benchmaxx by Dandandan · Pull Request #21365 · apache/datafusion

Dandandan · 2026-04-04T17:38:34Z

…ash table

Use create_hashes() to compute hashes for the entire batch at once instead of hashing one value at a time
Use two hash tables: a hot map (probed first, up to ~1MB) for cache locality, and a cold map for overflow in high cardinality scenarios
Store values inline in hash table entries as (group_index, T::Native) instead of separate (group_index, hash) table + Vec<T::Native>
Don't store hashes (recomputed on rehash), saving 8 bytes per entry
Build the output array from hash table entries during emit

Which issue does this PR close?

Closes #.

Rationale for this change

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

…ash table - Use create_hashes() to compute hashes for the entire batch at once instead of hashing one value at a time - Use two hash tables: a hot map (probed first, up to ~1MB) for cache locality, and a cold map for overflow in high cardinality scenarios - Store values inline in hash table entries as (group_index, T::Native) instead of separate (group_index, hash) table + Vec<T::Native> - Don't store hashes (recomputed on rehash), saving 8 bytes per entry - Build the output array from hash table entries during emit Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Dandandan · 2026-04-04T17:38:39Z

run benchmarks

adriangbot · 2026-04-04T17:40:39Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4187469437-800-fkrvj 6.12.55+ #1 SMP Sun Feb 1 08:59:41 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 speed-up-group-values-primitive (409f9de) to c17c87c (merge-base) diff using: clickbench_partitioned
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-04T17:40:51Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4187469437-801-bt446 6.12.55+ #1 SMP Sun Feb 1 08:59:41 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 speed-up-group-values-primitive (409f9de) to c17c87c (merge-base) diff using: tpcds
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-04T17:41:21Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4187469437-802-hdprk 6.12.55+ #1 SMP Sun Feb 1 08:59:41 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 speed-up-group-values-primitive (409f9de) to c17c87c (merge-base) diff using: tpch
Results will be posted here when complete

File an issue against this benchmark runner

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

adriangbot · 2026-04-04T17:57:32Z

🤖 Benchmark completed (GKE) | trigger

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

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 speed-up-group-values-primitive
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                     HEAD ┃          speed-up-group-values-primitive ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │              6.54 / 7.01 ±0.79 / 8.59 ms │              6.81 / 7.35 ±0.86 / 9.05 ms │     no change │
│ QQuery 2  │        145.23 / 145.63 ±0.75 / 147.13 ms │        149.37 / 150.33 ±0.79 / 151.46 ms │     no change │
│ QQuery 3  │        113.67 / 114.22 ±0.33 / 114.70 ms │        117.02 / 117.47 ±0.33 / 117.88 ms │     no change │
│ QQuery 4  │    1349.58 / 1365.51 ±11.79 / 1384.63 ms │    1381.61 / 1402.52 ±11.37 / 1415.31 ms │     no change │
│ QQuery 5  │        171.23 / 173.33 ±1.82 / 175.59 ms │        174.26 / 175.72 ±1.52 / 178.35 ms │     no change │
│ QQuery 6  │       831.10 / 861.99 ±17.23 / 879.16 ms │       843.17 / 863.82 ±15.34 / 886.65 ms │     no change │
│ QQuery 7  │        350.75 / 353.12 ±2.26 / 355.85 ms │        352.69 / 355.86 ±2.45 / 358.08 ms │     no change │
│ QQuery 8  │        115.80 / 116.91 ±0.72 / 118.02 ms │        115.97 / 118.30 ±1.55 / 120.61 ms │     no change │
│ QQuery 9  │       100.79 / 108.74 ±10.23 / 128.72 ms │        101.82 / 104.13 ±2.31 / 108.19 ms │     no change │
│ QQuery 10 │        105.68 / 107.08 ±0.93 / 108.50 ms │        108.45 / 109.26 ±0.83 / 110.70 ms │     no change │
│ QQuery 11 │      969.68 / 986.53 ±15.73 / 1006.10 ms │        965.55 / 977.03 ±6.73 / 983.03 ms │     no change │
│ QQuery 12 │           44.73 / 45.93 ±0.67 / 46.73 ms │           45.38 / 47.47 ±1.52 / 50.06 ms │     no change │
│ QQuery 13 │        398.84 / 402.82 ±2.30 / 405.10 ms │        401.36 / 405.08 ±3.20 / 409.63 ms │     no change │
│ QQuery 14 │     1015.18 / 1025.09 ±7.16 / 1034.28 ms │     1021.43 / 1030.22 ±8.62 / 1046.35 ms │     no change │
│ QQuery 15 │           15.89 / 16.96 ±1.01 / 18.46 ms │           16.21 / 17.34 ±0.98 / 18.81 ms │     no change │
│ QQuery 16 │              7.28 / 7.79 ±0.52 / 8.77 ms │             7.28 / 7.97 ±1.09 / 10.11 ms │     no change │
│ QQuery 17 │        240.72 / 242.83 ±1.81 / 245.67 ms │        241.09 / 243.21 ±3.08 / 249.31 ms │     no change │
│ QQuery 18 │        127.63 / 129.95 ±1.51 / 131.85 ms │        129.02 / 129.78 ±0.81 / 130.77 ms │     no change │
│ QQuery 19 │        153.92 / 156.02 ±1.53 / 157.86 ms │        156.22 / 157.98 ±2.38 / 162.44 ms │     no change │
│ QQuery 20 │           13.69 / 14.82 ±0.61 / 15.36 ms │           14.56 / 14.87 ±0.21 / 15.13 ms │     no change │
│ QQuery 21 │           20.14 / 20.57 ±0.36 / 21.00 ms │           19.47 / 19.92 ±0.43 / 20.70 ms │     no change │
│ QQuery 22 │        486.51 / 491.95 ±6.22 / 503.95 ms │        489.04 / 492.78 ±2.99 / 496.31 ms │     no change │
│ QQuery 23 │       900.09 / 909.16 ±10.27 / 923.05 ms │        903.20 / 909.60 ±3.29 / 912.23 ms │     no change │
│ QQuery 24 │        397.46 / 399.87 ±2.90 / 404.81 ms │        396.26 / 399.30 ±3.39 / 405.57 ms │     no change │
│ QQuery 25 │        349.06 / 351.21 ±2.18 / 354.84 ms │        355.50 / 356.09 ±0.73 / 357.50 ms │     no change │
│ QQuery 26 │           82.31 / 83.08 ±0.84 / 84.52 ms │           81.21 / 82.19 ±0.74 / 83.26 ms │     no change │
│ QQuery 27 │              6.85 / 7.05 ±0.15 / 7.27 ms │              6.93 / 7.51 ±0.73 / 8.94 ms │  1.07x slower │
│ QQuery 28 │        148.77 / 149.81 ±0.63 / 150.71 ms │        148.42 / 151.03 ±2.42 / 154.96 ms │     no change │
│ QQuery 29 │        293.23 / 296.48 ±2.61 / 299.61 ms │        294.49 / 297.11 ±1.92 / 299.99 ms │     no change │
│ QQuery 30 │           43.79 / 45.38 ±1.44 / 47.13 ms │           43.18 / 45.18 ±1.36 / 47.24 ms │     no change │
│ QQuery 31 │        170.42 / 173.32 ±2.69 / 176.93 ms │        173.55 / 174.70 ±0.94 / 176.35 ms │     no change │
│ QQuery 32 │           57.01 / 57.69 ±0.38 / 58.06 ms │           57.05 / 58.06 ±1.32 / 60.65 ms │     no change │
│ QQuery 33 │        140.69 / 142.39 ±1.39 / 144.89 ms │        141.46 / 142.28 ±0.81 / 143.79 ms │     no change │
│ QQuery 34 │              7.05 / 7.43 ±0.37 / 8.06 ms │              6.97 / 7.34 ±0.32 / 7.78 ms │     no change │
│ QQuery 35 │        105.92 / 109.37 ±1.99 / 112.09 ms │        108.62 / 109.95 ±1.04 / 111.51 ms │     no change │
│ QQuery 36 │              6.52 / 6.67 ±0.19 / 7.04 ms │              6.61 / 6.87 ±0.23 / 7.17 ms │     no change │
│ QQuery 37 │              8.34 / 8.81 ±0.27 / 9.12 ms │              8.22 / 8.69 ±0.36 / 9.11 ms │     no change │
│ QQuery 38 │           82.72 / 88.31 ±4.70 / 96.68 ms │           85.62 / 87.90 ±2.72 / 93.19 ms │     no change │
│ QQuery 39 │        125.06 / 127.64 ±1.45 / 129.44 ms │        126.95 / 129.94 ±3.51 / 136.55 ms │     no change │
│ QQuery 40 │        113.62 / 119.55 ±5.53 / 129.84 ms │        111.56 / 119.45 ±6.94 / 132.36 ms │     no change │
│ QQuery 41 │           15.27 / 15.80 ±0.31 / 16.24 ms │           14.22 / 15.51 ±0.74 / 16.29 ms │     no change │
