perf: Implement groups accumulator count distinct primitive types by coderfender · Pull Request #21561 · apache/datafusion

coderfender · 2026-04-11T23:30:31Z

Which issue does this PR close?

Evaluate perf with group accumulators for count distinct

Closes #.

Rationale for this change

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

coderfender · 2026-04-11T23:31:17Z

Requesting to run benchmarks TPCH , TPCDS , clickbench etc to see how groups accumulator impl performs in count(distinct) use case

Dandandan · 2026-04-12T04:54:50Z

datafusion/functions-aggregate-common/src/aggregate/count_distinct/groups.rs

+    T::Native: Eq + Hash,
+{
+    /// Count distinct per group.
+    values: Vec<HashSet<T::Native, RandomState>>,


It could be Hashset<(T::Native, usize)> to avoid individual hash sets.

Usize is the group index

Yup! I am evaluating that approach but it doesn't seem to be working out so far

Local testing did yield mixed results :| so far

Dandandan · 2026-04-12T04:57:34Z

run benchmark tpch tpcds clickbench_partitiomed clickbemch_extended

adriangbot · 2026-04-12T05:00:09Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4230786454-1098-ljnsh 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 implement_groups_accumulator_count_distinct_primitive_types (f0b2a4a) to 16e578d (merge-base) diff using: tpch
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-12T05:00:42Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4230786454-1099-gqfv6 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 implement_groups_accumulator_count_distinct_primitive_types (f0b2a4a) to 16e578d (merge-base) diff using: tpcds
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-12T05:00:42Z

🤖 Criterion benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4230786454-1101-4lf7q 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 implement_groups_accumulator_count_distinct_primitive_types (f0b2a4a) to 16e578d (merge-base) diff
BENCH_NAME=clickbemch_extended
BENCH_COMMAND=cargo bench --features=parquet --bench clickbemch_extended
BENCH_FILTER=
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-12T05:00:42Z

🤖 Criterion benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4230786454-1100-wf4xk 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 implement_groups_accumulator_count_distinct_primitive_types (f0b2a4a) to 16e578d (merge-base) diff
BENCH_NAME=clickbench_partitiomed
BENCH_COMMAND=cargo bench --features=parquet --bench clickbench_partitiomed
BENCH_FILTER=
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-12T05:00:48Z

Benchmark for this request failed.

Last 20 lines of output:

Click to expand

    struct_query_sql
    substr
    substr_index
    substring
    sum
    to_char
    to_hex
    to_local_time
    to_time
    to_timestamp
    topk_aggregate
    topk_repartition
    translate
    trim
    trunc
    unhex
    upper
    uuid
    window_query_sql
    with_hashes

File an issue against this benchmark runner

adriangbot · 2026-04-12T05:00:48Z

Benchmark for this request failed.

Last 20 lines of output:

Click to expand

    struct_query_sql
    substr
    substr_index
    substring
    sum
    to_char
    to_hex
    to_local_time
    to_time
    to_timestamp
    topk_aggregate
    topk_repartition
    translate
    trim
    trunc
    unhex
    upper
    uuid
    window_query_sql
    with_hashes

