Eliminate more redundant `ProjectionExec`s by Dandandan · Pull Request #21333 · apache/datafusion

Dandandan · 2026-04-03T08:20:12Z

Which issue does this PR close?

Rationale for this change

What changes are included in this PR?

How are these changes tested?

All existing sqllogictests pass (419 files) including TPC-H
Updated test expectations to reflect eliminated ProjectionExec nodes
Net reduction of 72 lines across test files (fewer plan operators)

Are there any user-facing changes?

No, only more efficient plans

🤖 Generated with Claude Code

Dandandan · 2026-04-03T08:21:31Z

run benchmarks

adriangbot · 2026-04-03T08:24:00Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4182489499-709-l255m 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 fix-projection-embedding-order (b0e0c07) to 1e93a67 (merge-base) diff using: tpch
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-03T08:24:01Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4182489499-708-6bkcm 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 fix-projection-embedding-order (b0e0c07) to 1e93a67 (merge-base) diff using: tpcds
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-03T08:24:14Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4182489499-707-nnbz6 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 fix-projection-embedding-order (b0e0c07) to 1e93a67 (merge-base) diff using: clickbench_partitioned
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-03T08:35:31Z

🤖 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 fix-projection-embedding-order
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Query     ┃                           HEAD ┃ fix-projection-embedding-order ┃    Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ QQuery 1  │ 45.38 / 46.13 ±0.71 / 47.33 ms │ 45.08 / 45.91 ±0.56 / 46.85 ms │ no change │
│ QQuery 2  │ 21.53 / 21.67 ±0.22 / 22.11 ms │ 20.47 / 20.79 ±0.23 / 21.17 ms │ no change │
│ QQuery 3  │ 31.77 / 32.17 ±0.22 / 32.37 ms │ 31.44 / 32.33 ±1.20 / 34.69 ms │ no change │
│ QQuery 4  │ 20.75 / 21.81 ±0.59 / 22.37 ms │ 20.23 / 20.78 ±0.59 / 21.89 ms │ no change │
│ QQuery 5  │ 48.03 / 51.53 ±2.11 / 54.32 ms │ 48.41 / 50.02 ±1.30 / 51.59 ms │ no change │
│ QQuery 6  │ 17.07 / 17.21 ±0.13 / 17.42 ms │ 16.90 / 17.16 ±0.18 / 17.37 ms │ no change │
│ QQuery 7  │ 53.87 / 55.26 ±1.16 / 57.11 ms │ 54.63 / 55.84 ±0.82 / 56.83 ms │ no change │
│ QQuery 8  │ 47.73 / 48.20 ±0.45 / 48.85 ms │ 47.80 / 48.08 ±0.31 / 48.55 ms │ no change │
│ QQuery 9  │ 54.20 / 55.99 ±1.21 / 57.97 ms │ 53.24 / 55.08 ±1.76 / 58.13 ms │ no change │
│ QQuery 10 │ 70.52 / 71.80 ±1.61 / 74.94 ms │ 70.21 / 71.51 ±0.82 / 72.35 ms │ no change │
│ QQuery 11 │ 14.08 / 14.41 ±0.37 / 15.11 ms │ 13.60 / 14.06 ±0.30 / 14.48 ms │ no change │
│ QQuery 12 │ 27.83 / 28.10 ±0.35 / 28.77 ms │ 27.34 / 27.77 ±0.30 / 28.15 ms │ no change │
│ QQuery 13 │ 37.75 / 39.21 ±0.75 / 39.75 ms │ 38.82 / 39.91 ±0.84 / 41.41 ms │ no change │
│ QQuery 14 │ 28.42 / 29.32 ±0.88 / 30.80 ms │ 28.39 / 28.69 ±0.21 / 28.97 ms │ no change │
│ QQuery 15 │ 33.61 / 34.34 ±1.01 / 36.25 ms │ 33.18 / 33.76 ±0.35 / 34.16 ms │ no change │
│ QQuery 16 │ 15.81 / 16.13 ±0.23 / 16.36 ms │ 16.08 / 16.25 ±0.15 / 16.50 ms │ no change │
│ QQuery 17 │ 72.43 / 73.45 ±0.76 / 74.49 ms │ 72.74 / 74.04 ±1.27 / 76.26 ms │ no change │
│ QQuery 18 │ 78.70 / 79.55 ±0.74 / 80.87 ms │ 77.80 / 79.17 ±0.69 / 79.66 ms │ no change │
│ QQuery 19 │ 37.40 / 38.32 ±0.70 / 39.56 ms │ 37.00 / 37.60 ±0.47 / 38.09 ms │ no change │
│ QQuery 20 │ 40.22 / 40.72 ±0.55 / 41.74 ms │ 40.45 / 40.66 ±0.19 / 41.00 ms │ no change │
│ QQuery 21 │ 64.04 / 66.65 ±1.78 / 68.57 ms │ 63.06 / 64.84 ±1.45 / 67.08 ms │ no change │
│ QQuery 22 │ 17.88 / 18.16 ±0.27 / 18.65 ms │ 17.91 / 18.15 ±0.32 / 18.78 ms │ no change │
└───────────┴────────────────────────────────┴────────────────────────────────┴───────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ Benchmark Summary                             ┃          ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━┩
│ Total Time (HEAD)                             │ 900.13ms │
│ Total Time (fix-projection-embedding-order)   │ 892.42ms │
│ Average Time (HEAD)                           │  40.92ms │
│ Average Time (fix-projection-embedding-order) │  40.56ms │
│ Queries Faster                                │        0 │
│ Queries Slower                                │        0 │
│ Queries with No Change                        │       22 │
│ Queries with Failure                          │        0 │
└───────────────────────────────────────────────┴──────────┘

Resource Usage

tpch — base (merge-base)

Metric	Value
Wall time	4.8s
Peak memory	4.1 GiB
Avg memory	3.6 GiB
CPU user	33.7s
CPU sys	2.7s
Disk read	0 B
Disk write	136.0 KiB

tpch — branch

Metric	Value
Wall time	4.7s
Peak memory	4.0 GiB
Avg memory	3.6 GiB
CPU user	33.6s
CPU sys	2.5s
Disk read	0 B
Disk write	72.0 KiB

