Intermediate result blocked approach to aggregation memory management by Rachelint · Pull Request #15591 · apache/datafusion

Rachelint · 2025-04-05T07:47:59Z

Which issue does this PR close?

Part of Improve aggregate performance with adaptive sizing in accumulators / avoiding reallocations in accumulators #7065

Rationale for this change

As mentioned in #7065 , we use a single Vec to manage aggregation intermediate results both in GroupAccumulator and GroupValues.

It is simple but not efficient enough in high-cardinality aggregation, because when Vec is not large enough, we need to allocate a new Vec and copy all data from the old one.

Copying a large amount of data(due to high-cardinality) is obviously expansive
And it is also not friendly to cpu (will refresh cache and tlb)

So this pr introduces a blocked approach to manage the aggregation intermediate results. We will never resize the Vec in the approach, and instead we split the data to blocks, when the capacity is not enough, we just allocate a new block. Detail can see #7065

What changes are included in this PR?

Implement the sketch for blocked approach
Implement blocked groups supporting PrimitiveGroupsAccumulator and GroupValuesPrimitive as the example

Are these changes tested?

Test by exist tests. And new unit tests, new fuzzy tests.

Are there any user-facing changes?

Two functions are added to GroupValues and GroupAccumulator trait.

But as you can see, there are default implementations for them, and users can choose to really support the blocked approach when wanting a better performance for their udafs.

    /// Returns `true` if this accumulator supports blocked groups.
    fn supports_blocked_groups(&self) -> bool {
        false
    }

    /// Alter the block size in the accumulator
    ///
    /// If the target block size is `None`, it will use a single big
    /// block(can think it a `Vec`) to manage the state.
    ///
    /// If the target block size` is `Some(blk_size)`, it will try to
    /// set the block size to `blk_size`, and the try will only success
    /// when the accumulator has supported blocked mode.
    ///
    /// NOTICE: After altering block size, all data in previous will be cleared.
    ///
    fn alter_block_size(&mut self, block_size: Option<usize>) -> Result<()> {
        if block_size.is_some() {
            return Err(DataFusionError::NotImplemented(
                "this accumulator doesn't support blocked mode yet".to_string(),
            ));
        }

        Ok(())
    }

Dandandan · 2025-04-08T07:36:20Z

Hi @Rachelint I think I have a alternative proposal that seems relatively easy to implement.
I'll share it with you once I have some time to validate the design (probably this evening).

Rachelint · 2025-04-08T07:54:02Z

Hi @Rachelint I think I have a alternative proposal that seems relatively easy to implement. I'll share it with you once I have some time to validate the design (probably this evening).

Really thanks. This design in pr indeed still introduces quite a few code changes...

I tried to not modify anythings about GroupAccumulator firstly:

Only implement the blocked logic in GroupValues
Then we reorder the input batch according to their block indices got from GroupValues
Apply input batch to related GroupAccumulator using slice
And when we found the new block is needed, create a new GroupAccumulator (one block one GroupAccumulator)

But I found this way will introduce too many extra cost...

Maybe we place the block indices into values in merge/update_batch as a Array?

Rachelint · 2025-04-17T12:03:17Z

Has finished development(and test) of all needed common structs!
Rest four things for this one:

Support blocked related logic in GroupedHashAggregateStream(we can copy it from Sketch for aggregation intermediate results blocked management #11943 )
Logic about deciding when we should enable this optimization
Example blocked version for GroupAccumulator and GroupValues
Unit test for blocked GroupValuesPrimitive, it is a bit complex
Fuzzy tests
Chore: fix docs, fix clippy, add more comments...

Rachelint · 2025-04-21T13:58:12Z

It is very close, just need to add more tests!

adriangbot · 2026-04-28T10:27:12Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4334353557-1862-xv64r 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 intermeidate-result-blocked-approach (a84d71c) to bbf67d9 (merge-base) diff using: tpch
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-28T10:27:15Z

Benchmark for this request failed.

Last 20 lines of output:

Click to expand

    |                                 ^^^^^^^ pattern `EmitTo::NextBlock` not covered
    |
note: `EmitTo` defined here
   --> datafusion/expr-common/src/groups_accumulator.rs:25:1
    |
 25 | pub enum EmitTo {
    | ^^^^^^^^^^^^^^^
...
 37 |     NextBlock,
    |     --------- not covered
    = note: the matched value is of type `EmitTo`
help: ensure that all possible cases are being handled by adding a match arm with a wildcard pattern or an explicit pattern as shown
    |
108 ~             EmitTo::First(n) => n,
109 ~             EmitTo::NextBlock => todo!(),
    |

For more information about this error, try `rustc --explain E0004`.
error: could not compile `datafusion-functions-aggregate-common` (lib) due to 2 previous errors
warning: build failed, waiting for other jobs to finish...

File an issue against this benchmark runner

Rachelint · 2026-04-28T10:31:04Z

Some build problems introduced by count distinct pr, fixing.
And after that, the pr seems can be ready.

Rachelint · 2026-04-28T11:10:50Z

I think it may be ready now.

Dandandan · 2026-04-28T12:02:21Z

run benchmarks

adriangbot · 2026-04-28T12:05:31Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4335018412-1864-w5tw2 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 intermeidate-result-blocked-approach (a818cd8) to 22bb4e6 (merge-base) diff using: clickbench_partitioned
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-28T12:06:01Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4335018412-1866-rzm7g 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 intermeidate-result-blocked-approach (a818cd8) to 22bb4e6 (merge-base) diff using: tpch
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-28T12:07:01Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4335018412-1865-d62wc 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 intermeidate-result-blocked-approach (a818cd8) to 22bb4e6 (merge-base) diff using: tpcds
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-28T12:20:50Z

🤖 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 intermeidate-result-blocked-approach
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃                           HEAD ┃ intermeidate-result-blocked-approach ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │ 40.72 / 41.57 ±1.06 / 43.63 ms │       40.37 / 41.16 ±1.00 / 43.12 ms │    no change │
│ QQuery 2  │ 20.72 / 21.51 ±0.62 / 22.55 ms │       21.04 / 21.34 ±0.25 / 21.62 ms │    no change │
│ QQuery 3  │ 38.35 / 40.09 ±1.37 / 41.87 ms │       38.43 / 39.34 ±1.06 / 41.26 ms │    no change │
│ QQuery 4  │ 18.38 / 18.58 ±0.20 / 18.95 ms │       18.27 / 18.60 ±0.20 / 18.86 ms │    no change │
│ QQuery 5  │ 47.70 / 50.67 ±2.39 / 53.39 ms │       48.15 / 50.52 ±1.69 / 52.22 ms │    no change │
│ QQuery 6  │ 17.39 / 17.67 ±0.19 / 17.90 ms │       17.48 / 17.60 ±0.11 / 17.78 ms │    no change │
│ QQuery 7  │ 54.37 / 55.31 ±0.66 / 56.38 ms │       53.61 / 55.88 ±1.27 / 57.34 ms │    no change │
│ QQuery 8  │ 48.41 / 48.93 ±0.52 / 49.84 ms │       48.46 / 48.71 ±0.27 / 49.18 ms │    no change │
│ QQuery 9  │ 53.68 / 55.26 ±1.43 / 57.73 ms │       54.43 / 54.68 ±0.32 / 55.29 ms │    no change │
│ QQuery 10 │ 65.52 / 66.01 ±0.36 / 66.56 ms │       65.57 / 66.14 ±0.63 / 67.30 ms │    no change │
│ QQuery 11 │ 14.25 / 14.47 ±0.12 / 14.60 ms │       13.93 / 14.20 ±0.19 / 14.40 ms │    no change │
│ QQuery 12 │ 28.14 / 28.25 ±0.10 / 28.39 ms │       27.65 / 28.07 ±0.34 / 28.64 ms │    no change │
│ QQuery 13 │ 37.26 / 38.28 ±0.65 / 39.13 ms │       37.99 / 38.44 ±0.55 / 39.43 ms │    no change │
│ QQuery 14 │ 27.86 / 28.24 ±0.21 / 28.44 ms │       28.04 / 28.13 ±0.06 / 28.19 ms │    no change │
│ QQuery 15 │ 33.52 / 33.86 ±0.20 / 34.06 ms │       33.62 / 34.41 ±0.83 / 35.81 ms │    no change │
│ QQuery 16 │ 15.25 / 15.41 ±0.13 / 15.65 ms │       15.33 / 15.41 ±0.07 / 15.50 ms │    no change │
│ QQuery 17 │ 77.75 / 78.94 ±0.78 / 80.14 ms │       82.48 / 83.06 ±0.48 / 83.70 ms │ 1.05x slower │
│ QQuery 18 │ 75.24 / 76.41 ±0.99 / 78.24 ms │       77.38 / 78.68 ±0.95 / 80.02 ms │    no change │
│ QQuery 19 │ 38.18 / 38.52 ±0.31 / 39.03 ms │       37.65 / 37.98 ±0.24 / 38.36 ms │    no change │
│ QQuery 20 │ 39.62 / 40.29 ±0.58 / 41.38 ms │       39.81 / 41.08 ±2.08 / 45.23 ms │    no change │
│ QQuery 21 │ 62.67 / 64.84 ±1.62 / 67.72 ms │       63.10 / 64.93 ±0.98 / 66.01 ms │    no change │
│ QQuery 22 │ 23.22 / 23.76 ±0.46 / 24.54 ms │       23.86 / 24.50 ±0.63 / 25.61 ms │    no change │
└───────────┴────────────────────────────────┴──────────────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ Benchmark Summary                                   ┃          ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━┩
│ Total Time (HEAD)                                   │ 896.89ms │
│ Total Time (intermeidate-result-blocked-approach)   │ 902.85ms │
│ Average Time (HEAD)                                 │  40.77ms │
│ Average Time (intermeidate-result-blocked-approach) │  41.04ms │
│ Queries Faster                                      │        0 │
│ Queries Slower                                      │        1 │
│ Queries with No Change                              │       21 │
│ Queries with Failure                                │        0 │
└─────────────────────────────────────────────────────┴──────────┘