File an issue against this benchmark runner

adriangbot · 2026-04-12T05:18:24Z

🤖 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 implement_groups_accumulator_count_distinct_primitive_types
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                     HEAD ┃ implement_groups_accumulator_count_distinct_primitive_types ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │              6.67 / 7.10 ±0.72 / 8.54 ms │                                 6.64 / 7.05 ±0.79 / 8.62 ms │     no change │
│ QQuery 2  │        145.52 / 146.06 ±0.55 / 146.84 ms │                           143.94 / 144.92 ±0.69 / 145.77 ms │     no change │
│ QQuery 3  │        113.10 / 113.85 ±0.57 / 114.51 ms │                           113.42 / 114.24 ±0.68 / 115.10 ms │     no change │
│ QQuery 4  │    1319.23 / 1330.98 ±11.07 / 1349.57 ms │                       1298.35 / 1327.61 ±27.90 / 1378.68 ms │     no change │
│ QQuery 5  │        172.99 / 174.44 ±1.15 / 176.35 ms │                           171.54 / 172.70 ±0.95 / 174.34 ms │     no change │
│ QQuery 6  │        851.09 / 863.36 ±6.48 / 870.22 ms │                          843.12 / 868.04 ±22.23 / 909.25 ms │     no change │
│ QQuery 7  │        340.10 / 341.59 ±1.36 / 344.13 ms │                           338.82 / 343.62 ±2.91 / 347.73 ms │     no change │
│ QQuery 8  │        115.78 / 117.00 ±0.99 / 117.90 ms │                           116.26 / 116.66 ±0.46 / 117.52 ms │     no change │
│ QQuery 9  │        101.00 / 104.78 ±3.16 / 108.12 ms │                           100.33 / 104.53 ±3.23 / 107.54 ms │     no change │
│ QQuery 10 │        104.75 / 106.03 ±0.71 / 106.81 ms │                           107.03 / 107.46 ±0.41 / 108.16 ms │     no change │
│ QQuery 11 │        899.26 / 911.99 ±7.96 / 921.52 ms │                          894.74 / 951.23 ±30.46 / 987.02 ms │     no change │
│ QQuery 12 │           44.96 / 46.16 ±0.96 / 47.38 ms │                              44.51 / 46.69 ±1.53 / 48.92 ms │     no change │
│ QQuery 13 │        399.04 / 400.05 ±0.87 / 401.68 ms │                           401.80 / 404.66 ±2.68 / 409.31 ms │     no change │
│ QQuery 14 │      997.99 / 1001.65 ±2.57 / 1004.77 ms │                          988.85 / 997.22 ±7.25 / 1009.32 ms │     no change │
│ QQuery 15 │           15.36 / 16.46 ±0.84 / 17.84 ms │                              15.01 / 16.80 ±1.12 / 18.28 ms │     no change │
│ QQuery 16 │              7.17 / 7.56 ±0.36 / 8.03 ms │                                 7.24 / 7.58 ±0.35 / 8.02 ms │     no change │
│ QQuery 17 │        227.77 / 229.20 ±0.94 / 230.58 ms │                           226.27 / 228.60 ±1.81 / 231.10 ms │     no change │
│ QQuery 18 │        126.99 / 127.64 ±0.65 / 128.49 ms │                           128.14 / 129.22 ±1.14 / 131.43 ms │     no change │
│ QQuery 19 │        155.20 / 156.58 ±1.07 / 157.69 ms │                           155.13 / 157.46 ±1.54 / 159.74 ms │     no change │
│ QQuery 20 │           13.85 / 14.29 ±0.26 / 14.59 ms │                              13.96 / 14.35 ±0.41 / 14.90 ms │     no change │
│ QQuery 21 │           19.10 / 19.48 ±0.33 / 19.88 ms │                              19.26 / 19.91 ±0.52 / 20.65 ms │     no change │
│ QQuery 22 │        484.06 / 487.99 ±2.62 / 492.04 ms │                           483.72 / 489.34 ±3.39 / 493.61 ms │     no change │
│ QQuery 23 │        852.27 / 864.37 ±7.39 / 873.54 ms │                           864.44 / 869.10 ±5.15 / 878.77 ms │     no change │
│ QQuery 24 │        374.54 / 378.04 ±3.53 / 384.26 ms │                           380.07 / 382.78 ±3.03 / 387.21 ms │     no change │
│ QQuery 25 │        337.29 / 340.29 ±2.27 / 343.81 ms │                           335.49 / 340.57 ±3.40 / 346.01 ms │     no change │
│ QQuery 26 │           80.06 / 81.81 ±2.20 / 86.02 ms │                              80.93 / 82.63 ±1.79 / 86.05 ms │     no change │
│ QQuery 27 │              7.00 / 7.08 ±0.05 / 7.14 ms │                                 6.78 / 7.25 ±0.41 / 7.98 ms │     no change │
│ QQuery 28 │        147.46 / 149.05 ±0.84 / 149.85 ms │                           149.12 / 150.47 ±1.22 / 152.38 ms │     no change │
│ QQuery 29 │        278.04 / 280.35 ±1.45 / 282.24 ms │                           278.34 / 281.08 ±1.50 / 282.75 ms │     no change │
│ QQuery 30 │           42.76 / 45.29 ±1.85 / 48.14 ms │                              43.36 / 44.93 ±1.35 / 47.38 ms │     no change │
│ QQuery 31 │        170.17 / 171.77 ±1.08 / 173.20 ms │                           169.10 / 171.27 ±1.54 / 172.88 ms │     no change │
│ QQuery 32 │           57.04 / 58.66 ±1.27 / 60.86 ms │                              56.69 / 57.35 ±0.70 / 58.22 ms │     no change │
│ QQuery 33 │        139.04 / 141.18 ±1.80 / 143.48 ms │                           139.62 / 141.68 ±2.04 / 145.24 ms │     no change │
│ QQuery 34 │              6.92 / 7.17 ±0.22 / 7.58 ms │                                 6.88 / 7.13 ±0.26 / 7.61 ms │     no change │
│ QQuery 35 │        108.29 / 108.76 ±0.68 / 110.08 ms │                           102.84 / 106.97 ±2.28 / 109.24 ms │     no change │
│ QQuery 36 │              6.39 / 6.65 ±0.26 / 7.15 ms │                                 6.36 / 6.57 ±0.29 / 7.14 ms │     no change │
│ QQuery 37 │              8.11 / 8.42 ±0.23 / 8.69 ms │                                 8.26 / 8.39 ±0.12 / 8.54 ms │     no change │
│ QQuery 38 │           82.84 / 86.11 ±3.37 / 92.29 ms │                              83.54 / 86.17 ±3.15 / 92.17 ms │     no change │
│ QQuery 39 │        122.37 / 124.22 ±1.69 / 127.37 ms │                           123.58 / 126.42 ±2.09 / 128.87 ms │     no change │
│ QQuery 40 │        108.73 / 113.81 ±8.15 / 130.04 ms │                           108.47 / 117.69 ±8.07 / 132.57 ms │     no change │
│ QQuery 41 │           14.86 / 15.71 ±0.76 / 17.12 ms │                              14.81 / 15.91 ±1.00 / 17.42 ms │     no change │
│ QQuery 42 │        107.91 / 109.56 ±0.99 / 110.87 ms │                           106.94 / 108.93 ±1.61 / 111.41 ms │     no change │
│ QQuery 43 │              5.85 / 6.03 ±0.18 / 6.37 ms │                                 6.18 / 6.44 ±0.23 / 6.74 ms │  1.07x slower │
│ QQuery 44 │           11.44 / 11.69 ±0.16 / 11.86 ms │                              11.94 / 12.67 ±0.99 / 14.54 ms │  1.08x slower │
│ QQuery 45 │           49.27 / 51.05 ±0.98 / 52.14 ms │                              49.28 / 51.37 ±1.57 / 53.21 ms │     no change │
│ QQuery 46 │              8.45 / 8.86 ±0.45 / 9.73 ms │                                 8.16 / 8.74 ±0.55 / 9.73 ms │     no change │
│ QQuery 47 │        703.68 / 717.37 ±7.63 / 724.32 ms │                          688.31 / 702.25 ±10.98 / 718.99 ms │     no change │
│ QQuery 48 │        283.48 / 290.31 ±4.54 / 297.33 ms │                           291.43 / 295.22 ±2.69 / 299.51 ms │     no change │
│ QQuery 49 │        247.34 / 251.04 ±2.12 / 253.23 ms │                           248.25 / 250.65 ±1.66 / 253.40 ms │     no change │
│ QQuery 50 │        221.95 / 225.40 ±3.91 / 232.82 ms │                           217.23 / 222.68 ±3.36 / 227.71 ms │     no change │
│ QQuery 51 │        179.02 / 180.58 ±1.30 / 182.02 ms │                           180.58 / 183.96 ±2.91 / 188.56 ms │     no change │
│ QQuery 52 │        108.24 / 109.22 ±0.78 / 110.44 ms │                           108.12 / 109.01 ±0.73 / 110.21 ms │     no change │
