute TPC-H across selected Cargo profiles, reporting timing and binary size


━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Supported Configuration (Environment Variables)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
DATA_DIR            directory to store datasets
CARGO_COMMAND       command that runs the benchmark binary
DATAFUSION_DIR      directory to use (default /workspace/datafusion-base)
RESULTS_NAME        folder where the benchmark files are stored
PREFER_HASH_JOIN    Prefer hash join algorithm (default true)
SIMULATE_LATENCY    Simulate object store latency to mimic S3 (default false)
DATAFUSION_*        Set the given datafusion configuration


stderr:

Kubernetes message

Job has reached the specified backoff limit

File an issue against this benchmark runner

adriangbot · 2026-10-08T20:25:24Z

🤖 Benchmark completed (GKE) | trigger

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

Comparing rich-T-kid/faster-hash-writes (5f5951c) to 409e3b8 (merge-base) diff

Run configuration

run benchmark tpch

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 rich-T-kid_faster-hash-writes
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Query     ┃     HEAD ┃ rich-T-kid_faster-hash-writes ┃    Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ QQuery 1  │ 40.42 ms │                      40.32 ms │ no change │
│ QQuery 2  │ 18.25 ms │                      18.28 ms │ no change │
│ QQuery 3  │ 28.05 ms │                      28.40 ms │ no change │
│ QQuery 4  │ 17.06 ms │                      17.04 ms │ no change │
│ QQuery 5  │ 34.92 ms │                      34.79 ms │ no change │
│ QQuery 6  │ 13.47 ms │                      13.41 ms │ no change │
│ QQuery 7  │ 42.00 ms │                      41.13 ms │ no change │
│ QQuery 8  │ 40.55 ms │                      40.55 ms │ no change │
│ QQuery 9  │ 50.22 ms │                      49.35 ms │ no change │
│ QQuery 10 │ 40.23 ms │                      40.17 ms │ no change │
│ QQuery 11 │ 13.16 ms │                      13.06 ms │ no change │
│ QQuery 12 │ 17.82 ms │                      17.74 ms │ no change │
│ QQuery 13 │ 38.62 ms │                      38.86 ms │ no change │
│ QQuery 14 │ 24.37 ms │                      24.39 ms │ no change │
│ QQuery 15 │ 30.43 ms │                      30.78 ms │ no change │
│ QQuery 16 │ 13.48 ms │                      13.41 ms │ no change │
│ QQuery 17 │ 74.31 ms │                      72.80 ms │ no change │
│ QQuery 18 │ 44.70 ms │                      45.61 ms │ no change │
│ QQuery 19 │ 31.32 ms │                      31.43 ms │ no change │
│ QQuery 20 │ 31.51 ms │                      31.38 ms │ no change │
│ QQuery 21 │ 55.32 ms │                      53.98 ms │ no change │
│ QQuery 22 │ 13.21 ms │                      13.25 ms │ no change │
└───────────┴──────────┴───────────────────────────────┴───────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ Benchmark Summary                            ┃          ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 713.42ms │
│ Total Time (rich-T-kid_faster-hash-writes)   │ 710.12ms │
│ Average Time (HEAD)                          │  32.43ms │
│ Average Time (rich-T-kid_faster-hash-writes) │  32.28ms │
│ Queries Faster                               │        0 │
│ Queries Slower                               │        0 │
│ Queries with No Change                       │       22 │
│ Queries with Failure                         │        0 │
└──────────────────────────────────────────────┴──────────┘