File an issue against this benchmark runner

adriangbot · 2026-04-03T08:40:59Z

🤖 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 fix-projection-embedding-order
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                     HEAD ┃           fix-projection-embedding-order ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │           42.90 / 43.62 ±0.81 / 45.16 ms │           43.71 / 44.70 ±0.60 / 45.52 ms │     no change │
│ QQuery 2  │        146.83 / 148.02 ±1.00 / 149.45 ms │        145.94 / 146.77 ±1.26 / 149.26 ms │     no change │
│ QQuery 3  │        113.34 / 115.01 ±1.07 / 116.23 ms │        115.29 / 115.48 ±0.27 / 116.01 ms │     no change │
│ QQuery 4  │    1320.77 / 1409.74 ±66.75 / 1487.08 ms │    1351.73 / 1394.41 ±34.73 / 1442.20 ms │     no change │
│ QQuery 5  │        174.33 / 176.59 ±1.71 / 179.56 ms │        170.93 / 173.86 ±1.57 / 175.65 ms │     no change │
│ QQuery 6  │     995.88 / 1028.58 ±30.24 / 1077.63 ms │       847.06 / 876.29 ±23.24 / 913.52 ms │ +1.17x faster │
│ QQuery 7  │        350.59 / 352.17 ±2.57 / 357.28 ms │        360.34 / 361.25 ±0.59 / 361.94 ms │     no change │
│ QQuery 8  │        116.24 / 117.35 ±0.65 / 118.21 ms │        118.05 / 118.57 ±0.63 / 119.70 ms │     no change │
│ QQuery 9  │       100.65 / 117.46 ±13.57 / 129.70 ms │        100.71 / 105.84 ±2.62 / 107.89 ms │ +1.11x faster │
│ QQuery 10 │        105.18 / 107.51 ±1.30 / 109.19 ms │        106.44 / 107.70 ±0.84 / 108.76 ms │     no change │
│ QQuery 11 │     926.37 / 1001.60 ±54.90 / 1057.78 ms │      923.04 / 955.64 ±32.61 / 1017.63 ms │     no change │
│ QQuery 12 │           45.18 / 46.60 ±1.32 / 48.37 ms │           44.70 / 46.18 ±1.10 / 47.99 ms │     no change │
│ QQuery 13 │        401.46 / 405.65 ±2.78 / 410.06 ms │        397.76 / 399.75 ±1.45 / 402.25 ms │     no change │
│ QQuery 14 │     1031.92 / 1034.73 ±2.67 / 1039.50 ms │     1013.28 / 1022.73 ±7.07 / 1035.22 ms │     no change │
│ QQuery 15 │           15.54 / 17.23 ±1.31 / 18.93 ms │           16.31 / 17.05 ±0.85 / 18.71 ms │     no change │
│ QQuery 16 │           40.88 / 41.70 ±0.49 / 42.41 ms │           41.33 / 43.77 ±2.18 / 47.83 ms │     no change │
│ QQuery 17 │        236.45 / 239.64 ±2.44 / 242.92 ms │        246.38 / 248.64 ±1.86 / 251.28 ms │     no change │
│ QQuery 18 │        128.63 / 129.95 ±1.50 / 132.87 ms │        129.22 / 131.89 ±1.61 / 133.52 ms │     no change │
│ QQuery 19 │        155.67 / 156.75 ±0.75 / 157.85 ms │        156.90 / 158.87 ±1.08 / 159.83 ms │     no change │
│ QQuery 20 │           14.31 / 14.96 ±0.39 / 15.42 ms │           14.28 / 14.71 ±0.51 / 15.70 ms │     no change │
│ QQuery 21 │           19.77 / 20.06 ±0.25 / 20.44 ms │           19.56 / 19.94 ±0.30 / 20.29 ms │     no change │
│ QQuery 22 │        500.58 / 504.89 ±4.20 / 512.42 ms │        495.35 / 500.48 ±4.39 / 506.63 ms │     no change │
│ QQuery 23 │       904.73 / 929.18 ±19.12 / 957.54 ms │        892.26 / 902.14 ±6.65 / 912.08 ms │     no change │
│ QQuery 24 │        408.97 / 415.26 ±4.61 / 423.12 ms │        412.28 / 416.93 ±3.64 / 423.37 ms │     no change │
│ QQuery 25 │        352.53 / 354.82 ±2.17 / 358.73 ms │        356.86 / 359.44 ±2.47 / 363.90 ms │     no change │
│ QQuery 26 │           81.38 / 83.44 ±1.32 / 85.10 ms │           80.97 / 82.81 ±1.15 / 84.51 ms │     no change │
│ QQuery 27 │        346.51 / 349.76 ±2.20 / 352.40 ms │        354.81 / 358.97 ±4.00 / 364.70 ms │     no change │
│ QQuery 28 │        147.40 / 149.34 ±1.47 / 151.40 ms │        148.68 / 152.00 ±2.72 / 156.59 ms │     no change │
│ QQuery 29 │        297.85 / 299.67 ±1.24 / 301.38 ms │        297.74 / 300.17 ±1.85 / 302.56 ms │     no change │
│ QQuery 30 │           45.02 / 46.96 ±1.06 / 47.89 ms │           44.40 / 46.90 ±1.51 / 48.57 ms │     no change │
│ QQuery 31 │        171.41 / 172.18 ±0.62 / 173.17 ms │        170.22 / 172.89 ±1.78 / 175.73 ms │     no change │
│ QQuery 32 │           58.16 / 59.58 ±1.39 / 62.11 ms │           56.84 / 58.05 ±1.10 / 60.10 ms │     no change │
│ QQuery 33 │        141.25 / 143.06 ±1.41 / 144.74 ms │        140.75 / 141.27 ±0.51 / 142.06 ms │     no change │
│ QQuery 34 │        107.05 / 107.96 ±0.63 / 108.87 ms │        105.31 / 106.62 ±0.88 / 107.90 ms │     no change │
│ QQuery 35 │        106.47 / 109.58 ±2.44 / 113.92 ms │        106.00 / 107.71 ±1.71 / 110.96 ms │     no change │
│ QQuery 36 │        223.10 / 225.92 ±2.92 / 231.30 ms │        208.04 / 214.82 ±5.41 / 220.76 ms │     no change │
│ QQuery 37 │        178.81 / 180.73 ±2.03 / 183.41 ms │        176.32 / 180.38 ±2.45 / 183.21 ms │     no change │
│ QQuery 38 │          87.37 / 92.97 ±5.20 / 102.80 ms │           83.74 / 89.06 ±3.41 / 92.38 ms │     no change │
│ QQuery 39 │        125.91 / 130.99 ±3.61 / 134.90 ms │        125.53 / 129.55 ±2.74 / 134.10 ms │     no change │
│ QQuery 40 │        113.24 / 122.49 ±5.55 / 130.30 ms │        113.44 / 117.12 ±5.12 / 127.25 ms │     no change │
│ QQuery 41 │           14.99 / 15.54 ±0.39 / 16.04 ms │           14.33 / 15.37 ±1.10 / 17.38 ms │     no change │
│ QQuery 42 │        110.30 / 111.65 ±1.14 / 113.50 ms │        106.52 / 108.72 ±1.32 / 110.35 ms │     no change │
│ QQuery 43 │           83.19 / 83.96 ±0.59 / 84.53 ms │           83.49 / 84.22 ±0.78 / 85.71 ms │     no change │
│ QQuery 44 │           11.65 / 12.45 ±0.98 / 14.29 ms │           11.15 / 12.06 ±0.94 / 13.88 ms │     no change │
│ QQuery 45 │           52.05 / 53.94 ±1.42 / 55.62 ms │           52.17 / 53.56 ±1.07 / 55.13 ms │     no change │
│ QQuery 46 │        230.41 / 231.92 ±1.23 / 233.26 ms │        232.21 / 234.26 ±1.41 / 236.01 ms │     no change │
│ QQuery 47 │        696.24 / 699.72 ±3.50 / 706.47 ms │        722.48 / 733.93 ±7.24 / 742.63 ms │     no change │
│ QQuery 48 │        283.10 / 288.74 ±5.02 / 297.99 ms │        283.54 / 291.12 ±4.54 / 297.32 ms │     no change │
│ QQuery 49 │        255.81 / 257.17 ±1.31 / 259.59 ms │        256.15 / 260.26 ±2.28 / 262.37 ms │     no change │
│ QQuery 50 │        237.68 / 244.63 ±5.04 / 249.98 ms │        237.35 / 240.49 ±2.99 / 244.82 ms │     no change │
│ QQuery 51 │        184.43 / 188.88 ±2.62 / 191.80 ms │        181.96 / 186.33 ±3.35 / 191.91 ms │     no change │
│ QQuery 52 │        108.23 / 110.53 ±1.74 / 113.25 ms │        106.78 / 107.32 ±0.33 / 107.71 ms │     no change │
│ QQuery 53 │        103.17 / 104.39 ±1.05 / 106.16 ms │        101.38 / 102.22 ±0.83 / 103.80 ms │     no change │
│ QQuery 54 │        147.17 / 149.88 ±1.64 / 151.80 ms │        145.51 / 147.21 ±1.18 / 149.08 ms │     no change │
│ QQuery 55 │        107.83 / 109.24 ±1.25 / 111.21 ms │        106.14 / 107.24 ±1.35 / 109.87 ms │     no change │
│ QQuery 56 │        141.27 / 142.06 ±0.77 / 143.27 ms │        138.74 / 141.07 ±2.27 / 145.40 ms │     no change │
│ QQuery 57 │        175.10 / 176.61 ±1.00 / 177.93 ms │        171.16 / 174.12 ±2.02 / 176.86 ms │     no change │
│ QQuery 58 │        294.65 / 299.32 ±2.58 / 302.34 ms │        290.06 / 295.10 ±5.05 / 304.14 ms │     no change │
│ QQuery 59 │        200.49 / 201.62 ±0.81 / 202.58 ms │        195.75 / 197.27 ±1.00 / 198.87 ms │     no change │
│ QQuery 60 │        143.77 / 144.49 ±1.11 / 146.70 ms │        142.47 / 144.42 ±1.50 / 146.91 ms │     no change │
│ QQuery 61 │        170.10 / 172.14 ±1.16 / 173.70 ms │        174.14 / 175.62 ±1.26 / 177.43 ms │     no change │
│ QQuery 62 │      905.14 / 974.99 ±49.90 / 1047.40 ms │       909.48 / 942.07 ±32.21 / 996.68 ms │     no change │
│ QQuery 63 │        103.75 / 105.49 ±1.69 / 108.61 ms │        102.76 / 104.26 ±1.15 / 105.95 ms │     no change │
│ QQuery 64 │        715.45 / 724.36 ±7.08 / 734.14 ms │        695.73 / 702.29 ±7.57 / 715.71 ms │     no change │
│ QQuery 65 │        257.76 / 261.03 ±3.57 / 267.92 ms │        246.88 / 248.58 ±1.13 / 250.37 ms │     no change │
│ QQuery 66 │        248.64 / 254.64 ±6.46 / 266.50 ms │        243.24 / 253.56 ±9.24 / 265.58 ms │     no change │
│ QQuery 67 │        304.73 / 311.17 ±4.66 / 318.56 ms │        309.05 / 314.18 ±4.18 / 321.13 ms │     no change │
│ QQuery 68 │        279.37 / 282.07 ±1.89 / 284.04 ms │        276.33 / 280.92 ±5.65 / 291.83 ms │     no change │
│ QQuery 69 │        100.07 / 103.55 ±2.72 / 108.34 ms │        100.45 / 101.43 ±0.67 / 102.47 ms │     no change │
│ QQuery 70 │       337.96 / 354.74 ±15.26 / 383.38 ms │       340.64 / 355.93 ±17.40 / 386.81 ms │     no change │
│ QQuery 71 │        134.10 / 136.19 ±3.54 / 143.27 ms │        134.70 / 138.22 ±1.98 / 140.43 ms │     no change │
│ QQuery 72 │       702.46 / 724.51 ±14.53 / 746.69 ms │        724.01 / 736.32 ±8.71 / 745.72 ms │     no change │
│ QQuery 73 │        107.14 / 108.19 ±0.77 / 109.51 ms │        101.10 / 103.72 ±2.45 / 108.32 ms │     no change │
│ QQuery 74 │       567.21 / 622.86 ±32.22 / 662.03 ms │        548.66 / 555.38 ±4.22 / 560.87 ms │ +1.12x faster │
│ QQuery 75 │        277.48 / 280.27 ±2.38 / 283.93 ms │        272.54 / 275.74 ±1.69 / 277.57 ms │     no change │
│ QQuery 76 │        131.71 / 133.44 ±1.19 / 135.07 ms │        132.10 / 134.15 ±1.56 / 136.29 ms │     no change │
│ QQuery 77 │        189.74 / 192.33 ±1.78 / 195.32 ms │        185.98 / 189.03 ±3.11 / 193.78 ms │     no change │
│ QQuery 78 │        356.58 / 358.70 ±1.74 / 361.09 ms │        343.60 / 349.89 ±3.43 / 352.77 ms │     no change │
│ QQuery 79 │        236.57 / 239.68 ±3.41 / 246.28 ms │        238.32 / 244.94 ±4.02 / 249.01 ms │     no change │
│ QQuery 80 │        329.82 / 334.82 ±2.88 / 338.62 ms │        329.39 / 332.78 ±2.80 / 336.50 ms │     no change │
│ QQuery 81 │           26.94 / 28.10 ±0.83 / 29.47 ms │           27.19 / 28.56 ±1.32 / 31.00 ms │     no change │
│ QQuery 82 │        201.69 / 203.64 ±1.77 / 206.89 ms │        198.42 / 202.86 ±3.24 / 206.74 ms │     no change │
│ QQuery 83 │           40.26 / 41.37 ±0.70 / 42.25 ms │           40.11 / 41.59 ±1.24 / 43.74 ms │     no change │
│ QQuery 84 │           49.62 / 50.44 ±0.97 / 52.18 ms │           47.49 / 49.98 ±1.65 / 52.03 ms │     no change │
│ QQuery 85 │        152.98 / 155.46 ±2.20 / 159.07 ms │        147.91 / 149.10 ±1.18 / 151.27 ms │     no change │
│ QQuery 86 │           40.35 / 41.16 ±0.65 / 42.06 ms │           38.95 / 40.86 ±1.28 / 42.33 ms │     no change │
│ QQuery 87 │           86.84 / 92.68 ±3.84 / 98.38 ms │           85.05 / 88.25 ±2.51 / 91.46 ms │     no change │
│ QQuery 88 │        102.81 / 103.74 ±1.29 / 106.29 ms │         99.81 / 100.61 ±0.58 / 101.56 ms │     no change │
│ QQuery 89 │        119.11 / 121.50 ±1.38 / 123.34 ms │        117.85 / 118.91 ±1.17 / 120.56 ms │     no change │
│ QQuery 90 │           24.64 / 25.14 ±0.58 / 26.20 ms │           22.87 / 23.87 ±0.54 / 24.50 ms │ +1.05x faster │
│ QQuery 91 │           64.03 / 66.26 ±1.40 / 68.20 ms │           62.37 / 63.77 ±1.01 / 65.09 ms │     no change │
│ QQuery 92 │           57.94 / 59.01 ±0.57 / 59.61 ms │           56.81 / 57.82 ±0.87 / 59.21 ms │     no change │
│ QQuery 93 │        200.35 / 201.79 ±1.38 / 204.04 ms │        188.80 / 191.99 ±2.52 / 196.18 ms │     no change │
│ QQuery 94 │           63.05 / 64.27 ±0.67 / 65.02 ms │           61.03 / 61.59 ±0.48 / 62.47 ms │     no change │
│ QQuery 95 │        137.10 / 138.63 ±1.56 / 141.48 ms │        132.65 / 134.45 ±1.51 / 137.11 ms │     no change │
│ QQuery 96 │           75.06 / 76.16 ±0.99 / 78.01 ms │           73.17 / 74.48 ±0.91 / 75.75 ms │     no change │
│ QQuery 97 │        131.53 / 133.93 ±2.77 / 138.99 ms │        129.92 / 131.87 ±1.42 / 133.90 ms │     no change │
│ QQuery 98 │        154.53 / 156.86 ±1.24 / 158.18 ms │        153.82 / 154.67 ±0.61 / 155.45 ms │     no change │
│ QQuery 99 │ 10746.22 / 10795.73 ±32.39 / 10844.10 ms │ 10787.49 / 10815.68 ±27.71 / 10854.16 ms │     no change │
└───────────┴──────────────────────────────────────────┴──────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                             ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                             │ 34061.46ms │
│ Total Time (fix-projection-embedding-order)   │ 33647.59ms │
│ Average Time (HEAD)                           │   344.06ms │
│ Average Time (fix-projection-embedding-order) │   339.87ms │
│ Queries Faster                                │          4 │
│ Queries Slower                                │          0 │
│ Queries with No Change                        │         95 │
│ Queries with Failure                          │          0 │
└───────────────────────────────────────────────┴────────────┘