│ QQuery 53 │        102.42 / 103.80 ±1.05 / 105.21 ms │                           101.58 / 103.28 ±1.38 / 105.02 ms │     no change │
│ QQuery 54 │        145.77 / 146.83 ±0.87 / 148.30 ms │                           145.73 / 147.44 ±1.20 / 149.27 ms │     no change │
│ QQuery 55 │        107.32 / 108.82 ±0.89 / 109.56 ms │                           106.60 / 107.46 ±0.71 / 108.29 ms │     no change │
│ QQuery 56 │        140.13 / 141.85 ±1.54 / 143.77 ms │                           138.89 / 141.09 ±1.37 / 142.88 ms │     no change │
│ QQuery 57 │        176.23 / 177.21 ±0.90 / 178.69 ms │                           174.08 / 175.85 ±1.38 / 178.30 ms │     no change │
│ QQuery 58 │        291.30 / 301.87 ±7.33 / 313.62 ms │                           300.86 / 303.47 ±2.65 / 308.42 ms │     no change │
│ QQuery 59 │        198.28 / 199.39 ±0.99 / 201.25 ms │                           200.47 / 204.11 ±3.12 / 209.38 ms │     no change │
│ QQuery 60 │        143.19 / 144.22 ±0.78 / 145.10 ms │                           143.39 / 144.69 ±0.98 / 146.32 ms │     no change │
│ QQuery 61 │           12.93 / 13.13 ±0.21 / 13.51 ms │                              13.23 / 13.58 ±0.34 / 14.17 ms │     no change │
│ QQuery 62 │       924.37 / 946.25 ±22.24 / 986.54 ms │                          890.22 / 919.10 ±18.71 / 938.71 ms │     no change │
│ QQuery 63 │        104.45 / 106.49 ±2.48 / 111.29 ms │                           102.29 / 103.86 ±1.04 / 105.10 ms │     no change │
│ QQuery 64 │        672.07 / 678.36 ±4.09 / 684.89 ms │                           679.64 / 685.80 ±4.04 / 691.82 ms │     no change │
│ QQuery 65 │        242.44 / 250.34 ±4.95 / 257.89 ms │                           248.45 / 251.95 ±3.53 / 258.50 ms │     no change │
│ QQuery 66 │        240.40 / 248.97 ±5.78 / 257.76 ms │                           241.48 / 254.93 ±9.99 / 266.99 ms │     no change │
│ QQuery 67 │        305.04 / 311.65 ±7.58 / 326.41 ms │                           312.94 / 321.02 ±7.59 / 332.20 ms │     no change │
│ QQuery 68 │           10.19 / 11.21 ±0.73 / 12.39 ms │                               8.68 / 10.29 ±1.69 / 13.12 ms │ +1.09x faster │
│ QQuery 69 │        101.77 / 103.42 ±1.75 / 106.28 ms │                           103.24 / 104.80 ±0.87 / 105.72 ms │     no change │
│ QQuery 70 │       336.64 / 347.16 ±11.48 / 368.52 ms │                          320.54 / 346.05 ±15.58 / 367.90 ms │     no change │
│ QQuery 71 │        133.27 / 136.10 ±1.84 / 138.54 ms │                           135.97 / 138.89 ±3.04 / 144.38 ms │     no change │
│ QQuery 72 │        611.94 / 617.05 ±4.48 / 624.93 ms │                           596.60 / 611.64 ±8.52 / 622.63 ms │     no change │
│ QQuery 73 │              6.60 / 7.92 ±0.72 / 8.74 ms │                                 7.03 / 7.94 ±0.87 / 9.27 ms │     no change │
│ QQuery 74 │        564.08 / 567.88 ±4.30 / 576.11 ms │                           553.35 / 563.86 ±9.53 / 579.02 ms │     no change │
│ QQuery 75 │        273.98 / 275.96 ±1.76 / 279.23 ms │                           275.31 / 277.50 ±1.68 / 279.83 ms │     no change │
│ QQuery 76 │        130.69 / 132.45 ±1.59 / 135.38 ms │                           131.03 / 131.88 ±0.65 / 132.88 ms │     no change │
│ QQuery 77 │        186.35 / 189.20 ±2.45 / 192.41 ms │                           188.41 / 190.03 ±0.84 / 190.80 ms │     no change │
│ QQuery 78 │        337.21 / 340.33 ±2.13 / 342.86 ms │                           335.72 / 337.13 ±1.56 / 339.94 ms │     no change │
│ QQuery 79 │        233.14 / 234.52 ±1.12 / 236.01 ms │                           230.24 / 234.02 ±1.99 / 235.88 ms │     no change │
│ QQuery 80 │        319.02 / 320.94 ±1.60 / 323.86 ms │                           319.97 / 324.30 ±3.86 / 329.96 ms │     no change │
│ QQuery 81 │           25.97 / 27.33 ±1.22 / 29.27 ms │                              26.11 / 26.62 ±0.42 / 27.30 ms │     no change │
│ QQuery 82 │        197.31 / 199.34 ±2.21 / 203.33 ms │                           197.15 / 199.69 ±2.47 / 204.18 ms │     no change │
│ QQuery 83 │           37.65 / 39.42 ±1.19 / 40.92 ms │                              38.74 / 39.71 ±0.86 / 41.22 ms │     no change │
│ QQuery 84 │           47.28 / 47.82 ±0.50 / 48.67 ms │                              47.82 / 48.99 ±0.79 / 50.20 ms │     no change │
│ QQuery 85 │        144.83 / 148.70 ±2.09 / 151.12 ms │                           148.28 / 150.12 ±1.76 / 153.48 ms │     no change │
│ QQuery 86 │           38.19 / 39.35 ±0.75 / 40.18 ms │                              39.43 / 40.19 ±0.48 / 40.64 ms │     no change │
│ QQuery 87 │           85.08 / 88.06 ±3.53 / 94.97 ms │                              80.11 / 86.50 ±4.50 / 93.88 ms │     no change │
│ QQuery 88 │         99.71 / 100.70 ±0.69 / 101.64 ms │                            99.49 / 100.48 ±1.16 / 102.71 ms │     no change │
│ QQuery 89 │        118.40 / 119.78 ±1.15 / 121.33 ms │                           118.44 / 119.35 ±0.94 / 120.99 ms │     no change │
│ QQuery 90 │           22.54 / 23.74 ±1.11 / 25.78 ms │                              23.26 / 24.18 ±0.68 / 25.36 ms │     no change │
│ QQuery 91 │           61.81 / 64.31 ±1.34 / 65.62 ms │                              64.07 / 64.48 ±0.39 / 64.96 ms │     no change │
│ QQuery 92 │           57.05 / 57.91 ±0.99 / 59.64 ms │                              56.16 / 57.74 ±1.00 / 59.07 ms │     no change │
│ QQuery 93 │        185.01 / 186.26 ±0.77 / 187.21 ms │                           183.50 / 186.60 ±1.61 / 188.15 ms │     no change │
│ QQuery 94 │           61.08 / 61.52 ±0.38 / 62.17 ms │                              61.35 / 62.19 ±0.77 / 63.53 ms │     no change │
│ QQuery 95 │        126.61 / 127.67 ±0.63 / 128.42 ms │                           126.61 / 128.23 ±1.52 / 131.03 ms │     no change │
│ QQuery 96 │           70.84 / 73.27 ±1.35 / 74.72 ms │                              73.14 / 74.07 ±1.03 / 75.65 ms │     no change │
│ QQuery 97 │        123.94 / 126.62 ±1.80 / 128.70 ms │                           125.18 / 127.21 ±1.56 / 129.10 ms │     no change │
│ QQuery 98 │        153.15 / 155.09 ±1.28 / 156.86 ms │                           151.91 / 153.64 ±1.54 / 156.30 ms │     no change │
│ QQuery 99 │ 10769.94 / 10822.64 ±39.76 / 10893.40 ms │                    10767.18 / 10783.97 ±13.36 / 10799.71 ms │     no change │
└───────────┴──────────────────────────────────────────┴─────────────────────────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                                                          ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                                                          │ 31346.92ms │
│ Total Time (implement_groups_accumulator_count_distinct_primitive_types)   │ 31356.03ms │
│ Average Time (HEAD)                                                        │   316.64ms │
│ Average Time (implement_groups_accumulator_count_distinct_primitive_types) │   316.73ms │
│ Queries Faster                                                             │          1 │
│ Queries Slower                                                             │          2 │
│ Queries with No Change                                                     │         96 │
│ Queries with Failure                                                       │          0 │
└────────────────────────────────────────────────────────────────────────────┴────────────┘