Resource Usage

tpch — base (merge-base)

Metric	Value
Wall time	5.0s
Peak memory	5.8 GiB
Avg memory	5.0 GiB
CPU user	33.9s
CPU sys	2.4s
Peak spill	0 B

tpch — branch

Metric	Value
Wall time	5.0s
Peak memory	5.7 GiB
Avg memory	5.0 GiB
CPU user	34.2s
CPU sys	2.3s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-04-28T12:26:26Z

🤖 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 intermeidate-result-blocked-approach
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                     HEAD ┃      intermeidate-result-blocked-approach ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │              7.01 / 7.51 ±0.83 / 9.16 ms │               6.97 / 7.43 ±0.80 / 9.02 ms │     no change │
│ QQuery 2  │        146.03 / 146.75 ±0.56 / 147.75 ms │         149.58 / 149.91 ±0.24 / 150.32 ms │     no change │
│ QQuery 3  │        113.71 / 114.62 ±0.73 / 115.67 ms │         114.49 / 115.04 ±0.61 / 115.95 ms │     no change │
│ QQuery 4  │     1320.29 / 1328.53 ±5.14 / 1335.27 ms │      1293.09 / 1302.21 ±6.87 / 1312.53 ms │     no change │
│ QQuery 5  │        172.47 / 174.00 ±1.21 / 175.96 ms │         175.76 / 176.60 ±1.20 / 178.98 ms │     no change │
│ QQuery 6  │        142.02 / 144.92 ±3.01 / 148.60 ms │         135.06 / 137.22 ±1.47 / 138.97 ms │ +1.06x faster │
│ QQuery 7  │        335.95 / 337.13 ±1.06 / 338.50 ms │         338.12 / 340.12 ±1.55 / 342.53 ms │     no change │
│ QQuery 8  │        114.08 / 115.03 ±0.61 / 115.97 ms │         114.34 / 115.20 ±0.71 / 116.48 ms │     no change │
│ QQuery 9  │         96.41 / 101.14 ±2.42 / 103.21 ms │          97.70 / 103.88 ±5.33 / 112.35 ms │     no change │
│ QQuery 10 │        102.58 / 102.88 ±0.22 / 103.25 ms │         103.94 / 104.66 ±0.69 / 105.57 ms │     no change │
│ QQuery 11 │        928.89 / 934.32 ±5.24 / 944.08 ms │         915.01 / 923.02 ±6.61 / 933.42 ms │     no change │
│ QQuery 12 │           43.37 / 43.66 ±0.25 / 44.07 ms │            43.62 / 43.96 ±0.31 / 44.43 ms │     no change │
│ QQuery 13 │        389.52 / 391.27 ±1.70 / 393.96 ms │         390.28 / 392.62 ±1.85 / 395.57 ms │     no change │
│ QQuery 14 │     1036.94 / 1047.09 ±8.48 / 1056.90 ms │      1040.20 / 1042.69 ±2.62 / 1046.02 ms │     no change │
│ QQuery 15 │           15.09 / 15.26 ±0.15 / 15.54 ms │            14.98 / 15.24 ±0.24 / 15.57 ms │     no change │
│ QQuery 16 │             7.47 / 9.32 ±3.34 / 15.99 ms │               7.56 / 7.72 ±0.21 / 8.14 ms │ +1.21x faster │
│ QQuery 17 │        224.18 / 225.51 ±0.92 / 226.59 ms │         222.56 / 224.13 ±0.95 / 225.47 ms │     no change │
│ QQuery 18 │        127.38 / 128.40 ±0.57 / 128.95 ms │         126.84 / 128.08 ±1.20 / 130.34 ms │     no change │
│ QQuery 19 │        156.88 / 157.80 ±1.26 / 160.27 ms │         155.69 / 156.47 ±0.79 / 157.46 ms │     no change │
│ QQuery 20 │           13.26 / 13.60 ±0.20 / 13.88 ms │            13.26 / 13.50 ±0.21 / 13.84 ms │     no change │
│ QQuery 21 │           19.76 / 19.96 ±0.20 / 20.30 ms │            19.21 / 19.55 ±0.23 / 19.87 ms │     no change │
│ QQuery 22 │        485.67 / 488.76 ±2.48 / 492.60 ms │         479.46 / 486.46 ±5.04 / 491.67 ms │     no change │
│ QQuery 23 │    1086.04 / 1101.64 ±11.47 / 1116.62 ms │      1099.98 / 1108.35 ±6.37 / 1115.87 ms │     no change │
│ QQuery 24 │        713.85 / 717.23 ±1.90 / 719.53 ms │         713.00 / 717.39 ±3.14 / 720.66 ms │     no change │
│ QQuery 25 │        334.21 / 335.63 ±1.42 / 338.11 ms │         334.46 / 336.95 ±2.09 / 340.74 ms │     no change │
│ QQuery 26 │           77.54 / 79.70 ±3.62 / 86.93 ms │            77.61 / 78.96 ±1.98 / 82.85 ms │     no change │
│ QQuery 27 │              7.01 / 7.17 ±0.17 / 7.50 ms │               7.00 / 7.11 ±0.18 / 7.47 ms │     no change │
│ QQuery 28 │        148.75 / 151.22 ±3.90 / 158.96 ms │         147.96 / 149.65 ±0.95 / 150.77 ms │     no change │
│ QQuery 29 │        271.71 / 274.31 ±2.21 / 278.18 ms │         274.00 / 277.74 ±4.18 / 283.98 ms │     no change │
│ QQuery 30 │           42.07 / 43.23 ±0.81 / 44.31 ms │            41.95 / 42.36 ±0.31 / 42.72 ms │     no change │
│ QQuery 31 │        166.48 / 169.74 ±3.76 / 177.11 ms │         165.94 / 168.50 ±4.05 / 176.58 ms │     no change │
│ QQuery 32 │           13.51 / 13.77 ±0.22 / 14.14 ms │            13.38 / 13.80 ±0.28 / 14.22 ms │     no change │
│ QQuery 33 │        139.21 / 140.40 ±1.59 / 143.54 ms │         139.88 / 140.79 ±0.60 / 141.54 ms │     no change │
│ QQuery 34 │              6.98 / 7.21 ±0.29 / 7.78 ms │               7.03 / 7.16 ±0.20 / 7.57 ms │     no change │
│ QQuery 35 │         99.93 / 101.74 ±1.12 / 103.44 ms │         100.33 / 102.39 ±1.16 / 103.65 ms │     no change │
│ QQuery 36 │              6.71 / 6.92 ±0.15 / 7.11 ms │               6.55 / 6.98 ±0.33 / 7.54 ms │     no change │
│ QQuery 37 │              8.35 / 8.43 ±0.08 / 8.53 ms │               8.23 / 8.37 ±0.08 / 8.45 ms │     no change │
│ QQuery 38 │          87.36 / 92.94 ±6.28 / 105.18 ms │            86.85 / 88.54 ±1.66 / 91.38 ms │     no change │
│ QQuery 39 │        117.84 / 118.68 ±0.52 / 119.33 ms │         116.65 / 117.89 ±1.00 / 119.40 ms │     no change │
│ QQuery 40 │        103.03 / 107.95 ±6.04 / 119.83 ms │         102.30 / 105.53 ±2.43 / 109.63 ms │     no change │
│ QQuery 41 │           14.82 / 15.06 ±0.16 / 15.28 ms │            14.70 / 14.81 ±0.10 / 14.99 ms │     no change │
│ QQuery 42 │        108.13 / 110.21 ±3.17 / 116.51 ms │         108.19 / 109.26 ±1.98 / 113.21 ms │     no change │
│ QQuery 43 │              5.79 / 5.90 ±0.16 / 6.21 ms │               5.77 / 5.88 ±0.17 / 6.21 ms │     no change │
│ QQuery 44 │           10.94 / 11.71 ±1.02 / 13.73 ms │            10.84 / 10.99 ±0.13 / 11.22 ms │ +1.07x faster │
│ QQuery 45 │           48.35 / 48.97 ±0.42 / 49.58 ms │            49.29 / 50.29 ±1.56 / 53.39 ms │     no change │
│ QQuery 46 │              8.53 / 8.72 ±0.23 / 9.08 ms │               8.48 / 8.75 ±0.23 / 9.07 ms │     no change │
│ QQuery 47 │        708.88 / 718.15 ±9.26 / 734.86 ms │         705.07 / 711.71 ±4.44 / 717.12 ms │     no change │
│ QQuery 48 │        273.42 / 279.44 ±5.45 / 289.65 ms │         275.43 / 279.45 ±4.00 / 285.09 ms │     no change │
│ QQuery 49 │        249.37 / 250.36 ±0.56 / 250.89 ms │         248.62 / 251.09 ±1.77 / 253.05 ms │     no change │
│ QQuery 50 │        205.38 / 211.59 ±5.78 / 222.10 ms │         205.12 / 207.51 ±1.65 / 209.88 ms │     no change │