│ QQuery 42 │        106.75 / 109.21 ±1.54 / 111.46 ms │        108.54 / 110.35 ±1.39 / 111.81 ms │     no change │
│ QQuery 43 │              5.93 / 6.06 ±0.14 / 6.32 ms │              5.97 / 6.13 ±0.24 / 6.59 ms │     no change │
│ QQuery 44 │           11.78 / 12.01 ±0.23 / 12.44 ms │           11.61 / 11.86 ±0.13 / 11.97 ms │     no change │
│ QQuery 45 │           52.98 / 54.33 ±0.93 / 55.70 ms │           52.81 / 53.92 ±0.94 / 55.17 ms │     no change │
│ QQuery 46 │              8.57 / 9.09 ±0.50 / 9.94 ms │              8.39 / 8.74 ±0.22 / 9.03 ms │     no change │
│ QQuery 47 │       714.99 / 733.30 ±18.23 / 763.31 ms │        692.14 / 702.51 ±6.67 / 711.49 ms │     no change │
│ QQuery 48 │        279.46 / 288.27 ±4.94 / 293.73 ms │        280.71 / 287.39 ±3.44 / 289.95 ms │     no change │
│ QQuery 49 │        253.81 / 256.34 ±1.91 / 259.49 ms │        253.05 / 256.87 ±2.14 / 259.42 ms │     no change │
│ QQuery 50 │        234.39 / 238.58 ±2.59 / 241.90 ms │        229.50 / 232.68 ±2.92 / 237.56 ms │     no change │
│ QQuery 51 │        182.72 / 184.51 ±1.15 / 186.09 ms │        184.51 / 187.50 ±2.68 / 191.69 ms │     no change │
│ QQuery 52 │        107.03 / 108.43 ±1.59 / 111.37 ms │        109.64 / 110.82 ±1.01 / 112.65 ms │     no change │
│ QQuery 53 │        102.26 / 102.87 ±0.56 / 103.71 ms │        103.72 / 104.67 ±0.79 / 105.67 ms │     no change │
│ QQuery 54 │        145.78 / 147.39 ±1.20 / 149.45 ms │        145.62 / 146.83 ±1.01 / 148.53 ms │     no change │
│ QQuery 55 │        107.08 / 107.86 ±0.93 / 109.67 ms │        108.08 / 108.94 ±0.98 / 110.81 ms │     no change │
│ QQuery 56 │        140.00 / 140.95 ±0.74 / 141.94 ms │        140.59 / 142.55 ±1.19 / 144.09 ms │     no change │
│ QQuery 57 │        175.56 / 176.63 ±0.82 / 177.53 ms │        174.30 / 175.89 ±0.86 / 176.92 ms │     no change │
│ QQuery 58 │        288.88 / 298.70 ±6.67 / 306.73 ms │        295.96 / 304.44 ±6.33 / 311.97 ms │     no change │
│ QQuery 59 │        199.06 / 201.47 ±1.61 / 204.14 ms │        200.04 / 201.82 ±1.61 / 204.80 ms │     no change │
│ QQuery 60 │        140.64 / 142.57 ±1.41 / 145.05 ms │        143.11 / 145.33 ±1.34 / 146.93 ms │     no change │
│ QQuery 61 │           12.98 / 13.38 ±0.32 / 13.70 ms │           13.13 / 13.33 ±0.27 / 13.87 ms │     no change │
│ QQuery 62 │       900.66 / 915.03 ±13.65 / 938.34 ms │       904.07 / 928.00 ±18.56 / 956.31 ms │     no change │
│ QQuery 63 │        103.53 / 106.93 ±2.79 / 111.06 ms │        103.68 / 106.53 ±2.27 / 110.16 ms │     no change │
│ QQuery 64 │        698.99 / 704.85 ±3.46 / 709.23 ms │        693.84 / 698.78 ±4.23 / 704.27 ms │     no change │
│ QQuery 65 │        253.07 / 256.82 ±3.04 / 262.38 ms │        245.71 / 253.64 ±5.99 / 262.74 ms │     no change │
│ QQuery 66 │        237.20 / 247.57 ±9.63 / 264.98 ms │        240.65 / 254.40 ±8.86 / 265.93 ms │     no change │
│ QQuery 67 │        312.09 / 320.60 ±7.52 / 334.06 ms │        314.28 / 315.63 ±0.87 / 316.63 ms │     no change │
│ QQuery 68 │             8.68 / 9.50 ±0.70 / 10.48 ms │            9.28 / 10.62 ±0.83 / 11.77 ms │  1.12x slower │
│ QQuery 69 │        102.99 / 104.30 ±0.90 / 105.42 ms │        103.87 / 106.12 ±1.83 / 108.63 ms │     no change │
│ QQuery 70 │       339.78 / 353.95 ±14.02 / 376.93 ms │        344.68 / 349.76 ±4.27 / 355.18 ms │     no change │
│ QQuery 71 │        134.67 / 138.72 ±4.84 / 148.06 ms │        133.65 / 137.73 ±2.95 / 141.89 ms │     no change │
│ QQuery 72 │        712.91 / 724.05 ±7.74 / 734.97 ms │        721.97 / 725.90 ±3.01 / 729.91 ms │     no change │
│ QQuery 73 │              7.21 / 8.06 ±0.58 / 8.78 ms │              6.69 / 7.66 ±0.53 / 8.15 ms │ +1.05x faster │
│ QQuery 74 │        598.55 / 603.80 ±4.75 / 609.60 ms │       595.64 / 609.98 ±10.71 / 627.71 ms │     no change │
│ QQuery 75 │        275.75 / 279.37 ±2.74 / 282.78 ms │        277.12 / 280.14 ±2.24 / 283.88 ms │     no change │
│ QQuery 76 │        130.82 / 133.52 ±1.51 / 135.32 ms │        133.88 / 136.15 ±1.30 / 137.27 ms │     no change │
│ QQuery 77 │        188.47 / 189.79 ±1.45 / 192.51 ms │        190.44 / 193.30 ±1.82 / 195.86 ms │     no change │
│ QQuery 78 │        348.19 / 354.72 ±3.52 / 358.85 ms │        356.05 / 360.37 ±3.83 / 366.75 ms │     no change │
│ QQuery 79 │        239.73 / 241.80 ±1.89 / 245.15 ms │        236.75 / 238.98 ±2.10 / 241.72 ms │     no change │
│ QQuery 80 │        326.61 / 328.72 ±2.84 / 334.19 ms │        329.03 / 332.00 ±1.92 / 334.56 ms │     no change │
│ QQuery 81 │           26.99 / 27.78 ±0.72 / 28.80 ms │           27.24 / 27.79 ±0.52 / 28.62 ms │     no change │
│ QQuery 82 │        198.76 / 201.25 ±2.08 / 203.70 ms │        198.75 / 200.96 ±2.02 / 203.50 ms │     no change │
│ QQuery 83 │           39.75 / 40.74 ±0.94 / 41.99 ms │           39.96 / 40.67 ±0.66 / 41.88 ms │     no change │
│ QQuery 84 │           47.89 / 48.54 ±0.41 / 48.95 ms │           47.79 / 48.76 ±0.93 / 50.36 ms │     no change │
│ QQuery 85 │        147.38 / 150.19 ±1.73 / 152.76 ms │        148.21 / 150.11 ±1.35 / 152.40 ms │     no change │
│ QQuery 86 │           39.69 / 40.52 ±0.96 / 42.36 ms │           39.49 / 40.82 ±1.13 / 42.85 ms │     no change │
│ QQuery 87 │           86.21 / 89.48 ±3.40 / 95.89 ms │           88.17 / 91.59 ±2.87 / 95.77 ms │     no change │
│ QQuery 88 │        100.70 / 102.12 ±0.93 / 103.65 ms │        102.20 / 102.55 ±0.25 / 102.91 ms │     no change │
│ QQuery 89 │        117.91 / 120.06 ±1.50 / 122.52 ms │        118.82 / 121.09 ±1.37 / 122.74 ms │     no change │
│ QQuery 90 │           23.88 / 24.52 ±0.43 / 25.13 ms │           24.59 / 25.03 ±0.30 / 25.53 ms │     no change │
│ QQuery 91 │           60.26 / 64.51 ±2.62 / 67.89 ms │           62.86 / 64.91 ±1.30 / 66.33 ms │     no change │
│ QQuery 92 │           57.31 / 58.13 ±0.63 / 58.83 ms │           59.07 / 59.47 ±0.52 / 60.48 ms │     no change │
│ QQuery 93 │        191.99 / 193.60 ±0.94 / 194.77 ms │        194.06 / 195.45 ±0.90 / 196.48 ms │     no change │
│ QQuery 94 │           60.10 / 60.90 ±0.61 / 61.87 ms │           62.11 / 62.87 ±0.70 / 64.19 ms │     no change │
│ QQuery 95 │        134.04 / 136.06 ±1.63 / 138.43 ms │        136.47 / 137.73 ±1.04 / 139.11 ms │     no change │
│ QQuery 96 │           74.03 / 74.84 ±0.86 / 76.30 ms │           73.42 / 75.58 ±1.37 / 76.99 ms │     no change │
│ QQuery 97 │        128.57 / 132.35 ±2.00 / 134.50 ms │        129.46 / 133.18 ±2.32 / 135.47 ms │     no change │
│ QQuery 98 │        155.24 / 157.57 ±1.29 / 159.19 ms │        152.69 / 155.69 ±2.09 / 158.28 ms │     no change │
│ QQuery 99 │ 10778.15 / 10805.64 ±21.30 / 10839.82 ms │ 10778.75 / 10804.20 ±19.82 / 10824.21 ms │     no change │
└───────────┴──────────────────────────────────────────┴──────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                              ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                              │ 31876.68ms │
│ Total Time (speed-up-group-values-primitive)   │ 31953.82ms │
│ Average Time (HEAD)                            │   321.99ms │
│ Average Time (speed-up-group-values-primitive) │   322.77ms │
│ Queries Faster                                 │          1 │
│ Queries Slower                                 │          2 │
│ Queries with No Change                         │         96 │
│ Queries with Failure                           │          0 │
└────────────────────────────────────────────────┴────────────┘