Resource Usage

tpcds — base (merge-base)

Metric	Value
Wall time	170.7s
Peak memory	5.6 GiB
Avg memory	4.5 GiB
CPU user	274.8s
CPU sys	18.2s
Disk read	0 B
Disk write	638.0 MiB

tpcds — branch

Metric	Value
Wall time	168.5s
Peak memory	5.2 GiB
Avg memory	4.5 GiB
CPU user	271.2s
CPU sys	17.9s
Disk read	0 B
Disk write	164.0 KiB

File an issue against this benchmark runner

adriangbot · 2026-04-03T08:42:13Z

🤖 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 fix-projection-embedding-order
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃        fix-projection-embedding-order ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │          1.37 / 4.76 ±6.58 / 17.92 ms │          1.35 / 4.73 ±6.55 / 17.83 ms │     no change │
│ QQuery 1  │        14.78 / 15.34 ±0.32 / 15.73 ms │        14.73 / 15.35 ±0.34 / 15.77 ms │     no change │
│ QQuery 2  │        44.53 / 45.25 ±0.61 / 46.25 ms │        44.77 / 45.21 ±0.46 / 45.94 ms │     no change │
│ QQuery 3  │        42.32 / 46.38 ±2.89 / 50.35 ms │        44.89 / 46.62 ±1.35 / 48.20 ms │     no change │
│ QQuery 4  │     333.43 / 341.02 ±5.72 / 350.22 ms │    325.98 / 338.70 ±10.86 / 356.52 ms │     no change │
│ QQuery 5  │     389.87 / 393.96 ±4.25 / 401.24 ms │     387.41 / 392.53 ±4.43 / 398.63 ms │     no change │
│ QQuery 6  │           7.17 / 7.98 ±0.79 / 9.31 ms │           5.60 / 6.89 ±0.83 / 7.92 ms │ +1.16x faster │
│ QQuery 7  │        17.93 / 18.24 ±0.17 / 18.43 ms │        17.64 / 18.94 ±1.32 / 20.61 ms │     no change │
│ QQuery 8  │    434.80 / 470.21 ±31.31 / 510.39 ms │    423.89 / 442.39 ±15.17 / 467.21 ms │ +1.06x faster │
│ QQuery 9  │    641.01 / 687.76 ±44.89 / 746.62 ms │    634.15 / 650.41 ±12.07 / 671.71 ms │ +1.06x faster │
│ QQuery 10 │      98.87 / 100.71 ±1.45 / 103.17 ms │        91.12 / 93.61 ±2.65 / 97.53 ms │ +1.08x faster │
│ QQuery 11 │     110.29 / 112.13 ±1.25 / 114.03 ms │     103.16 / 104.69 ±1.07 / 105.99 ms │ +1.07x faster │
│ QQuery 12 │     384.29 / 393.45 ±4.89 / 398.77 ms │     348.11 / 350.53 ±1.81 / 353.51 ms │ +1.12x faster │
│ QQuery 13 │    466.86 / 490.34 ±24.46 / 530.75 ms │     471.53 / 483.78 ±9.31 / 500.14 ms │     no change │
│ QQuery 14 │     354.05 / 360.20 ±6.33 / 372.15 ms │     351.35 / 361.61 ±6.10 / 370.51 ms │     no change │
│ QQuery 15 │    370.04 / 381.06 ±15.72 / 411.57 ms │    365.81 / 394.70 ±23.92 / 428.06 ms │     no change │
│ QQuery 16 │    727.63 / 750.64 ±25.34 / 796.49 ms │    736.72 / 749.26 ±12.91 / 771.98 ms │     no change │
│ QQuery 17 │    786.03 / 818.21 ±33.35 / 873.17 ms │    790.01 / 807.40 ±17.28 / 833.98 ms │     no change │
│ QQuery 18 │ 1478.09 / 1507.81 ±19.96 / 1534.50 ms │ 1497.12 / 1561.28 ±46.20 / 1640.18 ms │     no change │
│ QQuery 19 │       34.60 / 43.17 ±13.27 / 69.62 ms │        38.11 / 40.63 ±1.41 / 42.00 ms │ +1.06x faster │
│ QQuery 20 │    713.83 / 735.12 ±22.22 / 771.06 ms │    721.60 / 750.03 ±17.43 / 775.12 ms │     no change │
│ QQuery 21 │     767.57 / 769.76 ±1.68 / 772.44 ms │     761.15 / 769.34 ±8.41 / 782.07 ms │     no change │
│ QQuery 22 │ 1137.25 / 1154.02 ±18.43 / 1188.28 ms │ 1140.43 / 1160.76 ±23.52 / 1194.48 ms │     no change │
│ QQuery 23 │ 3099.28 / 3214.57 ±74.86 / 3303.94 ms │ 3088.18 / 3167.88 ±76.89 / 3262.02 ms │     no change │
│ QQuery 24 │     107.50 / 110.48 ±3.22 / 115.47 ms │     103.31 / 108.73 ±2.78 / 111.23 ms │     no change │
│ QQuery 25 │     143.96 / 147.53 ±3.34 / 153.86 ms │     143.03 / 146.19 ±2.71 / 149.85 ms │     no change │
│ QQuery 26 │     101.37 / 106.65 ±2.69 / 108.73 ms │     106.06 / 110.54 ±2.40 / 113.06 ms │     no change │
│ QQuery 27 │     865.61 / 875.04 ±5.13 / 880.33 ms │    845.74 / 857.16 ±16.63 / 890.22 ms │     no change │
│ QQuery 28 │ 7746.27 / 7834.16 ±54.88 / 7895.62 ms │ 7691.68 / 7824.59 ±69.44 / 7890.11 ms │     no change │
│ QQuery 29 │        49.50 / 54.88 ±6.70 / 66.75 ms │        52.01 / 55.97 ±4.50 / 64.05 ms │     no change │
│ QQuery 30 │     362.20 / 371.76 ±7.89 / 381.39 ms │    362.90 / 373.32 ±12.57 / 397.39 ms │     no change │
│ QQuery 31 │    369.25 / 384.58 ±11.70 / 402.46 ms │    360.17 / 378.80 ±10.37 / 391.24 ms │     no change │
│ QQuery 32 │ 1323.72 / 1358.70 ±28.84 / 1408.74 ms │ 1034.02 / 1067.59 ±28.28 / 1108.59 ms │ +1.27x faster │
│ QQuery 33 │ 1580.08 / 1646.20 ±77.76 / 1788.91 ms │ 1504.33 / 1619.72 ±62.99 / 1683.12 ms │     no change │
│ QQuery 34 │ 1492.07 / 1595.45 ±69.24 / 1675.28 ms │  1462.46 / 1474.07 ±7.99 / 1483.45 ms │ +1.08x faster │
│ QQuery 35 │    406.93 / 439.05 ±27.15 / 472.83 ms │     380.10 / 387.90 ±4.97 / 394.66 ms │ +1.13x faster │
│ QQuery 36 │     126.23 / 133.55 ±4.45 / 139.03 ms │     113.89 / 120.56 ±4.19 / 125.57 ms │ +1.11x faster │
│ QQuery 37 │        47.95 / 50.30 ±2.78 / 55.58 ms │        48.11 / 49.90 ±1.58 / 52.52 ms │     no change │
│ QQuery 38 │        85.17 / 86.17 ±0.86 / 87.36 ms │        75.27 / 77.65 ±1.28 / 78.87 ms │ +1.11x faster │
│ QQuery 39 │     217.86 / 233.31 ±8.74 / 244.46 ms │     216.76 / 225.26 ±6.02 / 232.66 ms │     no change │
│ QQuery 40 │        26.38 / 28.98 ±2.31 / 32.89 ms │        25.97 / 26.63 ±0.47 / 27.05 ms │ +1.09x faster │
│ QQuery 41 │        20.48 / 21.25 ±0.65 / 22.03 ms │        19.41 / 21.22 ±1.54 / 23.36 ms │     no change │
│ QQuery 42 │        19.48 / 20.32 ±0.72 / 21.62 ms │        18.84 / 20.12 ±1.01 / 21.75 ms │     no change │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                             ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                             │ 28360.45ms │
│ Total Time (fix-projection-embedding-order)   │ 27703.20ms │
│ Average Time (HEAD)                           │   659.55ms │
│ Average Time (fix-projection-embedding-order) │   644.26ms │
│ Queries Faster                                │         13 │
│ Queries Slower                                │          0 │
│ Queries with No Change                        │         30 │
│ Queries with Failure                          │          0 │
└───────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric	Value
Wall time	142.4s
Peak memory	35.8 GiB
Avg memory	26.4 GiB
CPU user	1325.9s
CPU sys	109.1s
Disk read	0 B
Disk write	3.7 GiB

clickbench_partitioned — branch

Metric	Value
Wall time	139.7s
Peak memory	42.9 GiB
Avg memory	32.3 GiB
CPU user	1314.6s
CPU sys	92.8s
Disk read	0 B
Disk write	120.0 KiB