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

Comparing HEAD and rich-T-kid_faster-hash-writes
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Query     ┃                           HEAD ┃  rich-T-kid_faster-hash-writes ┃    Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ QQuery 1  │ 40.42 / 41.12 ±1.04 / 43.16 ms │ 40.32 / 41.23 ±0.89 / 42.57 ms │ no change │
│ QQuery 2  │ 18.25 / 18.49 ±0.17 / 18.78 ms │ 18.28 / 18.80 ±0.48 / 19.56 ms │ no change │
│ QQuery 3  │ 28.05 / 28.32 ±0.18 / 28.61 ms │ 28.40 / 29.01 ±0.84 / 30.66 ms │ no change │
│ QQuery 4  │ 17.06 / 17.91 ±0.67 / 18.91 ms │ 17.04 / 17.82 ±0.62 / 18.71 ms │ no change │
│ QQuery 5  │ 34.92 / 35.19 ±0.14 / 35.32 ms │ 34.79 / 34.97 ±0.11 / 35.13 ms │ no change │
│ QQuery 6  │ 13.47 / 13.58 ±0.08 / 13.67 ms │ 13.41 / 13.74 ±0.29 / 14.27 ms │ no change │
│ QQuery 7  │ 42.00 / 43.08 ±1.62 / 46.28 ms │ 41.13 / 42.10 ±0.71 / 43.17 ms │ no change │
│ QQuery 8  │ 40.55 / 41.20 ±0.69 / 42.48 ms │ 40.55 / 41.04 ±0.59 / 42.17 ms │ no change │
│ QQuery 9  │ 50.22 / 51.40 ±1.10 / 52.77 ms │ 49.35 / 50.32 ±0.88 / 51.80 ms │ no change │
│ QQuery 10 │ 40.23 / 40.41 ±0.14 / 40.62 ms │ 40.17 / 40.91 ±0.78 / 42.42 ms │ no change │
│ QQuery 11 │ 13.16 / 13.61 ±0.53 / 14.60 ms │ 13.06 / 13.51 ±0.45 / 14.32 ms │ no change │
│ QQuery 12 │ 17.82 / 18.11 ±0.30 / 18.65 ms │ 17.74 / 18.14 ±0.31 / 18.62 ms │ no change │
│ QQuery 13 │ 38.62 / 39.47 ±1.10 / 41.57 ms │ 38.86 / 39.81 ±1.11 / 41.63 ms │ no change │
│ QQuery 14 │ 24.37 / 25.09 ±1.02 / 27.10 ms │ 24.39 / 24.55 ±0.14 / 24.81 ms │ no change │
│ QQuery 15 │ 30.43 / 30.71 ±0.23 / 31.11 ms │ 30.78 / 31.17 ±0.50 / 32.14 ms │ no change │
│ QQuery 16 │ 13.48 / 13.72 ±0.23 / 14.14 ms │ 13.41 / 13.64 ±0.13 / 13.79 ms │ no change │
│ QQuery 17 │ 74.31 / 76.00 ±1.61 / 78.13 ms │ 72.80 / 74.12 ±1.16 / 76.28 ms │ no change │
│ QQuery 18 │ 44.70 / 45.48 ±0.62 / 46.43 ms │ 45.61 / 46.50 ±1.41 / 49.30 ms │ no change │
│ QQuery 19 │ 31.32 / 32.64 ±2.00 / 36.50 ms │ 31.43 / 31.76 ±0.39 / 32.53 ms │ no change │
│ QQuery 20 │ 31.51 / 31.84 ±0.24 / 32.24 ms │ 31.38 / 31.90 ±0.29 / 32.20 ms │ no change │
│ QQuery 21 │ 55.32 / 56.02 ±0.47 / 56.56 ms │ 53.98 / 56.13 ±1.39 / 57.97 ms │ no change │
│ QQuery 22 │ 13.21 / 13.81 ±0.69 / 15.13 ms │ 13.25 / 13.48 ±0.20 / 13.75 ms │ no change │
└───────────┴────────────────────────────────┴────────────────────────────────┴───────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ Benchmark Summary                            ┃          ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 727.20ms │
│ Total Time (rich-T-kid_faster-hash-writes)   │ 724.65ms │
│ Average Time (HEAD)                          │  33.05ms │
│ Average Time (rich-T-kid_faster-hash-writes) │  32.94ms │
│ Queries Faster                               │        0 │
│ Queries Slower                               │        0 │
│ Queries with No Change                       │       22 │
│ Queries with Failure                         │        0 │
└──────────────────────────────────────────────┴──────────┘

Resource Usage

tpch — base (merge-base)

Metric	Value
Wall time	5.0s
Peak memory	1.1 GiB
Avg memory	476.3 MiB
CPU user	19.8s
CPU sys	1.7s
Peak spill	0 B

tpch — branch

Metric	Value
Wall time	5.0s
Peak memory	1.2 GiB
Avg memory	490.2 MiB
CPU user	19.9s
CPU sys	1.6s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-10-08T20:38:14Z