Resource Usage

tpcds — base (merge-base)

Metric	Value
Wall time	159.7s
Peak memory	5.3 GiB
Avg memory	4.4 GiB
CPU user	261.7s
CPU sys	16.9s
Peak spill	0 B

tpcds — branch

Metric	Value
Wall time	160.1s
Peak memory	5.7 GiB
Avg memory	4.7 GiB
CPU user	261.9s
CPU sys	17.2s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-04-04T17:58:07Z

🤖 Benchmark completed (GKE) | trigger

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

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 speed-up-group-values-primitive
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃       speed-up-group-values-primitive ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │          1.23 / 4.56 ±6.50 / 17.56 ms │          1.20 / 4.50 ±6.43 / 17.37 ms │     no change │
│ QQuery 1  │        14.31 / 14.76 ±0.25 / 15.06 ms │        14.18 / 14.77 ±0.38 / 15.22 ms │     no change │
│ QQuery 2  │        43.82 / 44.25 ±0.33 / 44.80 ms │        43.94 / 44.56 ±0.76 / 46.02 ms │     no change │
│ QQuery 3  │        41.48 / 45.06 ±2.17 / 47.75 ms │        42.69 / 44.42 ±1.04 / 45.44 ms │     no change │
│ QQuery 4  │     302.34 / 305.06 ±2.31 / 309.22 ms │     342.43 / 346.95 ±5.67 / 358.00 ms │  1.14x slower │
│ QQuery 5  │     354.95 / 357.33 ±1.94 / 360.65 ms │     360.69 / 363.44 ±1.75 / 365.77 ms │     no change │
│ QQuery 6  │           5.40 / 6.21 ±0.69 / 7.06 ms │        6.54 / 12.93 ±10.85 / 34.52 ms │  2.08x slower │
│ QQuery 7  │        16.61 / 17.07 ±0.41 / 17.72 ms │        17.36 / 19.06 ±1.11 / 20.68 ms │  1.12x slower │
│ QQuery 8  │     427.56 / 437.09 ±7.06 / 448.37 ms │    439.11 / 461.20 ±21.10 / 499.48 ms │  1.06x slower │
│ QQuery 9  │     657.06 / 663.40 ±4.43 / 670.48 ms │    655.42 / 679.59 ±14.14 / 698.70 ms │     no change │
│ QQuery 10 │        90.46 / 94.41 ±3.17 / 99.17 ms │        94.83 / 96.70 ±1.52 / 99.31 ms │     no change │
│ QQuery 11 │     104.30 / 105.70 ±0.87 / 106.78 ms │     104.80 / 105.50 ±0.51 / 106.21 ms │     no change │
│ QQuery 12 │     351.23 / 354.94 ±3.28 / 360.55 ms │    360.03 / 371.37 ±11.33 / 391.04 ms │     no change │
│ QQuery 13 │     470.46 / 480.46 ±8.43 / 491.81 ms │     485.91 / 492.28 ±6.00 / 503.31 ms │     no change │
│ QQuery 14 │     357.89 / 359.94 ±1.82 / 362.19 ms │     360.13 / 363.92 ±2.23 / 366.45 ms │     no change │
│ QQuery 15 │    387.04 / 407.89 ±22.10 / 435.72 ms │    414.17 / 440.28 ±30.88 / 500.46 ms │  1.08x slower │
│ QQuery 16 │    727.17 / 742.21 ±18.17 / 777.42 ms │     741.38 / 747.43 ±5.59 / 757.60 ms │     no change │
│ QQuery 17 │     723.32 / 736.38 ±6.98 / 743.46 ms │    723.18 / 739.70 ±10.83 / 755.82 ms │     no change │
│ QQuery 18 │ 1437.72 / 1493.98 ±33.35 / 1531.07 ms │ 1444.33 / 1515.05 ±48.40 / 1593.09 ms │     no change │
│ QQuery 19 │       35.39 / 43.52 ±15.45 / 74.42 ms │        35.85 / 38.87 ±4.38 / 47.59 ms │ +1.12x faster │
│ QQuery 20 │    730.59 / 748.62 ±18.58 / 771.63 ms │    721.09 / 739.49 ±22.14 / 779.87 ms │     no change │
│ QQuery 21 │    771.04 / 787.33 ±12.53 / 802.03 ms │     768.24 / 773.49 ±5.16 / 782.55 ms │     no change │
│ QQuery 22 │  1142.31 / 1146.81 ±3.99 / 1151.31 ms │  1137.28 / 1141.40 ±4.77 / 1150.06 ms │     no change │
│ QQuery 23 │ 3112.87 / 3133.39 ±18.57 / 3161.94 ms │ 3123.47 / 3146.16 ±19.52 / 3169.97 ms │     no change │
│ QQuery 24 │     100.58 / 103.70 ±2.56 / 106.49 ms │     100.51 / 103.62 ±3.10 / 109.13 ms │     no change │
│ QQuery 25 │     139.46 / 143.72 ±2.28 / 145.69 ms │     141.70 / 144.33 ±1.71 / 146.89 ms │     no change │
│ QQuery 26 │      99.49 / 102.14 ±2.31 / 105.59 ms │     101.38 / 103.99 ±2.00 / 106.67 ms │     no change │
│ QQuery 27 │     857.21 / 863.16 ±5.69 / 873.64 ms │     855.86 / 860.02 ±4.24 / 867.55 ms │     no change │
│ QQuery 28 │ 7742.13 / 7797.94 ±30.72 / 7835.00 ms │ 7716.85 / 7784.58 ±37.87 / 7825.03 ms │     no change │
│ QQuery 29 │        50.40 / 53.91 ±3.83 / 58.78 ms │        50.39 / 55.50 ±6.97 / 68.61 ms │     no change │
│ QQuery 30 │     373.01 / 381.84 ±6.00 / 388.23 ms │     370.11 / 375.52 ±4.21 / 381.69 ms │     no change │
│ QQuery 31 │    371.67 / 393.16 ±14.29 / 411.32 ms │    366.74 / 389.01 ±12.57 / 401.13 ms │     no change │
│ QQuery 32 │ 1208.90 / 1291.15 ±49.62 / 1352.08 ms │ 1174.43 / 1308.88 ±78.25 / 1392.66 ms │     no change │
│ QQuery 33 │ 1559.36 / 1625.70 ±60.97 / 1720.32 ms │ 1605.79 / 1626.04 ±15.58 / 1648.10 ms │     no change │
│ QQuery 34 │ 1604.07 / 1643.11 ±47.19 / 1720.30 ms │ 1513.03 / 1558.68 ±35.89 / 1603.21 ms │ +1.05x faster │
│ QQuery 35 │    426.50 / 490.15 ±94.09 / 674.27 ms │     382.67 / 387.47 ±3.46 / 391.76 ms │ +1.27x faster │
│ QQuery 36 │    132.62 / 148.10 ±27.94 / 203.96 ms │     113.14 / 119.98 ±4.89 / 125.85 ms │ +1.23x faster │
│ QQuery 37 │        50.22 / 51.31 ±0.70 / 52.37 ms │        48.64 / 49.77 ±0.87 / 51.24 ms │     no change │
│ QQuery 38 │        81.91 / 84.94 ±2.49 / 88.54 ms │        75.58 / 76.53 ±0.68 / 77.38 ms │ +1.11x faster │
│ QQuery 39 │     239.83 / 244.71 ±3.35 / 248.80 ms │     209.86 / 220.04 ±7.88 / 231.96 ms │ +1.11x faster │
│ QQuery 40 │        23.22 / 25.53 ±1.79 / 28.33 ms │        23.91 / 26.28 ±1.98 / 29.46 ms │     no change │
│ QQuery 41 │        21.34 / 21.96 ±0.44 / 22.52 ms │        20.68 / 21.71 ±0.85 / 22.91 ms │     no change │
│ QQuery 42 │        20.01 / 20.40 ±0.34 / 20.95 ms │        19.80 / 20.56 ±0.69 / 21.70 ms │     no change │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                              ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                              │ 28017.03ms │
│ Total Time (speed-up-group-values-primitive)   │ 27935.55ms │
│ Average Time (HEAD)                            │   651.56ms │
│ Average Time (speed-up-group-values-primitive) │   649.66ms │
│ Queries Faster                                 │          6 │
│ Queries Slower                                 │          5 │
│ Queries with No Change                         │         32 │
│ Queries with Failure                           │          0 │
└────────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric	Value
Wall time	140.6s
Peak memory	36.2 GiB
Avg memory	26.9 GiB
CPU user	1306.0s
CPU sys	106.8s
Peak spill	0 B

clickbench_partitioned — branch

Metric	Value
Wall time	140.9s
Peak memory	40.7 GiB
Avg memory	29.8 GiB
CPU user	1307.0s
CPU sys	108.9s
Peak spill	0 B

File an issue against this benchmark runner

Remove hot/cold two-level map approach in favor of a single map. Keep the other improvements: vectorized batch hashing via create_hashes, inline (group_index, value) entries, and no stored hashes. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Dandandan · 2026-04-04T18:04:55Z

run benchmarks

adriangbot · 2026-04-04T18:06:54Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4187508210-805-cghpp 6.12.55+ #1 SMP Sun Feb 1 08:59:41 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 speed-up-group-values-primitive (f70ea0a) to c17c87c (merge-base) diff using: tpch
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-04T18:06:56Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4187508210-803-28x5l 6.12.55+ #1 SMP Sun Feb 1 08:59:41 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 speed-up-group-values-primitive (f70ea0a) to c17c87c (merge-base) diff using: clickbench_partitioned
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-04T18:07:38Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4187508210-804-t6pwj 6.12.55+ #1 SMP Sun Feb 1 08:59:41 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 speed-up-group-values-primitive (f70ea0a) to c17c87c (merge-base) diff using: tpcds
Results will be posted here when complete

File an issue against this benchmark runner

Dandandan · 2026-04-04T18:15:11Z

run benchmarks

adriangbot · 2026-04-04T18:16:47Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4187523474-808-n9p95 6.12.55+ #1 SMP Sun Feb 1 08:59:41 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 speed-up-group-values-primitive (a8fde95) to c17c87c (merge-base) diff using: tpch
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-04T18:17:23Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4187523474-807-4n6pf 6.12.55+ #1 SMP Sun Feb 1 08:59:41 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 speed-up-group-values-primitive (a8fde95) to c17c87c (merge-base) diff using: tpcds
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-04T18:17:24Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4187523474-806-k4vp5 6.12.55+ #1 SMP Sun Feb 1 08:59:41 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 speed-up-group-values-primitive (a8fde95) to c17c87c (merge-base) diff using: clickbench_partitioned
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-04T18:23:53Z