File an issue against this benchmark runner

Dandandan · 2026-04-03T08:42:42Z

run benchmarks

adriangbot · 2026-04-03T08:43:11Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4182570328-710-5nvtx 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 fix-projection-embedding-order (6e943ae) to 1e93a67 (merge-base) diff using: clickbench_partitioned
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-03T08:45:52Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4182570328-711-fgj5n 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 fix-projection-embedding-order (50a9cb2) to 1e93a67 (merge-base) diff using: tpcds
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-03T08:45:56Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4182570328-712-j4x7c 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 fix-projection-embedding-order (50a9cb2) to 1e93a67 (merge-base) diff using: tpch
Results will be posted here when complete

File an issue against this benchmark runner

…ProjectionExec nodes When `try_embed_projection` embeds a projection into a HashJoinExec (or other operators), `collect_column_indices` previously collected column indices into a HashSet and sorted them, losing the original column ordering from the projection. This meant that when a projection simply reordered columns, the embedded projection would output columns in sorted index order instead of the desired order, requiring a residual ProjectionExec just to reorder them. This change preserves the insertion order of column indices: - For simple Column expressions: the index is taken directly in projection order - For complex expressions referencing multiple columns: indices are sorted for determinism (since `collect_columns` returns a HashSet) Additionally, the final embedding decision now checks whether the embedding actually reduced the number of columns. If the residual ProjectionExec is not removable AND no columns were eliminated, the embedding is skipped (it would only add an unnecessary column reorder). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

adriangbot · 2026-04-03T08:55:07Z

🤖 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 fix-projection-embedding-order
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                           HEAD ┃ fix-projection-embedding-order ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │ 46.90 / 47.65 ±0.67 / 48.62 ms │ 45.31 / 46.47 ±0.84 / 47.75 ms │     no change │
│ QQuery 2  │ 21.66 / 22.14 ±0.28 / 22.47 ms │ 20.75 / 20.97 ±0.13 / 21.08 ms │ +1.06x faster │
│ QQuery 3  │ 33.21 / 33.44 ±0.20 / 33.71 ms │ 31.48 / 31.79 ±0.19 / 32.05 ms │     no change │
│ QQuery 4  │ 21.93 / 23.06 ±0.62 / 23.82 ms │ 22.04 / 22.28 ±0.23 / 22.58 ms │     no change │
│ QQuery 5  │ 51.07 / 53.50 ±1.67 / 55.44 ms │ 47.30 / 50.10 ±1.46 / 51.44 ms │ +1.07x faster │
│ QQuery 6  │ 17.84 / 17.95 ±0.09 / 18.11 ms │ 17.02 / 17.16 ±0.18 / 17.50 ms │     no change │
│ QQuery 7  │ 58.45 / 60.61 ±1.89 / 64.12 ms │ 53.50 / 54.99 ±1.02 / 56.08 ms │ +1.10x faster │
│ QQuery 8  │ 50.14 / 50.53 ±0.32 / 50.93 ms │ 47.63 / 48.06 ±0.30 / 48.50 ms │     no change │
│ QQuery 9  │ 55.73 / 58.21 ±1.92 / 61.46 ms │ 53.04 / 55.18 ±1.43 / 56.95 ms │ +1.05x faster │
│ QQuery 10 │ 70.92 / 73.41 ±1.45 / 75.00 ms │ 71.01 / 71.82 ±0.87 / 73.44 ms │     no change │
│ QQuery 11 │ 13.78 / 14.04 ±0.22 / 14.40 ms │ 13.68 / 14.18 ±0.44 / 15.00 ms │     no change │
│ QQuery 12 │ 27.19 / 27.97 ±0.70 / 29.08 ms │ 28.30 / 29.21 ±0.79 / 30.62 ms │     no change │
│ QQuery 13 │ 38.61 / 39.65 ±0.84 / 40.65 ms │ 38.81 / 39.40 ±0.38 / 39.88 ms │     no change │
│ QQuery 14 │ 28.23 / 28.77 ±0.40 / 29.44 ms │ 28.32 / 28.60 ±0.31 / 29.18 ms │     no change │
│ QQuery 15 │ 33.08 / 33.63 ±0.62 / 34.81 ms │ 33.44 / 33.66 ±0.24 / 34.04 ms │     no change │
│ QQuery 16 │ 16.23 / 16.72 ±0.33 / 17.12 ms │ 15.81 / 16.53 ±0.58 / 17.57 ms │     no change │
│ QQuery 17 │ 72.11 / 74.32 ±1.37 / 75.58 ms │ 72.02 / 72.98 ±0.82 / 74.49 ms │     no change │
│ QQuery 18 │ 76.30 / 77.49 ±1.14 / 79.23 ms │ 76.72 / 78.39 ±1.51 / 80.98 ms │     no change │
│ QQuery 19 │ 36.92 / 37.51 ±0.57 / 38.37 ms │ 37.92 / 38.28 ±0.57 / 39.41 ms │     no change │
│ QQuery 20 │ 39.68 / 40.58 ±0.87 / 42.20 ms │ 41.06 / 42.13 ±1.14 / 44.08 ms │     no change │
│ QQuery 21 │ 63.16 / 64.69 ±0.88 / 65.70 ms │ 63.27 / 65.14 ±1.43 / 67.65 ms │     no change │
│ QQuery 22 │ 17.74 / 18.34 ±0.61 / 19.30 ms │ 17.37 / 17.82 ±0.42 / 18.44 ms │     no change │
└───────────┴────────────────────────────────┴────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ Benchmark Summary                             ┃          ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━┩
│ Total Time (HEAD)                             │ 914.21ms │
│ Total Time (fix-projection-embedding-order)   │ 895.16ms │
│ Average Time (HEAD)                           │  41.55ms │
│ Average Time (fix-projection-embedding-order) │  40.69ms │
│ Queries Faster                                │        4 │
│ Queries Slower                                │        0 │
│ Queries with No Change                        │       18 │
│ Queries with Failure                          │        0 │
└───────────────────────────────────────────────┴──────────┘

Resource Usage

tpch — base (merge-base)

Metric	Value
Wall time	4.9s
Peak memory	4.1 GiB
Avg memory	3.6 GiB
CPU user	34.3s
CPU sys	2.6s
Disk read	0 B
Disk write	152.0 KiB

tpch — branch

Metric	Value
Wall time	4.7s
Peak memory	4.1 GiB
Avg memory	3.6 GiB
CPU user	33.6s
CPU sys	2.5s
Disk read	0 B
Disk write	60.0 KiB

File an issue against this benchmark runner

adriangbot · 2026-04-03T09:00:07Z

🤖 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 fix-projection-embedding-order
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                     HEAD ┃           fix-projection-embedding-order ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │           42.72 / 43.74 ±0.98 / 45.30 ms │           42.52 / 43.20 ±0.80 / 44.74 ms │     no change │
│ QQuery 2  │        144.75 / 145.87 ±0.67 / 146.79 ms │        144.15 / 144.82 ±0.61 / 145.73 ms │     no change │
│ QQuery 3  │        112.75 / 113.95 ±1.26 / 116.24 ms │        111.87 / 112.69 ±0.68 / 113.86 ms │     no change │
│ QQuery 4  │     1320.38 / 1325.97 ±8.03 / 1341.91 ms │    1270.01 / 1300.38 ±18.19 / 1317.39 ms │     no change │
│ QQuery 5  │        172.49 / 174.86 ±1.32 / 176.43 ms │        170.43 / 171.75 ±1.08 / 173.59 ms │     no change │
│ QQuery 6  │     984.21 / 1037.28 ±28.87 / 1067.61 ms │       825.49 / 874.94 ±29.28 / 898.75 ms │ +1.19x faster │
│ QQuery 7  │        349.22 / 352.23 ±3.01 / 357.33 ms │        350.28 / 354.61 ±2.19 / 355.99 ms │     no change │
│ QQuery 8  │        114.49 / 115.85 ±1.03 / 116.96 ms │        114.09 / 115.60 ±0.89 / 116.67 ms │     no change │
│ QQuery 9  │        100.89 / 104.52 ±2.74 / 107.49 ms │        100.25 / 106.24 ±4.72 / 114.03 ms │     no change │
│ QQuery 10 │        105.26 / 106.84 ±1.14 / 108.78 ms │        106.62 / 107.59 ±0.82 / 108.77 ms │     no change │
│ QQuery 11 │        894.05 / 902.54 ±4.66 / 907.23 ms │       893.16 / 913.31 ±14.98 / 934.42 ms │     no change │
│ QQuery 12 │           44.84 / 45.77 ±0.50 / 46.33 ms │           44.13 / 44.93 ±0.95 / 46.75 ms │     no change │
│ QQuery 13 │        393.38 / 401.73 ±4.60 / 406.24 ms │        399.30 / 402.68 ±2.39 / 406.31 ms │     no change │
│ QQuery 14 │     1008.61 / 1015.09 ±5.31 / 1024.27 ms │     1010.76 / 1019.00 ±6.12 / 1026.40 ms │     no change │
│ QQuery 15 │           15.42 / 16.16 ±0.86 / 17.77 ms │           15.48 / 16.78 ±1.94 / 20.61 ms │     no change │
│ QQuery 16 │           40.62 / 42.19 ±1.28 / 44.21 ms │           40.88 / 41.56 ±0.65 / 42.47 ms │     no change │
│ QQuery 17 │        237.42 / 239.24 ±1.61 / 241.95 ms │        237.95 / 239.25 ±1.10 / 241.15 ms │     no change │
│ QQuery 18 │        128.02 / 129.17 ±0.93 / 130.59 ms │        126.36 / 128.30 ±1.47 / 130.52 ms │     no change │
│ QQuery 19 │        152.64 / 155.25 ±1.68 / 156.83 ms │        153.76 / 154.85 ±0.89 / 156.42 ms │     no change │
│ QQuery 20 │           13.53 / 13.84 ±0.30 / 14.31 ms │           13.64 / 14.40 ±0.63 / 15.40 ms │     no change │
│ QQuery 21 │           19.18 / 19.59 ±0.40 / 20.35 ms │           18.98 / 19.75 ±0.88 / 21.45 ms │     no change │
│ QQuery 22 │        490.51 / 493.91 ±2.47 / 497.61 ms │        482.37 / 488.50 ±4.51 / 495.60 ms │     no change │
│ QQuery 23 │        875.40 / 883.88 ±7.87 / 896.92 ms │       877.67 / 897.60 ±12.77 / 916.36 ms │     no change │
│ QQuery 24 │        411.93 / 415.07 ±3.10 / 421.03 ms │        411.77 / 415.63 ±2.16 / 418.42 ms │     no change │
│ QQuery 25 │        351.73 / 354.76 ±2.60 / 359.12 ms │        353.63 / 355.32 ±1.40 / 357.68 ms │     no change │
│ QQuery 26 │           81.78 / 82.36 ±0.69 / 83.67 ms │           83.06 / 84.85 ±1.66 / 87.43 ms │     no change │
│ QQuery 27 │        341.62 / 346.51 ±3.29 / 350.76 ms │        343.51 / 347.57 ±2.36 / 350.77 ms │     no change │
│ QQuery 28 │        147.43 / 149.09 ±0.92 / 150.06 ms │        147.29 / 149.15 ±1.54 / 151.64 ms │     no change │
│ QQuery 29 │        296.24 / 298.94 ±1.83 / 301.95 ms │        296.91 / 298.22 ±1.43 / 300.89 ms │     no change │
│ QQuery 30 │           44.81 / 46.77 ±2.31 / 51.22 ms │           44.64 / 47.03 ±2.01 / 50.19 ms │     no change │
│ QQuery 31 │        169.87 / 171.90 ±1.72 / 173.90 ms │        169.72 / 172.58 ±2.09 / 176.01 ms │     no change │
│ QQuery 32 │         56.55 / 65.88 ±17.26 / 100.36 ms │           56.79 / 57.99 ±0.93 / 59.27 ms │ +1.14x faster │
│ QQuery 33 │        137.66 / 139.77 ±1.56 / 142.35 ms │        138.28 / 140.30 ±1.34 / 141.80 ms │     no change │
│ QQuery 34 │        106.68 / 107.77 ±0.61 / 108.48 ms │        105.20 / 106.52 ±1.39 / 109.01 ms │     no change │
│ QQuery 35 │        106.86 / 108.12 ±1.13 / 109.96 ms │        106.48 / 107.57 ±0.96 / 108.79 ms │     no change │
│ QQuery 36 │        212.71 / 216.29 ±2.48 / 220.01 ms │        215.55 / 218.22 ±2.24 / 222.15 ms │     no change │
│ QQuery 37 │        171.73 / 175.87 ±3.04 / 179.63 ms │        175.49 / 176.85 ±1.48 / 179.38 ms │     no change │
│ QQuery 38 │           83.80 / 88.52 ±3.90 / 95.38 ms │           85.40 / 88.53 ±3.41 / 94.93 ms │     no change │
│ QQuery 39 │        124.34 / 126.76 ±2.11 / 130.18 ms │        123.61 / 126.84 ±2.06 / 129.94 ms │     no change │
│ QQuery 40 │        110.04 / 118.34 ±6.10 / 128.69 ms │        113.23 / 118.98 ±6.60 / 131.81 ms │     no change │
│ QQuery 41 │           14.72 / 15.69 ±1.23 / 17.91 ms │           14.39 / 15.34 ±0.83 / 16.72 ms │     no change │
│ QQuery 42 │        105.36 / 108.24 ±2.13 / 111.55 ms │        105.28 / 107.27 ±1.60 / 109.11 ms │     no change │
│ QQuery 43 │           83.66 / 84.27 ±0.47 / 85.10 ms │           82.76 / 84.41 ±1.37 / 86.18 ms │     no change │
│ QQuery 44 │           11.97 / 12.23 ±0.22 / 12.52 ms │           11.65 / 11.99 ±0.23 / 12.34 ms │     no change │
│ QQuery 45 │           51.90 / 53.59 ±1.58 / 56.41 ms │           50.87 / 52.73 ±1.29 / 54.45 ms │     no change │
│ QQuery 46 │        229.58 / 231.72 ±1.24 / 233.17 ms │        227.42 / 230.28 ±2.17 / 232.95 ms │     no change │
│ QQuery 47 │        688.02 / 693.44 ±3.04 / 696.59 ms │        684.15 / 689.34 ±3.74 / 695.74 ms │     no change │
│ QQuery 48 │        288.42 / 289.86 ±2.40 / 294.66 ms │        285.58 / 290.73 ±3.01 / 293.79 ms │     no change │
│ QQuery 49 │        253.98 / 257.18 ±1.91 / 259.64 ms │        252.11 / 254.60 ±1.58 / 256.65 ms │     no change │
│ QQuery 50 │        231.38 / 235.72 ±4.01 / 243.06 ms │        229.52 / 234.88 ±5.21 / 243.21 ms │     no change │
│ QQuery 51 │        180.46 / 183.85 ±3.21 / 189.30 ms │        178.19 / 184.48 ±5.33 / 194.29 ms │     no change │
│ QQuery 52 │        106.35 / 108.20 ±2.17 / 112.24 ms │        106.71 / 107.97 ±0.70 / 108.76 ms │     no change │
│ QQuery 53 │        101.04 / 102.30 ±1.04 / 103.70 ms │        101.32 / 101.75 ±0.34 / 102.19 ms │     no change │
│ QQuery 54 │        147.33 / 148.11 ±0.66 / 149.06 ms │        146.95 / 148.14 ±0.67 / 148.79 ms │     no change │
│ QQuery 55 │        106.19 / 106.81 ±0.62 / 107.86 ms │        106.26 / 107.60 ±1.59 / 110.03 ms │     no change │
│ QQuery 56 │        139.40 / 140.45 ±1.21 / 142.79 ms │        139.70 / 141.56 ±1.31 / 143.49 ms │     no change │
│ QQuery 57 │        173.35 / 175.65 ±1.81 / 178.21 ms │        168.08 / 172.40 ±2.89 / 176.01 ms │     no change │
│ QQuery 58 │        293.52 / 297.44 ±2.86 / 301.87 ms │       286.81 / 306.99 ±14.82 / 331.95 ms │     no change │
│ QQuery 59 │        199.46 / 200.73 ±0.75 / 201.42 ms │        195.85 / 198.83 ±2.04 / 201.67 ms │     no change │
│ QQuery 60 │        142.13 / 143.49 ±1.39 / 145.60 ms │        141.54 / 143.59 ±1.48 / 145.80 ms │     no change │
│ QQuery 61 │        170.51 / 171.63 ±0.66 / 172.17 ms │        169.42 / 171.03 ±1.09 / 172.22 ms │     no change │
│ QQuery 62 │       928.88 / 937.30 ±11.48 / 959.78 ms │      907.25 / 950.35 ±44.89 / 1007.60 ms │     no change │
│ QQuery 63 │        101.96 / 104.58 ±1.47 / 106.05 ms │        101.41 / 104.09 ±2.36 / 108.40 ms │     no change │
│ QQuery 64 │        693.70 / 700.38 ±6.47 / 712.11 ms │        687.64 / 693.98 ±5.08 / 703.19 ms │     no change │
│ QQuery 65 │        250.47 / 253.98 ±2.93 / 258.05 ms │        248.04 / 250.24 ±1.55 / 252.30 ms │     no change │
│ QQuery 66 │        242.04 / 257.29 ±8.46 / 266.41 ms │       243.83 / 258.62 ±13.34 / 280.21 ms │     no change │
│ QQuery 67 │        309.70 / 314.64 ±3.20 / 317.30 ms │        306.92 / 314.34 ±6.80 / 326.01 ms │     no change │
│ QQuery 68 │        279.15 / 282.30 ±1.79 / 284.58 ms │        274.92 / 276.86 ±2.40 / 281.53 ms │     no change │
│ QQuery 69 │        101.53 / 103.39 ±1.63 / 105.73 ms │        101.28 / 102.86 ±1.23 / 104.93 ms │     no change │
│ QQuery 70 │        342.54 / 350.34 ±7.05 / 363.15 ms │       347.71 / 358.46 ±11.11 / 374.82 ms │     no change │
│ QQuery 71 │        131.93 / 136.08 ±2.65 / 138.79 ms │        131.23 / 137.99 ±4.55 / 145.31 ms │     no change │
│ QQuery 72 │       707.57 / 720.74 ±12.88 / 741.19 ms │        701.04 / 712.92 ±7.31 / 723.78 ms │     no change │
│ QQuery 73 │        102.21 / 103.18 ±0.81 / 104.53 ms │        101.96 / 103.61 ±1.42 / 106.19 ms │     no change │
│ QQuery 74 │        542.48 / 552.53 ±5.12 / 556.89 ms │        547.24 / 554.44 ±3.97 / 558.98 ms │     no change │
│ QQuery 75 │        276.81 / 277.99 ±1.19 / 280.05 ms │        271.12 / 275.07 ±2.53 / 278.25 ms │     no change │
│ QQuery 76 │        128.79 / 131.99 ±1.77 / 134.09 ms │        131.62 / 133.04 ±1.17 / 134.41 ms │     no change │
│ QQuery 77 │        186.60 / 188.67 ±1.67 / 190.97 ms │        188.07 / 190.65 ±1.86 / 193.44 ms │     no change │
│ QQuery 78 │        345.35 / 353.13 ±4.12 / 357.47 ms │        345.62 / 349.35 ±3.67 / 354.70 ms │     no change │
│ QQuery 79 │        230.24 / 233.33 ±2.08 / 236.67 ms │        227.61 / 235.03 ±5.33 / 240.74 ms │     no change │
│ QQuery 80 │        328.40 / 330.94 ±1.87 / 332.84 ms │        326.12 / 328.24 ±1.90 / 331.73 ms │     no change │
│ QQuery 81 │           26.71 / 28.67 ±1.70 / 31.64 ms │           26.21 / 27.09 ±0.79 / 28.57 ms │ +1.06x faster │
│ QQuery 82 │        195.83 / 197.44 ±1.27 / 199.32 ms │        196.98 / 199.13 ±2.49 / 203.91 ms │     no change │
│ QQuery 83 │           39.34 / 40.06 ±0.92 / 41.77 ms │           39.65 / 41.16 ±1.36 / 43.53 ms │     no change │
│ QQuery 84 │           48.10 / 48.58 ±0.51 / 49.55 ms │           47.61 / 48.24 ±0.42 / 48.88 ms │     no change │
│ QQuery 85 │        147.30 / 149.54 ±1.91 / 151.92 ms │        147.08 / 148.82 ±1.14 / 150.29 ms │     no change │
│ QQuery 86 │           38.64 / 40.02 ±1.21 / 41.47 ms │           37.89 / 39.78 ±1.45 / 41.91 ms │     no change │
│ QQuery 87 │           84.49 / 88.15 ±3.76 / 93.20 ms │           87.27 / 91.10 ±2.98 / 96.29 ms │     no change │
│ QQuery 88 │         99.03 / 100.79 ±1.89 / 103.91 ms │          98.79 / 99.91 ±1.23 / 102.20 ms │     no change │
│ QQuery 89 │        116.43 / 117.97 ±1.04 / 119.33 ms │        118.09 / 119.11 ±0.83 / 120.09 ms │     no change │
│ QQuery 90 │           23.05 / 23.46 ±0.29 / 23.78 ms │           22.85 / 23.14 ±0.17 / 23.37 ms │     no change │
│ QQuery 91 │           61.57 / 63.09 ±1.14 / 64.61 ms │           63.80 / 65.08 ±0.81 / 66.16 ms │     no change │
│ QQuery 92 │           57.55 / 58.00 ±0.46 / 58.77 ms │           55.66 / 56.78 ±0.82 / 58.12 ms │     no change │
│ QQuery 93 │        188.47 / 191.94 ±2.38 / 195.16 ms │        188.37 / 192.17 ±2.35 / 195.64 ms │     no change │
│ QQuery 94 │           60.36 / 61.22 ±0.69 / 62.16 ms │           59.91 / 60.93 ±0.54 / 61.46 ms │     no change │
│ QQuery 95 │        134.15 / 136.41 ±1.74 / 137.94 ms │        134.25 / 136.46 ±2.26 / 139.97 ms │     no change │
│ QQuery 96 │           73.37 / 73.99 ±0.75 / 75.46 ms │           73.06 / 74.11 ±0.57 / 74.58 ms │     no change │
│ QQuery 97 │        129.19 / 131.80 ±2.13 / 134.56 ms │        130.96 / 131.15 ±0.16 / 131.32 ms │     no change │
│ QQuery 98 │        149.09 / 154.73 ±3.45 / 159.94 ms │        151.40 / 153.27 ±1.22 / 154.70 ms │     no change │
│ QQuery 99 │ 10709.37 / 10722.96 ±11.22 / 10741.57 ms │ 10709.98 / 10750.53 ±26.53 / 10780.98 ms │     no change │
└───────────┴──────────────────────────────────────────┴──────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                             ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                             │ 33414.32ms │
│ Total Time (fix-projection-embedding-order)   │ 33268.94ms │
│ Average Time (HEAD)                           │   337.52ms │
│ Average Time (fix-projection-embedding-order) │   336.05ms │
│ Queries Faster                                │          3 │
│ Queries Slower                                │          0 │
│ Queries with No Change                        │         96 │
│ Queries with Failure                          │          0 │
└───────────────────────────────────────────────┴────────────┘

Resource Usage

tpcds — base (merge-base)

Metric	Value
Wall time	167.4s
Peak memory	5.5 GiB
Avg memory	4.5 GiB
CPU user	268.9s
CPU sys	17.2s
Disk read	0 B
Disk write	703.6 MiB

tpcds — branch

Metric	Value
Wall time	166.6s
Peak memory	5.6 GiB
Avg memory	4.5 GiB
CPU user	266.9s
CPU sys	18.4s
Disk read	0 B
Disk write	252.0 KiB

File an issue against this benchmark runner

adriangbot · 2026-04-03T09:01:12Z

🤖 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 fix-projection-embedding-order
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃        fix-projection-embedding-order ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │          1.28 / 4.57 ±6.46 / 17.49 ms │          1.29 / 4.58 ±6.43 / 17.44 ms │     no change │
│ QQuery 1  │        14.44 / 14.69 ±0.15 / 14.89 ms │        14.29 / 14.65 ±0.19 / 14.78 ms │     no change │
│ QQuery 2  │        43.47 / 43.83 ±0.27 / 44.13 ms │        44.00 / 44.23 ±0.33 / 44.87 ms │     no change │
│ QQuery 3  │        43.85 / 44.73 ±0.71 / 45.87 ms │        41.59 / 45.91 ±2.93 / 48.79 ms │     no change │
│ QQuery 4  │    296.90 / 330.15 ±17.18 / 346.26 ms │    299.07 / 310.99 ±12.12 / 333.18 ms │ +1.06x faster │
│ QQuery 5  │     384.45 / 393.93 ±7.51 / 403.61 ms │    361.05 / 384.73 ±12.00 / 392.93 ms │     no change │
│ QQuery 6  │           6.31 / 7.79 ±1.32 / 9.82 ms │           5.69 / 7.51 ±1.33 / 9.48 ms │     no change │
│ QQuery 7  │        17.58 / 19.38 ±2.14 / 23.42 ms │        17.34 / 18.37 ±1.07 / 20.06 ms │ +1.06x faster │
│ QQuery 8  │    480.92 / 495.78 ±17.00 / 527.15 ms │    483.43 / 506.32 ±17.93 / 527.42 ms │     no change │
│ QQuery 9  │    654.36 / 683.03 ±31.14 / 741.26 ms │    651.48 / 701.50 ±44.61 / 762.56 ms │     no change │
│ QQuery 10 │        92.09 / 94.80 ±2.37 / 99.18 ms │       93.14 / 96.90 ±2.92 / 100.70 ms │     no change │
│ QQuery 11 │     106.07 / 106.87 ±0.58 / 107.51 ms │     106.40 / 107.84 ±1.34 / 109.94 ms │     no change │
│ QQuery 12 │    353.40 / 365.93 ±12.40 / 385.86 ms │    348.70 / 359.95 ±10.99 / 379.98 ms │     no change │
│ QQuery 13 │    473.34 / 509.69 ±21.73 / 539.91 ms │    444.52 / 475.45 ±16.39 / 489.70 ms │ +1.07x faster │
│ QQuery 14 │     391.63 / 395.62 ±4.41 / 403.89 ms │     359.28 / 361.62 ±1.82 / 363.75 ms │ +1.09x faster │
│ QQuery 15 │    367.78 / 407.65 ±42.11 / 465.34 ms │    369.30 / 394.54 ±24.96 / 429.46 ms │     no change │
│ QQuery 16 │    732.02 / 770.41 ±30.90 / 811.47 ms │    722.35 / 766.98 ±33.87 / 816.50 ms │     no change │
│ QQuery 17 │    727.47 / 751.83 ±22.15 / 779.35 ms │    746.70 / 790.54 ±28.99 / 822.56 ms │  1.05x slower │
│ QQuery 18 │ 1559.08 / 1627.74 ±63.64 / 1713.70 ms │ 1461.70 / 1556.84 ±68.42 / 1659.91 ms │     no change │
│ QQuery 19 │        35.54 / 36.46 ±0.82 / 37.95 ms │       35.80 / 42.12 ±12.44 / 67.00 ms │  1.16x slower │
│ QQuery 20 │    724.65 / 739.06 ±15.66 / 762.05 ms │    732.24 / 751.72 ±11.78 / 767.96 ms │     no change │
│ QQuery 21 │     770.10 / 776.41 ±7.29 / 790.25 ms │     786.14 / 798.76 ±7.40 / 805.44 ms │     no change │
│ QQuery 22 │  1138.03 / 1145.97 ±8.89 / 1162.81 ms │  1147.98 / 1154.01 ±7.83 / 1167.87 ms │     no change │
│ QQuery 23 │ 3095.14 / 3242.91 ±97.33 / 3366.19 ms │ 3144.43 / 3226.31 ±63.10 / 3300.77 ms │     no change │
│ QQuery 24 │     109.68 / 110.35 ±0.85 / 111.93 ms │     108.93 / 109.99 ±0.97 / 111.66 ms │     no change │
│ QQuery 25 │     147.76 / 148.69 ±1.19 / 151.02 ms │     146.78 / 147.86 ±1.16 / 149.94 ms │     no change │
│ QQuery 26 │     109.07 / 111.52 ±1.35 / 112.80 ms │     108.24 / 110.55 ±1.50 / 112.61 ms │     no change │
│ QQuery 27 │    855.50 / 873.70 ±15.12 / 893.69 ms │    857.96 / 874.30 ±13.77 / 892.53 ms │     no change │
│ QQuery 28 │ 7761.74 / 7795.07 ±33.13 / 7856.34 ms │ 7810.82 / 7851.60 ±38.93 / 7916.14 ms │     no change │
│ QQuery 29 │        52.16 / 56.58 ±3.68 / 61.60 ms │        52.50 / 55.83 ±2.69 / 59.78 ms │     no change │
│ QQuery 30 │     406.02 / 411.01 ±2.82 / 414.37 ms │     401.57 / 408.49 ±6.54 / 420.02 ms │     no change │
│ QQuery 31 │    415.64 / 429.45 ±11.15 / 448.29 ms │     404.00 / 414.57 ±7.21 / 422.85 ms │     no change │
│ QQuery 32 │ 1256.30 / 1308.08 ±47.30 / 1376.08 ms │ 1187.76 / 1260.94 ±82.95 / 1419.15 ms │     no change │
│ QQuery 33 │ 1558.90 / 1743.75 ±92.89 / 1799.44 ms │ 1570.51 / 1630.26 ±34.23 / 1662.42 ms │ +1.07x faster │
│ QQuery 34 │ 1513.86 / 1690.35 ±97.13 / 1801.42 ms │ 1484.56 / 1496.70 ±15.57 / 1527.25 ms │ +1.13x faster │
│ QQuery 35 │     392.73 / 400.19 ±8.55 / 416.84 ms │     382.89 / 392.90 ±6.52 / 402.95 ms │     no change │
│ QQuery 36 │     122.86 / 125.46 ±2.12 / 127.47 ms │     116.92 / 125.76 ±5.91 / 134.75 ms │     no change │
│ QQuery 37 │        48.03 / 51.31 ±1.68 / 52.62 ms │        47.46 / 49.29 ±1.41 / 51.66 ms │     no change │
│ QQuery 38 │        75.44 / 78.43 ±2.75 / 83.11 ms │        75.51 / 77.72 ±2.76 / 82.93 ms │     no change │
│ QQuery 39 │     205.09 / 214.93 ±6.68 / 220.65 ms │     204.01 / 217.99 ±7.30 / 225.45 ms │     no change │
│ QQuery 40 │        23.94 / 26.46 ±2.10 / 29.25 ms │        24.44 / 27.29 ±2.09 / 29.76 ms │     no change │
│ QQuery 41 │        21.08 / 21.83 ±1.15 / 24.10 ms │        20.72 / 22.78 ±1.18 / 24.37 ms │     no change │
│ QQuery 42 │        19.61 / 20.50 ±0.73 / 21.79 ms │        19.65 / 20.83 ±1.19 / 22.99 ms │     no change │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                             ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                             │ 28626.88ms │
│ Total Time (fix-projection-embedding-order)   │ 28218.02ms │
│ Average Time (HEAD)                           │   665.74ms │
│ Average Time (fix-projection-embedding-order) │   656.23ms │
│ Queries Faster                                │          6 │
│ Queries Slower                                │          2 │
│ Queries with No Change                        │         35 │
│ Queries with Failure                          │          0 │
└───────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric	Value
Wall time	144.4s
Peak memory	39.2 GiB
Avg memory	30.4 GiB
CPU user	1334.5s
CPU sys	118.1s
Disk read	0 B
Disk write	3.8 GiB

clickbench_partitioned — branch

Metric	Value
Wall time	142.1s
Peak memory	35.9 GiB
Avg memory	26.4 GiB
CPU user	1321.9s
CPU sys	108.6s
Disk read	0 B
Disk write	120.0 KiB

File an issue against this benchmark runner

Dandandan · 2026-04-03T09:10:03Z

run benchmark sql_planner

adriangbot · 2026-04-03T09:14:30Z

🤖 Criterion benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4182658342-713-446r8 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 fix-projection-embedding-order (15bbe4a) to 1e93a67 (merge-base) diff
BENCH_NAME=sql_planner
BENCH_COMMAND=cargo bench --features=parquet --bench sql_planner
BENCH_FILTER=
Results will be posted here when complete

File an issue against this benchmark runner

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

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

…mplex expression column collection" This reverts commit 68646e8.

…omplex expression column collection" This reverts commit 8037db5.

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

adriangbot · 2026-04-03T10:10:13Z

🤖 Criterion 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

group                                                 fix-projection-embedding-order         main
-----                                                 ------------------------------         ----
logical_aggregate_with_join                           1.00    434.5±1.02µs        ? ?/sec    1.01    436.9±1.29µs        ? ?/sec
logical_plan_struct_join_agg_sort                     1.00    167.5±0.65µs        ? ?/sec    1.02    170.1±0.82µs        ? ?/sec
logical_select_all_from_1000                          1.00      7.9±0.03ms        ? ?/sec    1.00      7.9±0.03ms        ? ?/sec
logical_select_one_from_700                           1.00    321.3±1.18µs        ? ?/sec    1.00    320.4±1.17µs        ? ?/sec
logical_trivial_join_high_numbered_columns            1.00    266.3±0.58µs        ? ?/sec    1.00    267.2±0.65µs        ? ?/sec
logical_trivial_join_low_numbered_columns             1.00    254.4±0.76µs        ? ?/sec    1.01    256.0±1.73µs        ? ?/sec
physical_intersection                                 1.00    586.7±1.65µs        ? ?/sec    1.00    586.1±6.99µs        ? ?/sec
physical_join_consider_sort                           1.01   1028.3±3.70µs        ? ?/sec    1.00   1021.0±4.93µs        ? ?/sec
physical_join_distinct                                1.00    249.2±0.78µs        ? ?/sec    1.00    248.7±0.83µs        ? ?/sec
physical_many_self_joins                              1.02      7.7±0.03ms        ? ?/sec    1.00      7.6±0.01ms        ? ?/sec
physical_plan_clickbench_all                          1.00    111.3±0.24ms        ? ?/sec    1.01    112.4±1.38ms        ? ?/sec
physical_plan_clickbench_q1                           1.01  1108.5±49.74µs        ? ?/sec    1.00   1102.8±4.38µs        ? ?/sec
physical_plan_clickbench_q10                          1.02  1792.0±12.93µs        ? ?/sec    1.00   1763.1±5.61µs        ? ?/sec
physical_plan_clickbench_q11                          1.02   1912.7±6.38µs        ? ?/sec    1.00   1881.5±5.45µs        ? ?/sec
physical_plan_clickbench_q12                          1.02   1996.2±6.21µs        ? ?/sec    1.00   1965.7±8.57µs        ? ?/sec
physical_plan_clickbench_q13                          1.02   1757.7±8.20µs        ? ?/sec    1.00   1721.6±4.23µs        ? ?/sec
physical_plan_clickbench_q14                          1.01   1912.3±5.97µs        ? ?/sec    1.00   1884.8±4.55µs        ? ?/sec
physical_plan_clickbench_q15                          1.01   1807.7±6.42µs        ? ?/sec    1.00   1795.0±6.26µs        ? ?/sec
physical_plan_clickbench_q16                          1.00   1505.3±5.40µs        ? ?/sec    1.00   1498.8±8.56µs        ? ?/sec
physical_plan_clickbench_q17                          1.00  1555.7±11.12µs        ? ?/sec    1.00  1552.1±10.47µs        ? ?/sec
physical_plan_clickbench_q18                          1.00   1374.8±4.23µs        ? ?/sec    1.01   1386.3±3.59µs        ? ?/sec
physical_plan_clickbench_q19                          1.00   1755.7±9.21µs        ? ?/sec    1.00   1763.6±4.52µs        ? ?/sec
physical_plan_clickbench_q2                           1.00   1473.9±4.89µs        ? ?/sec    1.01   1487.2±4.53µs        ? ?/sec
physical_plan_clickbench_q20                          1.00   1206.3±4.15µs        ? ?/sec    1.00   1211.9±5.49µs        ? ?/sec
physical_plan_clickbench_q21                          1.00   1467.3±3.94µs        ? ?/sec    1.00   1474.5±9.13µs        ? ?/sec
physical_plan_clickbench_q22                          1.00   1865.5±4.17µs        ? ?/sec    1.00   1865.8±4.14µs        ? ?/sec
physical_plan_clickbench_q23                          1.00      2.1±0.01ms        ? ?/sec    1.00      2.1±0.01ms        ? ?/sec
physical_plan_clickbench_q24                          1.00      3.0±0.01ms        ? ?/sec    1.00      3.0±0.01ms        ? ?/sec
physical_plan_clickbench_q25                          1.01   1578.0±4.28µs        ? ?/sec    1.00   1559.6±5.24µs        ? ?/sec
physical_plan_clickbench_q26                          1.01   1417.9±9.30µs        ? ?/sec    1.00   1406.0±4.08µs        ? ?/sec
physical_plan_clickbench_q27                          1.02  1609.3±12.00µs        ? ?/sec    1.00   1577.2±5.50µs        ? ?/sec
physical_plan_clickbench_q28                          1.00      2.1±0.02ms        ? ?/sec    1.00      2.1±0.01ms        ? ?/sec
physical_plan_clickbench_q29                          1.04      2.3±0.03ms        ? ?/sec    1.00      2.2±0.01ms        ? ?/sec
physical_plan_clickbench_q3                           1.01   1391.1±5.23µs        ? ?/sec    1.00   1378.3±3.43µs        ? ?/sec
physical_plan_clickbench_q30                          1.01     16.3±0.08ms        ? ?/sec    1.00     16.2±0.06ms        ? ?/sec
physical_plan_clickbench_q31                          1.01      2.1±0.01ms        ? ?/sec    1.00      2.1±0.00ms        ? ?/sec
physical_plan_clickbench_q32                          1.00      2.2±0.02ms        ? ?/sec    1.00      2.2±0.01ms        ? ?/sec
physical_plan_clickbench_q33                          1.00  1759.5±10.04µs        ? ?/sec    1.02   1797.3±4.44µs        ? ?/sec
physical_plan_clickbench_q34                          1.00   1495.3±4.16µs        ? ?/sec    1.03  1533.1±37.72µs        ? ?/sec
physical_plan_clickbench_q35                          1.00   1561.3±4.79µs        ? ?/sec    1.03   1610.9±3.90µs        ? ?/sec
physical_plan_clickbench_q36                          1.00   1874.4±7.00µs        ? ?/sec    1.03   1927.7±5.22µs        ? ?/sec
physical_plan_clickbench_q37                          1.00      2.3±0.01ms        ? ?/sec    1.01      2.3±0.00ms        ? ?/sec
physical_plan_clickbench_q38                          1.00      2.3±0.01ms        ? ?/sec    1.02      2.3±0.01ms        ? ?/sec
physical_plan_clickbench_q39                          1.00      2.1±0.01ms        ? ?/sec    1.02      2.2±0.01ms        ? ?/sec
physical_plan_clickbench_q4                           1.01   1205.3±4.07µs        ? ?/sec    1.00   1194.0±4.21µs        ? ?/sec
physical_plan_clickbench_q40                          1.00      2.7±0.01ms        ? ?/sec    1.05      2.8±0.04ms        ? ?/sec
physical_plan_clickbench_q41                          1.00      2.3±0.03ms        ? ?/sec    1.00      2.3±0.01ms        ? ?/sec
physical_plan_clickbench_q42                          1.00      2.3±0.01ms        ? ?/sec    1.00      2.3±0.01ms        ? ?/sec
physical_plan_clickbench_q43                          1.00      2.4±0.01ms        ? ?/sec    1.02      2.5±0.01ms        ? ?/sec
physical_plan_clickbench_q44                          1.00   1322.7±3.22µs        ? ?/sec    1.00   1318.3±4.63µs        ? ?/sec
physical_plan_clickbench_q45                          1.00   1322.0±5.17µs        ? ?/sec    1.00   1319.4±4.67µs        ? ?/sec
physical_plan_clickbench_q46                          1.00   1644.3±4.40µs        ? ?/sec    1.00   1646.8±4.90µs        ? ?/sec
physical_plan_clickbench_q47                          1.03      2.4±0.01ms        ? ?/sec    1.00      2.3±0.00ms        ? ?/sec
physical_plan_clickbench_q48                          1.03      2.5±0.01ms        ? ?/sec    1.00      2.4±0.01ms        ? ?/sec
physical_plan_clickbench_q49                          1.03      2.7±0.01ms        ? ?/sec    1.00      2.6±0.01ms        ? ?/sec
physical_plan_clickbench_q5                           1.01   1336.2±4.40µs        ? ?/sec    1.00   1323.0±3.96µs        ? ?/sec
physical_plan_clickbench_q50                          1.00      2.6±0.01ms        ? ?/sec    1.04      2.6±0.06ms        ? ?/sec
physical_plan_clickbench_q51                          1.00   1762.0±5.38µs        ? ?/sec    1.00   1767.3±8.13µs        ? ?/sec
physical_plan_clickbench_q6                           1.01  1351.9±12.28µs        ? ?/sec    1.00   1338.8±8.93µs        ? ?/sec
physical_plan_clickbench_q7                           1.00   1138.8±4.91µs        ? ?/sec    1.00   1133.4±4.54µs        ? ?/sec
physical_plan_clickbench_q8                           1.00   1604.2±6.00µs        ? ?/sec    1.00   1601.1±3.98µs        ? ?/sec
physical_plan_clickbench_q9                           1.03  1696.3±17.82µs        ? ?/sec    1.00   1640.2±3.98µs        ? ?/sec
physical_plan_struct_join_agg_sort                    1.00   1334.1±3.45µs        ? ?/sec    1.00   1339.2±2.32µs        ? ?/sec
physical_plan_tpcds_all                               1.00    756.0±4.98ms        ? ?/sec    1.01    759.9±5.25ms        ? ?/sec
physical_plan_tpch_all                                1.00     47.1±0.05ms        ? ?/sec    1.01     47.7±0.11ms        ? ?/sec
physical_plan_tpch_q1                                 1.00  1511.8±12.86µs        ? ?/sec    1.00   1516.1±4.20µs        ? ?/sec
physical_plan_tpch_q10                                1.00      2.8±0.00ms        ? ?/sec    1.01      2.9±0.00ms        ? ?/sec
physical_plan_tpch_q11                                1.00      2.5±0.00ms        ? ?/sec    1.01      2.5±0.00ms        ? ?/sec
physical_plan_tpch_q12                                1.00   1272.6±2.04µs        ? ?/sec    1.00   1270.7±2.25µs        ? ?/sec
physical_plan_tpch_q13                                1.00    985.2±1.85µs        ? ?/sec    1.00    985.7±2.01µs        ? ?/sec
physical_plan_tpch_q14                                1.00   1291.1±2.70µs        ? ?/sec    1.01   1299.7±3.19µs        ? ?/sec
physical_plan_tpch_q16                                1.00   1651.5±3.58µs        ? ?/sec    1.00   1647.5±3.66µs        ? ?/sec
physical_plan_tpch_q17                                1.00   1782.2±3.80µs        ? ?/sec    1.01   1792.9±7.80µs        ? ?/sec
physical_plan_tpch_q18                                1.00   1911.5±1.93µs        ? ?/sec    1.01   1935.1±2.62µs        ? ?/sec
physical_plan_tpch_q19                                1.00      2.4±0.00ms        ? ?/sec    1.01      2.5±0.00ms        ? ?/sec
physical_plan_tpch_q2                                 1.00      4.4±0.00ms        ? ?/sec    1.01      4.4±0.01ms        ? ?/sec
physical_plan_tpch_q20                                1.00      2.2±0.00ms        ? ?/sec    1.01      2.3±0.00ms        ? ?/sec
physical_plan_tpch_q21                                1.00      3.0±0.00ms        ? ?/sec    1.00      3.0±0.01ms        ? ?/sec
physical_plan_tpch_q22                                1.00      2.0±0.00ms        ? ?/sec    1.01      2.0±0.00ms        ? ?/sec
physical_plan_tpch_q3                                 1.00   1890.7±3.22µs        ? ?/sec    1.02   1931.9±8.43µs        ? ?/sec
physical_plan_tpch_q4                                 1.00   1019.5±2.19µs        ? ?/sec    1.00   1018.8±2.02µs        ? ?/sec
physical_plan_tpch_q5                                 1.00      2.4±0.00ms        ? ?/sec    1.02      2.5±0.01ms        ? ?/sec
physical_plan_tpch_q6                                 1.00    630.7±1.56µs        ? ?/sec    1.00    631.0±2.08µs        ? ?/sec
physical_plan_tpch_q7                                 1.00      3.0±0.01ms        ? ?/sec    1.01      3.1±0.00ms        ? ?/sec
physical_plan_tpch_q8                                 1.00      4.0±0.01ms        ? ?/sec    1.01      4.1±0.01ms        ? ?/sec
physical_plan_tpch_q9                                 1.00      2.9±0.01ms        ? ?/sec    1.01      3.0±0.01ms        ? ?/sec
physical_select_aggregates_from_200                   1.00     14.7±0.05ms        ? ?/sec    1.00     14.7±0.03ms        ? ?/sec
physical_select_all_from_1000                         1.00     17.5±0.04ms        ? ?/sec    1.01     17.7±0.06ms        ? ?/sec
physical_select_one_from_700                          1.00    785.2±6.31µs        ? ?/sec    1.00    782.0±3.44µs        ? ?/sec
physical_sorted_union_order_by_10_int64               1.00      4.8±0.02ms        ? ?/sec    1.00      4.8±0.01ms        ? ?/sec
physical_sorted_union_order_by_10_uint64              1.00     11.7±0.04ms        ? ?/sec    1.01     11.7±0.04ms        ? ?/sec
physical_sorted_union_order_by_50_int64               1.00    117.4±0.56ms        ? ?/sec    1.00    117.4±0.36ms        ? ?/sec
physical_sorted_union_order_by_50_uint64              1.00    617.4±2.66ms        ? ?/sec    1.01    621.8±3.00ms        ? ?/sec
physical_theta_join_consider_sort                     1.01   1065.1±3.76µs        ? ?/sec    1.00   1056.9±1.92µs        ? ?/sec
physical_unnest_to_join                               1.00   1224.3±2.82µs        ? ?/sec    1.00   1218.4±5.14µs        ? ?/sec
physical_window_function_partition_by_12_on_values    1.00    739.9±1.91µs        ? ?/sec    1.00    738.8±2.47µs        ? ?/sec
physical_window_function_partition_by_30_on_values    1.00   1502.7±4.58µs        ? ?/sec    1.00   1499.9±4.64µs        ? ?/sec
physical_window_function_partition_by_4_on_values     1.00    435.3±1.37µs        ? ?/sec    1.00    434.5±1.27µs        ? ?/sec
physical_window_function_partition_by_7_on_values     1.00    546.9±1.49µs        ? ?/sec    1.00    544.7±2.38µs        ? ?/sec
physical_window_function_partition_by_8_on_values     1.00    590.0±1.57µs        ? ?/sec    1.00    587.1±2.75µs        ? ?/sec
with_param_values_many_columns                        1.00    458.8±2.35µs        ? ?/sec    1.01    462.9±2.37µs        ? ?/sec

Resource Usage

base (merge-base)

Metric	Value
Wall time	1261.7s
Peak memory	18.5 GiB
Avg memory	18.5 GiB
CPU user	1507.4s
CPU sys	1.7s
Disk read	0 B
Disk write	630.2 MiB

branch

Metric	Value
Wall time	1257.6s
Peak memory	18.5 GiB
Avg memory	18.5 GiB
CPU user	1503.1s
CPU sys	1.3s
Disk read	0 B
Disk write	23.7 MiB

File an issue against this benchmark runner

github-actions bot added sqllogictest SQL Logic Tests (.slt) physical-plan Changes to the physical-plan crate labels Apr 3, 2026

Dandandan force-pushed the fix-projection-embedding-order branch from b0e0c07 to 6e943ae Compare April 3, 2026 08:34

Dandandan changed the title ~~Preserve column order in projection embedding to eliminate redundant ProjectionExec~~ Eliminate more redundant ProjectionExec Apr 3, 2026

Dandandan changed the title ~~Eliminate more redundant ProjectionExec~~ Eliminate more redundant ProjectionExecs Apr 3, 2026

Dandandan force-pushed the fix-projection-embedding-order branch from 6e943ae to 50a9cb2 Compare April 3, 2026 08:45

Dandandan force-pushed the fix-projection-embedding-order branch from 50a9cb2 to 8ad047c Compare April 3, 2026 08:52

Add insta tests

15bbe4a

github-actions bot added the core Core DataFusion crate label Apr 3, 2026

Dandandan and others added 2 commits April 3, 2026 11:33

Use expression tree traversal order instead of sorting for complex ex…

68646e8

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

Revert testing submodule to master

587064a

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

Dandandan force-pushed the fix-projection-embedding-order branch from 795fa3c to 587064a Compare April 3, 2026 09:37

Revert "Use expression tree traversal order instead of sorting for co…

8037db5

…mplex expression column collection" This reverts commit 68646e8.

Dandandan and others added 2 commits April 3, 2026 11:47

Reapply "Use expression tree traversal order instead of sorting for c…

54ac038

…omplex expression column collection" This reverts commit 8037db5.

Fix clippy collapsible_if warning

c9a91d5

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

Conversation

Dandandan commented Apr 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Which issue does this PR close?

Rationale for this change

What changes are included in this PR?

How are these changes tested?

Are there any user-facing changes?

Uh oh!

Dandandan commented Apr 3, 2026

Uh oh!

adriangbot commented Apr 3, 2026

Uh oh!

adriangbot commented Apr 3, 2026

Uh oh!

adriangbot commented Apr 3, 2026

Uh oh!

adriangbot commented Apr 3, 2026

Uh oh!

adriangbot commented Apr 3, 2026

Uh oh!

adriangbot commented Apr 3, 2026

Uh oh!

Dandandan commented Apr 3, 2026

Uh oh!

adriangbot commented Apr 3, 2026

Uh oh!

adriangbot commented Apr 3, 2026

Uh oh!

adriangbot commented Apr 3, 2026

Uh oh!

adriangbot commented Apr 3, 2026

Uh oh!

adriangbot commented Apr 3, 2026

Uh oh!

adriangbot commented Apr 3, 2026

Uh oh!

Dandandan commented Apr 3, 2026

Uh oh!

adriangbot commented Apr 3, 2026

Uh oh!

adriangbot commented Apr 3, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Dandandan commented Apr 3, 2026 •

edited

Loading