│ QQuery 51 │        176.05 / 178.44 ±1.65 / 180.82 ms │         176.03 / 178.83 ±2.03 / 182.22 ms │     no change │
│ QQuery 52 │        107.56 / 108.68 ±2.00 / 112.68 ms │         107.68 / 108.71 ±0.89 / 109.74 ms │     no change │
│ QQuery 53 │        102.57 / 103.84 ±0.94 / 105.12 ms │         103.38 / 103.99 ±0.88 / 105.70 ms │     no change │
│ QQuery 54 │        144.03 / 146.82 ±2.47 / 151.42 ms │         145.95 / 147.38 ±1.61 / 150.48 ms │     no change │
│ QQuery 55 │        106.16 / 106.79 ±0.76 / 108.28 ms │         107.26 / 108.01 ±0.63 / 108.83 ms │     no change │
│ QQuery 56 │        139.80 / 141.21 ±1.66 / 144.46 ms │         139.12 / 141.38 ±1.39 / 143.47 ms │     no change │
│ QQuery 57 │        165.18 / 168.79 ±2.60 / 172.69 ms │         164.41 / 166.79 ±1.92 / 168.69 ms │     no change │
│ QQuery 58 │        265.45 / 266.53 ±1.05 / 268.41 ms │         265.33 / 266.48 ±0.96 / 268.03 ms │     no change │
│ QQuery 59 │        196.01 / 197.37 ±1.24 / 199.32 ms │         195.12 / 196.74 ±1.69 / 199.92 ms │     no change │
│ QQuery 60 │        141.53 / 143.33 ±1.64 / 145.76 ms │         140.67 / 142.24 ±1.25 / 144.34 ms │     no change │
│ QQuery 61 │           13.80 / 13.89 ±0.11 / 14.11 ms │            13.38 / 13.55 ±0.22 / 13.98 ms │     no change │
│ QQuery 62 │        871.60 / 875.44 ±3.20 / 880.66 ms │         870.85 / 876.43 ±3.69 / 882.42 ms │     no change │
│ QQuery 63 │        103.07 / 105.73 ±2.73 / 109.06 ms │         103.95 / 106.04 ±3.00 / 111.98 ms │     no change │
│ QQuery 64 │        660.02 / 666.44 ±7.03 / 678.93 ms │         657.00 / 663.23 ±4.26 / 668.51 ms │     no change │
│ QQuery 65 │        248.22 / 251.28 ±3.78 / 258.65 ms │         245.17 / 248.04 ±3.08 / 253.55 ms │     no change │
│ QQuery 66 │        214.87 / 219.12 ±7.18 / 233.39 ms │         213.46 / 218.04 ±6.17 / 230.11 ms │     no change │
│ QQuery 67 │       294.52 / 307.49 ±15.48 / 335.82 ms │         292.83 / 299.79 ±5.77 / 307.03 ms │     no change │
│ QQuery 68 │              8.50 / 8.77 ±0.31 / 9.35 ms │             9.22 / 10.78 ±2.80 / 16.38 ms │  1.23x slower │
│ QQuery 69 │          96.78 / 98.93 ±2.46 / 103.69 ms │           97.32 / 98.69 ±0.99 / 100.28 ms │     no change │
│ QQuery 70 │       309.63 / 328.13 ±16.87 / 357.37 ms │         314.80 / 325.46 ±7.17 / 334.97 ms │     no change │
│ QQuery 71 │        133.23 / 136.15 ±2.62 / 139.87 ms │         133.94 / 136.13 ±2.35 / 140.24 ms │     no change │
│ QQuery 72 │       579.07 / 590.47 ±10.12 / 604.94 ms │        589.66 / 604.99 ±12.16 / 623.37 ms │     no change │
│ QQuery 73 │              6.83 / 6.92 ±0.16 / 7.24 ms │               6.83 / 7.05 ±0.20 / 7.42 ms │     no change │
│ QQuery 74 │        584.99 / 587.64 ±2.08 / 590.24 ms │        557.08 / 599.99 ±32.30 / 642.74 ms │     no change │
│ QQuery 75 │        268.04 / 270.56 ±3.32 / 277.10 ms │         266.77 / 267.65 ±0.49 / 268.22 ms │     no change │
│ QQuery 76 │        130.62 / 133.50 ±3.23 / 139.74 ms │         129.21 / 131.81 ±2.57 / 136.56 ms │     no change │
│ QQuery 77 │        187.12 / 189.29 ±2.41 / 193.86 ms │         188.64 / 189.93 ±1.27 / 192.36 ms │     no change │
│ QQuery 78 │        330.11 / 331.54 ±1.48 / 333.75 ms │         327.62 / 330.34 ±1.57 / 332.28 ms │     no change │
│ QQuery 79 │        232.21 / 236.10 ±2.98 / 239.51 ms │         228.19 / 231.54 ±4.17 / 239.22 ms │     no change │
│ QQuery 80 │        318.84 / 322.15 ±2.90 / 327.36 ms │         319.01 / 322.14 ±2.38 / 325.97 ms │     no change │
│ QQuery 81 │           26.04 / 26.39 ±0.27 / 26.76 ms │            25.71 / 26.15 ±0.38 / 26.72 ms │     no change │
│ QQuery 82 │           39.37 / 39.90 ±0.44 / 40.42 ms │            38.94 / 39.10 ±0.13 / 39.31 ms │     no change │
│ QQuery 83 │           37.32 / 38.74 ±1.93 / 42.57 ms │            36.85 / 37.89 ±1.53 / 40.93 ms │     no change │
│ QQuery 84 │           46.74 / 47.06 ±0.21 / 47.31 ms │            46.28 / 46.62 ±0.32 / 47.16 ms │     no change │
│ QQuery 85 │        141.56 / 143.20 ±2.26 / 147.65 ms │         140.80 / 141.18 ±0.36 / 141.72 ms │     no change │
│ QQuery 86 │           37.59 / 38.08 ±0.34 / 38.66 ms │            37.42 / 38.37 ±0.81 / 39.85 ms │     no change │
│ QQuery 87 │              3.56 / 3.69 ±0.21 / 4.10 ms │               3.55 / 3.67 ±0.15 / 3.96 ms │     no change │
│ QQuery 88 │         99.58 / 100.24 ±0.59 / 101.01 ms │          99.83 / 100.54 ±0.47 / 100.99 ms │     no change │
│ QQuery 89 │        117.19 / 121.00 ±4.84 / 130.34 ms │         116.94 / 121.24 ±6.83 / 134.79 ms │     no change │
│ QQuery 90 │           22.71 / 23.18 ±0.39 / 23.69 ms │            22.18 / 22.64 ±0.52 / 23.43 ms │     no change │
│ QQuery 91 │           58.68 / 59.53 ±0.69 / 60.42 ms │            58.13 / 59.39 ±0.96 / 60.55 ms │     no change │
│ QQuery 92 │           56.66 / 57.53 ±0.96 / 59.30 ms │            57.17 / 57.70 ±0.44 / 58.22 ms │     no change │
│ QQuery 93 │        180.23 / 181.45 ±0.73 / 182.41 ms │         179.62 / 180.68 ±1.47 / 183.58 ms │     no change │
│ QQuery 94 │           62.26 / 63.99 ±3.26 / 70.51 ms │            60.73 / 61.86 ±0.81 / 62.81 ms │     no change │
│ QQuery 95 │        126.59 / 127.36 ±0.69 / 128.40 ms │         125.90 / 126.90 ±0.87 / 128.44 ms │     no change │
│ QQuery 96 │           68.69 / 71.62 ±1.64 / 73.72 ms │            68.92 / 70.80 ±2.18 / 74.93 ms │     no change │
│ QQuery 97 │        116.96 / 118.53 ±0.89 / 119.48 ms │         117.75 / 118.44 ±0.48 / 119.12 ms │     no change │
│ QQuery 98 │        150.33 / 153.22 ±2.05 / 156.27 ms │         150.22 / 152.64 ±2.33 / 155.63 ms │     no change │
│ QQuery 99 │ 10759.39 / 10914.99 ±86.21 / 11011.33 ms │ 10695.39 / 10826.26 ±115.61 / 11010.63 ms │     no change │
└───────────┴──────────────────────────────────────────┴───────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                                   ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                                   │ 30756.49ms │
│ Total Time (intermeidate-result-blocked-approach)   │ 30608.12ms │
│ Average Time (HEAD)                                 │   310.67ms │
│ Average Time (intermeidate-result-blocked-approach) │   309.17ms │
│ Queries Faster                                      │          3 │
│ Queries Slower                                      │          1 │
│ Queries with No Change                              │         95 │
│ Queries with Failure                                │          0 │
└─────────────────────────────────────────────────────┴────────────┘