🤖 Benchmark completed (GKE) | trigger

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

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 speed-up-group-values-primitive
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                     HEAD ┃          speed-up-group-values-primitive ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │              6.85 / 7.33 ±0.78 / 8.87 ms │              6.76 / 7.23 ±0.80 / 8.83 ms │     no change │
│ QQuery 2  │        147.66 / 148.73 ±0.63 / 149.36 ms │        148.78 / 149.10 ±0.41 / 149.89 ms │     no change │
│ QQuery 3  │        114.07 / 114.52 ±0.31 / 115.00 ms │        115.70 / 116.39 ±0.58 / 117.21 ms │     no change │
│ QQuery 4  │    1354.57 / 1403.43 ±25.18 / 1424.51 ms │    1392.68 / 1406.25 ±13.94 / 1432.60 ms │     no change │
│ QQuery 5  │        171.25 / 173.43 ±1.33 / 175.29 ms │        174.43 / 175.58 ±1.03 / 177.22 ms │     no change │
│ QQuery 6  │       864.73 / 883.71 ±14.70 / 899.54 ms │       841.29 / 882.48 ±38.68 / 939.35 ms │     no change │
│ QQuery 7  │        352.07 / 358.75 ±3.54 / 362.69 ms │        355.76 / 357.89 ±2.30 / 362.33 ms │     no change │
│ QQuery 8  │        115.88 / 117.41 ±1.07 / 119.21 ms │        117.86 / 118.35 ±0.32 / 118.84 ms │     no change │
│ QQuery 9  │        100.67 / 104.79 ±3.01 / 107.87 ms │       102.13 / 110.74 ±10.03 / 128.60 ms │  1.06x slower │
│ QQuery 10 │        106.33 / 107.58 ±0.71 / 108.53 ms │        107.80 / 108.97 ±0.71 / 109.89 ms │     no change │
│ QQuery 11 │    1006.52 / 1013.72 ±10.47 / 1034.55 ms │      982.04 / 993.35 ±12.03 / 1014.87 ms │     no change │
│ QQuery 12 │           46.14 / 46.97 ±0.76 / 47.96 ms │           46.99 / 48.25 ±0.72 / 49.14 ms │     no change │
│ QQuery 13 │        402.12 / 406.82 ±2.84 / 411.00 ms │        402.07 / 405.84 ±2.84 / 408.52 ms │     no change │
│ QQuery 14 │     1023.76 / 1032.72 ±5.75 / 1040.28 ms │     1024.29 / 1032.30 ±4.94 / 1039.87 ms │     no change │
│ QQuery 15 │           16.44 / 18.15 ±1.90 / 21.61 ms │           16.00 / 17.33 ±1.05 / 19.12 ms │     no change │
│ QQuery 16 │              7.37 / 7.60 ±0.18 / 7.83 ms │              7.63 / 8.80 ±0.88 / 9.72 ms │  1.16x slower │
│ QQuery 17 │        243.18 / 244.95 ±1.83 / 248.23 ms │        239.62 / 243.41 ±2.54 / 247.38 ms │     no change │
│ QQuery 18 │        128.02 / 130.67 ±1.48 / 132.51 ms │        129.22 / 130.88 ±1.54 / 133.76 ms │     no change │
│ QQuery 19 │        156.94 / 158.88 ±1.06 / 160.16 ms │        158.57 / 160.04 ±1.21 / 161.82 ms │     no change │
│ QQuery 20 │           13.95 / 14.53 ±0.34 / 14.89 ms │           13.69 / 14.60 ±0.57 / 15.45 ms │     no change │
│ QQuery 21 │           19.75 / 20.32 ±0.32 / 20.66 ms │           20.23 / 21.10 ±0.64 / 22.03 ms │     no change │
│ QQuery 22 │        485.73 / 493.95 ±5.10 / 499.50 ms │        493.73 / 498.50 ±3.32 / 502.35 ms │     no change │
│ QQuery 23 │        908.37 / 917.50 ±7.76 / 929.22 ms │        939.30 / 946.93 ±5.63 / 953.81 ms │     no change │
│ QQuery 24 │        399.57 / 401.75 ±1.58 / 404.02 ms │        407.49 / 409.83 ±2.27 / 413.82 ms │     no change │
│ QQuery 25 │        355.96 / 358.61 ±2.03 / 362.00 ms │        356.69 / 359.86 ±2.99 / 364.65 ms │     no change │
│ QQuery 26 │           80.46 / 83.12 ±2.29 / 85.83 ms │           81.83 / 84.13 ±1.60 / 86.42 ms │     no change │
│ QQuery 27 │              7.14 / 8.23 ±0.97 / 9.68 ms │              7.29 / 7.55 ±0.24 / 7.87 ms │ +1.09x faster │
│ QQuery 28 │        151.30 / 152.06 ±0.68 / 152.92 ms │        149.40 / 151.05 ±1.40 / 152.91 ms │     no change │
│ QQuery 29 │        301.52 / 304.36 ±2.15 / 307.89 ms │        298.84 / 301.68 ±2.13 / 304.66 ms │     no change │
│ QQuery 30 │           42.96 / 45.25 ±1.67 / 48.15 ms │           46.81 / 48.56 ±1.64 / 50.92 ms │  1.07x slower │
│ QQuery 31 │        171.70 / 174.34 ±1.92 / 177.08 ms │        174.47 / 176.60 ±1.32 / 178.37 ms │     no change │
│ QQuery 32 │           57.35 / 58.88 ±1.20 / 60.72 ms │           57.08 / 58.40 ±0.82 / 59.13 ms │     no change │
│ QQuery 33 │        141.30 / 142.96 ±1.02 / 143.91 ms │        141.35 / 143.53 ±1.28 / 145.30 ms │     no change │
│ QQuery 34 │              7.28 / 7.44 ±0.22 / 7.85 ms │              7.20 / 7.59 ±0.33 / 8.06 ms │     no change │
│ QQuery 35 │        109.20 / 110.80 ±1.42 / 113.42 ms │        109.62 / 111.32 ±1.30 / 113.59 ms │     no change │
│ QQuery 36 │              6.75 / 6.92 ±0.12 / 7.09 ms │              6.63 / 6.80 ±0.14 / 6.98 ms │     no change │
│ QQuery 37 │              8.80 / 9.00 ±0.21 / 9.26 ms │             8.80 / 9.47 ±0.68 / 10.46 ms │  1.05x slower │
│ QQuery 38 │           83.64 / 87.75 ±3.31 / 93.55 ms │           87.78 / 90.61 ±3.10 / 96.32 ms │     no change │
│ QQuery 39 │        127.86 / 131.86 ±2.11 / 133.81 ms │        124.42 / 127.14 ±1.76 / 129.76 ms │     no change │
│ QQuery 40 │        116.57 / 122.17 ±9.02 / 140.17 ms │        113.15 / 120.63 ±5.16 / 129.13 ms │     no change │
│ QQuery 41 │           14.75 / 15.54 ±0.79 / 16.95 ms │           14.50 / 15.70 ±1.28 / 18.15 ms │     no change │
│ QQuery 42 │        107.52 / 110.27 ±1.61 / 112.22 ms │        109.06 / 110.47 ±1.13 / 112.33 ms │     no change │
│ QQuery 43 │              6.14 / 6.29 ±0.17 / 6.62 ms │              6.02 / 6.13 ±0.12 / 6.35 ms │     no change │
│ QQuery 44 │           12.03 / 12.47 ±0.49 / 13.26 ms │           11.88 / 12.38 ±0.51 / 13.32 ms │     no change │
│ QQuery 45 │           53.05 / 55.11 ±1.40 / 57.40 ms │           53.38 / 54.95 ±1.00 / 56.05 ms │     no change │
│ QQuery 46 │              8.98 / 9.53 ±0.38 / 9.98 ms │              8.81 / 9.04 ±0.12 / 9.15 ms │ +1.05x faster │
│ QQuery 47 │       748.49 / 763.14 ±10.01 / 776.38 ms │        750.88 / 760.06 ±8.79 / 770.96 ms │     no change │
│ QQuery 48 │        285.61 / 291.00 ±3.29 / 295.98 ms │        284.89 / 294.41 ±5.54 / 299.42 ms │     no change │
│ QQuery 49 │        255.94 / 258.60 ±1.59 / 260.16 ms │        259.21 / 260.70 ±1.63 / 263.34 ms │     no change │
│ QQuery 50 │        233.67 / 239.29 ±3.17 / 242.91 ms │        239.78 / 244.11 ±3.93 / 249.33 ms │     no change │
│ QQuery 51 │        185.24 / 189.69 ±3.14 / 193.47 ms │        186.90 / 191.87 ±5.24 / 201.41 ms │     no change │
│ QQuery 52 │        107.10 / 109.61 ±1.27 / 110.45 ms │        109.43 / 110.81 ±0.89 / 111.93 ms │     no change │
│ QQuery 53 │        103.06 / 104.30 ±1.35 / 106.80 ms │        104.33 / 105.26 ±0.71 / 106.42 ms │     no change │
│ QQuery 54 │        145.87 / 147.45 ±1.44 / 149.88 ms │        149.14 / 151.56 ±2.07 / 155.38 ms │     no change │
│ QQuery 55 │        106.45 / 107.24 ±0.55 / 107.82 ms │        108.15 / 110.04 ±1.77 / 113.14 ms │     no change │
│ QQuery 56 │        140.27 / 141.86 ±0.85 / 142.64 ms │        143.12 / 144.37 ±0.99 / 146.14 ms │     no change │
│ QQuery 57 │        178.55 / 179.01 ±0.46 / 179.74 ms │        178.18 / 179.38 ±0.90 / 180.74 ms │     no change │
│ QQuery 58 │        295.83 / 302.34 ±4.16 / 307.21 ms │        295.79 / 305.30 ±5.88 / 311.86 ms │     no change │
│ QQuery 59 │        199.06 / 200.54 ±1.31 / 202.32 ms │        201.57 / 202.48 ±0.62 / 203.44 ms │     no change │
│ QQuery 60 │        142.87 / 144.89 ±1.07 / 145.75 ms │        145.02 / 147.00 ±1.49 / 149.10 ms │     no change │
│ QQuery 61 │           13.37 / 13.79 ±0.26 / 14.14 ms │           13.39 / 13.65 ±0.24 / 14.11 ms │     no change │
│ QQuery 62 │      939.98 / 991.95 ±44.77 / 1064.74 ms │     926.99 / 1053.96 ±75.85 / 1150.67 ms │  1.06x slower │
│ QQuery 63 │        104.15 / 106.70 ±1.99 / 110.18 ms │        105.53 / 109.85 ±4.38 / 117.81 ms │     no change │
│ QQuery 64 │        701.58 / 710.22 ±5.88 / 718.30 ms │        728.52 / 733.82 ±2.78 / 736.21 ms │     no change │
│ QQuery 65 │        259.12 / 263.31 ±4.29 / 271.20 ms │        269.14 / 274.34 ±4.16 / 279.42 ms │     no change │
│ QQuery 66 │        250.65 / 259.10 ±7.74 / 271.76 ms │        249.81 / 259.44 ±9.50 / 273.08 ms │     no change │
│ QQuery 67 │        317.61 / 327.32 ±5.40 / 332.72 ms │        320.15 / 328.99 ±6.72 / 338.06 ms │     no change │
│ QQuery 68 │            8.84 / 11.46 ±3.30 / 17.78 ms │           10.19 / 10.74 ±0.49 / 11.41 ms │ +1.07x faster │
│ QQuery 69 │        103.46 / 105.46 ±1.40 / 107.55 ms │        103.66 / 104.13 ±0.44 / 104.82 ms │     no change │
│ QQuery 70 │       332.82 / 352.28 ±14.44 / 376.86 ms │        342.16 / 351.11 ±8.75 / 366.09 ms │     no change │
│ QQuery 71 │        135.26 / 137.97 ±1.58 / 139.71 ms │        135.51 / 139.26 ±3.02 / 143.39 ms │     no change │
│ QQuery 72 │       721.04 / 736.18 ±11.22 / 752.01 ms │        728.20 / 735.95 ±6.30 / 745.87 ms │     no change │
│ QQuery 73 │             7.49 / 9.26 ±1.38 / 10.92 ms │              6.79 / 7.56 ±0.77 / 8.85 ms │ +1.22x faster │
│ QQuery 74 │        622.17 / 635.00 ±8.04 / 645.05 ms │        598.03 / 609.03 ±7.86 / 619.99 ms │     no change │
│ QQuery 75 │        277.24 / 281.54 ±2.84 / 285.12 ms │        279.66 / 282.16 ±2.13 / 285.43 ms │     no change │
│ QQuery 76 │        132.32 / 134.44 ±1.55 / 136.41 ms │        133.67 / 134.91 ±0.82 / 136.09 ms │     no change │
│ QQuery 77 │        188.87 / 190.08 ±0.79 / 191.27 ms │        188.86 / 190.62 ±1.45 / 192.99 ms │     no change │
│ QQuery 78 │        359.52 / 362.77 ±1.76 / 364.89 ms │        358.87 / 363.10 ±3.15 / 367.52 ms │     no change │
│ QQuery 79 │        236.27 / 239.65 ±2.81 / 244.49 ms │        241.25 / 247.13 ±3.37 / 250.04 ms │     no change │
│ QQuery 80 │        331.05 / 333.73 ±1.80 / 336.68 ms │        334.29 / 336.20 ±1.95 / 339.84 ms │     no change │
│ QQuery 81 │           26.71 / 27.98 ±1.21 / 30.21 ms │           27.27 / 28.47 ±1.51 / 31.35 ms │     no change │
│ QQuery 82 │        203.07 / 205.64 ±2.06 / 208.16 ms │        201.13 / 202.39 ±0.71 / 203.23 ms │     no change │
│ QQuery 83 │           41.41 / 43.10 ±1.05 / 44.01 ms │           40.14 / 40.44 ±0.25 / 40.87 ms │ +1.07x faster │
│ QQuery 84 │           49.87 / 50.75 ±0.77 / 51.60 ms │           48.18 / 49.46 ±1.13 / 51.36 ms │     no change │
│ QQuery 85 │        150.51 / 153.19 ±2.25 / 157.28 ms │        150.34 / 151.35 ±1.35 / 153.89 ms │     no change │
│ QQuery 86 │           40.23 / 40.88 ±0.50 / 41.53 ms │           40.39 / 40.85 ±0.39 / 41.39 ms │     no change │
│ QQuery 87 │           89.91 / 93.21 ±3.10 / 98.34 ms │           88.37 / 90.68 ±2.44 / 94.95 ms │     no change │
│ QQuery 88 │        102.66 / 103.48 ±0.84 / 105.09 ms │        101.92 / 102.53 ±0.71 / 103.90 ms │     no change │
│ QQuery 89 │        119.92 / 123.38 ±2.23 / 126.90 ms │        118.21 / 119.86 ±0.92 / 120.77 ms │     no change │
│ QQuery 90 │           23.71 / 24.79 ±1.20 / 27.12 ms │           24.19 / 25.39 ±0.85 / 26.50 ms │     no change │
│ QQuery 91 │           62.65 / 65.12 ±1.33 / 66.55 ms │           63.83 / 66.04 ±1.49 / 67.58 ms │     no change │
│ QQuery 92 │           59.13 / 59.36 ±0.18 / 59.59 ms │           57.89 / 59.80 ±1.29 / 61.51 ms │     no change │
│ QQuery 93 │        200.66 / 205.11 ±4.19 / 211.52 ms │        193.40 / 198.02 ±3.01 / 202.76 ms │     no change │
│ QQuery 94 │           61.85 / 63.49 ±1.34 / 65.48 ms │           62.22 / 62.49 ±0.28 / 62.90 ms │     no change │
│ QQuery 95 │        137.41 / 139.24 ±1.21 / 140.87 ms │        135.88 / 138.71 ±1.80 / 140.95 ms │     no change │
│ QQuery 96 │           75.40 / 78.59 ±2.20 / 81.74 ms │           76.07 / 77.12 ±1.08 / 79.21 ms │     no change │
│ QQuery 97 │        136.61 / 138.94 ±2.70 / 144.20 ms │        132.76 / 134.05 ±1.15 / 135.52 ms │     no change │
│ QQuery 98 │        157.59 / 162.54 ±2.71 / 165.55 ms │        155.71 / 158.55 ±1.86 / 161.16 ms │     no change │
│ QQuery 99 │ 10821.91 / 10851.04 ±24.97 / 10888.27 ms │ 10820.25 / 10853.45 ±28.30 / 10903.92 ms │     no change │
└───────────┴──────────────────────────────────────────┴──────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                              ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                              │ 32352.15ms │
│ Total Time (speed-up-group-values-primitive)   │ 32470.69ms │
│ Average Time (HEAD)                            │   326.79ms │
│ Average Time (speed-up-group-values-primitive) │   327.99ms │
│ Queries Faster                                 │          5 │
│ Queries Slower                                 │          5 │
│ Queries with No Change                         │         89 │
│ Queries with Failure                           │          0 │
└────────────────────────────────────────────────┴────────────┘