Resource Usage

tpcds — base (merge-base)

Metric	Value
Wall time	157.1s
Peak memory	5.7 GiB
Avg memory	4.7 GiB
CPU user	258.2s
CPU sys	16.8s
Peak spill	0 B

tpcds — branch

Metric	Value
Wall time	157.0s
Peak memory	5.4 GiB
Avg memory	4.4 GiB
CPU user	258.4s
CPU sys	17.0s
Peak spill	0 B

File an issue against this benchmark runner

coderfender · 2026-04-12T06:15:27Z

@Dandandan could you retrigger the benchmarks again please?

Dandandan · 2026-04-12T08:51:34Z

run benchmark tpch tpcds clickbench_partitioned clickbench_extended

adriangbot · 2026-04-12T08:54:08Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4231151449-1102-g9w92 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 implement_groups_accumulator_count_distinct_primitive_types (ee0c865) to 16e578d (merge-base) diff using: tpch
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-12T08:54:11Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4231151449-1105-xpttn 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 implement_groups_accumulator_count_distinct_primitive_types (ee0c865) to 16e578d (merge-base) diff using: clickbench_extended
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-12T08:54:12Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4231151449-1104-k4vbw 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 implement_groups_accumulator_count_distinct_primitive_types (ee0c865) to 16e578d (merge-base) diff using: clickbench_partitioned
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-12T08:54:16Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4231151449-1103-pfrtl 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 implement_groups_accumulator_count_distinct_primitive_types (ee0c865) to 16e578d (merge-base) diff using: tpcds
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-12T09:12:41Z