Resource Usage

tpcds — base (merge-base)

Metric	Value
Wall time	155.0s
Peak memory	6.4 GiB
Avg memory	5.7 GiB
CPU user	242.9s
CPU sys	8.5s
Peak spill	0 B

tpcds — branch

Metric	Value
Wall time	155.0s
Peak memory	6.3 GiB
Avg memory	5.7 GiB
CPU user	242.2s
CPU sys	8.1s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-04-28T12:26:26Z

🤖 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 intermeidate-result-blocked-approach
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃  intermeidate-result-blocked-approach ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │          1.21 / 4.69 ±6.82 / 18.34 ms │          1.25 / 4.74 ±6.87 / 18.48 ms │     no change │
│ QQuery 1  │        12.42 / 12.98 ±0.31 / 13.28 ms │        13.04 / 13.41 ±0.24 / 13.74 ms │     no change │
│ QQuery 2  │        36.77 / 37.00 ±0.28 / 37.55 ms │        36.72 / 37.01 ±0.21 / 37.27 ms │     no change │
│ QQuery 3  │        31.82 / 32.55 ±0.61 / 33.29 ms │        31.66 / 32.03 ±0.21 / 32.24 ms │     no change │
│ QQuery 4  │     246.13 / 252.06 ±5.38 / 261.04 ms │     288.46 / 293.26 ±2.85 / 296.74 ms │  1.16x slower │
│ QQuery 5  │     287.55 / 289.80 ±1.47 / 291.18 ms │     290.72 / 294.74 ±2.80 / 298.27 ms │     no change │
│ QQuery 6  │          6.44 / 8.32 ±2.41 / 13.07 ms │           6.69 / 7.41 ±0.44 / 7.79 ms │ +1.12x faster │
│ QQuery 7  │        13.76 / 13.95 ±0.13 / 14.16 ms │        14.49 / 14.70 ±0.20 / 15.07 ms │  1.05x slower │
│ QQuery 8  │     337.06 / 340.72 ±2.75 / 344.90 ms │     339.33 / 345.93 ±4.24 / 351.57 ms │     no change │
│ QQuery 9  │     467.45 / 471.13 ±4.79 / 480.04 ms │     482.13 / 493.24 ±9.53 / 509.78 ms │     no change │
│ QQuery 10 │        73.96 / 75.46 ±0.77 / 76.18 ms │        75.11 / 75.84 ±0.72 / 77.15 ms │     no change │
│ QQuery 11 │        85.74 / 86.49 ±1.01 / 88.48 ms │        87.51 / 88.88 ±2.41 / 93.69 ms │     no change │
│ QQuery 12 │     283.78 / 286.86 ±3.08 / 291.09 ms │     284.25 / 292.10 ±6.74 / 302.54 ms │     no change │
│ QQuery 13 │     406.46 / 420.25 ±8.86 / 430.15 ms │     413.33 / 423.15 ±8.11 / 432.17 ms │     no change │
│ QQuery 14 │     292.45 / 301.14 ±7.98 / 316.27 ms │     295.63 / 299.22 ±2.46 / 303.32 ms │     no change │
│ QQuery 15 │     305.37 / 309.09 ±3.65 / 315.83 ms │     332.46 / 345.11 ±9.11 / 360.85 ms │  1.12x slower │
│ QQuery 16 │     648.86 / 651.03 ±1.24 / 652.66 ms │    634.06 / 650.45 ±12.92 / 673.27 ms │     no change │
│ QQuery 17 │     651.63 / 652.81 ±0.80 / 653.90 ms │     645.18 / 655.54 ±7.56 / 665.60 ms │     no change │
│ QQuery 18 │ 1308.89 / 1318.44 ±11.32 / 1338.43 ms │ 1298.23 / 1322.43 ±17.61 / 1342.60 ms │     no change │
│ QQuery 19 │        28.78 / 29.46 ±0.74 / 30.71 ms │        29.42 / 31.13 ±1.91 / 33.96 ms │  1.06x slower │
│ QQuery 20 │     516.43 / 525.56 ±6.74 / 534.53 ms │     520.66 / 531.41 ±7.26 / 539.58 ms │     no change │
│ QQuery 21 │     597.45 / 603.19 ±4.52 / 611.30 ms │     598.63 / 602.55 ±3.35 / 607.56 ms │     no change │
│ QQuery 22 │  1069.91 / 1078.43 ±6.11 / 1084.80 ms │  1071.85 / 1077.67 ±4.60 / 1083.29 ms │     no change │
│ QQuery 23 │ 3384.14 / 3413.94 ±28.36 / 3459.00 ms │ 3349.35 / 3377.43 ±17.97 / 3404.89 ms │     no change │
│ QQuery 24 │        42.01 / 48.94 ±7.81 / 62.88 ms │        41.90 / 42.85 ±0.52 / 43.38 ms │ +1.14x faster │
│ QQuery 25 │     115.62 / 117.15 ±1.29 / 118.94 ms │     115.26 / 116.65 ±1.25 / 118.64 ms │     no change │
│ QQuery 26 │        42.93 / 44.05 ±1.25 / 45.65 ms │        42.90 / 43.47 ±0.39 / 43.99 ms │     no change │
│ QQuery 27 │     676.09 / 685.03 ±7.26 / 696.46 ms │     682.12 / 687.49 ±4.86 / 696.48 ms │     no change │
│ QQuery 28 │ 3054.14 / 3069.82 ±19.42 / 3106.66 ms │ 3047.09 / 3062.80 ±12.41 / 3081.02 ms │     no change │
│ QQuery 29 │        43.50 / 48.21 ±5.58 / 55.92 ms │       43.09 / 48.72 ±10.62 / 69.95 ms │     no change │
│ QQuery 30 │     318.52 / 321.71 ±3.29 / 326.37 ms │     323.61 / 324.85 ±1.09 / 326.49 ms │     no change │
│ QQuery 31 │     308.45 / 316.54 ±5.71 / 325.25 ms │     313.78 / 319.24 ±3.69 / 323.63 ms │     no change │
│ QQuery 32 │ 1038.99 / 1049.02 ±10.88 / 1068.44 ms │ 1053.87 / 1068.04 ±12.20 / 1087.34 ms │     no change │
│ QQuery 33 │ 1486.15 / 1515.97 ±24.06 / 1552.98 ms │ 1481.92 / 1502.10 ±23.35 / 1546.35 ms │     no change │
│ QQuery 34 │  1514.35 / 1522.62 ±7.98 / 1535.04 ms │ 1496.49 / 1506.89 ±10.55 / 1526.49 ms │     no change │
│ QQuery 35 │    305.85 / 317.75 ±17.70 / 352.89 ms │    345.22 / 355.32 ±15.40 / 385.82 ms │  1.12x slower │
│ QQuery 36 │        63.97 / 71.87 ±6.22 / 79.36 ms │        63.14 / 70.24 ±5.16 / 77.65 ms │     no change │
│ QQuery 37 │        36.62 / 39.88 ±2.25 / 43.30 ms │        36.77 / 41.69 ±3.79 / 47.37 ms │     no change │
│ QQuery 38 │        41.44 / 45.41 ±4.60 / 54.23 ms │        40.78 / 42.38 ±1.48 / 44.71 ms │ +1.07x faster │
│ QQuery 39 │     128.25 / 135.32 ±5.73 / 142.96 ms │     127.25 / 138.93 ±8.91 / 152.38 ms │     no change │
│ QQuery 40 │        15.33 / 19.61 ±5.91 / 31.19 ms │        15.62 / 15.81 ±0.13 / 15.98 ms │ +1.24x faster │
│ QQuery 41 │        14.70 / 14.77 ±0.05 / 14.85 ms │        14.50 / 14.96 ±0.38 / 15.52 ms │     no change │
│ QQuery 42 │        14.12 / 14.27 ±0.15 / 14.54 ms │        14.75 / 14.83 ±0.07 / 14.95 ms │     no change │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                                   ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                                   │ 20613.32ms │
│ Total Time (intermeidate-result-blocked-approach)   │ 20726.64ms │
│ Average Time (HEAD)                                 │   479.38ms │
│ Average Time (intermeidate-result-blocked-approach) │   482.01ms │
│ Queries Faster                                      │          4 │
│ Queries Slower                                      │          5 │
│ Queries with No Change                              │         34 │
│ Queries with Failure                                │          0 │
└─────────────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric	Value
Wall time	105.0s
Peak memory	30.0 GiB
Avg memory	23.1 GiB
CPU user	1091.5s
CPU sys	64.8s
Peak spill	0 B

clickbench_partitioned — branch

Metric	Value
Wall time	105.0s
Peak memory	30.2 GiB
Avg memory	23.1 GiB
CPU user	1097.9s
CPU sys	65.7s
Peak spill	0 B

File an issue against this benchmark runner

Rachelint · 2026-04-29T07:58:21Z

@Dandandan hi, can help trigger the benchmark again?