Resource Usage

tpcds — base (merge-base)

Metric	Value
Wall time	162.1s
Peak memory	5.5 GiB
Avg memory	4.7 GiB
CPU user	265.0s
CPU sys	17.8s
Peak spill	0 B

tpcds — branch

Metric	Value
Wall time	162.6s
Peak memory	5.5 GiB
Avg memory	4.4 GiB
CPU user	265.4s
CPU sys	18.6s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-04-04T18:24:36Z

🤖 Benchmark completed (GKE) | trigger

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

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 speed-up-group-values-primitive
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃       speed-up-group-values-primitive ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │          1.19 / 4.51 ±6.48 / 17.46 ms │          1.19 / 4.49 ±6.44 / 17.36 ms │     no change │
│ QQuery 1  │        14.17 / 14.78 ±0.33 / 15.09 ms │        14.08 / 14.54 ±0.25 / 14.84 ms │     no change │
│ QQuery 2  │        44.72 / 45.68 ±1.10 / 47.79 ms │        43.64 / 43.88 ±0.21 / 44.25 ms │     no change │
│ QQuery 3  │        44.04 / 45.27 ±0.72 / 45.98 ms │        40.19 / 44.55 ±2.33 / 46.66 ms │     no change │
│ QQuery 4  │     293.83 / 300.38 ±5.91 / 309.14 ms │     288.73 / 301.84 ±9.80 / 312.79 ms │     no change │
│ QQuery 5  │     344.95 / 347.69 ±1.93 / 350.27 ms │     351.07 / 356.32 ±5.41 / 365.14 ms │     no change │
│ QQuery 6  │           5.14 / 5.63 ±0.42 / 6.37 ms │           5.67 / 7.11 ±0.88 / 8.18 ms │  1.26x slower │
│ QQuery 7  │        17.13 / 19.80 ±3.41 / 25.96 ms │        16.71 / 16.92 ±0.30 / 17.51 ms │ +1.17x faster │
│ QQuery 8  │     427.03 / 437.99 ±5.57 / 442.64 ms │     430.22 / 440.07 ±5.44 / 446.96 ms │     no change │
│ QQuery 9  │     643.97 / 654.47 ±6.55 / 663.16 ms │    653.69 / 670.43 ±10.11 / 684.54 ms │     no change │
│ QQuery 10 │        93.68 / 95.88 ±1.59 / 98.41 ms │       92.39 / 95.52 ±2.91 / 101.01 ms │     no change │
│ QQuery 11 │     106.87 / 107.96 ±1.56 / 111.06 ms │     104.08 / 105.80 ±1.62 / 108.53 ms │     no change │
│ QQuery 12 │     350.03 / 359.54 ±8.55 / 374.63 ms │     353.17 / 358.14 ±3.67 / 363.69 ms │     no change │
│ QQuery 13 │    463.21 / 482.11 ±11.19 / 493.04 ms │     471.84 / 476.26 ±7.39 / 491.00 ms │     no change │
│ QQuery 14 │     359.14 / 365.59 ±6.84 / 378.68 ms │     355.03 / 360.24 ±3.88 / 364.72 ms │     no change │
│ QQuery 15 │    356.38 / 374.02 ±17.12 / 404.69 ms │    356.05 / 400.46 ±35.05 / 433.39 ms │  1.07x slower │
│ QQuery 16 │    730.18 / 746.09 ±12.24 / 762.66 ms │    732.96 / 751.75 ±16.74 / 782.11 ms │     no change │
│ QQuery 17 │     723.59 / 730.61 ±4.18 / 735.32 ms │    748.46 / 771.80 ±26.80 / 819.26 ms │  1.06x slower │
│ QQuery 18 │ 1397.83 / 1483.87 ±52.08 / 1544.82 ms │ 1462.08 / 1503.08 ±36.53 / 1555.16 ms │     no change │
│ QQuery 19 │        34.33 / 35.48 ±0.86 / 36.74 ms │      36.52 / 52.00 ±29.89 / 111.78 ms │  1.47x slower │
│ QQuery 20 │    714.92 / 736.49 ±17.72 / 763.75 ms │    723.45 / 744.57 ±21.28 / 782.43 ms │     no change │
│ QQuery 21 │     766.51 / 767.62 ±0.76 / 768.79 ms │     764.67 / 772.26 ±4.61 / 777.69 ms │     no change │
│ QQuery 22 │  1130.58 / 1139.69 ±6.29 / 1148.77 ms │  1129.77 / 1134.55 ±3.85 / 1139.10 ms │     no change │
│ QQuery 23 │ 3071.72 / 3099.17 ±18.32 / 3125.06 ms │ 3105.36 / 3148.57 ±28.67 / 3186.43 ms │     no change │
│ QQuery 24 │     101.21 / 105.54 ±2.72 / 109.35 ms │     102.82 / 105.71 ±1.77 / 107.84 ms │     no change │
│ QQuery 25 │     139.59 / 141.84 ±1.59 / 143.49 ms │     142.38 / 143.20 ±0.77 / 144.60 ms │     no change │
│ QQuery 26 │     102.64 / 105.17 ±2.12 / 109.06 ms │     103.92 / 105.47 ±1.41 / 107.81 ms │     no change │
│ QQuery 27 │     858.74 / 863.76 ±9.08 / 881.91 ms │     858.53 / 864.95 ±4.89 / 872.85 ms │     no change │
│ QQuery 28 │ 7742.04 / 7799.73 ±32.09 / 7830.60 ms │ 7749.96 / 7805.49 ±31.59 / 7835.17 ms │     no change │
│ QQuery 29 │        51.08 / 55.49 ±3.60 / 60.99 ms │        50.47 / 54.81 ±5.33 / 64.12 ms │     no change │
│ QQuery 30 │     369.26 / 374.06 ±4.40 / 381.37 ms │     369.28 / 374.68 ±5.68 / 385.01 ms │     no change │
│ QQuery 31 │    365.59 / 387.04 ±11.15 / 397.90 ms │     385.09 / 390.41 ±4.79 / 397.95 ms │     no change │
│ QQuery 32 │ 1040.82 / 1073.30 ±40.24 / 1146.08 ms │ 1215.32 / 1263.57 ±47.37 / 1352.30 ms │  1.18x slower │
│ QQuery 33 │  1471.57 / 1485.40 ±9.11 / 1497.00 ms │ 1554.57 / 1603.09 ±35.17 / 1645.02 ms │  1.08x slower │
│ QQuery 34 │  1485.29 / 1490.72 ±4.51 / 1498.67 ms │ 1535.00 / 1557.95 ±19.94 / 1590.58 ms │     no change │
│ QQuery 35 │    396.12 / 409.65 ±11.25 / 429.60 ms │     330.48 / 333.21 ±2.81 / 338.55 ms │ +1.23x faster │
│ QQuery 36 │     117.97 / 123.76 ±3.42 / 128.41 ms │     121.56 / 124.74 ±2.44 / 127.90 ms │     no change │
│ QQuery 37 │        48.68 / 51.25 ±1.97 / 53.79 ms │        49.55 / 50.31 ±1.17 / 52.64 ms │     no change │
│ QQuery 38 │        77.17 / 79.47 ±1.58 / 81.33 ms │        77.39 / 78.43 ±0.85 / 79.51 ms │     no change │
│ QQuery 39 │     222.87 / 231.05 ±7.05 / 239.46 ms │     214.53 / 225.45 ±7.93 / 237.62 ms │     no change │
│ QQuery 40 │        22.77 / 26.44 ±2.22 / 28.54 ms │        23.35 / 25.73 ±1.77 / 27.52 ms │     no change │
│ QQuery 41 │        20.94 / 22.04 ±1.23 / 24.27 ms │        20.45 / 21.80 ±1.10 / 23.34 ms │     no change │
│ QQuery 42 │        19.98 / 20.70 ±0.65 / 21.80 ms │        19.35 / 19.92 ±0.59 / 21.02 ms │     no change │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                              ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                              │ 27246.75ms │
│ Total Time (speed-up-group-values-primitive)   │ 27724.05ms │
│ Average Time (HEAD)                            │   633.65ms │
│ Average Time (speed-up-group-values-primitive) │   644.75ms │
│ Queries Faster                                 │          2 │
│ Queries Slower                                 │          6 │
│ Queries with No Change                         │         35 │
│ Queries with Failure                           │          0 │
└────────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric	Value
Wall time	137.4s
Peak memory	40.6 GiB
Avg memory	32.6 GiB
CPU user	1300.1s
CPU sys	85.9s
Peak spill	0 B