🤖 Benchmark completed (GKE) | trigger

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

Comparing rich-T-kid/faster-hash-writes (5f5951c) to 409e3b8 (merge-base) diff

Run configuration

run benchmark with_hashes

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

group                                         HEAD                                   rich-T-kid_faster-hash-writes
-----                                         ----                                   -----------------------------
dense_union: multiple, no nulls               1.05     43.8±0.06µs        ? ?/sec    1.00     41.5±0.07µs        ? ?/sec
dense_union: single, no nulls                 1.06     14.9±0.02µs        ? ?/sec    1.00     14.1±0.02µs        ? ?/sec
dense_union_sliced: 1/10 of 81920 rows        1.04     22.5±0.02µs        ? ?/sec    1.00     21.6±0.02µs        ? ?/sec
dense_union_sliced: 1/2 of 16384 rows         1.04     22.5±0.02µs        ? ?/sec    1.00     21.7±0.03µs        ? ?/sec
dense_union_sliced: 1/5 of 40960 rows         1.05     22.5±0.02µs        ? ?/sec    1.00     21.5±0.02µs        ? ?/sec
dictionary_utf8_int32: multiple, no nulls     1.00     14.8±0.15µs        ? ?/sec    1.00     14.8±0.07µs        ? ?/sec
dictionary_utf8_int32: multiple, nulls        1.00     29.0±0.04µs        ? ?/sec    1.01     29.2±0.06µs        ? ?/sec
dictionary_utf8_int32: single, no nulls       1.00      4.6±0.01µs        ? ?/sec    1.00      4.6±0.01µs        ? ?/sec
dictionary_utf8_int32: single, nulls          1.00      9.4±0.01µs        ? ?/sec    1.00      9.4±0.02µs        ? ?/sec
fixed_size_list_array: multiple, no nulls     1.13    110.2±1.67µs        ? ?/sec    1.00     97.9±0.17µs        ? ?/sec
fixed_size_list_array: multiple, nulls        1.09    119.2±0.25µs        ? ?/sec    1.00    109.4±0.11µs        ? ?/sec
fixed_size_list_array: single, no nulls       1.10     36.4±0.07µs        ? ?/sec    1.00     33.1±0.30µs        ? ?/sec
fixed_size_list_array: single, nulls          1.18     43.3±0.12µs        ? ?/sec    1.00     36.8±0.03µs        ? ?/sec
int64: multiple, no nulls                     1.42     16.9±0.01µs        ? ?/sec    1.00     11.9±0.01µs        ? ?/sec
int64: multiple, nulls                        1.00     26.7±0.02µs        ? ?/sec    1.12     29.9±0.01µs        ? ?/sec
int64: single, no nulls                       1.22      4.6±0.00µs        ? ?/sec    1.00      3.7±0.00µs        ? ?/sec
int64: single, nulls                          1.00      8.7±0.01µs        ? ?/sec    1.10      9.6±0.00µs        ? ?/sec
large_utf8: multiple, no nulls                1.16    116.7±0.52µs        ? ?/sec    1.00    101.1±0.39µs        ? ?/sec
large_utf8: multiple, nulls                   1.04    144.6±0.74µs        ? ?/sec    1.00    138.6±0.66µs        ? ?/sec
large_utf8: single, no nulls                  1.38     39.1±0.32µs        ? ?/sec    1.00     28.3±0.22µs        ? ?/sec
large_utf8: single, nulls                     1.00     34.1±0.12µs        ? ?/sec    1.04     35.6±0.11µs        ? ?/sec
list_array: multiple, no nulls                1.09    140.4±0.09µs        ? ?/sec    1.00    128.6±0.07µs        ? ?/sec
list_array: multiple, nulls                   1.08    168.9±0.84µs        ? ?/sec    1.00    157.0±0.10µs        ? ?/sec
list_array: single, no nulls                  1.09     47.0±0.03µs        ? ?/sec    1.00     43.0±0.03µs        ? ?/sec
list_array: single, nulls                     1.07     56.4±0.05µs        ? ?/sec    1.00     52.6±0.05µs        ? ?/sec
list_array_sliced: 1/10 of 81920 rows         1.11     48.1±1.26µs        ? ?/sec    1.00     43.3±0.73µs        ? ?/sec
list_array_sliced: 1/2 of 16384 rows          1.10     47.3±0.05µs        ? ?/sec    1.00     43.1±0.04µs        ? ?/sec
list_array_sliced: 1/5 of 40960 rows          1.09     47.3±0.03µs        ? ?/sec    1.00     43.4±0.04µs        ? ?/sec
list_view_array: multiple, no nulls           1.05    135.5±0.46µs        ? ?/sec    1.00    129.5±0.36µs        ? ?/sec
list_view_array: multiple, nulls              1.09    143.2±0.23µs        ? ?/sec    1.00    131.8±0.21µs        ? ?/sec
list_view_array: single, no nulls             1.10     45.6±0.42µs        ? ?/sec    1.00     41.5±0.12µs        ? ?/sec
list_view_array: single, nulls                1.10     48.4±0.04µs        ? ?/sec    1.00     44.2±0.06µs        ? ?/sec
map_array: multiple, no nulls                 1.24    241.7±0.21µs        ? ?/sec    1.00    195.6±0.17µs        ? ?/sec
map_array: multiple, nulls                    1.23    248.3±0.18µs        ? ?/sec    1.00    202.6±0.57µs        ? ?/sec
map_array: single, no nulls                   1.23     80.7±0.06µs        ? ?/sec    1.00     65.5±0.03µs        ? ?/sec
map_array: single, nulls                      1.23     83.2±0.11µs        ? ?/sec    1.00     67.8±0.20µs        ? ?/sec
map_array_sliced: 1/10 of 81920 rows          1.22     80.3±0.15µs        ? ?/sec    1.00     65.8±0.04µs        ? ?/sec
map_array_sliced: 1/2 of 16384 rows           1.23     80.8±0.14µs        ? ?/sec    1.00     65.6±0.06µs        ? ?/sec
map_array_sliced: 1/5 of 40960 rows           1.22     80.7±0.08µs        ? ?/sec    1.00     66.0±0.08µs        ? ?/sec
mixed: 3 columns (int64, utf8, utf8_view)     1.24     81.9±0.21µs        ? ?/sec    1.00     66.1±0.12µs        ? ?/sec
null: multiple, no nulls                      1.00      2.5±0.00µs        ? ?/sec    1.00      2.5±0.00µs        ? ?/sec
null: single, no nulls                        1.00   1111.1±2.60ns        ? ?/sec    1.00   1111.9±2.32ns        ? ?/sec
run_array_int32: multiple, no nulls           1.02      5.1±0.02µs        ? ?/sec    1.00      4.9±0.01µs        ? ?/sec
run_array_int32: multiple, nulls              1.00      6.8±0.01µs        ? ?/sec    1.01      6.8±0.01µs        ? ?/sec
run_array_int32: single, no nulls             1.02   1902.0±2.92ns        ? ?/sec    1.00   1870.4±3.46ns        ? ?/sec
run_array_int32: single, nulls                1.00      2.4±0.01µs        ? ?/sec    1.02      2.4±0.00µs        ? ?/sec
sparse_union (2 types): multiple, no nulls    1.03    102.4±0.42µs        ? ?/sec    1.00     99.8±0.35µs        ? ?/sec
sparse_union (2 types): single, no nulls      1.06     35.5±1.55µs        ? ?/sec    1.00     33.5±0.13µs        ? ?/sec
sparse_union (utf8): multiple, no nulls       1.00    238.3±1.40µs        ? ?/sec    1.03    245.4±0.90µs        ? ?/sec
sparse_union (utf8): single, no nulls         1.00     79.8±0.47µs        ? ?/sec    1.03     82.0±0.41µs        ? ?/sec
sparse_union: multiple, no nulls              1.01     98.2±0.62µs        ? ?/sec    1.00     97.2±1.42µs        ? ?/sec
sparse_union: single, no nulls                1.04     33.1±0.21µs        ? ?/sec    1.00     31.8±0.18µs        ? ?/sec
sparse_union_sliced: 1/10 of 81920 rows       1.03     33.6±0.13µs        ? ?/sec    1.00     32.6±0.10µs        ? ?/sec
sparse_union_sliced: 1/2 of 16384 rows        1.05     33.6±0.16µs        ? ?/sec    1.00     32.1±0.15µs        ? ?/sec
sparse_union_sliced: 1/5 of 40960 rows        1.04     33.3±0.22µs        ? ?/sec    1.00     32.2±0.29µs        ? ?/sec
struct_array: multiple, no nulls              1.20    163.7±0.64µs        ? ?/sec    1.00    136.3±0.34µs        ? ?/sec
struct_array: multiple, nulls                 1.17    176.6±0.32µs        ? ?/sec    1.00    150.9±0.08µs        ? ?/sec
struct_array: single, no nulls                1.19     54.1±0.17µs        ? ?/sec    1.00     45.6±0.11µs        ? ?/sec
struct_array: single, nulls                   1.18     59.6±0.41µs        ? ?/sec    1.00     50.6±0.03µs        ? ?/sec
utf8: multiple, no nulls                      1.27    114.9±0.45µs        ? ?/sec    1.00     90.2±0.20µs        ? ?/sec
utf8: multiple, nulls                         1.01    161.2±1.23µs        ? ?/sec    1.00    159.8±0.80µs        ? ?/sec
utf8: single, no nulls                        1.14     29.6±0.18µs        ? ?/sec    1.00     25.9±0.03µs        ? ?/sec
utf8: single, nulls                           1.00     44.0±0.16µs        ? ?/sec    1.08     47.5±0.27µs        ? ?/sec
utf8_view (small): multiple, no nulls         1.35     19.0±0.01µs        ? ?/sec    1.00     14.1±0.00µs        ? ?/sec
utf8_view (small): multiple, nulls            1.01     29.3±0.04µs        ? ?/sec    1.00     29.1±0.05µs        ? ?/sec
utf8_view (small): single, no nulls           1.18      5.1±0.00µs        ? ?/sec    1.00      4.3±0.00µs        ? ?/sec
utf8_view (small): single, nulls              1.00      9.6±0.03µs        ? ?/sec    1.00      9.7±0.01µs        ? ?/sec
utf8_view: multiple, no nulls                 1.05    102.4±0.21µs        ? ?/sec    1.00     97.2±0.28µs        ? ?/sec
utf8_view: multiple, nulls                    1.00    123.7±0.35µs        ? ?/sec    1.00    123.8±0.28µs        ? ?/sec
utf8_view: single, no nulls                   1.17     35.6±0.12µs        ? ?/sec    1.00     30.3±0.06µs        ? ?/sec
utf8_view: single, nulls                      1.00     36.7±0.13µs        ? ?/sec    1.00     36.6±0.08µs        ? ?/sec

Resource Usage

with_hashes — base (merge-base)

Metric	Value
Wall time	770.2s
Peak memory	40.4 MiB
Avg memory	33.8 MiB
CPU user	887.2s
CPU sys	0.8s
Peak spill	0 B

with_hashes — branch

Metric	Value
Wall time	755.2s
Peak memory	45.9 MiB
Avg memory	37.9 MiB
CPU user	880.4s
CPU sys	0.8s
Peak spill	0 B

File an issue against this benchmark runner

Rich-T-kid · 2026-10-09T00:35:43Z

run benchmark with_hashes
env:
BENCH_FILTER: int

Rich-T-kid · 2026-10-09T00:35:56Z

run benchmark with_hashes
env:
BENCH_FILTER: int

adriangbot · 2026-10-09T00:40:37Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highcpu-48 (12 vCPU / 65 GiB) | Linux bench-c6071833939-3220-c2tmm 6.12.94+ #1 SMP Fri Aug 21 08:00:16 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu)

Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  48
On-line CPU(s) list:                     0-47
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     48
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:                               3 MiB (48 instances)
L1i cache:                               3 MiB (48 instances)
L2 cache:                                96 MiB (48 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-47
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 rich-T-kid/faster-hash-writes (b0dff4a) to 409e3b8 (merge-base) diff

Run configuration

run benchmark with_hashes
env:
  BENCH_FILTER: "int"

Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-10-09T00:40:48Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highcpu-48 (12 vCPU / 65 GiB) | Linux bench-c6071831545-3219-m5pvc 6.12.94+ #1 SMP Fri Aug 21 08:00:16 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu)

Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  48
On-line CPU(s) list:                     0-47
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     48
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:                               3 MiB (48 instances)
L1i cache:                               3 MiB (48 instances)
L2 cache:                                96 MiB (48 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-47
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 rich-T-kid/faster-hash-writes (b0dff4a) to 409e3b8 (merge-base) diff

Run configuration

run benchmark with_hashes
env:
  BENCH_FILTER: "int"

Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-10-09T00:46:48Z

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highcpu-48 (12 vCPU / 65 GiB)

Comparing rich-T-kid/faster-hash-writes (b0dff4a) to 409e3b8 (merge-base) diff

Run configuration

run benchmark with_hashes
env:
  BENCH_FILTER: "int"

CPU Details (lscpu)

Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  48
On-line CPU(s) list:                     0-47
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     48
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:                               3 MiB (48 instances)
L1i cache:                               3 MiB (48 instances)
L2 cache:                                96 MiB (48 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-47
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

group                                        HEAD                                   rich-T-kid_faster-hash-writes
-----                                        ----                                   -----------------------------
dictionary_utf8_int32: multiple, no nulls    1.00     14.8±0.01µs        ? ?/sec    1.23     18.2±1.33µs        ? ?/sec
dictionary_utf8_int32: multiple, nulls       1.00     29.1±0.15µs        ? ?/sec    1.00     29.2±0.12µs        ? ?/sec
dictionary_utf8_int32: single, no nulls      1.18      4.6±0.00µs        ? ?/sec    1.00      3.9±0.04µs        ? ?/sec
dictionary_utf8_int32: single, nulls         1.00      9.5±0.01µs        ? ?/sec    1.01      9.6±0.01µs        ? ?/sec
int64: multiple, no nulls                    1.42     16.9±0.01µs        ? ?/sec    1.00     11.9±0.01µs        ? ?/sec
int64: multiple, nulls                       1.72     26.7±0.01µs        ? ?/sec    1.00     15.6±0.02µs        ? ?/sec
int64: single, no nulls                      1.18      4.6±0.00µs        ? ?/sec    1.00      3.9±0.00µs        ? ?/sec
int64: single, nulls                         1.72      8.7±0.01µs        ? ?/sec    1.00      5.1±0.02µs        ? ?/sec
mixed: 3 columns (int64, utf8, utf8_view)    1.00     81.4±0.97µs        ? ?/sec    1.03     83.7±1.40µs        ? ?/sec
run_array_int32: multiple, no nulls          1.03      5.1±0.01µs        ? ?/sec    1.00      4.9±0.01µs        ? ?/sec
run_array_int32: multiple, nulls             1.03      6.8±0.02µs        ? ?/sec    1.00      6.6±0.02µs        ? ?/sec
run_array_int32: single, no nulls            1.04   1900.7±3.48ns        ? ?/sec    1.00   1822.9±4.79ns        ? ?/sec
run_array_int32: single, nulls               1.00      2.4±0.00µs        ? ?/sec    1.01      2.4±0.00µs        ? ?/sec

Resource Usage

with_hashes — base (merge-base)

Metric	Value
Wall time	175.0s
Peak memory	43.8 MiB
Avg memory	26.0 MiB
CPU user	157.1s
CPU sys	0.1s
Peak spill	0 B

with_hashes — branch

Metric	Value
Wall time	185.0s
Peak memory	39.4 MiB
Avg memory	22.3 MiB
CPU user	156.4s
CPU sys	0.2s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-10-09T00:46:49Z

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highcpu-48 (12 vCPU / 65 GiB)

Comparing rich-T-kid/faster-hash-writes (b0dff4a) to 409e3b8 (merge-base) diff

Run configuration

run benchmark with_hashes
env:
  BENCH_FILTER: "int"

CPU Details (lscpu)

Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  48
On-line CPU(s) list:                     0-47
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     48
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:                               3 MiB (48 instances)
L1i cache:                               3 MiB (48 instances)
L2 cache:                                96 MiB (48 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-47
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

group                                        HEAD                                   rich-T-kid_faster-hash-writes
-----                                        ----                                   -----------------------------
dictionary_utf8_int32: multiple, no nulls    1.06     14.8±0.08µs        ? ?/sec    1.00     14.0±0.01µs        ? ?/sec
dictionary_utf8_int32: multiple, nulls       1.01     29.1±0.10µs        ? ?/sec    1.00     28.9±0.02µs        ? ?/sec
dictionary_utf8_int32: single, no nulls      1.20      4.7±0.01µs        ? ?/sec    1.00      3.9±0.01µs        ? ?/sec
dictionary_utf8_int32: single, nulls         1.00      9.4±0.02µs        ? ?/sec    1.02      9.6±0.11µs        ? ?/sec
int64: multiple, no nulls                    1.42     16.9±0.01µs        ? ?/sec    1.00     11.9±0.01µs        ? ?/sec
int64: multiple, nulls                       1.72     26.7±0.06µs        ? ?/sec    1.00     15.6±0.02µs        ? ?/sec
int64: single, no nulls                      1.18      4.6±0.00µs        ? ?/sec    1.00      3.9±0.00µs        ? ?/sec
int64: single, nulls                         1.74      8.7±0.02µs        ? ?/sec    1.00      5.0±0.02µs        ? ?/sec
mixed: 3 columns (int64, utf8, utf8_view)    1.00     81.8±0.34µs        ? ?/sec    1.00     82.0±0.49µs        ? ?/sec
run_array_int32: multiple, no nulls          1.02      5.1±0.01µs        ? ?/sec    1.00      5.0±0.01µs        ? ?/sec
run_array_int32: multiple, nulls             1.03      6.8±0.02µs        ? ?/sec    1.00      6.6±0.01µs        ? ?/sec
run_array_int32: single, no nulls            1.04   1891.6±3.86ns        ? ?/sec    1.00   1821.9±4.20ns        ? ?/sec
run_array_int32: single, nulls               1.00      2.4±0.01µs        ? ?/sec    1.01      2.4±0.00µs        ? ?/sec

Resource Usage

with_hashes — base (merge-base)

Metric	Value
Wall time	170.0s
Peak memory	39.5 MiB
Avg memory	24.1 MiB
CPU user	156.9s
CPU sys	0.1s
Peak spill	0 B

with_hashes — branch

Metric	Value
Wall time	180.0s
Peak memory	39.3 MiB
Avg memory	22.6 MiB
CPU user	155.2s
CPU sys	0.1s
Peak spill	0 B

File an issue against this benchmark runner

Splits the single scalar loop in `hash_array_primitive`'s no-null branch into two const-generic-free helpers that batch independent hashes so the pipeline isn't serialized on store-to-load deps: - `hash_prim_fresh_dense`: overwrite mode, 16 rows per iteration. - `hash_prim_rehash_dense`: fold-with-prev mode, 8 rows per iteration (narrower because each lane holds both `prev` and `value` live). Both use `get_unchecked` / `get_unchecked_mut` with proven invariants (hashes.len() == values.len()) and a scalar tail for the remainder. Benchmarks (int64, 8192 rows, no nulls) on Apple M4 Max: single, no nulls : neutral to small regression (<+7%) multiple, no nulls : ~-14% The nullable path is unchanged -- see PR apache#26143 for the sibling change.

tmp

5f5951c

github-actions Bot added the common Related to common crate label Oct 8, 2026

primtives only

b0dff4a

clean up

926cd62

Rich-T-kid changed the title ~~WIP speed up hashes~~ WIP speed up hashes (Primitive) Oct 9, 2026

Rich-T-kid closed this Oct 10, 2026

Rich-T-kid deleted the rich-T-kid/faster-hash-writes branch October 10, 2026 18:36

Rich-T-kid mentioned this pull request Oct 10, 2026

perf: unroll PrimitiveArray hashing for the no-null path #26191

Draft

Conversation

Rich-T-kid commented Oct 8, 2026

Which issue does this PR close?

Rationale for this change

What changes are included in this PR?

What is the testing strategy for this PR?

Are there any user-facing changes?

Uh oh!

Rich-T-kid commented Oct 8, 2026

Uh oh!

adriangbot commented Oct 8, 2026

Uh oh!

Rich-T-kid commented Oct 8, 2026

Uh oh!

Rich-T-kid commented Oct 8, 2026

Uh oh!

Rich-T-kid commented Oct 8, 2026

Uh oh!

adriangbot commented Oct 8, 2026

Uh oh!

adriangbot commented Oct 8, 2026

Uh oh!

adriangbot commented Oct 8, 2026

Uh oh!

adriangbot commented Oct 8, 2026

Uh oh!

adriangbot commented Oct 8, 2026

Uh oh!

codecov-commenter commented Oct 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

adriangbot commented Oct 8, 2026

Uh oh!

adriangbot commented Oct 8, 2026

Uh oh!

adriangbot commented Oct 8, 2026

Uh oh!

adriangbot commented Oct 8, 2026

Uh oh!

adriangbot commented Oct 8, 2026

Uh oh!

Rich-T-kid commented Oct 9, 2026

Uh oh!

Rich-T-kid commented Oct 9, 2026

Uh oh!

adriangbot commented Oct 9, 2026

Uh oh!

adriangbot commented Oct 9, 2026

Uh oh!

adriangbot commented Oct 9, 2026

Uh oh!

adriangbot commented Oct 9, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

codecov-commenter commented Oct 8, 2026 •

edited

Loading