🤖 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 implement_groups_accumulator_count_distinct_primitive_types
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                     HEAD ┃ implement_groups_accumulator_count_distinct_primitive_types ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │              6.56 / 7.04 ±0.81 / 8.66 ms │                                 6.75 / 7.24 ±0.73 / 8.68 ms │     no change │
│ QQuery 2  │        146.79 / 147.78 ±0.80 / 148.83 ms │                           145.30 / 146.68 ±1.17 / 148.38 ms │     no change │
│ QQuery 3  │        112.83 / 113.80 ±0.64 / 114.70 ms │                           114.32 / 115.24 ±0.88 / 116.68 ms │     no change │
│ QQuery 4  │    1355.07 / 1386.38 ±22.59 / 1418.44 ms │                       1387.67 / 1407.95 ±14.22 / 1427.10 ms │     no change │
│ QQuery 5  │        171.43 / 172.73 ±1.42 / 175.49 ms │                           173.06 / 175.02 ±2.22 / 178.69 ms │     no change │
│ QQuery 6  │       822.98 / 861.15 ±24.96 / 892.09 ms │                          855.97 / 879.77 ±17.26 / 902.44 ms │     no change │
│ QQuery 7  │        336.33 / 344.51 ±4.31 / 348.02 ms │                           343.05 / 345.13 ±2.14 / 348.89 ms │     no change │
│ QQuery 8  │        115.20 / 116.50 ±1.08 / 118.21 ms │                           118.51 / 119.15 ±0.50 / 119.82 ms │     no change │
│ QQuery 9  │        101.46 / 110.46 ±9.91 / 129.78 ms │                           100.83 / 103.72 ±3.07 / 107.51 ms │ +1.06x faster │
│ QQuery 10 │        107.15 / 108.22 ±0.81 / 109.14 ms │                           106.52 / 108.09 ±1.01 / 109.41 ms │     no change │
│ QQuery 11 │       963.92 / 976.73 ±10.67 / 995.75 ms │                          982.23 / 992.76 ±8.92 / 1005.70 ms │     no change │
│ QQuery 12 │           44.90 / 46.66 ±1.40 / 48.80 ms │                              46.89 / 47.49 ±0.56 / 48.51 ms │     no change │
│ QQuery 13 │        402.02 / 405.58 ±2.08 / 407.31 ms │                           402.47 / 404.05 ±1.32 / 406.28 ms │     no change │
│ QQuery 14 │     1018.49 / 1020.69 ±2.09 / 1024.38 ms │                        1003.19 / 1012.03 ±6.41 / 1020.19 ms │     no change │
│ QQuery 15 │           15.95 / 16.64 ±0.56 / 17.40 ms │                              16.40 / 17.23 ±0.81 / 18.62 ms │     no change │
│ QQuery 16 │              7.20 / 7.54 ±0.25 / 7.87 ms │                                 7.66 / 8.22 ±0.53 / 9.20 ms │  1.09x slower │
│ QQuery 17 │       229.34 / 247.33 ±34.46 / 316.25 ms │                           226.53 / 229.77 ±3.28 / 235.79 ms │ +1.08x faster │
│ QQuery 18 │        126.95 / 129.19 ±1.47 / 131.24 ms │                           126.98 / 128.99 ±1.21 / 130.34 ms │     no change │
│ QQuery 19 │        154.37 / 157.10 ±1.67 / 158.90 ms │                           156.33 / 158.27 ±1.72 / 161.28 ms │     no change │
│ QQuery 20 │           14.11 / 14.64 ±0.39 / 15.30 ms │                              14.05 / 14.66 ±0.37 / 15.10 ms │     no change │
│ QQuery 21 │           19.34 / 19.69 ±0.27 / 20.04 ms │                              19.88 / 20.40 ±0.34 / 20.88 ms │     no change │
│ QQuery 22 │        486.43 / 491.90 ±3.12 / 495.01 ms │                           489.28 / 495.17 ±4.93 / 501.65 ms │     no change │
│ QQuery 23 │        890.26 / 901.68 ±7.33 / 910.76 ms │                          905.74 / 918.25 ±10.42 / 933.82 ms │     no change │
│ QQuery 24 │        385.36 / 386.52 ±0.84 / 387.33 ms │                           391.48 / 392.49 ±1.08 / 394.38 ms │     no change │
│ QQuery 25 │        341.38 / 344.73 ±2.30 / 348.60 ms │                           347.63 / 350.84 ±2.54 / 354.31 ms │     no change │
│ QQuery 26 │           80.94 / 81.96 ±0.70 / 83.07 ms │                              81.84 / 84.34 ±2.04 / 87.71 ms │     no change │
│ QQuery 27 │              7.39 / 8.00 ±0.70 / 9.33 ms │                                 7.34 / 8.02 ±0.67 / 9.06 ms │     no change │
│ QQuery 28 │        147.68 / 149.53 ±1.24 / 151.11 ms │                           152.99 / 154.48 ±0.94 / 155.83 ms │     no change │
│ QQuery 29 │        282.96 / 284.16 ±0.89 / 285.42 ms │                           287.39 / 291.18 ±2.42 / 294.05 ms │     no change │
│ QQuery 30 │           43.79 / 46.13 ±1.77 / 48.27 ms │                              44.52 / 46.55 ±2.30 / 50.98 ms │     no change │
│ QQuery 31 │        173.05 / 174.28 ±1.25 / 176.62 ms │                           172.16 / 174.64 ±1.25 / 175.40 ms │     no change │
│ QQuery 32 │           57.09 / 58.61 ±0.86 / 59.67 ms │                              58.03 / 58.87 ±1.03 / 60.89 ms │     no change │
│ QQuery 33 │        140.94 / 141.80 ±0.82 / 143.04 ms │                           140.84 / 143.32 ±1.59 / 145.86 ms │     no change │
│ QQuery 34 │              6.94 / 7.11 ±0.23 / 7.57 ms │                                 7.29 / 7.50 ±0.22 / 7.90 ms │  1.06x slower │
│ QQuery 35 │        109.60 / 110.36 ±0.97 / 112.18 ms │                           108.19 / 109.71 ±0.88 / 110.93 ms │     no change │
│ QQuery 36 │              6.73 / 6.91 ±0.19 / 7.22 ms │                                 6.81 / 7.03 ±0.14 / 7.23 ms │     no change │
│ QQuery 37 │              8.54 / 8.85 ±0.28 / 9.29 ms │                                8.69 / 9.22 ±0.55 / 10.21 ms │     no change │
│ QQuery 38 │           83.80 / 87.89 ±4.64 / 95.78 ms │                              86.78 / 90.82 ±2.86 / 95.20 ms │     no change │
│ QQuery 39 │        126.74 / 129.75 ±1.91 / 132.09 ms │                           130.69 / 132.81 ±1.20 / 134.36 ms │     no change │
│ QQuery 40 │        109.11 / 115.65 ±8.54 / 132.38 ms │                           110.99 / 118.36 ±7.56 / 132.66 ms │     no change │
│ QQuery 41 │           14.74 / 16.11 ±0.92 / 16.98 ms │                              14.72 / 15.91 ±0.81 / 17.09 ms │     no change │
│ QQuery 42 │        107.00 / 109.18 ±1.35 / 110.73 ms │                           109.50 / 111.60 ±1.10 / 112.60 ms │     no change │
│ QQuery 43 │              6.02 / 6.16 ±0.14 / 6.42 ms │                                 6.19 / 6.81 ±0.74 / 8.23 ms │  1.11x slower │
│ QQuery 44 │           11.86 / 12.01 ±0.09 / 12.15 ms │                              11.83 / 12.06 ±0.18 / 12.38 ms │     no change │
│ QQuery 45 │           50.30 / 51.50 ±1.12 / 53.40 ms │                              51.51 / 52.30 ±0.45 / 52.75 ms │     no change │
│ QQuery 46 │             8.56 / 9.40 ±1.01 / 10.97 ms │                                 9.00 / 9.33 ±0.28 / 9.67 ms │     no change │
│ QQuery 47 │        768.50 / 773.33 ±2.80 / 777.07 ms │                          752.92 / 775.27 ±19.36 / 800.53 ms │     no change │
│ QQuery 48 │        293.06 / 297.40 ±2.65 / 300.94 ms │                           287.55 / 292.91 ±4.07 / 297.77 ms │     no change │
│ QQuery 49 │        253.31 / 256.07 ±2.92 / 260.77 ms │                           250.51 / 253.13 ±1.95 / 255.39 ms │     no change │
│ QQuery 50 │        222.76 / 230.07 ±4.14 / 234.72 ms │                           222.90 / 227.82 ±5.25 / 235.67 ms │     no change │
│ QQuery 51 │        179.81 / 183.93 ±4.16 / 191.83 ms │                           182.58 / 184.11 ±1.73 / 187.33 ms │     no change │
│ QQuery 52 │        107.73 / 109.29 ±1.45 / 112.01 ms │                           108.63 / 109.37 ±0.54 / 110.01 ms │     no change │