clickbench_partitioned — branch

Metric	Value
Wall time	139.7s
Peak memory	38.7 GiB
Avg memory	28.9 GiB
CPU user	1298.8s
CPU sys	104.5s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-04-04T18:33:53Z

🤖 Benchmark completed (GKE) | trigger

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

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 speed-up-group-values-primitive
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                     HEAD ┃          speed-up-group-values-primitive ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │              6.70 / 7.22 ±0.75 / 8.70 ms │              6.65 / 7.11 ±0.80 / 8.70 ms │     no change │
│ QQuery 2  │        146.27 / 147.06 ±0.66 / 148.09 ms │        146.75 / 147.95 ±0.92 / 149.41 ms │     no change │
│ QQuery 3  │        114.87 / 115.19 ±0.35 / 115.69 ms │        114.98 / 115.64 ±0.38 / 116.07 ms │     no change │
│ QQuery 4  │    1375.26 / 1392.88 ±13.50 / 1408.92 ms │    1369.72 / 1390.88 ±14.60 / 1408.64 ms │     no change │
│ QQuery 5  │        172.56 / 173.98 ±1.09 / 175.87 ms │        172.99 / 174.35 ±1.27 / 176.43 ms │     no change │
│ QQuery 6  │       846.21 / 869.37 ±20.72 / 902.47 ms │       840.03 / 873.76 ±32.64 / 920.97 ms │     no change │
│ QQuery 7  │        352.22 / 357.95 ±3.86 / 363.52 ms │        355.90 / 357.09 ±1.04 / 358.44 ms │     no change │
│ QQuery 8  │        115.23 / 116.10 ±0.65 / 116.99 ms │        115.42 / 117.84 ±1.28 / 118.91 ms │     no change │
│ QQuery 9  │        101.88 / 106.55 ±4.38 / 113.65 ms │       101.67 / 109.55 ±10.07 / 129.09 ms │     no change │
│ QQuery 10 │        106.45 / 107.70 ±0.90 / 108.84 ms │        107.76 / 108.94 ±0.82 / 110.05 ms │     no change │
│ QQuery 11 │      984.65 / 992.26 ±11.29 / 1014.58 ms │       957.45 / 972.01 ±10.81 / 990.27 ms │     no change │
│ QQuery 12 │           46.01 / 47.15 ±0.62 / 47.69 ms │           45.71 / 46.96 ±0.96 / 48.43 ms │     no change │
│ QQuery 13 │        400.14 / 404.02 ±2.41 / 407.12 ms │        400.73 / 403.15 ±1.70 / 405.84 ms │     no change │
│ QQuery 14 │     1027.17 / 1033.54 ±4.12 / 1039.33 ms │     1021.11 / 1025.76 ±2.61 / 1028.20 ms │     no change │
│ QQuery 15 │           16.57 / 17.73 ±1.07 / 19.65 ms │           15.56 / 16.90 ±0.94 / 17.94 ms │     no change │
│ QQuery 16 │              7.21 / 8.07 ±0.70 / 9.06 ms │              7.22 / 7.77 ±0.76 / 9.27 ms │     no change │
│ QQuery 17 │        238.35 / 241.81 ±2.04 / 243.67 ms │        239.60 / 242.03 ±1.99 / 244.41 ms │     no change │
│ QQuery 18 │        129.00 / 130.00 ±0.72 / 130.84 ms │        128.71 / 130.40 ±1.09 / 131.73 ms │     no change │
│ QQuery 19 │        156.92 / 158.39 ±1.43 / 160.96 ms │        158.07 / 158.67 ±0.45 / 159.29 ms │     no change │
│ QQuery 20 │           13.27 / 14.20 ±0.62 / 15.09 ms │           14.09 / 15.18 ±0.95 / 16.73 ms │  1.07x slower │
│ QQuery 21 │           19.64 / 20.18 ±0.44 / 20.91 ms │           19.74 / 20.31 ±0.45 / 20.89 ms │     no change │
│ QQuery 22 │        488.60 / 492.74 ±3.41 / 498.80 ms │        485.79 / 490.52 ±3.04 / 493.81 ms │     no change │
│ QQuery 23 │       908.28 / 925.81 ±11.75 / 943.04 ms │        895.75 / 911.86 ±9.42 / 924.53 ms │     no change │
│ QQuery 24 │        397.08 / 399.22 ±2.37 / 402.62 ms │        396.04 / 398.38 ±2.08 / 402.23 ms │     no change │
│ QQuery 25 │        350.45 / 353.24 ±2.59 / 357.59 ms │        350.88 / 352.16 ±1.06 / 353.46 ms │     no change │
│ QQuery 26 │           82.51 / 83.20 ±0.56 / 83.95 ms │           81.82 / 83.29 ±1.37 / 85.06 ms │     no change │
│ QQuery 27 │              6.84 / 7.39 ±0.91 / 9.20 ms │              6.97 / 7.22 ±0.28 / 7.72 ms │     no change │
│ QQuery 28 │        149.20 / 150.26 ±0.84 / 151.68 ms │        148.76 / 149.55 ±0.74 / 150.81 ms │     no change │
│ QQuery 29 │        298.89 / 300.06 ±1.05 / 301.52 ms │        292.48 / 295.30 ±2.13 / 298.87 ms │     no change │
│ QQuery 30 │           42.50 / 45.60 ±1.58 / 46.86 ms │           43.33 / 45.56 ±1.54 / 47.48 ms │     no change │
│ QQuery 31 │        171.04 / 172.43 ±1.03 / 173.92 ms │        172.87 / 173.48 ±0.50 / 173.98 ms │     no change │
│ QQuery 32 │           57.61 / 58.57 ±0.57 / 59.18 ms │           56.75 / 57.52 ±0.66 / 58.50 ms │     no change │
│ QQuery 33 │        141.85 / 143.47 ±1.07 / 145.10 ms │        140.51 / 142.65 ±1.76 / 145.42 ms │     no change │
│ QQuery 34 │              7.15 / 7.28 ±0.19 / 7.65 ms │              7.31 / 7.65 ±0.62 / 8.88 ms │  1.05x slower │
│ QQuery 35 │        109.61 / 110.49 ±0.73 / 111.60 ms │        106.55 / 109.15 ±1.51 / 111.19 ms │     no change │
│ QQuery 36 │              6.56 / 6.87 ±0.29 / 7.33 ms │              6.67 / 6.90 ±0.27 / 7.39 ms │     no change │
│ QQuery 37 │             8.78 / 9.34 ±0.57 / 10.43 ms │              8.65 / 9.00 ±0.22 / 9.33 ms │     no change │
│ QQuery 38 │           86.50 / 89.37 ±2.97 / 94.80 ms │           84.96 / 89.78 ±3.63 / 95.91 ms │     no change │
│ QQuery 39 │        126.11 / 129.62 ±2.48 / 133.49 ms │        127.09 / 128.30 ±0.90 / 129.85 ms │     no change │
│ QQuery 40 │        112.03 / 117.26 ±5.95 / 128.51 ms │        113.17 / 118.08 ±5.72 / 129.18 ms │     no change │
│ QQuery 41 │           14.45 / 15.79 ±0.96 / 16.84 ms │           14.24 / 15.58 ±0.84 / 16.45 ms │     no change │
│ QQuery 42 │        107.29 / 109.96 ±1.43 / 111.53 ms │        107.38 / 110.63 ±2.09 / 113.65 ms │     no change │
│ QQuery 43 │              5.87 / 6.05 ±0.17 / 6.36 ms │              5.94 / 6.16 ±0.36 / 6.89 ms │     no change │
│ QQuery 44 │           11.49 / 12.46 ±1.24 / 14.86 ms │           11.96 / 12.17 ±0.16 / 12.39 ms │     no change │
│ QQuery 45 │           53.10 / 54.05 ±0.80 / 55.27 ms │           52.88 / 54.43 ±1.54 / 57.13 ms │     no change │
│ QQuery 46 │              8.48 / 8.83 ±0.21 / 9.12 ms │              8.69 / 9.00 ±0.24 / 9.42 ms │     no change │
│ QQuery 47 │        731.41 / 741.67 ±5.98 / 748.01 ms │        718.62 / 727.20 ±5.65 / 736.46 ms │     no change │
│ QQuery 48 │        288.57 / 291.36 ±2.36 / 295.10 ms │        291.05 / 294.48 ±2.26 / 297.29 ms │     no change │
│ QQuery 49 │        253.20 / 257.07 ±2.70 / 260.38 ms │        257.68 / 260.53 ±3.74 / 267.87 ms │     no change │
│ QQuery 50 │        229.81 / 233.93 ±3.69 / 240.35 ms │        230.91 / 237.90 ±3.75 / 241.53 ms │     no change │