Dandandan · 2026-04-29T12:22:15Z

run benchmarks

adriangbot · 2026-04-29T12:25:03Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4343604373-1909-n4jjp 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 intermeidate-result-blocked-approach (47b6976) to 8f033e4 (merge-base) diff using: tpcds
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-29T12:25:44Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4343604373-1908-2hjmh 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 intermeidate-result-blocked-approach (47b6976) to 8f033e4 (merge-base) diff using: clickbench_partitioned
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-29T12:26:42Z

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4343604373-1910-rgfv2 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 intermeidate-result-blocked-approach (47b6976) to 8f033e4 (merge-base) diff using: tpch
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-29T12:40:39Z

🤖 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 intermeidate-result-blocked-approach
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃                           HEAD ┃ intermeidate-result-blocked-approach ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │ 40.80 / 42.01 ±1.13 / 43.54 ms │       40.34 / 41.03 ±1.08 / 43.17 ms │    no change │
│ QQuery 2  │ 21.45 / 21.62 ±0.09 / 21.73 ms │       21.46 / 21.80 ±0.51 / 22.81 ms │    no change │
│ QQuery 3  │ 38.57 / 40.03 ±1.59 / 42.27 ms │       38.29 / 40.24 ±1.18 / 41.68 ms │    no change │
│ QQuery 4  │ 18.47 / 18.90 ±0.75 / 20.40 ms │       18.43 / 18.76 ±0.40 / 19.53 ms │    no change │
│ QQuery 5  │ 49.41 / 52.85 ±2.22 / 55.90 ms │       49.47 / 52.69 ±2.74 / 56.93 ms │    no change │
│ QQuery 6  │ 18.25 / 18.62 ±0.57 / 19.75 ms │       17.95 / 18.23 ±0.20 / 18.55 ms │    no change │
│ QQuery 7  │ 54.57 / 55.57 ±0.56 / 56.16 ms │       53.37 / 55.60 ±2.05 / 58.28 ms │    no change │
│ QQuery 8  │ 48.55 / 48.94 ±0.23 / 49.17 ms │       47.93 / 48.38 ±0.28 / 48.78 ms │    no change │
│ QQuery 9  │ 54.81 / 57.25 ±2.08 / 60.30 ms │       54.04 / 55.16 ±1.41 / 57.89 ms │    no change │
│ QQuery 10 │ 66.09 / 68.44 ±1.77 / 70.86 ms │       66.52 / 68.64 ±1.75 / 71.53 ms │    no change │
│ QQuery 11 │ 14.20 / 14.35 ±0.12 / 14.53 ms │       14.32 / 14.52 ±0.25 / 14.96 ms │    no change │
│ QQuery 12 │ 27.75 / 28.23 ±0.38 / 28.68 ms │       27.29 / 27.56 ±0.23 / 27.83 ms │    no change │
│ QQuery 13 │ 37.73 / 40.20 ±3.16 / 46.41 ms │       38.48 / 39.24 ±0.54 / 40.10 ms │    no change │
│ QQuery 14 │ 28.02 / 28.48 ±0.46 / 29.34 ms │       28.09 / 28.28 ±0.13 / 28.47 ms │    no change │
│ QQuery 15 │ 35.84 / 36.57 ±0.53 / 37.18 ms │       35.61 / 36.11 ±0.32 / 36.46 ms │    no change │
│ QQuery 16 │ 15.53 / 16.27 ±0.41 / 16.69 ms │       16.05 / 16.40 ±0.22 / 16.71 ms │    no change │
│ QQuery 17 │ 78.74 / 80.64 ±1.77 / 83.42 ms │       84.26 / 85.25 ±0.80 / 86.15 ms │ 1.06x slower │
│ QQuery 18 │ 76.83 / 78.67 ±2.05 / 82.50 ms │       76.34 / 78.26 ±1.63 / 81.20 ms │    no change │
│ QQuery 19 │ 38.44 / 40.05 ±0.87 / 41.03 ms │       38.80 / 40.39 ±0.89 / 41.50 ms │    no change │
│ QQuery 20 │ 39.88 / 40.68 ±1.50 / 43.69 ms │       40.01 / 40.22 ±0.18 / 40.55 ms │    no change │
│ QQuery 21 │ 63.73 / 65.15 ±1.50 / 68.04 ms │       62.87 / 64.26 ±1.10 / 65.87 ms │    no change │
│ QQuery 22 │ 24.02 / 24.80 ±0.94 / 26.54 ms │       23.97 / 24.28 ±0.19 / 24.56 ms │    no change │
└───────────┴────────────────────────────────┴──────────────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ Benchmark Summary                                   ┃          ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━┩
│ Total Time (HEAD)                                   │ 918.32ms │
│ Total Time (intermeidate-result-blocked-approach)   │ 915.29ms │
│ Average Time (HEAD)                                 │  41.74ms │
│ Average Time (intermeidate-result-blocked-approach) │  41.60ms │
│ Queries Faster                                      │        0 │
│ Queries Slower                                      │        1 │
│ Queries with No Change                              │       21 │
│ Queries with Failure                                │        0 │
└─────────────────────────────────────────────────────┴──────────┘

Resource Usage

tpch — base (merge-base)

Metric	Value
Wall time	5.0s
Peak memory	5.7 GiB
Avg memory	5.0 GiB
CPU user	34.2s
CPU sys	2.7s
Peak spill	0 B

tpch — branch

Metric	Value
Wall time	5.0s
Peak memory	5.7 GiB
Avg memory	5.0 GiB
CPU user	34.5s
CPU sys	2.5s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-04-29T12:45:06Z

🤖 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 intermeidate-result-blocked-approach
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                     HEAD ┃      intermeidate-result-blocked-approach ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │              7.00 / 7.68 ±0.86 / 9.38 ms │               6.89 / 7.34 ±0.80 / 8.94 ms │     no change │
│ QQuery 2  │        147.94 / 148.31 ±0.33 / 148.87 ms │         150.27 / 150.81 ±0.35 / 151.34 ms │     no change │
│ QQuery 3  │        113.77 / 114.52 ±1.04 / 116.44 ms │         112.57 / 114.06 ±0.93 / 115.49 ms │     no change │
│ QQuery 4  │     1272.35 / 1289.93 ±9.47 / 1297.51 ms │      1279.06 / 1287.61 ±7.47 / 1300.99 ms │     no change │
│ QQuery 5  │        172.36 / 173.77 ±1.11 / 175.21 ms │         172.17 / 173.72 ±0.99 / 175.27 ms │     no change │
│ QQuery 6  │        132.86 / 135.21 ±1.96 / 138.81 ms │         134.27 / 135.56 ±1.48 / 138.38 ms │     no change │
│ QQuery 7  │        331.97 / 336.06 ±2.26 / 338.52 ms │         329.35 / 332.45 ±2.18 / 334.35 ms │     no change │
│ QQuery 8  │        114.21 / 115.08 ±0.89 / 116.78 ms │         113.60 / 114.44 ±0.68 / 115.53 ms │     no change │
│ QQuery 9  │         96.72 / 100.31 ±2.88 / 103.05 ms │          96.54 / 100.04 ±4.64 / 108.63 ms │     no change │
│ QQuery 10 │        103.15 / 103.86 ±0.59 / 104.90 ms │         102.04 / 102.96 ±0.66 / 103.86 ms │     no change │
│ QQuery 11 │        879.95 / 890.92 ±6.51 / 898.80 ms │         873.33 / 885.27 ±7.12 / 891.83 ms │     no change │
│ QQuery 12 │           43.37 / 44.06 ±0.50 / 44.85 ms │            43.27 / 44.66 ±1.27 / 46.61 ms │     no change │
│ QQuery 13 │        387.44 / 390.62 ±3.05 / 394.81 ms │         385.70 / 390.07 ±2.94 / 394.16 ms │     no change │
│ QQuery 14 │     1035.29 / 1041.95 ±3.73 / 1046.72 ms │      1033.68 / 1036.02 ±2.80 / 1041.51 ms │     no change │
│ QQuery 15 │           15.13 / 15.70 ±0.31 / 16.08 ms │            14.94 / 15.76 ±0.70 / 16.86 ms │     no change │
│ QQuery 16 │              7.61 / 7.75 ±0.20 / 8.14 ms │               7.29 / 7.45 ±0.23 / 7.90 ms │     no change │
│ QQuery 17 │        221.34 / 222.78 ±1.01 / 224.21 ms │         218.09 / 221.31 ±1.63 / 222.40 ms │     no change │
│ QQuery 18 │        126.29 / 127.09 ±0.73 / 128.19 ms │         125.42 / 126.53 ±0.80 / 127.90 ms │     no change │
│ QQuery 19 │        154.15 / 155.01 ±1.18 / 157.36 ms │         152.93 / 153.92 ±0.66 / 155.00 ms │     no change │
│ QQuery 20 │           13.28 / 13.64 ±0.38 / 14.33 ms │            12.95 / 13.23 ±0.25 / 13.60 ms │     no change │
│ QQuery 21 │           19.38 / 19.71 ±0.36 / 20.38 ms │            19.09 / 19.69 ±1.00 / 21.69 ms │     no change │
│ QQuery 22 │        474.17 / 481.32 ±6.53 / 492.97 ms │         471.26 / 477.65 ±5.56 / 484.10 ms │     no change │
│ QQuery 23 │     1089.14 / 1095.14 ±4.66 / 1102.99 ms │      1083.49 / 1087.13 ±4.69 / 1095.88 ms │     no change │
│ QQuery 24 │        713.12 / 715.91 ±3.47 / 722.73 ms │         700.60 / 708.53 ±5.74 / 715.25 ms │     no change │
│ QQuery 25 │        332.76 / 335.08 ±1.62 / 337.59 ms │         329.47 / 332.35 ±1.89 / 334.65 ms │     no change │
│ QQuery 26 │           78.59 / 78.95 ±0.29 / 79.45 ms │            77.25 / 77.97 ±0.45 / 78.54 ms │     no change │
│ QQuery 27 │              7.01 / 7.29 ±0.28 / 7.75 ms │               7.02 / 7.13 ±0.18 / 7.49 ms │     no change │
│ QQuery 28 │        149.11 / 150.10 ±1.31 / 152.60 ms │         146.72 / 148.75 ±1.73 / 151.64 ms │     no change │
│ QQuery 29 │        272.26 / 275.71 ±4.54 / 284.70 ms │         270.94 / 273.02 ±1.80 / 275.81 ms │     no change │
│ QQuery 30 │           41.51 / 42.07 ±0.32 / 42.45 ms │            40.69 / 41.52 ±0.58 / 42.22 ms │     no change │
│ QQuery 31 │        165.98 / 167.33 ±0.91 / 168.55 ms │         164.89 / 168.19 ±2.68 / 172.95 ms │     no change │
│ QQuery 32 │           13.52 / 13.81 ±0.34 / 14.48 ms │            13.34 / 13.53 ±0.15 / 13.75 ms │     no change │
│ QQuery 33 │        139.05 / 141.23 ±2.13 / 144.38 ms │         138.72 / 141.17 ±2.31 / 145.35 ms │     no change │
│ QQuery 34 │              7.13 / 7.32 ±0.26 / 7.83 ms │               7.14 / 7.27 ±0.14 / 7.55 ms │     no change │
│ QQuery 35 │        101.18 / 102.82 ±1.08 / 104.58 ms │         100.00 / 100.60 ±0.46 / 101.39 ms │     no change │
│ QQuery 36 │              7.13 / 7.26 ±0.07 / 7.35 ms │               6.76 / 6.95 ±0.11 / 7.11 ms │     no change │
│ QQuery 37 │              8.46 / 8.54 ±0.08 / 8.67 ms │              8.37 / 9.39 ±1.46 / 12.28 ms │  1.10x slower │
│ QQuery 38 │           86.54 / 87.63 ±1.02 / 88.94 ms │            85.31 / 87.11 ±2.36 / 91.64 ms │     no change │
│ QQuery 39 │        118.80 / 122.53 ±4.97 / 132.15 ms │         115.77 / 119.89 ±4.67 / 128.90 ms │     no change │
│ QQuery 40 │        102.34 / 108.67 ±3.60 / 113.16 ms │         104.20 / 105.78 ±2.27 / 110.21 ms │     no change │
│ QQuery 41 │           14.60 / 14.78 ±0.18 / 15.10 ms │            14.45 / 14.70 ±0.26 / 15.20 ms │     no change │
│ QQuery 42 │        106.33 / 107.75 ±1.30 / 109.51 ms │         106.93 / 109.37 ±2.43 / 113.47 ms │     no change │
│ QQuery 43 │              5.90 / 6.05 ±0.17 / 6.37 ms │               5.64 / 5.76 ±0.11 / 5.96 ms │     no change │
│ QQuery 44 │           11.01 / 12.56 ±2.68 / 17.91 ms │            11.00 / 11.20 ±0.11 / 11.33 ms │ +1.12x faster │
│ QQuery 45 │           48.47 / 48.99 ±0.44 / 49.61 ms │            48.87 / 49.13 ±0.17 / 49.36 ms │     no change │
│ QQuery 46 │              8.73 / 8.86 ±0.14 / 9.11 ms │               8.40 / 8.53 ±0.16 / 8.84 ms │     no change │
│ QQuery 47 │        690.52 / 697.24 ±7.48 / 707.46 ms │         684.67 / 694.47 ±7.61 / 707.87 ms │     no change │
│ QQuery 48 │        273.79 / 276.80 ±2.35 / 280.10 ms │         274.36 / 278.26 ±3.19 / 281.88 ms │     no change │
│ QQuery 49 │        249.77 / 250.38 ±0.71 / 251.77 ms │         247.13 / 248.86 ±2.02 / 251.68 ms │     no change │
│ QQuery 50 │        203.92 / 208.89 ±5.40 / 219.00 ms │         201.63 / 206.69 ±5.96 / 218.05 ms │     no change │
│ QQuery 51 │        176.48 / 178.47 ±1.64 / 180.90 ms │         176.92 / 178.69 ±1.80 / 181.94 ms │     no change │
│ QQuery 52 │        107.06 / 108.70 ±2.30 / 113.19 ms │         106.91 / 108.43 ±1.72 / 111.76 ms │     no change │
│ QQuery 53 │        102.10 / 102.45 ±0.36 / 103.10 ms │         102.69 / 103.13 ±0.32 / 103.67 ms │     no change │
│ QQuery 54 │        145.32 / 146.85 ±2.23 / 151.21 ms │         145.03 / 147.76 ±2.82 / 153.09 ms │     no change │
│ QQuery 55 │        105.91 / 106.72 ±0.85 / 107.97 ms │         105.78 / 106.75 ±0.73 / 107.71 ms │     no change │
│ QQuery 56 │        139.36 / 139.99 ±0.65 / 140.91 ms │         139.93 / 141.00 ±1.03 / 142.26 ms │     no change │
│ QQuery 57 │        166.40 / 167.91 ±1.65 / 171.12 ms │         166.59 / 168.71 ±2.07 / 172.68 ms │     no change │
│ QQuery 58 │        265.07 / 267.43 ±1.33 / 268.84 ms │         264.50 / 265.44 ±0.66 / 266.41 ms │     no change │
│ QQuery 59 │        196.64 / 198.55 ±2.36 / 203.09 ms │         195.43 / 197.95 ±1.49 / 199.77 ms │     no change │
│ QQuery 60 │        141.00 / 142.00 ±0.81 / 142.88 ms │         140.10 / 141.38 ±1.09 / 143.22 ms │     no change │
│ QQuery 61 │           13.62 / 13.75 ±0.20 / 14.15 ms │            13.34 / 13.47 ±0.20 / 13.87 ms │     no change │
│ QQuery 62 │        864.82 / 871.16 ±3.80 / 875.97 ms │         857.81 / 863.66 ±5.99 / 873.88 ms │     no change │
│ QQuery 63 │        102.51 / 105.46 ±3.63 / 112.11 ms │         101.81 / 103.52 ±1.47 / 105.73 ms │     no change │
│ QQuery 64 │        660.64 / 666.47 ±3.88 / 672.36 ms │         655.08 / 661.50 ±5.23 / 670.81 ms │     no change │
│ QQuery 65 │        245.34 / 247.38 ±3.24 / 253.82 ms │         242.15 / 247.07 ±3.18 / 252.16 ms │     no change │
│ QQuery 66 │       215.12 / 224.49 ±12.64 / 249.50 ms │        214.34 / 223.14 ±10.22 / 241.20 ms │     no change │
│ QQuery 67 │       294.46 / 308.03 ±15.10 / 333.37 ms │         290.38 / 299.10 ±7.70 / 311.26 ms │     no change │
│ QQuery 68 │              8.90 / 9.09 ±0.28 / 9.64 ms │               8.79 / 9.02 ±0.21 / 9.42 ms │     no change │
│ QQuery 69 │         99.38 / 102.65 ±3.24 / 108.88 ms │            96.89 / 97.59 ±0.42 / 98.16 ms │     no change │
│ QQuery 70 │        317.32 / 329.13 ±9.17 / 341.99 ms │         311.39 / 320.77 ±7.32 / 330.29 ms │     no change │
│ QQuery 71 │        135.54 / 136.94 ±1.38 / 139.50 ms │         132.97 / 135.82 ±3.75 / 143.24 ms │     no change │
│ QQuery 72 │        593.07 / 598.81 ±3.20 / 602.94 ms │         578.99 / 589.68 ±8.80 / 605.61 ms │     no change │
│ QQuery 73 │              6.90 / 7.08 ±0.24 / 7.56 ms │               6.68 / 6.86 ±0.18 / 7.17 ms │     no change │
│ QQuery 74 │        553.70 / 564.78 ±6.99 / 575.31 ms │        545.71 / 563.15 ±15.99 / 590.98 ms │     no change │
│ QQuery 75 │        268.27 / 271.04 ±2.47 / 275.31 ms │         264.18 / 267.29 ±2.90 / 271.55 ms │     no change │
│ QQuery 76 │        129.49 / 131.12 ±2.25 / 135.56 ms │         128.38 / 128.83 ±0.45 / 129.65 ms │     no change │
│ QQuery 77 │        187.39 / 188.10 ±0.48 / 188.64 ms │         186.79 / 189.13 ±3.07 / 194.97 ms │     no change │
│ QQuery 78 │        330.92 / 332.48 ±1.11 / 334.23 ms │         326.54 / 329.54 ±2.05 / 332.11 ms │     no change │
│ QQuery 79 │        227.68 / 230.77 ±2.20 / 233.58 ms │         227.07 / 231.02 ±4.71 / 240.09 ms │     no change │
│ QQuery 80 │        318.82 / 320.86 ±1.69 / 322.93 ms │         315.83 / 319.58 ±2.99 / 324.81 ms │     no change │
│ QQuery 81 │           26.08 / 26.50 ±0.48 / 27.25 ms │            25.85 / 26.24 ±0.21 / 26.44 ms │     no change │
│ QQuery 82 │           38.99 / 39.36 ±0.31 / 39.86 ms │            38.54 / 38.96 ±0.44 / 39.70 ms │     no change │
│ QQuery 83 │           37.32 / 38.16 ±1.13 / 40.36 ms │            36.67 / 36.79 ±0.13 / 37.04 ms │     no change │
│ QQuery 84 │           46.19 / 46.41 ±0.16 / 46.66 ms │            46.41 / 46.52 ±0.13 / 46.76 ms │     no change │
│ QQuery 85 │        142.18 / 143.91 ±2.31 / 148.47 ms │         139.51 / 142.40 ±2.73 / 147.38 ms │     no change │
│ QQuery 86 │           37.65 / 38.19 ±0.38 / 38.82 ms │            37.09 / 37.48 ±0.61 / 38.70 ms │     no change │
│ QQuery 87 │              3.59 / 3.71 ±0.16 / 4.03 ms │               3.50 / 3.62 ±0.15 / 3.92 ms │     no change │
│ QQuery 88 │        100.75 / 101.06 ±0.22 / 101.39 ms │          99.48 / 100.82 ±1.74 / 104.25 ms │     no change │
│ QQuery 89 │        116.12 / 119.59 ±4.85 / 128.95 ms │         116.53 / 119.39 ±4.81 / 128.95 ms │     no change │
│ QQuery 90 │           22.93 / 23.33 ±0.23 / 23.55 ms │            22.84 / 23.70 ±0.47 / 24.14 ms │     no change │
│ QQuery 91 │           58.43 / 59.35 ±0.74 / 60.49 ms │            58.12 / 59.01 ±0.59 / 59.96 ms │     no change │
│ QQuery 92 │           56.87 / 58.00 ±1.82 / 61.60 ms │            56.48 / 57.02 ±0.45 / 57.80 ms │     no change │
│ QQuery 93 │        179.52 / 180.80 ±1.01 / 182.64 ms │         177.44 / 181.30 ±3.77 / 188.27 ms │     no change │
│ QQuery 94 │           60.88 / 62.00 ±1.38 / 64.66 ms │            60.72 / 61.06 ±0.23 / 61.39 ms │     no change │
│ QQuery 95 │        126.06 / 126.85 ±0.62 / 127.95 ms │         125.09 / 126.08 ±0.85 / 127.60 ms │     no change │
│ QQuery 96 │           70.70 / 72.68 ±1.40 / 75.04 ms │            68.74 / 69.52 ±0.56 / 70.17 ms │     no change │
│ QQuery 97 │        116.86 / 118.08 ±1.01 / 119.28 ms │         116.88 / 118.33 ±1.68 / 121.58 ms │     no change │
│ QQuery 98 │        150.71 / 152.14 ±1.19 / 153.54 ms │         148.32 / 150.96 ±2.47 / 155.34 ms │     no change │
│ QQuery 99 │ 10729.03 / 10820.23 ±69.00 / 10906.36 ms │ 10684.37 / 10794.78 ±110.72 / 10962.14 ms │     no change │
└───────────┴──────────────────────────────────────────┴───────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                                   ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                                   │ 30484.02ms │
│ Total Time (intermeidate-result-blocked-approach)   │ 30326.75ms │
│ Average Time (HEAD)                                 │   307.92ms │
│ Average Time (intermeidate-result-blocked-approach) │   306.33ms │
│ Queries Faster                                      │          1 │
│ Queries Slower                                      │          1 │
│ Queries with No Change                              │         97 │
│ Queries with Failure                                │          0 │
└─────────────────────────────────────────────────────┴────────────┘

Resource Usage

tpcds — base (merge-base)

Metric	Value
Wall time	155.0s
Peak memory	6.4 GiB
Avg memory	5.7 GiB
CPU user	242.0s
CPU sys	7.7s
Peak spill	0 B

tpcds — branch

Metric	Value
Wall time	155.0s
Peak memory	6.4 GiB
Avg memory	5.7 GiB
CPU user	240.7s
CPU sys	7.6s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-04-29T12:45:58Z

🤖 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 intermeidate-result-blocked-approach
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃  intermeidate-result-blocked-approach ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │          1.20 / 4.69 ±6.81 / 18.31 ms │          1.24 / 4.81 ±6.91 / 18.63 ms │     no change │
│ QQuery 1  │        12.31 / 12.90 ±0.32 / 13.19 ms │        13.31 / 13.63 ±0.24 / 13.98 ms │  1.06x slower │
│ QQuery 2  │        37.22 / 37.65 ±0.25 / 37.95 ms │        36.69 / 36.97 ±0.33 / 37.61 ms │     no change │
│ QQuery 3  │        32.18 / 33.68 ±1.39 / 35.77 ms │        32.71 / 33.30 ±0.64 / 34.13 ms │     no change │
│ QQuery 4  │     248.07 / 258.84 ±8.10 / 268.82 ms │     273.71 / 282.89 ±9.01 / 299.80 ms │  1.09x slower │
│ QQuery 5  │     286.89 / 297.68 ±9.46 / 312.31 ms │     286.98 / 298.35 ±6.28 / 305.42 ms │     no change │
│ QQuery 6  │           7.52 / 7.75 ±0.24 / 8.16 ms │           6.82 / 7.50 ±0.36 / 7.82 ms │     no change │
│ QQuery 7  │        14.90 / 15.08 ±0.12 / 15.20 ms │        14.12 / 14.32 ±0.12 / 14.46 ms │ +1.05x faster │
│ QQuery 8  │     337.53 / 345.55 ±9.01 / 362.21 ms │     333.37 / 351.74 ±9.95 / 363.07 ms │     no change │
│ QQuery 9  │    464.60 / 483.70 ±16.12 / 510.06 ms │    461.18 / 474.85 ±16.45 / 505.45 ms │     no change │
│ QQuery 10 │        75.17 / 76.56 ±0.87 / 77.89 ms │        74.67 / 76.74 ±3.02 / 82.73 ms │     no change │
│ QQuery 11 │        86.06 / 87.43 ±1.05 / 88.94 ms │        87.27 / 88.03 ±0.48 / 88.76 ms │     no change │
│ QQuery 12 │    283.93 / 296.27 ±12.32 / 317.43 ms │     299.81 / 303.52 ±3.50 / 308.05 ms │     no change │
│ QQuery 13 │     408.64 / 420.85 ±9.93 / 437.20 ms │     433.94 / 444.02 ±9.34 / 459.35 ms │  1.06x slower │
│ QQuery 14 │     291.08 / 300.86 ±7.46 / 307.85 ms │    286.68 / 306.20 ±10.34 / 316.37 ms │     no change │
│ QQuery 15 │     292.78 / 302.90 ±8.71 / 313.91 ms │    316.43 / 342.96 ±17.00 / 359.50 ms │  1.13x slower │
│ QQuery 16 │    650.21 / 660.55 ±13.01 / 685.91 ms │    620.71 / 645.65 ±16.13 / 663.58 ms │     no change │
│ QQuery 17 │    639.90 / 655.32 ±10.36 / 669.82 ms │    633.90 / 654.99 ±14.13 / 676.69 ms │     no change │
│ QQuery 18 │ 1327.57 / 1355.57 ±16.41 / 1376.44 ms │  1317.82 / 1321.98 ±5.75 / 1333.21 ms │     no change │
│ QQuery 19 │        30.08 / 32.87 ±4.67 / 42.16 ms │       28.83 / 36.34 ±10.06 / 54.79 ms │  1.11x slower │
│ QQuery 20 │    520.14 / 533.95 ±12.12 / 548.47 ms │     523.39 / 529.56 ±7.76 / 544.75 ms │     no change │
│ QQuery 21 │     599.06 / 608.53 ±6.16 / 616.52 ms │     602.97 / 611.65 ±6.15 / 620.66 ms │     no change │
│ QQuery 22 │ 1080.59 / 1093.18 ±12.40 / 1114.58 ms │  1085.22 / 1096.94 ±6.83 / 1104.52 ms │     no change │
│ QQuery 23 │ 3399.00 / 3476.23 ±51.02 / 3533.61 ms │ 3417.06 / 3452.18 ±38.39 / 3524.14 ms │     no change │
│ QQuery 24 │        42.58 / 43.01 ±0.50 / 43.98 ms │        44.93 / 46.95 ±2.69 / 51.88 ms │  1.09x slower │
│ QQuery 25 │     114.67 / 117.89 ±2.13 / 120.31 ms │     114.49 / 116.49 ±1.96 / 119.04 ms │     no change │
│ QQuery 26 │        44.09 / 46.08 ±2.23 / 50.34 ms │        43.42 / 44.42 ±1.80 / 48.02 ms │     no change │
│ QQuery 27 │     671.04 / 684.12 ±9.27 / 694.03 ms │    691.05 / 701.46 ±11.09 / 722.79 ms │     no change │
│ QQuery 28 │ 3053.45 / 3071.24 ±19.18 / 3105.73 ms │ 3068.83 / 3108.85 ±39.04 / 3180.45 ms │     no change │
│ QQuery 29 │        43.02 / 44.37 ±1.62 / 47.54 ms │       42.37 / 50.09 ±14.53 / 79.14 ms │  1.13x slower │
│ QQuery 30 │    307.50 / 315.86 ±10.46 / 335.97 ms │     313.14 / 325.33 ±8.08 / 335.93 ms │     no change │
│ QQuery 31 │    306.97 / 321.06 ±10.68 / 336.95 ms │    310.16 / 324.41 ±11.33 / 338.17 ms │     no change │
│ QQuery 32 │ 1066.60 / 1077.99 ±13.14 / 1103.24 ms │ 1070.57 / 1099.60 ±29.96 / 1152.43 ms │     no change │
│ QQuery 33 │ 1495.03 / 1532.43 ±41.29 / 1605.16 ms │ 1484.13 / 1504.08 ±17.25 / 1534.34 ms │     no change │
│ QQuery 34 │ 1519.36 / 1536.75 ±22.40 / 1580.33 ms │ 1518.67 / 1563.30 ±32.55 / 1609.63 ms │     no change │
│ QQuery 35 │    304.44 / 325.27 ±11.11 / 336.38 ms │    368.67 / 411.91 ±46.46 / 498.51 ms │  1.27x slower │
│ QQuery 36 │        62.39 / 67.07 ±4.41 / 74.08 ms │        66.32 / 68.84 ±2.05 / 71.17 ms │     no change │
│ QQuery 37 │        36.64 / 38.24 ±1.12 / 39.64 ms │        35.45 / 37.60 ±2.82 / 43.07 ms │     no change │
│ QQuery 38 │        41.88 / 47.12 ±6.11 / 58.64 ms │        40.89 / 45.55 ±7.32 / 59.97 ms │     no change │
│ QQuery 39 │     130.92 / 137.60 ±5.43 / 143.97 ms │    122.17 / 139.96 ±14.48 / 159.49 ms │     no change │
│ QQuery 40 │        15.72 / 16.03 ±0.30 / 16.41 ms │        14.24 / 14.49 ±0.19 / 14.77 ms │ +1.11x faster │
│ QQuery 41 │        14.69 / 14.83 ±0.09 / 14.93 ms │        13.64 / 14.42 ±1.16 / 16.71 ms │     no change │
│ QQuery 42 │        13.96 / 16.36 ±4.32 / 24.99 ms │        14.26 / 20.25 ±8.14 / 36.17 ms │  1.24x slower │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                                   ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                                   │ 20851.93ms │
│ Total Time (intermeidate-result-blocked-approach)   │ 21067.12ms │
│ Average Time (HEAD)                                 │   484.93ms │
│ Average Time (intermeidate-result-blocked-approach) │   489.93ms │
│ Queries Faster                                      │          2 │
│ Queries Slower                                      │          9 │
│ Queries with No Change                              │         32 │
│ Queries with Failure                                │          0 │
└─────────────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric	Value
Wall time	105.0s
Peak memory	29.9 GiB
Avg memory	23.0 GiB
CPU user	1101.0s
CPU sys	69.4s
Peak spill	0 B

clickbench_partitioned — branch

Metric	Value
Wall time	110.0s
Peak memory	30.9 GiB
Avg memory	23.4 GiB
CPU user	1106.9s
CPU sys	71.1s
Peak spill	0 B

File an issue against this benchmark runner

Rachelint · 2026-04-29T17:32:16Z

Refactoring for zero cost when disabling the blocked approach

ariel-miculas · 2026-04-30T15:27:55Z

                if let Some(emit_to) = self.group_ordering.oom_emit_to(n)
                    && let Some(batch) = self.emit(emit_to, false)?
                {
                    return Ok(Some(ExecutionState::ProducingOutput(batch)));


should there also be a transition to ProducingBlocks here instead of ProducingOutput if enable_blocked_groups is set?

Could be outside of this PR's scope, but another question I have is why isn't there a transition to ProducingBlocks as soon as we have a batch ready. Instead, all the input is consumed first before producing the first batch (assuming there's no group ordering)

Rachelint changed the title ~~Impl Intermeidate result blocked approach framework~~ Impl intermeidate result blocked approach framework Apr 5, 2025

Rachelint changed the title ~~Impl intermeidate result blocked approach framework~~ Impl intermeidate result blocked approach sketch Apr 5, 2025

Rachelint mentioned this pull request Apr 5, 2025

Improve aggregate performance with adaptive sizing in accumulators / avoiding reallocations in accumulators #7065

Open

2 tasks

github-actions Bot added the logical-expr Logical plan and expressions label Apr 5, 2025

Rachelint mentioned this pull request Apr 8, 2025

Implement PoC block allocation for count accumulator #15642

Closed

Rachelint force-pushed the intermeidate-result-blocked-approach branch 2 times, most recently from cc37eba to f690940 Compare April 9, 2025 14:37

Rachelint mentioned this pull request Apr 9, 2025

Sketch for aggregation intermediate results blocked management #11943

Closed

github-actions Bot added the functions Changes to functions implementation label Apr 10, 2025

Rachelint force-pushed the intermeidate-result-blocked-approach branch from 95c6a36 to a4c6f42 Compare April 10, 2025 11:10

github-actions Bot added the physical-expr Changes to the physical-expr crates label Apr 10, 2025

Rachelint force-pushed the intermeidate-result-blocked-approach branch 6 times, most recently from 2100a5b to 0ee951c Compare April 17, 2025 11:56

Rachelint force-pushed the intermeidate-result-blocked-approach branch 2 times, most recently from c51d409 to 2863809 Compare April 20, 2025 14:46

github-actions Bot added execution Related to the execution crate common Related to common crate sqllogictest SQL Logic Tests (.slt) labels Apr 21, 2025

Rachelint force-pushed the intermeidate-result-blocked-approach branch 3 times, most recently from 31d660d to 2b8dd1e Compare April 22, 2025 18:52

define the needed methods in GroupAccumulator and GroupValues.

4353748

Fix blocked groups default regressions

4891fa5

Rachelint added 2 commits April 28, 2026 19:11

remove some useless codes.

9a7527d

Merge branch 'main' into intermeidate-result-blocked-approach

a818cd8

Rachelint added 2 commits April 29, 2026 08:35

Merge branch 'main' into intermeidate-result-blocked-approach

902972f

init with 128 cap.

47b6976

rluvaton mentioned this pull request Apr 29, 2026

build: allow posting comments on PRs made from forks and fix missing protobuf #21913

Merged

Add block store abstraction

f0ba361

Rachelint added 2 commits April 30, 2026 01:32

Refactor group accumulator block stores

d1f96f1

Encapsulate primitive group state

48078ec

ariel-miculas reviewed Apr 30, 2026

View reviewed changes

Conversation

Rachelint commented Apr 5, 2025 • 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?

Are these changes tested?

Are there any user-facing changes?

Uh oh!

Dandandan commented Apr 8, 2025

Uh oh!

Rachelint commented Apr 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Rachelint commented Apr 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Rachelint commented Apr 21, 2025

Uh oh!

adriangbot commented Apr 28, 2026

Uh oh!

adriangbot commented Apr 28, 2026

Uh oh!

Rachelint commented Apr 28, 2026

Uh oh!

Rachelint commented Apr 28, 2026

Uh oh!

Dandandan commented Apr 28, 2026

Uh oh!

adriangbot commented Apr 28, 2026

Uh oh!

adriangbot commented Apr 28, 2026

Uh oh!

adriangbot commented Apr 28, 2026

Uh oh!

adriangbot commented Apr 28, 2026

Uh oh!

adriangbot commented Apr 28, 2026

Uh oh!

adriangbot commented Apr 28, 2026

Uh oh!

Rachelint commented Apr 29, 2026

Uh oh!

Dandandan commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

Rachelint commented Apr 29, 2026

Uh oh!

ariel-miculas Apr 30, 2026

Choose a reason for hiding this comment

Uh oh!

ariel-miculas Apr 30, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

9 participants

Rachelint commented Apr 5, 2025 •

edited

Loading

Rachelint commented Apr 8, 2025 •

edited

Loading

Rachelint commented Apr 17, 2025 •

edited

Loading