│ QQuery 53 │        102.19 / 102.78 ±0.49 / 103.41 ms │                           103.18 / 104.74 ±1.18 / 106.24 ms │     no change │
│ QQuery 54 │        147.94 / 149.30 ±1.45 / 152.08 ms │                           148.08 / 149.37 ±1.28 / 151.10 ms │     no change │
│ QQuery 55 │        107.05 / 108.83 ±1.65 / 111.72 ms │                           108.14 / 108.85 ±0.48 / 109.46 ms │     no change │
│ QQuery 56 │        141.63 / 143.61 ±1.64 / 145.88 ms │                           141.42 / 144.51 ±1.73 / 146.48 ms │     no change │
│ QQuery 57 │        177.53 / 179.24 ±1.83 / 182.61 ms │                           177.62 / 179.04 ±1.01 / 180.34 ms │     no change │
│ QQuery 58 │       290.03 / 305.28 ±11.56 / 324.22 ms │                           305.71 / 308.03 ±3.06 / 313.94 ms │     no change │
│ QQuery 59 │        201.59 / 204.43 ±2.33 / 208.66 ms │                           199.37 / 202.00 ±1.96 / 204.23 ms │     no change │
│ QQuery 60 │        143.53 / 146.09 ±1.72 / 148.19 ms │                           143.10 / 145.99 ±1.67 / 148.14 ms │     no change │
│ QQuery 61 │           13.41 / 13.61 ±0.23 / 14.05 ms │                              13.40 / 13.59 ±0.13 / 13.79 ms │     no change │
│ QQuery 62 │      903.50 / 953.44 ±49.24 / 1047.66 ms │                          914.91 / 927.28 ±12.48 / 950.56 ms │     no change │
│ QQuery 63 │        102.20 / 104.71 ±1.95 / 107.82 ms │                           103.30 / 105.81 ±2.47 / 109.89 ms │     no change │
│ QQuery 64 │        705.67 / 712.59 ±4.83 / 720.36 ms │                           683.32 / 688.30 ±3.81 / 693.97 ms │     no change │
│ QQuery 65 │        267.45 / 273.67 ±4.93 / 279.94 ms │                           252.66 / 259.09 ±3.85 / 263.97 ms │ +1.06x faster │
│ QQuery 66 │        259.19 / 268.79 ±7.35 / 281.14 ms │                           246.04 / 254.69 ±5.76 / 262.51 ms │ +1.06x faster │
│ QQuery 67 │        327.73 / 332.44 ±5.09 / 342.22 ms │                           313.20 / 323.17 ±7.81 / 336.86 ms │     no change │
│ QQuery 68 │            9.24 / 10.87 ±2.07 / 14.67 ms │                               9.22 / 10.46 ±0.96 / 11.85 ms │     no change │
│ QQuery 69 │        105.20 / 107.13 ±2.23 / 110.42 ms │                           103.26 / 104.67 ±1.03 / 105.87 ms │     no change │
│ QQuery 70 │        348.44 / 356.46 ±8.54 / 371.33 ms │                          337.04 / 349.46 ±10.31 / 367.83 ms │     no change │
│ QQuery 71 │        137.25 / 140.12 ±2.22 / 142.78 ms │                           135.23 / 139.57 ±4.90 / 148.72 ms │     no change │
│ QQuery 72 │       623.65 / 639.42 ±12.11 / 659.85 ms │                           613.78 / 624.57 ±8.53 / 637.95 ms │     no change │
│ QQuery 73 │             7.63 / 8.96 ±1.58 / 11.99 ms │                                 7.66 / 8.64 ±0.83 / 9.89 ms │     no change │
│ QQuery 74 │        656.28 / 660.71 ±4.14 / 667.54 ms │                           613.47 / 618.70 ±3.59 / 623.77 ms │ +1.07x faster │
│ QQuery 75 │        277.46 / 281.99 ±3.21 / 286.41 ms │                           275.09 / 278.62 ±3.47 / 284.19 ms │     no change │
│ QQuery 76 │        133.88 / 135.43 ±1.20 / 137.16 ms │                           131.75 / 133.00 ±1.04 / 134.15 ms │     no change │
│ QQuery 77 │        189.77 / 191.55 ±1.07 / 192.67 ms │                           188.97 / 190.48 ±1.55 / 192.93 ms │     no change │
│ QQuery 78 │        349.76 / 353.37 ±2.26 / 356.73 ms │                           340.58 / 343.64 ±2.40 / 346.68 ms │     no change │
│ QQuery 79 │        244.40 / 250.25 ±3.09 / 253.65 ms │                           236.43 / 240.86 ±3.40 / 245.58 ms │     no change │
│ QQuery 80 │        320.86 / 324.67 ±2.69 / 328.74 ms │                           320.20 / 321.85 ±1.09 / 323.32 ms │     no change │
│ QQuery 81 │           26.94 / 28.46 ±1.73 / 31.70 ms │                              26.34 / 27.17 ±0.80 / 28.21 ms │     no change │
│ QQuery 82 │        197.05 / 200.03 ±2.61 / 204.23 ms │                           198.51 / 201.49 ±2.08 / 203.59 ms │     no change │
│ QQuery 83 │           39.34 / 40.54 ±1.04 / 42.09 ms │                              38.29 / 40.61 ±1.22 / 41.59 ms │     no change │
│ QQuery 84 │           49.78 / 50.49 ±0.40 / 51.01 ms │                              48.06 / 49.10 ±0.79 / 50.41 ms │     no change │
│ QQuery 85 │        148.88 / 151.06 ±1.42 / 153.22 ms │                           146.31 / 148.69 ±1.41 / 150.57 ms │     no change │
│ QQuery 86 │           40.10 / 40.54 ±0.46 / 41.27 ms │                              39.31 / 40.77 ±0.98 / 42.31 ms │     no change │
│ QQuery 87 │           83.85 / 89.44 ±3.64 / 95.11 ms │                              87.71 / 92.19 ±3.33 / 96.11 ms │     no change │
│ QQuery 88 │        101.60 / 103.05 ±1.18 / 104.85 ms │                           101.14 / 102.34 ±0.82 / 103.52 ms │     no change │
│ QQuery 89 │        119.34 / 121.14 ±1.24 / 123.05 ms │                           119.26 / 120.85 ±1.38 / 122.75 ms │     no change │
│ QQuery 90 │           23.87 / 24.33 ±0.35 / 24.70 ms │                              23.74 / 24.48 ±0.73 / 25.71 ms │     no change │
│ QQuery 91 │           63.81 / 64.62 ±0.70 / 65.92 ms │                              63.43 / 64.76 ±0.97 / 66.28 ms │     no change │
│ QQuery 92 │           58.08 / 59.78 ±1.20 / 60.84 ms │                              57.97 / 59.04 ±0.72 / 60.04 ms │     no change │
│ QQuery 93 │        190.00 / 191.89 ±1.35 / 193.80 ms │                           187.83 / 188.57 ±0.72 / 189.83 ms │     no change │
│ QQuery 94 │           62.48 / 62.95 ±0.37 / 63.46 ms │                              61.97 / 62.46 ±0.39 / 63.09 ms │     no change │
│ QQuery 95 │        129.61 / 130.74 ±1.06 / 132.73 ms │                           129.73 / 131.13 ±1.48 / 133.76 ms │     no change │
│ QQuery 96 │           74.50 / 75.63 ±0.83 / 76.32 ms │                              75.08 / 75.89 ±0.82 / 77.45 ms │     no change │
│ QQuery 97 │        126.31 / 129.58 ±1.75 / 131.47 ms │                           126.86 / 128.48 ±1.37 / 130.20 ms │     no change │
│ QQuery 98 │        155.17 / 159.14 ±2.16 / 161.08 ms │                           155.06 / 156.75 ±1.81 / 159.68 ms │     no change │
│ QQuery 99 │ 10825.07 / 10905.21 ±44.63 / 10953.01 ms │                    10740.37 / 10800.10 ±38.46 / 10842.77 ms │     no change │
└───────────┴──────────────────────────────────────────┴─────────────────────────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                                                          ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                                                          │ 32088.92ms │
│ Total Time (implement_groups_accumulator_count_distinct_primitive_types)   │ 31879.92ms │
│ Average Time (HEAD)                                                        │   324.13ms │
│ Average Time (implement_groups_accumulator_count_distinct_primitive_types) │   322.02ms │
│ Queries Faster                                                             │          5 │
│ Queries Slower                                                             │          3 │
│ Queries with No Change                                                     │         91 │
│ Queries with Failure                                                       │          0 │
└────────────────────────────────────────────────────────────────────────────┴────────────┘

Resource Usage

tpcds — base (merge-base)

Metric	Value
Wall time	160.8s
Peak memory	5.6 GiB
Avg memory	4.6 GiB
CPU user	266.1s
CPU sys	17.6s
Peak spill	0 B

tpcds — branch

Metric	Value
Wall time	159.7s
Peak memory	5.6 GiB
Avg memory	4.7 GiB
CPU user	265.0s
CPU sys	17.0s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-04-12T09:13:14Z

🤖 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 implement_groups_accumulator_count_distinct_primitive_types
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                   HEAD ┃ implement_groups_accumulator_count_distinct_primitive_types ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │           1.18 / 4.43 ±6.40 / 17.24 ms │                                1.22 / 4.42 ±6.33 / 17.08 ms │     no change │
│ QQuery 1  │         14.14 / 14.35 ±0.12 / 14.46 ms │                              14.19 / 14.40 ±0.14 / 14.60 ms │     no change │
│ QQuery 2  │         42.96 / 43.33 ±0.26 / 43.76 ms │                              43.29 / 43.61 ±0.19 / 43.86 ms │     no change │
│ QQuery 3  │         43.20 / 46.58 ±2.17 / 49.34 ms │                              43.61 / 45.22 ±0.84 / 45.94 ms │     no change │
│ QQuery 4  │      279.62 / 285.18 ±3.85 / 291.38 ms │                           282.80 / 286.99 ±3.55 / 293.04 ms │     no change │
│ QQuery 5  │      335.55 / 339.28 ±3.26 / 343.68 ms │                           339.51 / 342.90 ±3.60 / 347.58 ms │     no change │
│ QQuery 6  │            5.41 / 5.84 ±0.32 / 6.15 ms │                                5.35 / 7.00 ±2.10 / 11.10 ms │  1.20x slower │
│ QQuery 7  │         16.62 / 16.75 ±0.11 / 16.89 ms │                              16.31 / 18.35 ±2.80 / 23.79 ms │  1.10x slower │
│ QQuery 8  │      404.05 / 415.63 ±8.26 / 423.57 ms │                           408.26 / 415.35 ±6.32 / 426.82 ms │     no change │
│ QQuery 9  │     619.95 / 637.18 ±12.54 / 654.86 ms │                           589.90 / 596.08 ±3.37 / 599.61 ms │ +1.07x faster │
│ QQuery 10 │         91.77 / 94.44 ±2.32 / 98.04 ms │                              90.59 / 92.41 ±2.02 / 96.15 ms │     no change │
│ QQuery 11 │      103.91 / 105.27 ±1.43 / 107.90 ms │                           101.47 / 102.56 ±0.71 / 103.60 ms │     no change │
│ QQuery 12 │      336.25 / 340.52 ±3.39 / 346.20 ms │                          327.77 / 345.68 ±19.23 / 381.56 ms │     no change │
│ QQuery 13 │      442.42 / 458.16 ±8.64 / 466.84 ms │                          448.85 / 473.02 ±20.35 / 508.42 ms │     no change │
│ QQuery 14 │      333.66 / 342.27 ±4.75 / 348.25 ms │                           341.73 / 347.10 ±4.12 / 353.42 ms │     no change │
│ QQuery 15 │      343.71 / 355.88 ±8.85 / 366.32 ms │                          349.27 / 359.30 ±13.24 / 382.96 ms │     no change │
│ QQuery 16 │      720.11 / 730.36 ±8.27 / 743.79 ms │                          728.67 / 747.74 ±14.48 / 772.58 ms │     no change │
│ QQuery 17 │      706.95 / 715.76 ±7.32 / 728.39 ms │                          695.30 / 727.76 ±23.16 / 750.61 ms │     no change │
│ QQuery 18 │  1437.57 / 1488.96 ±29.05 / 1521.91 ms │                       1338.52 / 1352.15 ±11.03 / 1367.01 ms │ +1.10x faster │
│ QQuery 19 │         35.00 / 37.63 ±2.10 / 41.46 ms │                              35.29 / 37.06 ±1.37 / 39.40 ms │     no change │
│ QQuery 20 │     718.48 / 735.78 ±17.63 / 757.90 ms │                          704.57 / 713.90 ±13.53 / 740.83 ms │     no change │
│ QQuery 21 │      759.40 / 766.12 ±4.15 / 772.20 ms │                           752.14 / 759.23 ±4.94 / 765.71 ms │     no change │
│ QQuery 22 │   1134.45 / 1140.74 ±6.13 / 1150.62 ms │                        1123.10 / 1130.92 ±4.50 / 1135.36 ms │     no change │
│ QQuery 23 │  3057.42 / 3077.54 ±16.97 / 3102.84 ms │                       3050.07 / 3067.41 ±11.77 / 3083.38 ms │     no change │
│ QQuery 24 │      100.04 / 102.33 ±1.21 / 103.48 ms │                            98.14 / 101.89 ±3.20 / 106.95 ms │     no change │
│ QQuery 25 │      137.55 / 138.63 ±1.09 / 139.99 ms │                           137.75 / 139.73 ±1.63 / 142.16 ms │     no change │
│ QQuery 26 │      100.87 / 101.88 ±0.98 / 103.75 ms │                            99.43 / 101.34 ±1.42 / 102.80 ms │     no change │
│ QQuery 27 │      846.69 / 852.87 ±4.71 / 860.44 ms │                           846.08 / 850.91 ±5.36 / 860.84 ms │     no change │
│ QQuery 28 │  3262.91 / 3290.00 ±19.98 / 3321.98 ms │                       3241.62 / 3279.85 ±20.94 / 3302.28 ms │     no change │
│ QQuery 29 │        50.34 / 61.98 ±16.80 / 95.30 ms │                              49.28 / 52.83 ±2.41 / 55.41 ms │ +1.17x faster │
│ QQuery 30 │      350.63 / 362.40 ±8.87 / 372.66 ms │                           349.23 / 357.21 ±4.84 / 361.45 ms │     no change │
│ QQuery 31 │      352.81 / 365.75 ±9.01 / 379.52 ms │                          359.74 / 373.17 ±10.94 / 390.17 ms │     no change │
│ QQuery 32 │ 1021.09 / 1134.61 ±114.10 / 1326.90 ms │                       1020.55 / 1035.70 ±15.61 / 1059.72 ms │ +1.10x faster │
│ QQuery 33 │   1439.08 / 1446.58 ±5.57 / 1453.43 ms │                       1417.77 / 1444.08 ±17.16 / 1469.70 ms │     no change │
│ QQuery 34 │  1452.12 / 1467.95 ±10.61 / 1483.12 ms │                       1422.70 / 1446.76 ±25.07 / 1492.31 ms │     no change │
│ QQuery 35 │      381.35 / 383.43 ±1.57 / 385.86 ms │                          373.26 / 385.90 ±11.36 / 400.74 ms │     no change │
│ QQuery 36 │      116.10 / 121.74 ±3.65 / 125.44 ms │                           109.99 / 117.32 ±5.10 / 123.26 ms │     no change │
│ QQuery 37 │         46.62 / 49.31 ±1.68 / 51.76 ms │                              47.54 / 49.77 ±1.61 / 51.77 ms │     no change │
│ QQuery 38 │         74.88 / 76.25 ±0.93 / 77.18 ms │                              73.80 / 76.49 ±1.92 / 79.73 ms │     no change │
│ QQuery 39 │      207.95 / 218.31 ±5.25 / 222.54 ms │                           209.61 / 214.34 ±3.98 / 218.99 ms │     no change │
│ QQuery 40 │         24.19 / 25.49 ±1.30 / 27.77 ms │                              21.12 / 24.09 ±1.88 / 26.86 ms │ +1.06x faster │
│ QQuery 41 │         18.86 / 20.76 ±1.18 / 21.93 ms │                              18.84 / 20.39 ±1.19 / 22.51 ms │     no change │
│ QQuery 42 │         19.18 / 19.56 ±0.31 / 20.02 ms │                              19.29 / 19.53 ±0.18 / 19.82 ms │     no change │
└───────────┴────────────────────────────────────────┴─────────────────────────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                                                          ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                                                          │ 22437.80ms │
│ Total Time (implement_groups_accumulator_count_distinct_primitive_types)   │ 22121.88ms │
│ Average Time (HEAD)                                                        │   521.81ms │
│ Average Time (implement_groups_accumulator_count_distinct_primitive_types) │   514.46ms │
│ Queries Faster                                                             │          5 │
│ Queries Slower                                                             │          2 │
│ Queries with No Change                                                     │         36 │
│ Queries with Failure                                                       │          0 │
└────────────────────────────────────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric	Value
Wall time	113.3s
Peak memory	40.5 GiB
Avg memory	30.1 GiB
CPU user	1068.2s
CPU sys	84.2s
Peak spill	0 B

clickbench_partitioned — branch

Metric	Value
Wall time	111.7s
Peak memory	41.3 GiB
Avg memory	33.0 GiB
CPU user	1056.9s
CPU sys	78.8s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-04-12T09:15:54Z

🤖 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 implement_groups_accumulator_count_distinct_primitive_types
--------------------
Benchmark clickbench_extended.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃                                    HEAD ┃ implement_groups_accumulator_count_distinct_primitive_types ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 0  │      821.46 / 832.96 ±11.19 / 852.43 ms │                          818.15 / 832.59 ±10.24 / 844.30 ms │    no change │
│ QQuery 1  │       210.26 / 210.95 ±0.86 / 212.60 ms │                           207.38 / 209.25 ±1.01 / 210.20 ms │    no change │
│ QQuery 2  │       499.88 / 502.56 ±1.92 / 505.88 ms │                           499.74 / 502.04 ±1.98 / 505.60 ms │    no change │
│ QQuery 3  │       314.20 / 316.27 ±1.34 / 318.27 ms │                           315.85 / 317.00 ±0.69 / 317.96 ms │    no change │
│ QQuery 4  │      659.30 / 669.99 ±10.42 / 688.02 ms │                           649.63 / 662.11 ±9.56 / 672.86 ms │    no change │
│ QQuery 5  │ 9837.93 / 9974.65 ±119.24 / 10120.51 ms │                       9623.91 / 9747.29 ±91.15 / 9882.31 ms │    no change │
│ QQuery 6  │    995.83 / 1010.24 ±18.78 / 1047.21 ms │                         992.44 / 1001.14 ±7.30 / 1012.59 ms │    no change │
│ QQuery 7  │      752.69 / 766.98 ±11.39 / 784.94 ms │                          828.62 / 901.53 ±42.20 / 955.19 ms │ 1.18x slower │
│ QQuery 8  │       397.40 / 403.44 ±6.58 / 415.95 ms │                           403.08 / 410.36 ±3.96 / 415.19 ms │    no change │
│ QQuery 9  │   2780.05 / 2825.71 ±41.79 / 2903.89 ms │                       2809.19 / 2836.50 ±25.10 / 2881.82 ms │    no change │
│ QQuery 10 │      639.23 / 649.24 ±10.65 / 666.89 ms │                           630.96 / 640.78 ±6.27 / 650.18 ms │    no change │
│ QQuery 11 │   2039.54 / 2059.48 ±13.79 / 2077.65 ms │                        2056.40 / 2068.64 ±8.85 / 2079.13 ms │    no change │
│ QQuery 12 │       197.12 / 203.80 ±6.72 / 215.86 ms │                           196.73 / 199.63 ±2.32 / 203.35 ms │    no change │
└───────────┴─────────────────────────────────────────┴─────────────────────────────────────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                                                          ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                                                          │ 20426.28ms │
│ Total Time (implement_groups_accumulator_count_distinct_primitive_types)   │ 20328.87ms │
│ Average Time (HEAD)                                                        │  1571.25ms │
│ Average Time (implement_groups_accumulator_count_distinct_primitive_types) │  1563.76ms │
│ Queries Faster                                                             │          0 │
│ Queries Slower                                                             │          1 │
│ Queries with No Change                                                     │         12 │
│ Queries with Failure                                                       │          0 │
└────────────────────────────────────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_extended — base (merge-base)

Metric	Value
Wall time	103.0s
Peak memory	33.3 GiB
Avg memory	29.6 GiB
CPU user	995.5s
CPU sys	45.8s
Peak spill	0 B

clickbench_extended — branch

Metric	Value
Wall time	102.4s
Peak memory	32.0 GiB
Avg memory	27.6 GiB
CPU user	984.1s
CPU sys	53.5s
Peak spill	0 B

File an issue against this benchmark runner

coderfender · 2026-04-12T19:08:48Z

Comparison with main


group                                                        impl_groupacc                           main
-----                                                        -----------                            ----
count_distinct_groups i64 10 groups                          1.00     85.6±1.80µs        ? ?/sec    4.97    425.0±5.72µs        ? ?/sec
count_distinct_groups i64 10 groups (GroupsAccumulator)                                             1.00     82.6±0.83µs        ? ?/sec
count_distinct_groups i64 100 groups                         1.00     86.1±1.41µs        ? ?/sec    5.18    445.9±5.79µs        ? ?/sec
count_distinct_groups i64 100 groups (GroupsAccumulator)                                            1.00     83.1±1.25µs        ? ?/sec
count_distinct_groups i64 1000 groups                        1.00     86.7±2.55µs        ? ?/sec    7.06    612.7±4.72µs        ? ?/sec
count_distinct_groups i64 1000 groups (GroupsAccumulator)                                           1.00     84.1±2.01µs        ? ?/sec

github-actions bot added the functions Changes to functions implementation label Apr 11, 2026

Dandandan reviewed Apr 12, 2026

View reviewed changes

coderfender force-pushed the implement_groups_accumulator_count_distinct_primitive_types branch from 5692893 to ee0c865 Compare April 12, 2026 05:09

coderfender mentioned this pull request Apr 12, 2026

chore: add count distinct group benchmarks #21575

Open

coderfender added 8 commits April 12, 2026 11:12

add count distinct group benchmarks

bacee66

add count distinct group benchmarks

c4461b7

add count distinct group benchmarks

45a19b0

init

7d5f2d0

implement_group_accumulators_count_distinct

ea30579

implement_group_accumulators_count_distinct

31a9f0f

implement_group_accumulators_count_distinct

3b1fb55

implement_group_accumulators_count_distinct

9a11396

coderfender added 4 commits April 12, 2026 11:58

implement_group_accumulators_count_distinct_use_hashtable

f2524a7

implement_group_accumulators_count_distinct_use_hashtable

2595e56

add group benches

f5ccfbb

Use same benchmark names for comparison

067a6b6

coderfender force-pushed the implement_groups_accumulator_count_distinct_primitive_types branch from 5b7f593 to 067a6b6 Compare April 12, 2026 18:59

count group benchmark check

4352265

count group benchmark check

651acf3

Conversation

coderfender commented Apr 11, 2026

Which issue does this PR close?

Rationale for this change

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

Uh oh!

coderfender commented Apr 11, 2026

Uh oh!

Dandandan Apr 12, 2026

Choose a reason for hiding this comment

Uh oh!

Dandandan Apr 12, 2026

Choose a reason for hiding this comment

Uh oh!

coderfender Apr 12, 2026

Choose a reason for hiding this comment

Uh oh!

coderfender Apr 12, 2026

Choose a reason for hiding this comment

Uh oh!

Dandandan commented Apr 12, 2026

Uh oh!

adriangbot commented Apr 12, 2026

Uh oh!

adriangbot commented Apr 12, 2026

Uh oh!

adriangbot commented Apr 12, 2026

Uh oh!

adriangbot commented Apr 12, 2026

Uh oh!

adriangbot commented Apr 12, 2026

Uh oh!

adriangbot commented Apr 12, 2026

Uh oh!

adriangbot commented Apr 12, 2026

Uh oh!

coderfender commented Apr 12, 2026

Uh oh!

Dandandan commented Apr 12, 2026

Uh oh!

adriangbot commented Apr 12, 2026

Uh oh!

adriangbot commented Apr 12, 2026

Uh oh!

adriangbot commented Apr 12, 2026

Uh oh!

adriangbot commented Apr 12, 2026

Uh oh!

adriangbot commented Apr 12, 2026

Uh oh!

adriangbot commented Apr 12, 2026

Uh oh!

adriangbot commented Apr 12, 2026

Uh oh!

coderfender commented Apr 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

coderfender commented Apr 12, 2026 •

edited

Loading