│ QQuery 51 │        185.32 / 188.56 ±2.39 / 192.38 ms │        182.38 / 188.22 ±3.45 / 191.80 ms │     no change │
│ QQuery 52 │        108.79 / 109.43 ±0.75 / 110.90 ms │        107.94 / 108.93 ±0.63 / 109.79 ms │     no change │
│ QQuery 53 │        104.35 / 104.89 ±0.55 / 105.79 ms │        102.28 / 104.50 ±1.36 / 106.43 ms │     no change │
│ QQuery 54 │        147.12 / 148.01 ±0.59 / 148.92 ms │        147.63 / 151.60 ±2.02 / 152.93 ms │     no change │
│ QQuery 55 │        107.47 / 109.04 ±2.15 / 113.26 ms │        108.30 / 110.15 ±1.20 / 111.89 ms │     no change │
│ QQuery 56 │        142.12 / 143.62 ±1.55 / 146.12 ms │        140.61 / 142.38 ±1.46 / 144.32 ms │     no change │
│ QQuery 57 │        175.31 / 178.10 ±1.76 / 180.84 ms │        173.75 / 176.10 ±1.96 / 178.81 ms │     no change │
│ QQuery 58 │        289.46 / 298.54 ±7.16 / 310.29 ms │        297.32 / 304.57 ±5.01 / 311.98 ms │     no change │
│ QQuery 59 │        198.08 / 200.08 ±1.79 / 203.23 ms │        201.32 / 202.25 ±0.81 / 203.36 ms │     no change │
│ QQuery 60 │        144.25 / 145.63 ±1.19 / 147.18 ms │        143.08 / 144.49 ±1.02 / 146.05 ms │     no change │
│ QQuery 61 │           13.11 / 13.77 ±0.86 / 15.40 ms │           13.63 / 13.77 ±0.08 / 13.87 ms │     no change │
│ QQuery 62 │      915.15 / 959.64 ±39.47 / 1012.82 ms │       912.56 / 951.65 ±25.50 / 979.14 ms │     no change │
│ QQuery 63 │        104.98 / 106.74 ±1.06 / 108.16 ms │        103.15 / 108.70 ±4.39 / 114.90 ms │     no change │
│ QQuery 64 │        700.05 / 704.98 ±5.22 / 714.27 ms │        695.95 / 706.82 ±6.23 / 715.31 ms │     no change │
│ QQuery 65 │        257.60 / 261.08 ±2.51 / 265.32 ms │        263.91 / 265.66 ±1.31 / 267.36 ms │     no change │
│ QQuery 66 │        257.20 / 262.00 ±6.20 / 274.13 ms │        252.43 / 265.51 ±7.30 / 273.02 ms │     no change │
│ QQuery 67 │        308.56 / 319.21 ±6.23 / 327.19 ms │        327.26 / 330.33 ±2.29 / 333.20 ms │     no change │
│ QQuery 68 │           10.40 / 12.10 ±1.41 / 14.69 ms │            9.26 / 11.11 ±1.33 / 13.27 ms │ +1.09x faster │
│ QQuery 69 │        104.09 / 105.63 ±1.58 / 107.90 ms │        103.30 / 104.97 ±1.58 / 107.68 ms │     no change │
│ QQuery 70 │       347.23 / 362.48 ±13.37 / 385.14 ms │       338.47 / 355.58 ±13.64 / 379.52 ms │     no change │
│ QQuery 71 │        135.27 / 136.67 ±1.69 / 139.52 ms │        135.24 / 138.20 ±2.65 / 142.88 ms │     no change │
│ QQuery 72 │        718.02 / 731.19 ±9.13 / 743.88 ms │       712.59 / 734.41 ±14.78 / 751.61 ms │     no change │
│ QQuery 73 │              6.82 / 7.51 ±0.65 / 8.63 ms │              7.63 / 8.73 ±0.76 / 9.78 ms │  1.16x slower │
│ QQuery 74 │        595.70 / 601.45 ±6.14 / 611.71 ms │        627.56 / 638.50 ±9.05 / 648.65 ms │  1.06x slower │
│ QQuery 75 │        275.93 / 279.53 ±2.52 / 281.74 ms │        278.17 / 281.12 ±1.74 / 283.61 ms │     no change │
│ QQuery 76 │        132.63 / 135.12 ±1.60 / 137.19 ms │        136.53 / 137.50 ±0.83 / 138.59 ms │     no change │
│ QQuery 77 │        189.67 / 191.24 ±0.89 / 192.29 ms │        189.45 / 191.76 ±1.26 / 193.16 ms │     no change │
│ QQuery 78 │        352.05 / 355.36 ±3.14 / 360.64 ms │        356.13 / 361.37 ±3.35 / 365.10 ms │     no change │
│ QQuery 79 │        233.73 / 237.64 ±2.25 / 240.57 ms │        238.97 / 243.97 ±2.92 / 246.28 ms │     no change │
│ QQuery 80 │        326.16 / 329.80 ±3.29 / 335.17 ms │        331.40 / 332.36 ±1.21 / 334.59 ms │     no change │
│ QQuery 81 │           26.55 / 27.19 ±0.51 / 27.84 ms │           26.66 / 28.15 ±1.42 / 30.69 ms │     no change │
│ QQuery 82 │        198.85 / 203.35 ±3.34 / 208.06 ms │        198.75 / 201.44 ±1.73 / 203.92 ms │     no change │
│ QQuery 83 │           40.33 / 41.60 ±1.00 / 42.81 ms │           39.87 / 41.71 ±1.62 / 44.51 ms │     no change │
│ QQuery 84 │           49.23 / 50.15 ±1.03 / 51.57 ms │           49.05 / 49.48 ±0.26 / 49.78 ms │     no change │
│ QQuery 85 │        146.79 / 149.76 ±1.67 / 151.69 ms │        147.70 / 150.12 ±1.30 / 151.59 ms │     no change │
│ QQuery 86 │           40.37 / 41.00 ±0.93 / 42.84 ms │           40.49 / 41.01 ±0.57 / 42.05 ms │     no change │
│ QQuery 87 │           85.53 / 90.62 ±3.18 / 95.51 ms │           88.20 / 91.62 ±3.11 / 97.16 ms │     no change │
│ QQuery 88 │         99.81 / 101.52 ±1.17 / 103.06 ms │        100.96 / 102.12 ±0.86 / 103.29 ms │     no change │
│ QQuery 89 │        119.86 / 121.59 ±1.09 / 122.88 ms │        118.49 / 120.27 ±1.13 / 121.78 ms │     no change │
│ QQuery 90 │           23.80 / 24.35 ±0.45 / 25.04 ms │           24.21 / 25.16 ±0.80 / 26.50 ms │     no change │
│ QQuery 91 │           62.59 / 64.77 ±1.67 / 67.72 ms │           64.42 / 65.45 ±0.82 / 66.60 ms │     no change │
│ QQuery 92 │           57.95 / 59.36 ±1.68 / 62.66 ms │           57.53 / 58.44 ±0.69 / 59.18 ms │     no change │
│ QQuery 93 │        198.42 / 201.63 ±2.64 / 205.93 ms │        192.28 / 195.52 ±2.63 / 198.38 ms │     no change │
│ QQuery 94 │           61.84 / 62.50 ±0.43 / 63.12 ms │           62.29 / 62.85 ±0.54 / 63.83 ms │     no change │
│ QQuery 95 │        136.90 / 137.28 ±0.34 / 137.77 ms │        136.14 / 138.47 ±1.91 / 140.86 ms │     no change │
│ QQuery 96 │           71.30 / 74.29 ±2.32 / 76.59 ms │           74.48 / 76.99 ±3.00 / 82.78 ms │     no change │
│ QQuery 97 │        133.91 / 135.76 ±1.19 / 137.63 ms │        131.15 / 135.46 ±2.32 / 137.99 ms │     no change │
│ QQuery 98 │        156.39 / 159.00 ±1.65 / 160.86 ms │        155.45 / 157.40 ±1.40 / 159.69 ms │     no change │
│ QQuery 99 │ 10807.71 / 10843.57 ±23.13 / 10873.43 ms │ 10874.35 / 10907.38 ±22.41 / 10941.65 ms │     no change │
└───────────┴──────────────────────────────────────────┴──────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                              ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                              │ 32123.02ms │
│ Total Time (speed-up-group-values-primitive)   │ 32210.99ms │
│ Average Time (HEAD)                            │   324.47ms │
│ Average Time (speed-up-group-values-primitive) │   325.36ms │
│ Queries Faster                                 │          1 │
│ Queries Slower                                 │          4 │
│ Queries with No Change                         │         94 │
│ Queries with Failure                           │          0 │
└────────────────────────────────────────────────┴────────────┘

Resource Usage

tpcds — base (merge-base)

Metric	Value
Wall time	161.0s
Peak memory	5.2 GiB
Avg memory	4.5 GiB
CPU user	263.0s
CPU sys	17.5s
Peak spill	0 B

tpcds — branch

Metric	Value
Wall time	161.3s
Peak memory	5.3 GiB
Avg memory	4.5 GiB
CPU user	263.6s
CPU sys	18.0s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-04-04T18:34:21Z

🤖 Benchmark completed (GKE) | trigger

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

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 speed-up-group-values-primitive
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃       speed-up-group-values-primitive ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │          1.20 / 4.51 ±6.48 / 17.46 ms │          1.20 / 4.46 ±6.37 / 17.20 ms │     no change │
│ QQuery 1  │        14.24 / 14.64 ±0.24 / 14.97 ms │        14.74 / 15.07 ±0.24 / 15.34 ms │     no change │
│ QQuery 2  │        44.19 / 44.49 ±0.28 / 45.01 ms │        43.19 / 43.43 ±0.29 / 44.00 ms │     no change │
│ QQuery 3  │        41.12 / 44.53 ±2.26 / 47.31 ms │        41.06 / 43.26 ±2.14 / 46.54 ms │     no change │
│ QQuery 4  │     301.04 / 303.06 ±2.03 / 306.86 ms │     284.49 / 290.57 ±5.02 / 298.98 ms │     no change │
│ QQuery 5  │     347.04 / 353.07 ±5.01 / 360.47 ms │     350.67 / 354.39 ±2.27 / 356.95 ms │     no change │
│ QQuery 6  │           5.19 / 6.04 ±0.65 / 7.08 ms │           6.79 / 7.46 ±0.67 / 8.65 ms │  1.24x slower │
│ QQuery 7  │        16.58 / 16.93 ±0.26 / 17.36 ms │        17.31 / 17.58 ±0.18 / 17.87 ms │     no change │
│ QQuery 8  │     431.20 / 434.87 ±2.76 / 437.90 ms │     413.44 / 426.36 ±7.44 / 436.45 ms │     no change │
│ QQuery 9  │    650.65 / 663.10 ±11.22 / 679.36 ms │     651.75 / 655.35 ±2.44 / 658.84 ms │     no change │
│ QQuery 10 │        91.56 / 94.80 ±2.13 / 97.90 ms │       91.39 / 94.94 ±3.66 / 101.93 ms │     no change │
│ QQuery 11 │     102.80 / 103.96 ±0.74 / 105.11 ms │     105.36 / 106.88 ±1.48 / 109.46 ms │     no change │
│ QQuery 12 │     344.69 / 352.95 ±4.81 / 358.62 ms │     350.69 / 357.84 ±5.76 / 366.14 ms │     no change │
│ QQuery 13 │    454.98 / 470.09 ±13.75 / 494.90 ms │     463.24 / 467.88 ±3.69 / 472.85 ms │     no change │
│ QQuery 14 │     352.26 / 357.84 ±5.75 / 368.82 ms │     355.25 / 360.60 ±3.07 / 364.02 ms │     no change │
│ QQuery 15 │    363.86 / 390.86 ±22.09 / 422.94 ms │    341.13 / 368.39 ±28.74 / 421.58 ms │ +1.06x faster │
│ QQuery 16 │    732.23 / 762.05 ±24.40 / 806.74 ms │    734.14 / 752.56 ±26.10 / 802.16 ms │     no change │
│ QQuery 17 │    729.19 / 754.17 ±28.25 / 804.08 ms │     731.07 / 738.01 ±5.98 / 747.49 ms │     no change │
│ QQuery 18 │ 1460.65 / 1505.45 ±33.63 / 1536.22 ms │ 1411.29 / 1474.07 ±40.42 / 1522.16 ms │     no change │
│ QQuery 19 │       36.71 / 44.26 ±12.35 / 68.90 ms │      36.05 / 69.93 ±53.13 / 173.42 ms │  1.58x slower │
│ QQuery 20 │    718.89 / 737.75 ±20.65 / 765.35 ms │    716.54 / 737.77 ±23.50 / 782.56 ms │     no change │
│ QQuery 21 │     757.59 / 761.64 ±3.13 / 764.66 ms │     763.20 / 765.32 ±2.17 / 768.65 ms │     no change │
│ QQuery 22 │  1139.86 / 1148.12 ±8.36 / 1163.87 ms │  1143.34 / 1148.83 ±4.12 / 1154.98 ms │     no change │
│ QQuery 23 │ 3144.30 / 3171.20 ±20.36 / 3195.80 ms │ 3117.83 / 3137.03 ±16.67 / 3163.79 ms │     no change │
│ QQuery 24 │     100.63 / 103.42 ±2.38 / 107.55 ms │     102.59 / 105.71 ±1.68 / 107.14 ms │     no change │
│ QQuery 25 │     139.82 / 141.63 ±1.56 / 143.93 ms │     138.54 / 141.05 ±2.13 / 143.97 ms │     no change │
│ QQuery 26 │     100.54 / 103.19 ±1.61 / 105.45 ms │     100.08 / 103.46 ±2.21 / 106.58 ms │     no change │
│ QQuery 27 │     847.63 / 854.49 ±4.28 / 859.64 ms │     852.62 / 858.99 ±7.58 / 873.26 ms │     no change │
│ QQuery 28 │ 7715.27 / 7773.41 ±33.47 / 7805.51 ms │ 7751.82 / 7796.74 ±27.72 / 7837.32 ms │     no change │
│ QQuery 29 │        50.80 / 56.09 ±5.75 / 67.08 ms │        50.39 / 56.88 ±7.81 / 71.69 ms │     no change │
│ QQuery 30 │     367.84 / 373.89 ±3.85 / 379.52 ms │     373.22 / 379.67 ±4.59 / 385.94 ms │     no change │
│ QQuery 31 │    361.24 / 379.85 ±14.89 / 404.34 ms │    375.92 / 391.76 ±16.96 / 421.78 ms │     no change │
│ QQuery 32 │ 1184.16 / 1236.79 ±54.32 / 1335.44 ms │ 1207.69 / 1273.60 ±54.65 / 1371.88 ms │     no change │
│ QQuery 33 │ 1520.36 / 1563.98 ±41.22 / 1627.54 ms │ 1507.31 / 1580.81 ±56.24 / 1647.21 ms │     no change │
│ QQuery 34 │ 1488.12 / 1516.58 ±20.55 / 1547.84 ms │  1515.60 / 1527.79 ±6.81 / 1535.14 ms │     no change │
│ QQuery 35 │     390.19 / 399.04 ±5.80 / 406.70 ms │     326.73 / 328.07 ±1.30 / 330.43 ms │ +1.22x faster │
│ QQuery 36 │     113.59 / 121.48 ±5.54 / 127.52 ms │     121.01 / 125.21 ±2.40 / 127.58 ms │     no change │
│ QQuery 37 │        48.26 / 50.07 ±1.46 / 52.17 ms │        48.18 / 50.45 ±1.43 / 51.98 ms │     no change │
│ QQuery 38 │        77.20 / 78.19 ±0.96 / 79.93 ms │        75.73 / 76.79 ±1.05 / 78.33 ms │     no change │
│ QQuery 39 │     213.93 / 222.68 ±7.34 / 235.98 ms │     212.03 / 224.42 ±9.61 / 237.91 ms │     no change │
│ QQuery 40 │        23.19 / 25.53 ±2.72 / 30.65 ms │        24.68 / 27.27 ±2.52 / 31.41 ms │  1.07x slower │
│ QQuery 41 │        21.15 / 21.67 ±0.65 / 22.87 ms │        19.97 / 21.30 ±0.91 / 22.62 ms │     no change │
│ QQuery 42 │        20.04 / 20.50 ±0.50 / 21.43 ms │        19.66 / 19.99 ±0.37 / 20.62 ms │     no change │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                              ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                              │ 27582.88ms │
│ Total Time (speed-up-group-values-primitive)   │ 27527.94ms │
│ Average Time (HEAD)                            │   641.46ms │
│ Average Time (speed-up-group-values-primitive) │   640.18ms │
│ Queries Faster                                 │          2 │
│ Queries Slower                                 │          3 │
│ Queries with No Change                         │         38 │
│ Queries with Failure                           │          0 │
└────────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric	Value
Wall time	139.0s
Peak memory	39.1 GiB
Avg memory	30.1 GiB
CPU user	1301.1s
CPU sys	99.3s
Peak spill	0 B

clickbench_partitioned — branch

Metric	Value
Wall time	138.7s
Peak memory	39.4 GiB
Avg memory	27.9 GiB
CPU user	1297.2s
CPU sys	97.6s
Peak spill	0 B

File an issue against this benchmark runner

Dandandan · 2026-04-04T19:09:46Z

run benchmark clickbench_extended

adriangbot · 2026-04-04T19:11:44Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4187598795-809-6vrzp 6.12.55+ #1 SMP Sun Feb 1 08:59:41 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 speed-up-group-values-primitive (a8fde95) to c17c87c (merge-base) diff using: clickbench_extended
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-04T19:29:12Z

🤖 Benchmark completed (GKE) | trigger

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

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 speed-up-group-values-primitive
--------------------
Benchmark clickbench_extended.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                     HEAD ┃          speed-up-group-values-primitive ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │       814.65 / 835.60 ±17.61 / 864.71 ms │        814.91 / 820.21 ±3.73 / 826.17 ms │     no change │
│ QQuery 1  │        206.12 / 207.58 ±1.35 / 209.78 ms │        207.60 / 208.63 ±1.78 / 212.19 ms │     no change │
│ QQuery 2  │        499.02 / 500.46 ±0.84 / 501.61 ms │        497.22 / 499.22 ±1.07 / 500.29 ms │     no change │
│ QQuery 3  │        312.22 / 313.28 ±0.81 / 314.66 ms │        311.82 / 315.14 ±2.28 / 318.00 ms │     no change │
│ QQuery 4  │       664.31 / 675.36 ±11.08 / 695.51 ms │       678.19 / 691.28 ±11.09 / 711.07 ms │     no change │
│ QQuery 5  │ 9860.74 / 10054.82 ±160.42 / 10259.77 ms │ 9771.07 / 10076.89 ±192.72 / 10316.73 ms │     no change │
│ QQuery 6  │     986.43 / 1002.33 ±18.03 / 1036.22 ms │       988.23 / 994.12 ±4.02 / 1000.40 ms │     no change │
│ QQuery 7  │      819.80 / 927.15 ±82.18 / 1071.40 ms │       809.79 / 848.18 ±48.48 / 936.72 ms │ +1.09x faster │
│ QQuery 8  │        395.75 / 401.35 ±6.46 / 413.48 ms │        394.61 / 400.85 ±5.47 / 409.29 ms │     no change │
│ QQuery 9  │    2850.57 / 2959.88 ±95.00 / 3125.13 ms │    2819.41 / 2853.72 ±22.08 / 2880.17 ms │     no change │
│ QQuery 10 │        633.78 / 641.51 ±4.20 / 645.31 ms │       631.09 / 643.75 ±12.08 / 666.79 ms │     no change │
│ QQuery 11 │    2267.53 / 2290.72 ±13.07 / 2305.72 ms │    2235.84 / 2255.15 ±15.02 / 2278.38 ms │     no change │
│ QQuery 12 │        192.88 / 196.42 ±6.11 / 208.63 ms │        189.55 / 196.08 ±6.90 / 208.04 ms │     no change │
└───────────┴──────────────────────────────────────────┴──────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                              ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                              │ 21006.46ms │
│ Total Time (speed-up-group-values-primitive)   │ 20803.22ms │
│ Average Time (HEAD)                            │  1615.88ms │
│ Average Time (speed-up-group-values-primitive) │  1600.25ms │
│ Queries Faster                                 │          1 │
│ Queries Slower                                 │          0 │
│ Queries with No Change                         │         12 │
│ Queries with Failure                           │          0 │
└────────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_extended — base (merge-base)

Metric	Value
Wall time	105.8s
Peak memory	31.7 GiB
Avg memory	26.2 GiB
CPU user	995.7s
CPU sys	56.0s
Peak spill	0 B

clickbench_extended — branch

Metric	Value
Wall time	104.8s
Peak memory	32.3 GiB
Avg memory	28.2 GiB
CPU user	995.4s
CPU sys	47.3s
Peak spill	0 B

File an issue against this benchmark runner

github-actions Bot added the physical-plan Changes to the physical-plan crate label Apr 4, 2026

Extract find_or_insert helper and short-circuit empty cold map probe

71f0ebb

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Push

a8fde95

Dandandan closed this Apr 4, 2026

Conversation

Dandandan commented Apr 4, 2026

Which issue does this PR close?

Rationale for this change

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

Uh oh!

Dandandan commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

Dandandan commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

Dandandan commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

Dandandan commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

adriangbot commented Apr 4, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants