You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@datafusion.apache.org by "adriangb (via GitHub)" <gi...@apache.org> on 2026/09/24 18:03:25 UTC
[PR] perf: push hash join dynamic filter bounds and membership as separate optional filters [datafusion]
adriangb opened a new pull request, #25713:
URL: https://github.com/apache/datafusion/pull/25713
## Which issue does this PR close?
- Part of #22883 (experimental "optional filters" stack, design notes: https://claude.ai/artifact/SSz7t6hPyhFWp1MDPecVqt). Builds on the prototype #24235.
- **Depends on #25673, #25681, #25674, #25682 and #25677.** Review only the top two commits (the change and the benchmark).
## Rationale for this change
A partitioned hash join pushes one dynamic filter into the probe side scan:
```
DynamicFilter [ CASE hash_repartition % N WHEN i THEN bounds_i AND membership_i ... END ]
```
Statistics pruning cannot use a `CASE` expression. Thus in partitioned mode the scan cannot use the join key bounds to skip row groups, even when the build side matches only a small, ordered slice of the probe key (for example one day of time-ordered events). The whole probe table is read.
Also, the bounds and the membership check have very different cost and selectivity, but a consumer can only keep or pause the combined filter.
Example (`hj_ordered_subset` Q01: `events` has 20M rows sorted by `event_id` in 200 row groups; the build side matches one day, 1% of the key range; partitioned join):
| | Probe row groups read | Bytes scanned | Join probe rows | Time (min) |
|---|---|---|---|---|
| before | 200 of 200 | 134 MB | 15.3 M | 257β271 ms |
| after | 2 of 200 | 1.45 MB | 200 K | 19β20 ms |
## What changes are included in this PR?
`HashJoinExec` pushes two optional dynamic filters instead of one:
```
Optional(DynamicFilter [ event_id >= 8400000 AND event_id <= 8599999 ]) -- bounds
AND Optional(DynamicFilter [ CASE hash_repartition % 12 WHEN 0 THEN hash_lookup ... END ]) -- membership
```
- **Bounds:** the union of the bounds of all build partitions, one range per key column. Statistics pruning can use it in all join modes. For range partitioning, the disjoint ranges of the partitions are kept (up to 8 per column).
- **Membership:** the routed `CASE` without the per-partition bounds (they never reject a row that the membership check of that partition accepts), or the collapsed `IN` list from #25677.
- **CollectLeft** is split the same way, for consistency.
- Each filter has its own expression id, gets `update()` / `mark_complete()`, is transferred across join keys as an optional view, and round trips through proto. A proto plan without the new field decodes to the old single filter.
- Fallbacks: with a canceled partition or no usable bounds, the bounds filter stays `true` and the bounds stay in the `CASE`. An empty build side sets both filters to `false`. The null-equal / null-aware `IS NULL OR` escape wraps each filter.
- Because each part is a separate optional filter, the Parquet gate (#25682) measures and pauses each part on its own. On TPC-H Q3 the bounds reject 0.01% of rows and are paused, while the membership check (95.5% rejected) stays on.
New benchmark suite `hj_ordered_subset` (`./bench.sh run hj_ordered_subset`): time-ordered probe table, build side matching 1 day, 10 days, or a scattered 1% of keys (control), in partitioned and CollectLeft mode.
## What is the testing strategy for this PR?
- 12 new unit tests for the split filters (canceled and empty partitions, range partitioning, null-equal, bounds only, `mark_complete` on both) and 11 for the bounds union.
- The proto round trip checks that both filters stay linked to their probe side copies.
- `hash_join` tests (also with `force_hash_collisions`), `physical-plan` lib tests, `filter_pushdown`, proto and `parquet_integration` tests pass. Snapshot diffs change only the dynamic filter text and derived pruning predicates.
- The benchmark suite checks the join mode with `expect_plan` and asserts that the join count equals the build count. Results are identical before and after in all combinations.
### Benchmark
`hj_ordered_subset`, `optional_filter_mode = adaptive`, min time in ms, two runs (A: n=15, B: n=24), base = the same stack without this PR. The machine was busy (load average 40β58 on 12 cores): a no-join control query varied up to Β±25%, so only large differences are significant.
| pushdown_filters | Join mode | Query | Base A / B | This PR A / B |
|---|---|---|---|---|
| true | Partitioned | Q01 ordered, 1 day | 271 / 257 | **19 / 20** |
| true | Partitioned | Q02 ordered, 10 days | 384 / 332 | **188 / 134** |
| true | Partitioned | Q03 scattered (control) | 298 / 273 | 244 / 238 |
| true | CollectLeft | Q04 / Q05 / Q06 | 13 / 102 / 253 | 17 / 140 / 229 (noise) |
| false | Partitioned | Q01 ordered, 1 day | 177 / 193 | **15 / 13** |
| false | Partitioned | Q02 ordered, 10 days | 263 / 355 | **102 / 170** |
| false | Partitioned | Q03 scattered (control) | 254 / 270 | 289 / 246 |
| false | CollectLeft | Q04 / Q05 / Q06 | 12 / 106 / 200 | 14 / 132 / 257 (noise) |
The win does not depend on `pushdown_filters`: it comes from statistics pruning. CollectLeft already pruned before this PR (its filter is a plain conjunction), so it does not change. On TPC-H and TPC-DS SF1 this PR is neutral within noise.
Known limitations:
- A paused optional filter is still a separate row filter stage in the scan, so splitting adds a small cost when the bounds are useless (for example `predicate_cache_records` doubles). A follow-up can drop paused filters from the `RowFilter` at the next row group.
- In the scattered CollectLeft control with pushdown on, the gate also paused the membership part for part of the scan (the join probed 447K rows instead of 226K). The time effect was within noise, but the gate may pause too early there.
- Q01 needs the mode forced to Partitioned (`hash_join_single_partition_threshold[_rows] = 0`); with the defaults the 1-day build side plans as CollectLeft. Numbers should be confirmed on a quiet machine or the benchmark bot.
## Are there any user-facing changes?
The text of hash join dynamic filters in `EXPLAIN` changes (two filters instead of one). New proto field for the second filter (backward compatible for decoding). New benchmark suite.
π€ Generated with [Claude Code](https://claude.com/claude-code)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push hash join dynamic filter bounds and membership as separate optional filters [datafusion]
Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5819703349
<!-- semver-check-comment -->
Thank you for opening this pull request!
Reviewer note: [cargo-semver-checks](https://github.com/obi1kenobi/cargo-semver-checks) reported the current version number is not SemVer-compatible with the changes in this pull request (compared against the base branch).
<details>
<summary>Details</summary>
```
Cloning apache/main
Building datafusion v55.1.0 (current)
Built [ 62.621s] (current)
Parsing datafusion v55.1.0 (current)
Parsed [ 0.035s] (current)
Building datafusion v55.1.0 (baseline)
Built [ 61.811s] (baseline)
Parsing datafusion v55.1.0 (baseline)
Parsed [ 0.036s] (baseline)
Checking datafusion v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 0.596s] 223 checks: 223 pass, 31 skip
Summary no semver update required
Finished [ 126.739s] datafusion
Building datafusion-common v55.1.0 (current)
Built [ 35.299s] (current)
Parsing datafusion-common v55.1.0 (current)
Parsed [ 0.063s] (current)
Building datafusion-common v55.1.0 (baseline)
Built [ 35.275s] (baseline)
Parsing datafusion-common v55.1.0 (baseline)
Parsed [ 0.064s] (baseline)
Checking datafusion-common v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 0.660s] 223 checks: 222 pass, 1 fail, 0 warn, 31 skip
--- failure constructible_struct_adds_field: struct exhaustively constructible through public API adds field ---
Description:
A pub struct that could be exhaustively constructed with a literal using only public API has a new pub field, breaking existing exhaustive literals.
ref: https://doc.rust-lang.org/reference/expressions/struct-expr.html
impl: https://github.com/obi1kenobi/cargo-semver-checks/tree/v0.50.0/src/lints/constructible_struct_adds_field.ron
Failed in:
field ExecutionOptions.optional_filter_mode in /home/runner/work/datafusion/datafusion/datafusion/common/src/config.rs:949
field ExecutionOptions.optional_filter_max_pass_ratio in /home/runner/work/datafusion/datafusion/datafusion/common/src/config.rs:949
field ExecutionOptions.optional_filter_min_saving_ns_per_row in /home/runner/work/datafusion/datafusion/datafusion/common/src/config.rs:949
Summary semver requires new major version: 1 major and 0 minor checks failed
Finished [ 72.673s] datafusion-common
Building datafusion-datasource-parquet v55.1.0 (current)
Built [ 52.317s] (current)
Parsing datafusion-datasource-parquet v55.1.0 (current)
Parsed [ 0.047s] (current)
Building datafusion-datasource-parquet v55.1.0 (baseline)
Built [ 52.359s] (baseline)
Parsing datafusion-datasource-parquet v55.1.0 (baseline)
Parsed [ 0.035s] (baseline)
Checking datafusion-datasource-parquet v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 0.150s] 223 checks: 223 pass, 31 skip
Summary no semver update required
Finished [ 106.227s] datafusion-datasource-parquet
Building datafusion-physical-expr v55.1.0 (current)
Built [ 30.844s] (current)
Parsing datafusion-physical-expr v55.1.0 (current)
Parsed [ 0.052s] (current)
Building datafusion-physical-expr v55.1.0 (baseline)
Built [ 30.536s] (baseline)
Parsing datafusion-physical-expr v55.1.0 (baseline)
Parsed [ 0.052s] (baseline)
Checking datafusion-physical-expr v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 0.363s] 223 checks: 223 pass, 31 skip
Summary no semver update required
Finished [ 62.796s] datafusion-physical-expr
Building datafusion-physical-plan v55.1.0 (current)
Built [ 40.954s] (current)
Parsing datafusion-physical-plan v55.1.0 (current)
Parsed [ 0.173s] (current)
Building datafusion-physical-plan v55.1.0 (baseline)
Built [ 40.760s] (baseline)
Parsing datafusion-physical-plan v55.1.0 (baseline)
Parsed [ 0.171s] (baseline)
Checking datafusion-physical-plan v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 0.650s] 223 checks: 223 pass, 31 skip
Summary no semver update required
Finished [ 83.890s] datafusion-physical-plan
Building datafusion-proto v55.1.0 (current)
Built [ 58.132s] (current)
Parsing datafusion-proto v55.1.0 (current)
Parsed [ 0.017s] (current)
Building datafusion-proto v55.1.0 (baseline)
Built [ 57.692s] (baseline)
Parsing datafusion-proto v55.1.0 (baseline)
Parsed [ 0.018s] (baseline)
Checking datafusion-proto v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 0.119s] 223 checks: 223 pass, 31 skip
Summary no semver update required
Finished [ 117.333s] datafusion-proto
Building datafusion-proto-models v55.1.0 (current)
Built [ 26.505s] (current)
Parsing datafusion-proto-models v55.1.0 (current)
Parsed [ 0.133s] (current)
Building datafusion-proto-models v55.1.0 (baseline)
Built [ 26.297s] (baseline)
Parsing datafusion-proto-models v55.1.0 (baseline)
Parsed [ 0.136s] (baseline)
Checking datafusion-proto-models v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 1.811s] 223 checks: 221 pass, 2 fail, 0 warn, 31 skip
--- failure constructible_struct_adds_field: struct exhaustively constructible through public API adds field ---
Description:
A pub struct that could be exhaustively constructed with a literal using only public API has a new pub field, breaking existing exhaustive literals.
ref: https://doc.rust-lang.org/reference/expressions/struct-expr.html
impl: https://github.com/obi1kenobi/cargo-semver-checks/tree/v0.50.0/src/lints/constructible_struct_adds_field.ron
Failed in:
field HashJoinExecNode.dynamic_filter_bounds in /home/runner/work/datafusion/datafusion/datafusion/proto-models/src/generated/prost.rs:2160
field HashJoinExecNode.dynamic_filter_bounds in /home/runner/work/datafusion/datafusion/datafusion/proto-models/src/generated/prost.rs:2160
--- failure enum_variant_added: enum variant added on exhaustive enum ---
Description:
A publicly-visible enum without #[non_exhaustive] has a new variant.
ref: https://doc.rust-lang.org/cargo/reference/semver.html#enum-variant-new
impl: https://github.com/obi1kenobi/cargo-semver-checks/tree/v0.50.0/src/lints/enum_variant_added.ron
Failed in:
variant ExprType:OptionalFilter in /home/runner/work/datafusion/datafusion/datafusion/proto-models/src/generated/prost.rs:1686
variant ExprType:OptionalFilter in /home/runner/work/datafusion/datafusion/datafusion/proto-models/src/generated/prost.rs:1686
Summary semver requires new major version: 2 major and 0 minor checks failed
Finished [ 55.976s] datafusion-proto-models
Building datafusion-pruning v55.1.0 (current)
Built [ 42.606s] (current)
Parsing datafusion-pruning v55.1.0 (current)
Parsed [ 0.020s] (current)
Building datafusion-pruning v55.1.0 (baseline)
Built [ 42.595s] (baseline)
Parsing datafusion-pruning v55.1.0 (baseline)
Parsed [ 0.016s] (baseline)
Checking datafusion-pruning v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 0.079s] 223 checks: 223 pass, 31 skip
Summary no semver update required
Finished [ 86.350s] datafusion-pruning
Building datafusion-sqllogictest v55.1.0 (current)
Built [ 106.209s] (current)
Parsing datafusion-sqllogictest v55.1.0 (current)
Parsed [ 0.020s] (current)
Building datafusion-sqllogictest v55.1.0 (baseline)
Built [ 106.442s] (baseline)
Parsing datafusion-sqllogictest v55.1.0 (baseline)
Parsed [ 0.033s] (baseline)
Checking datafusion-sqllogictest v55.1.0 -> v55.1.0 (no change; assume patch)
Checked [ 0.094s] 223 checks: 223 pass, 31 skip
Summary no semver update required
Finished [ 215.460s] datafusion-sqllogictest
```
</details>
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push partitioned hash join dynamic filter bounds and membership as separate filters [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5821670744
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/25713#issuecomment-5821636144)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c5821636144-2701-p2s6c 6.12.94+ #1 SMP Wed Aug 19 07:47:20 UTC 2026 aarch64 GNU/Linux`
<details><summary>CPU Details (lscpu)</summary>
```
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 hash-join-split-optional-filter (83d19600b4a207db882058b35ec5126b1a031f61) to 95bb0a0 (merge-base) [diff](https://github.com/apache/datafusion/compare/95bb0a0dfa48ca20d017faa55b2fc96ee2f300fe..83d19600b4a207db882058b35ec5126b1a031f61)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpcds
env:
DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
```
</details>
Results will be posted here when complete
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push partitioned hash join dynamic filter bounds and membership as separate filters [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5821758448
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/25713#issuecomment-5821635949)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing hash-join-split-optional-filter (83d19600b4a207db882058b35ec5126b1a031f61) to 95bb0a0 (merge-base) [diff](https://github.com/apache/datafusion/compare/95bb0a0dfa48ca20d017faa55b2fc96ee2f300fe..83d19600b4a207db882058b35ec5126b1a031f61)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpch
```
</details>
<details><summary>CPU Details (lscpu)</summary>
```
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>
<details><summary>Details</summary>
<p>
```
Comparing HEAD and hash-join-split-optional-filter
--------------------
Benchmark tpch_sf1.json
--------------------
βββββββββββββ³βββββββββββ³ββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β hash-join-split-optional-filter β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 40.79 ms β 40.72 ms β no change β
β QQuery 2 β 18.63 ms β 18.26 ms β no change β
β QQuery 3 β 28.01 ms β 28.15 ms β no change β
β QQuery 4 β 17.33 ms β 16.79 ms β no change β
β QQuery 5 β 34.95 ms β 34.98 ms β no change β
β QQuery 6 β 16.16 ms β 16.10 ms β no change β
β QQuery 7 β 41.25 ms β 41.66 ms β no change β
β QQuery 8 β 40.16 ms β 40.18 ms β no change β
β QQuery 9 β 48.31 ms β 48.30 ms β no change β
β QQuery 10 β 40.45 ms β 40.27 ms β no change β
β QQuery 11 β 13.42 ms β 13.33 ms β no change β
β QQuery 12 β 20.81 ms β 20.88 ms β no change β
β QQuery 13 β 39.30 ms β 39.19 ms β no change β
β QQuery 14 β 24.55 ms β 24.59 ms β no change β
β QQuery 15 β 30.92 ms β 30.68 ms β no change β
β QQuery 16 β 14.15 ms β 13.84 ms β no change β
β QQuery 17 β 72.42 ms β 67.51 ms β +1.07x faster β
β QQuery 18 β 60.33 ms β 60.43 ms β no change β
β QQuery 19 β 31.93 ms β 31.47 ms β no change β
β QQuery 20 β 31.92 ms β 31.96 ms β no change β
β QQuery 21 β 54.99 ms β 54.46 ms β no change β
β QQuery 22 β 13.59 ms β 13.52 ms β no change β
βββββββββββββ΄βββββββββββ΄ββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββ
β Benchmark Summary β β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 734.39ms β
β Total Time (hash-join-split-optional-filter) β 727.27ms β
β Average Time (HEAD) β 33.38ms β
β Average Time (hash-join-split-optional-filter) β 33.06ms β
β Queries Faster β 1 β
β Queries Slower β 0 β
β Queries with No Change β 21 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββ
Distribution per query (min / mean Β±stddev / max):
Comparing HEAD and hash-join-split-optional-filter
--------------------
Benchmark tpch_sf1.json
--------------------
βββββββββββββ³βββββββββββββββββββββββββββββββββ³ββββββββββββββββββββββββββββββββββ³ββββββββββββ
β Query β HEAD β hash-join-split-optional-filter β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 40.79 / 41.66 Β±1.03 / 43.58 ms β 40.72 / 41.44 Β±1.04 / 43.46 ms β no change β
β QQuery 2 β 18.63 / 19.11 Β±0.53 / 19.88 ms β 18.26 / 18.86 Β±0.33 / 19.17 ms β no change β
β QQuery 3 β 28.01 / 28.43 Β±0.36 / 29.00 ms β 28.15 / 28.92 Β±1.19 / 31.29 ms β no change β
β QQuery 4 β 17.33 / 17.65 Β±0.44 / 18.51 ms β 16.79 / 17.07 Β±0.21 / 17.41 ms β no change β
β QQuery 5 β 34.95 / 35.37 Β±0.35 / 36.02 ms β 34.98 / 35.15 Β±0.30 / 35.74 ms β no change β
β QQuery 6 β 16.16 / 16.24 Β±0.05 / 16.30 ms β 16.10 / 16.36 Β±0.21 / 16.73 ms β no change β
β QQuery 7 β 41.25 / 42.79 Β±1.50 / 45.32 ms β 41.66 / 42.72 Β±0.88 / 43.86 ms β no change β
β QQuery 8 β 40.16 / 41.45 Β±1.55 / 44.42 ms β 40.18 / 40.65 Β±0.58 / 41.66 ms β no change β
β QQuery 9 β 48.31 / 49.74 Β±0.75 / 50.54 ms β 48.30 / 48.79 Β±0.44 / 49.58 ms β no change β
β QQuery 10 β 40.45 / 40.66 Β±0.22 / 41.07 ms β 40.27 / 40.36 Β±0.07 / 40.48 ms β no change β
β QQuery 11 β 13.42 / 13.83 Β±0.42 / 14.60 ms β 13.33 / 13.51 Β±0.16 / 13.71 ms β no change β
β QQuery 12 β 20.81 / 21.44 Β±0.51 / 22.29 ms β 20.88 / 21.32 Β±0.34 / 21.89 ms β no change β
β QQuery 13 β 39.30 / 40.69 Β±1.30 / 42.82 ms β 39.19 / 39.68 Β±0.69 / 41.04 ms β no change β
β QQuery 14 β 24.55 / 24.75 Β±0.17 / 25.00 ms β 24.59 / 25.36 Β±1.16 / 27.67 ms β no change β
β QQuery 15 β 30.92 / 31.70 Β±0.51 / 32.43 ms β 30.68 / 31.30 Β±0.69 / 32.61 ms β no change β
β QQuery 16 β 14.15 / 14.30 Β±0.14 / 14.51 ms β 13.84 / 14.22 Β±0.37 / 14.87 ms β no change β
β QQuery 17 β 72.42 / 72.82 Β±0.38 / 73.37 ms β 67.51 / 69.31 Β±2.14 / 72.94 ms β no change β
β QQuery 18 β 60.33 / 61.62 Β±1.08 / 62.96 ms β 60.43 / 63.05 Β±2.89 / 68.53 ms β no change β
β QQuery 19 β 31.93 / 32.63 Β±0.56 / 33.55 ms β 31.47 / 31.75 Β±0.23 / 32.10 ms β no change β
β QQuery 20 β 31.92 / 32.15 Β±0.20 / 32.42 ms β 31.96 / 32.10 Β±0.11 / 32.28 ms β no change β
β QQuery 21 β 54.99 / 55.46 Β±0.54 / 56.48 ms β 54.46 / 55.12 Β±0.59 / 56.04 ms β no change β
β QQuery 22 β 13.59 / 14.08 Β±0.41 / 14.71 ms β 13.52 / 13.93 Β±0.23 / 14.21 ms β no change β
βββββββββββββ΄βββββββββββββββββββββββββββββββββ΄ββββββββββββββββββββββββββββββββββ΄ββββββββββββ
ββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββ
β Benchmark Summary β β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 748.56ms β
β Total Time (hash-join-split-optional-filter) β 740.96ms β
β Average Time (HEAD) β 34.03ms β
β Average Time (hash-join-split-optional-filter) β 33.68ms β
β Queries Faster β 0 β
β Queries Slower β 0 β
β Queries with No Change β 22 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**tpch β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 5.0s |
| Peak memory | 1.3 GiB |
| Avg memory | 512.6 MiB |
| CPU user | 20.9s |
| CPU sys | 1.7s |
| Peak spill | 0 B |
**tpch β branch**
| Metric | Value |
|--------|-------|
| Wall time | 5.0s |
| Peak memory | 1.3 GiB |
| Avg memory | 500.3 MiB |
| CPU user | 20.5s |
| CPU sys | 1.7s |
| Peak spill | 0 B |
</details>
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push partitioned hash join dynamic filter bounds and membership as separate filters [datafusion]
Posted by "adriangb (via GitHub)" <gi...@apache.org>.
adriangb commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5821816529
## Benchmark summary (83d19600b4)
Base is merge-base 95bb0a0, the same base as the [#25677 results](https://github.com/apache/datafusion/pull/25677#issuecomment-5809027081). The measured change is therefore #25677 + split. Times are the minimum of 5 iterations. Change = PR / base - 1.
### Totals
| Suite | Config | Result | Base | PR | Change | #25677 change |
|---|---|---|---|---|---|---|
| TPC-H SF1 | default | [link](https://github.com/apache/datafusion/pull/25713#issuecomment-5821758448) | 734.37 ms | 727.27 ms | -1.0% | -0.4% |
| TPC-DS SF1 | default | [link](https://github.com/apache/datafusion/pull/25713#issuecomment-5821787004) | 9138.20 ms | 9022.41 ms | -1.3% | +1.7% |
| TPC-H SF1 | `pushdown_filters=true` | [link](https://github.com/apache/datafusion/pull/25713#issuecomment-5821761074) | 847.72 ms | 815.06 ms | -3.9% | -4.1% |
| TPC-H SF10 | `pushdown_filters=true` | [link](https://github.com/apache/datafusion/pull/25713#issuecomment-5821783995) | 8012.40 ms | 7903.48 ms | -1.4% | -3.2% |
| TPC-DS SF1 | `pushdown_filters=true` | [link](https://github.com/apache/datafusion/pull/25713#issuecomment-5821775779) | 7826.07 ms | 7319.96 ms | **-6.5%** | -7.4% |
### Queries beyond ~7%
| Suite | Config | Query | Base | PR | Change | #25677 change | Cause |
|---|---|---|---|---|---|---|---|
| TPC-DS SF1 | default | Q6 | 35.59 ms | 38.42 ms | +8.0% | +5.4% | noise (mean +6%) |
| TPC-DS SF1 | default | Q34 | 10.75 ms | 9.57 ms | -11.0% | +1.5% | split or noise (mean -7%, pushdown run flat) |
| TPC-DS SF1 | default | Q65 | 133.17 ms | 67.72 ms | **-49.1%** | -45.9% | #25677 |
| TPC-DS SF1 | pushdown | Q17 | 114.94 ms | 68.54 ms | **-40.4%** | -36.0% | #25677 |
| TPC-DS SF1 | pushdown | Q18 | 161.18 ms | 106.79 ms | **-33.7%** | -34.2% | #25677 |
| TPC-DS SF1 | pushdown | Q25 | 115.21 ms | 66.47 ms | **-42.3%** | -37.7% | #25677 |
| TPC-DS SF1 | pushdown | Q29 | 145.75 ms | 92.97 ms | **-36.2%** | -39.6% | #25677 |
| TPC-DS SF1 | pushdown | Q50 | 124.75 ms | 106.40 ms | **-14.7%** | +1.6% | split |
| TPC-DS SF1 | pushdown | Q64 | 510.41 ms | 337.43 ms | **-33.9%** | -33.1% | #25677 |
| TPC-DS SF1 | pushdown | Q65 | 159.03 ms | 88.47 ms | **-44.4%** | -45.4% | #25677 |
| TPC-DS SF1 | pushdown | Q85 | 164.45 ms | 128.29 ms | **-22.0%** | -3.1% | split |
| TPC-DS SF1 | pushdown | Q95 | 92.42 ms | 55.75 ms | **-39.7%** | -41.4% | #25677 |
| TPC-H SF1 | pushdown | Q17 | 71.08 ms | 44.62 ms | **-37.2%** | -39.5% | #25677 |
| TPC-H SF1 | pushdown | Q18 | 74.25 ms | 66.47 ms | -10.5% | -10.5% | #25677 |
| TPC-H SF10 | pushdown | Q4 | 296.72 ms | 321.48 ms | **+8.3%** | -1.3% | split (spread Β±3-4 ms, not noise) |
| TPC-H SF10 | pushdown | Q18 | 666.71 ms | 431.64 ms | **-35.3%** | -35.9% | #25677 |
The bot cannot run `hj_ordered_subset`: it runs `bench.sh` from `main`, and the merge base has no suite files.
**Conclusion:** the split adds TPC-DS Q50 (-15%) and Q85 (-22%) with pushdown on and keeps all #25677 gains, but TPC-H SF10 Q4 with pushdown on is 8% slower; suite totals are within noise of #25677.
π€ Generated with [Claude Code](https://claude.com/claude-code)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push partitioned hash join dynamic filter bounds and membership as separate filters [datafusion]
Posted by "adriangb (via GitHub)" <gi...@apache.org>.
adriangb commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5821635949
run benchmark tpch tpcds
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push partitioned hash join dynamic filter bounds and membership as separate filters [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5821761074
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/25713#issuecomment-5821636144)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing hash-join-split-optional-filter (83d19600b4a207db882058b35ec5126b1a031f61) to 95bb0a0 (merge-base) [diff](https://github.com/apache/datafusion/compare/95bb0a0dfa48ca20d017faa55b2fc96ee2f300fe..83d19600b4a207db882058b35ec5126b1a031f61)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpch
env:
DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
```
</details>
<details><summary>CPU Details (lscpu)</summary>
```
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>
<details><summary>Details</summary>
<p>
```
Comparing HEAD and hash-join-split-optional-filter
--------------------
Benchmark tpch_sf1.json
--------------------
βββββββββββββ³βββββββββββ³ββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β hash-join-split-optional-filter β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 41.27 ms β 41.03 ms β no change β
β QQuery 2 β 18.88 ms β 19.18 ms β no change β
β QQuery 3 β 31.21 ms β 31.44 ms β no change β
β QQuery 4 β 25.25 ms β 25.34 ms β no change β
β QQuery 5 β 42.89 ms β 43.34 ms β no change β
β QQuery 6 β 22.38 ms β 22.32 ms β no change β
β QQuery 7 β 44.00 ms β 44.73 ms β no change β
β QQuery 8 β 46.10 ms β 46.26 ms β no change β
β QQuery 9 β 63.94 ms β 65.28 ms β no change β
β QQuery 10 β 46.25 ms β 46.08 ms β no change β
β QQuery 11 β 14.83 ms β 14.82 ms β no change β
β QQuery 12 β 39.77 ms β 39.48 ms β no change β
β QQuery 13 β 39.37 ms β 39.38 ms β no change β
β QQuery 14 β 26.20 ms β 26.13 ms β no change β
β QQuery 15 β 41.18 ms β 40.13 ms β no change β
β QQuery 16 β 15.46 ms β 15.69 ms β no change β
β QQuery 17 β 71.08 ms β 44.62 ms β +1.59x faster β
β QQuery 18 β 74.25 ms β 66.47 ms β +1.12x faster β
β QQuery 19 β 32.12 ms β 31.76 ms β no change β
β QQuery 20 β 30.19 ms β 30.18 ms β no change β
β QQuery 21 β 64.96 ms β 65.43 ms β no change β
β QQuery 22 β 16.14 ms β 15.97 ms β no change β
βββββββββββββ΄βββββββββββ΄ββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββ
β Benchmark Summary β β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 847.70ms β
β Total Time (hash-join-split-optional-filter) β 815.05ms β
β Average Time (HEAD) β 38.53ms β
β Average Time (hash-join-split-optional-filter) β 37.05ms β
β Queries Faster β 2 β
β Queries Slower β 0 β
β Queries with No Change β 20 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββ
Distribution per query (min / mean Β±stddev / max):
Comparing HEAD and hash-join-split-optional-filter
--------------------
Benchmark tpch_sf1.json
--------------------
βββββββββββββ³βββββββββββββββββββββββββββββββββ³ββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β hash-join-split-optional-filter β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 41.27 / 42.18 Β±1.08 / 44.23 ms β 41.03 / 41.71 Β±1.16 / 44.03 ms β no change β
β QQuery 2 β 18.88 / 19.56 Β±0.56 / 20.59 ms β 19.18 / 19.50 Β±0.33 / 20.14 ms β no change β
β QQuery 3 β 31.21 / 31.75 Β±0.57 / 32.51 ms β 31.44 / 31.74 Β±0.31 / 32.12 ms β no change β
β QQuery 4 β 25.25 / 26.60 Β±1.05 / 27.88 ms β 25.34 / 25.62 Β±0.21 / 25.87 ms β no change β
β QQuery 5 β 42.89 / 43.55 Β±0.39 / 43.94 ms β 43.34 / 43.57 Β±0.21 / 43.92 ms β no change β
β QQuery 6 β 22.38 / 22.56 Β±0.31 / 23.18 ms β 22.32 / 22.41 Β±0.10 / 22.54 ms β no change β
β QQuery 7 β 44.00 / 44.69 Β±0.53 / 45.62 ms β 44.73 / 45.37 Β±0.41 / 45.92 ms β no change β
β QQuery 8 β 46.10 / 46.93 Β±0.87 / 48.38 ms β 46.26 / 46.45 Β±0.27 / 46.97 ms β no change β
β QQuery 9 β 63.94 / 64.71 Β±0.41 / 65.06 ms β 65.28 / 66.31 Β±0.66 / 67.01 ms β no change β
β QQuery 10 β 46.25 / 46.62 Β±0.49 / 47.59 ms β 46.08 / 47.09 Β±0.63 / 47.80 ms β no change β
β QQuery 11 β 14.83 / 15.65 Β±1.01 / 17.60 ms β 14.82 / 15.67 Β±1.49 / 18.64 ms β no change β
β QQuery 12 β 39.77 / 40.06 Β±0.15 / 40.21 ms β 39.48 / 39.91 Β±0.31 / 40.40 ms β no change β
β QQuery 13 β 39.37 / 40.26 Β±1.08 / 42.30 ms β 39.38 / 40.36 Β±1.06 / 41.98 ms β no change β
β QQuery 14 β 26.20 / 26.35 Β±0.16 / 26.63 ms β 26.13 / 26.28 Β±0.09 / 26.40 ms β no change β
β QQuery 15 β 41.18 / 41.95 Β±0.81 / 43.42 ms β 40.13 / 40.86 Β±0.64 / 41.75 ms β no change β
β QQuery 16 β 15.46 / 15.72 Β±0.16 / 15.90 ms β 15.69 / 16.00 Β±0.30 / 16.56 ms β no change β
β QQuery 17 β 71.08 / 71.62 Β±0.53 / 72.64 ms β 44.62 / 46.00 Β±1.26 / 48.19 ms β +1.56x faster β
β QQuery 18 β 74.25 / 74.69 Β±0.47 / 75.44 ms β 66.47 / 67.64 Β±0.71 / 68.52 ms β +1.10x faster β
β QQuery 19 β 32.12 / 32.89 Β±1.35 / 35.60 ms β 31.76 / 32.36 Β±0.49 / 33.21 ms β no change β
β QQuery 20 β 30.19 / 30.42 Β±0.23 / 30.83 ms β 30.18 / 30.68 Β±0.32 / 31.08 ms β no change β
β QQuery 21 β 64.96 / 66.83 Β±2.07 / 70.68 ms β 65.43 / 66.78 Β±1.13 / 68.86 ms β no change β
β QQuery 22 β 16.14 / 16.32 Β±0.17 / 16.62 ms β 15.97 / 16.20 Β±0.14 / 16.35 ms β no change β
βββββββββββββ΄βββββββββββββββββββββββββββββββββ΄ββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββ
β Benchmark Summary β β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 861.91ms β
β Total Time (hash-join-split-optional-filter) β 828.51ms β
β Average Time (HEAD) β 39.18ms β
β Average Time (hash-join-split-optional-filter) β 37.66ms β
β Queries Faster β 2 β
β Queries Slower β 0 β
β Queries with No Change β 20 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**tpch β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 5.0s |
| Peak memory | 1.1 GiB |
| Avg memory | 709.5 MiB |
| CPU user | 23.3s |
| CPU sys | 1.7s |
| Peak spill | 0 B |
**tpch β branch**
| Metric | Value |
|--------|-------|
| Wall time | 5.0s |
| Peak memory | 1.1 GiB |
| Avg memory | 684.9 MiB |
| CPU user | 23.0s |
| CPU sys | 1.8s |
| Peak spill | 0 B |
</details>
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push partitioned hash join dynamic filter bounds and membership as separate filters [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5821679861
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/25713#issuecomment-5821636144)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c5821636144-2699-rhd48 6.12.94+ #1 SMP Wed Aug 19 07:47:20 UTC 2026 aarch64 GNU/Linux`
<details><summary>CPU Details (lscpu)</summary>
```
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 hash-join-split-optional-filter (83d19600b4a207db882058b35ec5126b1a031f61) to 95bb0a0 (merge-base) [diff](https://github.com/apache/datafusion/compare/95bb0a0dfa48ca20d017faa55b2fc96ee2f300fe..83d19600b4a207db882058b35ec5126b1a031f61)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpch
env:
DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
```
</details>
Results will be posted here when complete
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push partitioned hash join dynamic filter bounds and membership as separate filters [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5821775779
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/25713#issuecomment-5821636144)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing hash-join-split-optional-filter (83d19600b4a207db882058b35ec5126b1a031f61) to 95bb0a0 (merge-base) [diff](https://github.com/apache/datafusion/compare/95bb0a0dfa48ca20d017faa55b2fc96ee2f300fe..83d19600b4a207db882058b35ec5126b1a031f61)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpcds
env:
DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
```
</details>
<details><summary>CPU Details (lscpu)</summary>
```
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>
<details><summary>Details</summary>
<p>
```
Comparing HEAD and hash-join-split-optional-filter
--------------------
Benchmark tpcds_sf1.json
--------------------
βββββββββββββ³ββββββββββββ³ββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β hash-join-split-optional-filter β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 5.01 ms β 5.08 ms β no change β
β QQuery 2 β 45.32 ms β 44.24 ms β no change β
β QQuery 3 β 26.65 ms β 26.94 ms β no change β
β QQuery 4 β 353.91 ms β 352.67 ms β no change β
β QQuery 5 β 47.79 ms β 47.55 ms β no change β
β QQuery 6 β 36.21 ms β 36.33 ms β no change β
β QQuery 7 β 82.91 ms β 83.67 ms β no change β
β QQuery 8 β 14.80 ms β 15.24 ms β no change β
β QQuery 9 β 116.34 ms β 115.09 ms β no change β
β QQuery 10 β 72.56 ms β 73.03 ms β no change β
β QQuery 11 β 214.89 ms β 218.36 ms β no change β
β QQuery 12 β 28.11 ms β 28.38 ms β no change β
β QQuery 13 β 114.12 ms β 116.48 ms β no change β
β QQuery 14 β 429.62 ms β 427.91 ms β no change β
β QQuery 15 β 23.70 ms β 23.65 ms β no change β
β QQuery 16 β 5.86 ms β 6.03 ms β no change β
β QQuery 17 β 114.94 ms β 68.54 ms β +1.68x faster β
β QQuery 18 β 161.18 ms β 106.79 ms β +1.51x faster β
β QQuery 19 β 39.32 ms β 39.85 ms β no change β
β QQuery 20 β 31.24 ms β 31.47 ms β no change β
β QQuery 21 β 19.43 ms β 19.38 ms β no change β
β QQuery 22 β 62.47 ms β 62.75 ms β no change β
β QQuery 23 β 376.84 ms β 378.31 ms β no change β
β QQuery 24 β 166.18 ms β 169.70 ms β no change β
β QQuery 25 β 115.21 ms β 66.47 ms β +1.73x faster β
β QQuery 26 β 63.79 ms β 65.49 ms β no change β
β QQuery 27 β 5.69 ms β 5.94 ms β no change β
β QQuery 28 β 73.67 ms β 74.25 ms β no change β
β QQuery 29 β 145.75 ms β 92.97 ms β +1.57x faster β
β QQuery 30 β 32.11 ms β 32.33 ms β no change β
β QQuery 31 β 123.06 ms β 124.33 ms β no change β
β QQuery 32 β 21.59 ms β 21.67 ms β no change β
β QQuery 33 β 39.60 ms β 39.24 ms β no change β
β QQuery 34 β 10.72 ms β 10.67 ms β no change β
β QQuery 35 β 82.64 ms β 82.84 ms β no change β
β QQuery 36 β 5.41 ms β 5.58 ms β no change β
β QQuery 37 β 7.02 ms β 7.04 ms β no change β
β QQuery 38 β 77.42 ms β 77.75 ms β no change β
β QQuery 39 β 70.45 ms β 71.11 ms β no change β
β QQuery 40 β 20.28 ms β 20.34 ms β no change β
β QQuery 41 β 11.90 ms β 12.12 ms β no change β
β QQuery 42 β 23.80 ms β 23.82 ms β no change β
β QQuery 43 β 4.73 ms β 4.79 ms β no change β
β QQuery 44 β 8.92 ms β 9.03 ms β no change β
β QQuery 45 β 26.37 ms β 26.66 ms β no change β
β QQuery 46 β 12.33 ms β 12.15 ms β no change β
β QQuery 47 β 222.79 ms β 224.43 ms β no change β
β QQuery 48 β 121.94 ms β 126.85 ms β no change β
β QQuery 49 β 87.03 ms β 87.11 ms β no change β
β QQuery 50 β 124.75 ms β 106.40 ms β +1.17x faster β
β QQuery 51 β 101.69 ms β 100.48 ms β no change β
β QQuery 52 β 24.12 ms β 24.19 ms β no change β
β QQuery 53 β 30.01 ms β 30.27 ms β no change β
β QQuery 54 β 29.16 ms β 28.37 ms β no change β
β QQuery 55 β 23.20 ms β 23.63 ms β no change β
β QQuery 56 β 34.23 ms β 35.16 ms β no change β
β QQuery 57 β 135.57 ms β 137.07 ms β no change β
β QQuery 58 β 72.30 ms β 72.17 ms β no change β
β QQuery 59 β 80.38 ms β 80.47 ms β no change β
β QQuery 60 β 37.78 ms β 38.37 ms β no change β
β QQuery 61 β 10.99 ms β 11.56 ms β 1.05x slower β
β QQuery 62 β 42.12 ms β 42.45 ms β no change β
β QQuery 63 β 30.40 ms β 30.99 ms β no change β
β QQuery 64 β 510.41 ms β 337.43 ms β +1.51x faster β
β QQuery 65 β 159.03 ms β 88.47 ms β +1.80x faster β
β QQuery 66 β 64.29 ms β 64.62 ms β no change β
β QQuery 67 β 259.26 ms β 264.04 ms β no change β
β QQuery 68 β 12.18 ms β 12.05 ms β no change β
β QQuery 69 β 69.66 ms β 69.98 ms β no change β
β QQuery 70 β 118.37 ms β 117.94 ms β no change β
β QQuery 71 β 32.09 ms β 32.23 ms β no change β
β QQuery 72 β 170.91 ms β 169.89 ms β no change β
β QQuery 73 β 10.25 ms β 10.34 ms β no change β
β QQuery 74 β 144.83 ms β 148.30 ms β no change β
β QQuery 75 β 180.19 ms β 178.03 ms β no change β
β QQuery 76 β 41.32 ms β 40.98 ms β no change β
β QQuery 77 β 61.49 ms β 62.09 ms β no change β
β QQuery 78 β 118.28 ms β 120.05 ms β no change β
β QQuery 79 β 67.29 ms β 67.03 ms β no change β
β QQuery 80 β 75.98 ms β 75.73 ms β no change β
β QQuery 81 β 26.90 ms β 27.13 ms β no change β
β QQuery 82 β 30.88 ms β 31.08 ms β no change β
β QQuery 83 β 29.21 ms β 29.41 ms β no change β
β QQuery 84 β 30.12 ms β 30.62 ms β no change β
β QQuery 85 β 164.45 ms β 128.29 ms β +1.28x faster β
β QQuery 86 β 29.80 ms β 30.33 ms β no change β
β QQuery 87 β 78.75 ms β 77.88 ms β no change β
β QQuery 88 β 64.57 ms β 63.90 ms β no change β
β QQuery 89 β 43.48 ms β 42.78 ms β no change β
β QQuery 90 β 18.24 ms β 18.73 ms β no change β
β QQuery 91 β 46.77 ms β 47.36 ms β no change β
β QQuery 92 β 29.49 ms β 30.12 ms β no change β
β QQuery 93 β 48.98 ms β 48.42 ms β no change β
β QQuery 94 β 36.51 ms β 36.43 ms β no change β
β QQuery 95 β 92.42 ms β 55.75 ms β +1.66x faster β
β QQuery 96 β 26.32 ms β 25.95 ms β no change β
β QQuery 97 β 53.56 ms β 53.59 ms β no change β
β QQuery 98 β 38.86 ms β 38.24 ms β no change β
β QQuery 99 β 60.61 ms β 60.75 ms β no change β
βββββββββββββ΄ββββββββββββ΄ββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 7826.06ms β
β Total Time (hash-join-split-optional-filter) β 7319.91ms β
β Average Time (HEAD) β 79.05ms β
β Average Time (hash-join-split-optional-filter) β 73.94ms β
β Queries Faster β 9 β
β Queries Slower β 1 β
β Queries with No Change β 89 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββ
Distribution per query (min / mean Β±stddev / max):
Comparing HEAD and hash-join-split-optional-filter
--------------------
Benchmark tpcds_sf1.json
--------------------
βββββββββββββ³ββββββββββββββββββββββββββββββββββββ³ββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β hash-join-split-optional-filter β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 5.01 / 5.57 Β±0.96 / 7.49 ms β 5.08 / 5.68 Β±1.03 / 7.75 ms β no change β
β QQuery 2 β 45.32 / 45.67 Β±0.27 / 46.09 ms β 44.24 / 44.85 Β±0.41 / 45.34 ms β no change β
β QQuery 3 β 26.65 / 27.11 Β±0.29 / 27.45 ms β 26.94 / 27.18 Β±0.26 / 27.61 ms β no change β
β QQuery 4 β 353.91 / 355.15 Β±1.08 / 357.03 ms β 352.67 / 357.44 Β±3.49 / 362.27 ms β no change β
β QQuery 5 β 47.79 / 48.10 Β±0.31 / 48.62 ms β 47.55 / 48.31 Β±0.46 / 48.86 ms β no change β
β QQuery 6 β 36.21 / 36.56 Β±0.33 / 37.13 ms β 36.33 / 36.68 Β±0.23 / 37.02 ms β no change β
β QQuery 7 β 82.91 / 83.83 Β±1.59 / 87.01 ms β 83.67 / 85.07 Β±2.35 / 89.76 ms β no change β
β QQuery 8 β 14.80 / 15.11 Β±0.30 / 15.66 ms β 15.24 / 15.42 Β±0.18 / 15.71 ms β no change β
β QQuery 9 β 116.34 / 117.87 Β±1.42 / 119.93 ms β 115.09 / 118.07 Β±2.11 / 120.39 ms β no change β
β QQuery 10 β 72.56 / 73.34 Β±1.22 / 75.76 ms β 73.03 / 76.93 Β±6.91 / 90.73 ms β no change β
β QQuery 11 β 214.89 / 220.37 Β±5.43 / 230.18 ms β 218.36 / 221.64 Β±2.91 / 226.50 ms β no change β
β QQuery 12 β 28.11 / 28.53 Β±0.52 / 29.52 ms β 28.38 / 28.73 Β±0.51 / 29.71 ms β no change β
β QQuery 13 β 114.12 / 114.83 Β±1.02 / 116.85 ms β 116.48 / 117.42 Β±0.99 / 119.28 ms β no change β
β QQuery 14 β 429.62 / 434.00 Β±4.25 / 440.85 ms β 427.91 / 435.10 Β±7.67 / 449.28 ms β no change β
β QQuery 15 β 23.70 / 25.22 Β±2.40 / 30.00 ms β 23.65 / 23.81 Β±0.13 / 24.01 ms β +1.06x faster β
β QQuery 16 β 5.86 / 6.04 Β±0.23 / 6.48 ms β 6.03 / 6.18 Β±0.19 / 6.55 ms β no change β
β QQuery 17 β 114.94 / 115.38 Β±0.29 / 115.84 ms β 68.54 / 68.78 Β±0.16 / 69.03 ms β +1.68x faster β
β QQuery 18 β 161.18 / 164.24 Β±2.45 / 166.82 ms β 106.79 / 108.95 Β±1.30 / 110.21 ms β +1.51x faster β
β QQuery 19 β 39.32 / 40.02 Β±0.72 / 41.30 ms β 39.85 / 40.66 Β±0.75 / 41.60 ms β no change β
β QQuery 20 β 31.24 / 31.65 Β±0.39 / 32.14 ms β 31.47 / 32.01 Β±0.53 / 33.01 ms β no change β
β QQuery 21 β 19.43 / 20.22 Β±1.26 / 22.72 ms β 19.38 / 19.63 Β±0.17 / 19.86 ms β no change β
β QQuery 22 β 62.47 / 62.75 Β±0.50 / 63.74 ms β 62.75 / 63.55 Β±0.83 / 64.92 ms β no change β
β QQuery 23 β 376.84 / 380.20 Β±2.80 / 383.93 ms β 378.31 / 383.69 Β±4.29 / 388.51 ms β no change β
β QQuery 24 β 166.18 / 169.20 Β±3.92 / 176.95 ms β 169.70 / 175.85 Β±5.52 / 183.02 ms β no change β
β QQuery 25 β 115.21 / 116.94 Β±1.44 / 119.55 ms β 66.47 / 67.94 Β±2.18 / 72.28 ms β +1.72x faster β
β QQuery 26 β 63.79 / 63.95 Β±0.13 / 64.17 ms β 65.49 / 67.59 Β±1.88 / 70.50 ms β 1.06x slower β
β QQuery 27 β 5.69 / 5.95 Β±0.16 / 6.13 ms β 5.94 / 6.09 Β±0.09 / 6.19 ms β no change β
β QQuery 28 β 73.67 / 74.87 Β±1.30 / 77.29 ms β 74.25 / 75.90 Β±2.95 / 81.80 ms β no change β
β QQuery 29 β 145.75 / 147.87 Β±3.47 / 154.78 ms β 92.97 / 96.70 Β±4.96 / 106.09 ms β +1.53x faster β
β QQuery 30 β 32.11 / 34.14 Β±3.28 / 40.69 ms β 32.33 / 32.86 Β±0.53 / 33.84 ms β no change β
β QQuery 31 β 123.06 / 123.98 Β±1.16 / 126.20 ms β 124.33 / 125.93 Β±1.34 / 128.36 ms β no change β
β QQuery 32 β 21.59 / 22.23 Β±0.57 / 23.12 ms β 21.67 / 22.51 Β±0.54 / 23.22 ms β no change β
β QQuery 33 β 39.60 / 40.84 Β±1.13 / 42.60 ms β 39.24 / 39.90 Β±0.60 / 40.89 ms β no change β
β QQuery 34 β 10.72 / 11.50 Β±0.87 / 13.01 ms β 10.67 / 10.99 Β±0.20 / 11.24 ms β no change β
β QQuery 35 β 82.64 / 83.25 Β±0.59 / 84.16 ms β 82.84 / 83.08 Β±0.17 / 83.33 ms β no change β
β QQuery 36 β 5.41 / 5.61 Β±0.23 / 6.06 ms β 5.58 / 5.76 Β±0.22 / 6.17 ms β no change β
β QQuery 37 β 7.02 / 7.07 Β±0.04 / 7.11 ms β 7.04 / 7.16 Β±0.08 / 7.27 ms β no change β
β QQuery 38 β 77.42 / 79.17 Β±1.86 / 82.42 ms β 77.75 / 81.11 Β±3.58 / 86.06 ms β no change β
β QQuery 39 β 70.45 / 72.40 Β±2.45 / 77.23 ms β 71.11 / 71.75 Β±0.51 / 72.50 ms β no change β
β QQuery 40 β 20.28 / 20.52 Β±0.17 / 20.73 ms β 20.34 / 20.70 Β±0.43 / 21.29 ms β no change β
β QQuery 41 β 11.90 / 12.13 Β±0.22 / 12.52 ms β 12.12 / 12.29 Β±0.16 / 12.55 ms β no change β
β QQuery 42 β 23.80 / 23.93 Β±0.10 / 24.05 ms β 23.82 / 24.30 Β±0.28 / 24.63 ms β no change β
β QQuery 43 β 4.73 / 4.89 Β±0.26 / 5.39 ms β 4.79 / 4.92 Β±0.24 / 5.40 ms β no change β
β QQuery 44 β 8.92 / 9.01 Β±0.05 / 9.07 ms β 9.03 / 10.28 Β±2.23 / 14.73 ms β 1.14x slower β
β QQuery 45 β 26.37 / 27.67 Β±2.03 / 31.66 ms β 26.66 / 28.43 Β±1.33 / 30.65 ms β no change β
β QQuery 46 β 12.33 / 12.89 Β±0.49 / 13.74 ms β 12.15 / 12.98 Β±0.85 / 14.54 ms β no change β
β QQuery 47 β 222.79 / 230.09 Β±7.58 / 241.98 ms β 224.43 / 228.01 Β±3.66 / 233.50 ms β no change β
β QQuery 48 β 121.94 / 125.15 Β±3.93 / 132.55 ms β 126.85 / 130.07 Β±3.38 / 136.14 ms β no change β
β QQuery 49 β 87.03 / 91.55 Β±7.53 / 106.51 ms β 87.11 / 88.70 Β±1.09 / 90.21 ms β no change β
β QQuery 50 β 124.75 / 125.69 Β±0.87 / 126.96 ms β 106.40 / 107.01 Β±0.49 / 107.61 ms β +1.17x faster β
β QQuery 51 β 101.69 / 104.56 Β±1.92 / 107.22 ms β 100.48 / 105.61 Β±4.44 / 110.97 ms β no change β
β QQuery 52 β 24.12 / 24.52 Β±0.31 / 24.86 ms β 24.19 / 24.55 Β±0.23 / 24.89 ms β no change β
β QQuery 53 β 30.01 / 30.13 Β±0.11 / 30.31 ms β 30.27 / 30.53 Β±0.20 / 30.85 ms β no change β
β QQuery 54 β 29.16 / 29.91 Β±0.70 / 31.06 ms β 28.37 / 29.84 Β±1.06 / 31.07 ms β no change β
β QQuery 55 β 23.20 / 23.55 Β±0.24 / 23.81 ms β 23.63 / 24.40 Β±1.06 / 26.50 ms β no change β
β QQuery 56 β 34.23 / 34.70 Β±0.33 / 35.22 ms β 35.16 / 35.71 Β±0.92 / 37.55 ms β no change β
β QQuery 57 β 135.57 / 138.14 Β±2.31 / 141.98 ms β 137.07 / 138.95 Β±1.33 / 140.39 ms β no change β
β QQuery 58 β 72.30 / 73.77 Β±1.23 / 75.44 ms β 72.17 / 73.30 Β±1.09 / 75.31 ms β no change β
β QQuery 59 β 80.38 / 80.89 Β±0.47 / 81.74 ms β 80.47 / 80.76 Β±0.35 / 81.45 ms β no change β
β QQuery 60 β 37.78 / 38.30 Β±0.42 / 38.99 ms β 38.37 / 38.77 Β±0.39 / 39.50 ms β no change β
β QQuery 61 β 10.99 / 11.23 Β±0.21 / 11.62 ms β 11.56 / 11.68 Β±0.18 / 12.03 ms β no change β
β QQuery 62 β 42.12 / 43.83 Β±1.71 / 47.06 ms β 42.45 / 44.58 Β±1.94 / 47.57 ms β no change β
β QQuery 63 β 30.40 / 32.09 Β±2.05 / 36.11 ms β 30.99 / 31.31 Β±0.39 / 32.07 ms β no change β
β QQuery 64 β 510.41 / 514.28 Β±2.24 / 517.37 ms β 337.43 / 343.30 Β±5.68 / 352.58 ms β +1.50x faster β
β QQuery 65 β 159.03 / 165.49 Β±5.46 / 174.58 ms β 88.47 / 91.65 Β±4.73 / 101.07 ms β +1.81x faster β
β QQuery 66 β 64.29 / 64.59 Β±0.19 / 64.87 ms β 64.62 / 64.97 Β±0.32 / 65.48 ms β no change β
β QQuery 67 β 259.26 / 267.37 Β±7.67 / 281.05 ms β 264.04 / 270.32 Β±5.00 / 277.01 ms β no change β
β QQuery 68 β 12.18 / 12.63 Β±0.54 / 13.60 ms β 12.05 / 12.33 Β±0.17 / 12.57 ms β no change β
β QQuery 69 β 69.66 / 70.86 Β±1.38 / 73.42 ms β 69.98 / 76.18 Β±10.98 / 98.11 ms β 1.08x slower β
β QQuery 70 β 118.37 / 123.85 Β±6.84 / 135.45 ms β 117.94 / 119.02 Β±0.63 / 119.86 ms β no change β
β QQuery 71 β 32.09 / 32.40 Β±0.21 / 32.62 ms β 32.23 / 35.77 Β±5.58 / 46.87 ms β 1.10x slower β
β QQuery 72 β 170.91 / 173.01 Β±1.65 / 175.02 ms β 169.89 / 171.72 Β±2.26 / 176.19 ms β no change β
β QQuery 73 β 10.25 / 10.53 Β±0.29 / 11.00 ms β 10.34 / 15.43 Β±8.97 / 33.32 ms β 1.47x slower β
β QQuery 74 β 144.83 / 148.20 Β±4.00 / 155.96 ms β 148.30 / 150.46 Β±2.09 / 153.70 ms β no change β
β QQuery 75 β 180.19 / 182.93 Β±2.49 / 186.66 ms β 178.03 / 183.36 Β±7.91 / 199.05 ms β no change β
β QQuery 76 β 41.32 / 42.18 Β±1.21 / 44.58 ms β 40.98 / 42.64 Β±2.75 / 48.12 ms β no change β
β QQuery 77 β 61.49 / 61.97 Β±0.56 / 62.99 ms β 62.09 / 63.26 Β±1.73 / 66.69 ms β no change β
β QQuery 78 β 118.28 / 120.99 Β±2.44 / 124.54 ms β 120.05 / 122.26 Β±1.41 / 123.94 ms β no change β
β QQuery 79 β 67.29 / 70.95 Β±3.34 / 76.58 ms β 67.03 / 70.01 Β±4.78 / 79.51 ms β no change β
β QQuery 80 β 75.98 / 76.75 Β±0.39 / 77.00 ms β 75.73 / 76.55 Β±0.71 / 77.61 ms β no change β
β QQuery 81 β 26.90 / 28.73 Β±2.78 / 34.27 ms β 27.13 / 28.62 Β±2.82 / 34.25 ms β no change β
β QQuery 82 β 30.88 / 31.22 Β±0.31 / 31.78 ms β 31.08 / 31.57 Β±0.56 / 32.66 ms β no change β
β QQuery 83 β 29.21 / 32.90 Β±6.23 / 45.34 ms β 29.41 / 32.81 Β±6.01 / 44.83 ms β no change β
β QQuery 84 β 30.12 / 30.70 Β±0.47 / 31.45 ms β 30.62 / 30.90 Β±0.36 / 31.59 ms β no change β
β QQuery 85 β 164.45 / 169.31 Β±3.71 / 174.12 ms β 128.29 / 131.92 Β±3.06 / 137.31 ms β +1.28x faster β
β QQuery 86 β 29.80 / 30.17 Β±0.29 / 30.47 ms β 30.33 / 30.70 Β±0.38 / 31.43 ms β no change β
β QQuery 87 β 78.75 / 79.12 Β±0.31 / 79.68 ms β 77.88 / 82.60 Β±6.38 / 95.21 ms β no change β
β QQuery 88 β 64.57 / 71.58 Β±5.83 / 80.21 ms β 63.90 / 68.95 Β±4.38 / 76.08 ms β no change β
β QQuery 89 β 43.48 / 44.34 Β±1.07 / 46.14 ms β 42.78 / 43.03 Β±0.14 / 43.18 ms β no change β
β QQuery 90 β 18.24 / 18.66 Β±0.24 / 18.95 ms β 18.73 / 21.98 Β±5.38 / 32.68 ms β 1.18x slower β
β QQuery 91 β 46.77 / 48.78 Β±1.85 / 51.32 ms β 47.36 / 50.26 Β±3.15 / 56.21 ms β no change β
β QQuery 92 β 29.49 / 30.18 Β±0.64 / 31.32 ms β 30.12 / 31.06 Β±0.63 / 31.96 ms β no change β
β QQuery 93 β 48.98 / 49.46 Β±0.46 / 50.31 ms β 48.42 / 49.32 Β±0.59 / 50.26 ms β no change β
β QQuery 94 β 36.51 / 39.76 Β±4.15 / 47.68 ms β 36.43 / 37.26 Β±0.87 / 38.92 ms β +1.07x faster β
β QQuery 95 β 92.42 / 92.97 Β±0.42 / 93.60 ms β 55.75 / 58.12 Β±3.24 / 64.52 ms β +1.60x faster β
β QQuery 96 β 26.32 / 27.99 Β±3.16 / 34.31 ms β 25.95 / 26.21 Β±0.15 / 26.36 ms β +1.07x faster β
β QQuery 97 β 53.56 / 56.24 Β±2.10 / 59.97 ms β 53.59 / 55.34 Β±0.89 / 55.99 ms β no change β
β QQuery 98 β 38.86 / 39.54 Β±0.78 / 40.98 ms β 38.24 / 39.95 Β±1.62 / 42.62 ms β no change β
β QQuery 99 β 60.61 / 61.33 Β±0.48 / 62.02 ms β 60.75 / 63.36 Β±4.63 / 72.61 ms β no change β
βββββββββββββ΄ββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 7981.78ms β
β Total Time (hash-join-split-optional-filter) β 7497.79ms β
β Average Time (HEAD) β 80.62ms β
β Average Time (hash-join-split-optional-filter) β 75.74ms β
β Queries Faster β 12 β
β Queries Slower β 6 β
β Queries with No Change β 81 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**tpcds β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 45.0s |
| Peak memory | 1.7 GiB |
| Avg memory | 1.2 GiB |
| CPU user | 116.6s |
| CPU sys | 5.8s |
| Peak spill | 0 B |
**tpcds β branch**
| Metric | Value |
|--------|-------|
| Wall time | 40.0s |
| Peak memory | 1.9 GiB |
| Avg memory | 1.3 GiB |
| CPU user | 109.3s |
| CPU sys | 5.6s |
| Peak spill | 0 B |
</details>
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push partitioned hash join dynamic filter bounds and membership as separate filters [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5821684309
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/25713#issuecomment-5821636144)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c5821636144-2700-jcxqw 6.12.94+ #1 SMP Wed Aug 19 07:47:20 UTC 2026 aarch64 GNU/Linux`
<details><summary>CPU Details (lscpu)</summary>
```
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 hash-join-split-optional-filter (83d19600b4a207db882058b35ec5126b1a031f61) to 95bb0a0 (merge-base) [diff](https://github.com/apache/datafusion/compare/95bb0a0dfa48ca20d017faa55b2fc96ee2f300fe..83d19600b4a207db882058b35ec5126b1a031f61)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpch10
env:
DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
```
</details>
Results will be posted here when complete
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push partitioned hash join dynamic filter bounds and membership as separate filters [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5821787004
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/25713#issuecomment-5821635949)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing hash-join-split-optional-filter (83d19600b4a207db882058b35ec5126b1a031f61) to 95bb0a0 (merge-base) [diff](https://github.com/apache/datafusion/compare/95bb0a0dfa48ca20d017faa55b2fc96ee2f300fe..83d19600b4a207db882058b35ec5126b1a031f61)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpcds
```
</details>
<details><summary>CPU Details (lscpu)</summary>
```
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>
<details><summary>Details</summary>
<p>
```
Comparing HEAD and hash-join-split-optional-filter
--------------------
Benchmark tpcds_sf1.json
--------------------
βββββββββββββ³βββββββββββββ³ββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β hash-join-split-optional-filter β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 5.73 ms β 5.55 ms β no change β
β QQuery 2 β 81.28 ms β 81.45 ms β no change β
β QQuery 3 β 28.82 ms β 28.56 ms β no change β
β QQuery 4 β 447.19 ms β 455.46 ms β no change β
β QQuery 5 β 52.18 ms β 52.86 ms β no change β
β QQuery 6 β 35.59 ms β 38.42 ms β 1.08x slower β
β QQuery 7 β 75.36 ms β 74.87 ms β no change β
β QQuery 8 β 36.18 ms β 36.84 ms β no change β
β QQuery 9 β 53.97 ms β 53.08 ms β no change β
β QQuery 10 β 63.55 ms β 62.66 ms β no change β
β QQuery 11 β 266.97 ms β 269.30 ms β no change β
β QQuery 12 β 29.52 ms β 28.48 ms β no change β
β QQuery 13 β 120.13 ms β 117.22 ms β no change β
β QQuery 14 β 402.61 ms β 396.19 ms β no change β
β QQuery 15 β 55.10 ms β 52.98 ms β no change β
β QQuery 16 β 6.65 ms β 7.01 ms β 1.05x slower β
β QQuery 17 β 78.88 ms β 77.70 ms β no change β
β QQuery 18 β 105.31 ms β 103.32 ms β no change β
β QQuery 19 β 41.55 ms β 40.52 ms β no change β
β QQuery 20 β 35.98 ms β 36.82 ms β no change β
β QQuery 21 β 16.76 ms β 17.11 ms β no change β
β QQuery 22 β 64.36 ms β 63.36 ms β no change β
β QQuery 23 β 318.47 ms β 311.76 ms β no change β
β QQuery 24 β 188.81 ms β 189.86 ms β no change β
β QQuery 25 β 107.66 ms β 108.91 ms β no change β
β QQuery 26 β 48.72 ms β 48.22 ms β no change β
β QQuery 27 β 6.30 ms β 5.95 ms β +1.06x faster β
β QQuery 28 β 56.02 ms β 57.94 ms β no change β
β QQuery 29 β 98.00 ms β 94.60 ms β no change β
β QQuery 30 β 33.76 ms β 34.30 ms β no change β
β QQuery 31 β 112.70 ms β 110.74 ms β no change β
β QQuery 32 β 21.39 ms β 20.96 ms β no change β
β QQuery 33 β 38.20 ms β 37.16 ms β no change β
β QQuery 34 β 10.75 ms β 9.57 ms β +1.12x faster β
β QQuery 35 β 75.63 ms β 72.19 ms β no change β
β QQuery 36 β 6.22 ms β 6.33 ms β no change β
β QQuery 37 β 7.58 ms β 7.39 ms β no change β
β QQuery 38 β 60.84 ms β 61.80 ms β no change β
β QQuery 39 β 72.96 ms β 74.70 ms β no change β
β QQuery 40 β 24.94 ms β 25.23 ms β no change β
β QQuery 41 β 11.04 ms β 11.73 ms β 1.06x slower β
β QQuery 42 β 23.85 ms β 24.07 ms β no change β
β QQuery 43 β 5.09 ms β 5.01 ms β no change β
β QQuery 44 β 9.13 ms β 9.11 ms β no change β
β QQuery 45 β 38.72 ms β 38.90 ms β no change β
β QQuery 46 β 11.63 ms β 11.91 ms β no change β
β QQuery 47 β 224.71 ms β 235.48 ms β no change β
β QQuery 48 β 96.75 ms β 95.01 ms β no change β
β QQuery 49 β 70.77 ms β 70.24 ms β no change β
β QQuery 50 β 57.84 ms β 58.48 ms β no change β
β QQuery 51 β 93.46 ms β 94.46 ms β no change β
β QQuery 52 β 24.56 ms β 24.09 ms β no change β
β QQuery 53 β 29.11 ms β 29.00 ms β no change β
β QQuery 54 β 53.57 ms β 52.53 ms β no change β
β QQuery 55 β 23.59 ms β 23.25 ms β no change β
β QQuery 56 β 38.87 ms β 38.11 ms β no change β
β QQuery 57 β 169.27 ms β 168.85 ms β no change β
β QQuery 58 β 111.94 ms β 109.92 ms β no change β
β QQuery 59 β 118.15 ms β 117.89 ms β no change β
β QQuery 60 β 38.98 ms β 39.11 ms β no change β
β QQuery 61 β 12.56 ms β 12.03 ms β no change β
β QQuery 62 β 44.81 ms β 45.81 ms β no change β
β QQuery 63 β 29.18 ms β 29.36 ms β no change β
β QQuery 64 β 359.37 ms β 372.84 ms β no change β
β QQuery 65 β 133.17 ms β 67.72 ms β +1.97x faster β
β QQuery 66 β 77.60 ms β 77.64 ms β no change β
β QQuery 67 β 252.44 ms β 249.39 ms β no change β
β QQuery 68 β 12.33 ms β 11.54 ms β +1.07x faster β
β QQuery 69 β 58.10 ms β 56.02 ms β no change β
β QQuery 70 β 104.32 ms β 105.15 ms β no change β
β QQuery 71 β 35.81 ms β 35.32 ms β no change β
β QQuery 72 β 1800.28 ms β 1743.17 ms β no change β
β QQuery 73 β 10.05 ms β 9.90 ms β no change β
β QQuery 74 β 161.61 ms β 161.79 ms β no change β
β QQuery 75 β 141.31 ms β 143.60 ms β no change β
β QQuery 76 β 35.39 ms β 34.52 ms β no change β
β QQuery 77 β 60.34 ms β 60.85 ms β no change β
β QQuery 78 β 164.38 ms β 164.37 ms β no change β
β QQuery 79 β 66.79 ms β 68.04 ms β no change β
β QQuery 80 β 95.00 ms β 96.31 ms β no change β
β QQuery 81 β 27.59 ms β 26.17 ms β +1.05x faster β
β QQuery 82 β 16.78 ms β 17.36 ms β no change β
β QQuery 83 β 32.70 ms β 33.60 ms β no change β
β QQuery 84 β 29.89 ms β 30.16 ms β no change β
β QQuery 85 β 104.11 ms β 104.52 ms β no change β
β QQuery 86 β 25.78 ms β 25.58 ms β no change β
β QQuery 87 β 60.84 ms β 62.12 ms β no change β
β QQuery 88 β 60.75 ms β 61.16 ms β no change β
β QQuery 89 β 35.72 ms β 35.17 ms β no change β
β QQuery 90 β 16.49 ms β 17.52 ms β 1.06x slower β
β QQuery 91 β 44.49 ms β 45.05 ms β no change β
β QQuery 92 β 29.50 ms β 29.35 ms β no change β
β QQuery 93 β 51.24 ms β 48.01 ms β +1.07x faster β
β QQuery 94 β 39.06 ms β 38.36 ms β no change β
β QQuery 95 β 79.35 ms β 81.59 ms β no change β
β QQuery 96 β 24.52 ms β 24.16 ms β no change β
β QQuery 97 β 50.93 ms β 51.95 ms β no change β
β QQuery 98 β 43.58 ms β 43.90 ms β no change β
β QQuery 99 β 66.39 ms β 66.79 ms β no change β
βββββββββββββ΄βββββββββββββ΄ββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 9138.26ms β
β Total Time (hash-join-split-optional-filter) β 9022.43ms β
β Average Time (HEAD) β 92.31ms β
β Average Time (hash-join-split-optional-filter) β 91.14ms β
β Queries Faster β 6 β
β Queries Slower β 4 β
β Queries with No Change β 89 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββ
Distribution per query (min / mean Β±stddev / max):
Comparing HEAD and hash-join-split-optional-filter
--------------------
Benchmark tpcds_sf1.json
--------------------
βββββββββββββ³ββββββββββββββββββββββββββββββββββββββββ³βββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β hash-join-split-optional-filter β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 5.73 / 6.39 Β±0.91 / 8.19 ms β 5.55 / 6.11 Β±1.04 / 8.19 ms β no change β
β QQuery 2 β 81.28 / 81.77 Β±0.58 / 82.91 ms β 81.45 / 81.86 Β±0.33 / 82.36 ms β no change β
β QQuery 3 β 28.82 / 29.20 Β±0.29 / 29.55 ms β 28.56 / 28.85 Β±0.25 / 29.23 ms β no change β
β QQuery 4 β 447.19 / 468.53 Β±16.92 / 486.18 ms β 455.46 / 478.54 Β±26.25 / 528.90 ms β no change β
β QQuery 5 β 52.18 / 54.06 Β±2.65 / 59.24 ms β 52.86 / 54.49 Β±2.28 / 59.00 ms β no change β
β QQuery 6 β 35.59 / 36.45 Β±0.80 / 37.83 ms β 38.42 / 38.74 Β±0.43 / 39.53 ms β 1.06x slower β
β QQuery 7 β 75.36 / 76.93 Β±1.18 / 78.38 ms β 74.87 / 75.81 Β±0.64 / 76.81 ms β no change β
β QQuery 8 β 36.18 / 36.38 Β±0.18 / 36.61 ms β 36.84 / 36.98 Β±0.13 / 37.22 ms β no change β
β QQuery 9 β 53.97 / 56.40 Β±3.49 / 63.30 ms β 53.08 / 55.18 Β±1.30 / 56.60 ms β no change β
β QQuery 10 β 63.55 / 63.81 Β±0.22 / 64.19 ms β 62.66 / 63.06 Β±0.29 / 63.41 ms β no change β
β QQuery 11 β 266.97 / 292.05 Β±18.10 / 322.49 ms β 269.30 / 284.80 Β±12.04 / 301.86 ms β no change β
β QQuery 12 β 29.52 / 30.84 Β±0.83 / 32.06 ms β 28.48 / 30.09 Β±2.32 / 34.70 ms β no change β
β QQuery 13 β 120.13 / 122.27 Β±1.50 / 124.20 ms β 117.22 / 120.27 Β±2.13 / 122.49 ms β no change β
β QQuery 14 β 402.61 / 408.49 Β±4.76 / 416.95 ms β 396.19 / 401.22 Β±3.20 / 404.08 ms β no change β
β QQuery 15 β 55.10 / 59.04 Β±2.75 / 62.75 ms β 52.98 / 55.30 Β±1.71 / 57.03 ms β +1.07x faster β
β QQuery 16 β 6.65 / 6.88 Β±0.20 / 7.23 ms β 7.01 / 7.28 Β±0.16 / 7.50 ms β 1.06x slower β
β QQuery 17 β 78.88 / 80.32 Β±1.15 / 82.40 ms β 77.70 / 79.38 Β±2.63 / 84.62 ms β no change β
β QQuery 18 β 105.31 / 107.72 Β±2.40 / 112.05 ms β 103.32 / 106.22 Β±2.25 / 109.05 ms β no change β
β QQuery 19 β 41.55 / 42.26 Β±0.74 / 43.37 ms β 40.52 / 42.63 Β±1.87 / 46.08 ms β no change β
β QQuery 20 β 35.98 / 36.28 Β±0.25 / 36.72 ms β 36.82 / 37.48 Β±0.42 / 37.93 ms β no change β
β QQuery 21 β 16.76 / 17.13 Β±0.19 / 17.29 ms β 17.11 / 17.94 Β±0.94 / 19.76 ms β no change β
β QQuery 22 β 64.36 / 67.16 Β±2.44 / 70.81 ms β 63.36 / 67.76 Β±3.09 / 72.57 ms β no change β
β QQuery 23 β 318.47 / 327.90 Β±7.28 / 340.94 ms β 311.76 / 318.20 Β±5.80 / 326.13 ms β no change β
β QQuery 24 β 188.81 / 202.43 Β±13.47 / 227.64 ms β 189.86 / 198.81 Β±6.51 / 205.39 ms β no change β
β QQuery 25 β 107.66 / 111.05 Β±5.31 / 121.56 ms β 108.91 / 110.78 Β±1.33 / 112.65 ms β no change β
β QQuery 26 β 48.72 / 49.79 Β±0.78 / 50.74 ms β 48.22 / 49.50 Β±1.14 / 51.06 ms β no change β
β QQuery 27 β 6.30 / 6.61 Β±0.29 / 7.16 ms β 5.95 / 6.16 Β±0.25 / 6.63 ms β +1.07x faster β
β QQuery 28 β 56.02 / 62.65 Β±4.89 / 71.21 ms β 57.94 / 60.67 Β±1.43 / 62.02 ms β no change β
β QQuery 29 β 98.00 / 100.72 Β±2.79 / 106.06 ms β 94.60 / 98.98 Β±4.73 / 106.42 ms β no change β
β QQuery 30 β 33.76 / 34.52 Β±0.59 / 35.31 ms β 34.30 / 35.32 Β±1.19 / 37.47 ms β no change β
β QQuery 31 β 112.70 / 115.87 Β±3.78 / 123.13 ms β 110.74 / 113.84 Β±2.47 / 118.24 ms β no change β
β QQuery 32 β 21.39 / 22.32 Β±1.06 / 24.34 ms β 20.96 / 21.35 Β±0.22 / 21.63 ms β no change β
β QQuery 33 β 38.20 / 38.65 Β±0.46 / 39.42 ms β 37.16 / 38.67 Β±1.58 / 41.57 ms β no change β
β QQuery 34 β 10.75 / 10.93 Β±0.14 / 11.15 ms β 9.57 / 10.15 Β±0.33 / 10.58 ms β +1.08x faster β
β QQuery 35 β 75.63 / 78.12 Β±2.36 / 81.85 ms β 72.19 / 77.90 Β±2.94 / 80.18 ms β no change β
β QQuery 36 β 6.22 / 6.38 Β±0.18 / 6.72 ms β 6.33 / 6.47 Β±0.14 / 6.73 ms β no change β
β QQuery 37 β 7.58 / 7.85 Β±0.15 / 8.04 ms β 7.39 / 7.56 Β±0.13 / 7.80 ms β no change β
β QQuery 38 β 60.84 / 63.21 Β±2.40 / 67.50 ms β 61.80 / 63.18 Β±0.77 / 63.99 ms β no change β
β QQuery 39 β 72.96 / 74.08 Β±0.71 / 74.95 ms β 74.70 / 76.68 Β±1.40 / 78.80 ms β no change β
β QQuery 40 β 24.94 / 25.51 Β±0.31 / 25.78 ms β 25.23 / 25.86 Β±0.48 / 26.64 ms β no change β
β QQuery 41 β 11.04 / 11.28 Β±0.25 / 11.75 ms β 11.73 / 11.89 Β±0.22 / 12.32 ms β 1.05x slower β
β QQuery 42 β 23.85 / 26.03 Β±1.99 / 29.77 ms β 24.07 / 24.38 Β±0.19 / 24.57 ms β +1.07x faster β
β QQuery 43 β 5.09 / 5.36 Β±0.22 / 5.68 ms β 5.01 / 5.22 Β±0.22 / 5.63 ms β no change β
β QQuery 44 β 9.13 / 9.39 Β±0.16 / 9.59 ms β 9.11 / 9.36 Β±0.16 / 9.53 ms β no change β
β QQuery 45 β 38.72 / 42.47 Β±2.47 / 45.51 ms β 38.90 / 41.22 Β±3.71 / 48.49 ms β no change β
β QQuery 46 β 11.63 / 12.22 Β±0.38 / 12.73 ms β 11.91 / 12.41 Β±0.33 / 12.88 ms β no change β
β QQuery 47 β 224.71 / 239.90 Β±12.09 / 255.43 ms β 235.48 / 243.41 Β±5.23 / 251.68 ms β no change β
β QQuery 48 β 96.75 / 99.15 Β±1.82 / 101.37 ms β 95.01 / 96.16 Β±1.28 / 98.48 ms β no change β
β QQuery 49 β 70.77 / 72.92 Β±2.33 / 77.32 ms β 70.24 / 73.82 Β±4.27 / 82.20 ms β no change β
β QQuery 50 β 57.84 / 58.45 Β±0.32 / 58.76 ms β 58.48 / 59.59 Β±1.32 / 61.89 ms β no change β
β QQuery 51 β 93.46 / 97.46 Β±5.69 / 108.55 ms β 94.46 / 97.59 Β±3.17 / 103.64 ms β no change β
β QQuery 52 β 24.56 / 25.57 Β±1.35 / 28.04 ms β 24.09 / 24.70 Β±0.41 / 25.19 ms β no change β
β QQuery 53 β 29.11 / 29.44 Β±0.25 / 29.85 ms β 29.00 / 29.43 Β±0.31 / 29.95 ms β no change β
β QQuery 54 β 53.57 / 54.34 Β±0.55 / 54.96 ms β 52.53 / 53.34 Β±0.62 / 54.40 ms β no change β
β QQuery 55 β 23.59 / 23.96 Β±0.32 / 24.50 ms β 23.25 / 23.88 Β±0.32 / 24.13 ms β no change β
β QQuery 56 β 38.87 / 41.69 Β±3.02 / 47.44 ms β 38.11 / 38.70 Β±0.40 / 39.35 ms β +1.08x faster β
β QQuery 57 β 169.27 / 170.27 Β±1.48 / 173.20 ms β 168.85 / 175.31 Β±6.11 / 186.22 ms β no change β
β QQuery 58 β 111.94 / 114.95 Β±3.53 / 121.70 ms β 109.92 / 112.62 Β±2.29 / 116.54 ms β no change β
β QQuery 59 β 118.15 / 120.23 Β±2.07 / 124.15 ms β 117.89 / 118.31 Β±0.34 / 118.93 ms β no change β
β QQuery 60 β 38.98 / 40.28 Β±0.74 / 41.01 ms β 39.11 / 40.38 Β±0.97 / 42.13 ms β no change β
β QQuery 61 β 12.56 / 13.08 Β±0.51 / 14.04 ms β 12.03 / 12.16 Β±0.13 / 12.40 ms β +1.08x faster β
β QQuery 62 β 44.81 / 45.62 Β±0.56 / 46.28 ms β 45.81 / 46.51 Β±0.85 / 48.16 ms β no change β
β QQuery 63 β 29.18 / 29.50 Β±0.26 / 29.77 ms β 29.36 / 29.80 Β±0.53 / 30.84 ms β no change β
β QQuery 64 β 359.37 / 368.89 Β±5.81 / 377.33 ms β 372.84 / 379.95 Β±7.53 / 393.65 ms β no change β
β QQuery 65 β 133.17 / 134.33 Β±1.13 / 136.30 ms β 67.72 / 71.24 Β±3.75 / 75.83 ms β +1.89x faster β
β QQuery 66 β 77.60 / 78.98 Β±1.44 / 81.50 ms β 77.64 / 80.78 Β±4.21 / 89.04 ms β no change β
β QQuery 67 β 252.44 / 263.97 Β±8.58 / 274.38 ms β 249.39 / 265.05 Β±10.74 / 282.60 ms β no change β
β QQuery 68 β 12.33 / 15.61 Β±5.19 / 25.95 ms β 11.54 / 11.95 Β±0.27 / 12.36 ms β +1.31x faster β
β QQuery 69 β 58.10 / 58.84 Β±0.49 / 59.48 ms β 56.02 / 56.64 Β±0.80 / 58.21 ms β no change β
β QQuery 70 β 104.32 / 105.04 Β±0.66 / 105.85 ms β 105.15 / 109.59 Β±4.95 / 119.25 ms β no change β
β QQuery 71 β 35.81 / 39.19 Β±3.69 / 45.33 ms β 35.32 / 35.79 Β±0.45 / 36.55 ms β +1.10x faster β
β QQuery 72 β 1800.28 / 1904.07 Β±93.53 / 2070.31 ms β 1743.17 / 1893.85 Β±115.01 / 2088.26 ms β no change β
β QQuery 73 β 10.05 / 10.47 Β±0.33 / 10.94 ms β 9.90 / 10.53 Β±0.74 / 11.62 ms β no change β
β QQuery 74 β 161.61 / 173.34 Β±9.99 / 185.84 ms β 161.79 / 174.31 Β±8.58 / 183.65 ms β no change β
β QQuery 75 β 141.31 / 143.52 Β±1.59 / 145.57 ms β 143.60 / 144.98 Β±1.15 / 146.49 ms β no change β
β QQuery 76 β 35.39 / 36.01 Β±0.33 / 36.29 ms β 34.52 / 34.97 Β±0.37 / 35.60 ms β no change β
β QQuery 77 β 60.34 / 60.54 Β±0.18 / 60.79 ms β 60.85 / 64.10 Β±5.62 / 75.29 ms β 1.06x slower β
β QQuery 78 β 164.38 / 172.12 Β±4.82 / 177.52 ms β 164.37 / 169.56 Β±5.42 / 178.44 ms β no change β
β QQuery 79 β 66.79 / 67.34 Β±0.36 / 67.83 ms β 68.04 / 69.33 Β±1.38 / 71.71 ms β no change β
β QQuery 80 β 95.00 / 97.60 Β±1.37 / 99.03 ms β 96.31 / 98.18 Β±1.13 / 99.65 ms β no change β
β QQuery 81 β 27.59 / 27.97 Β±0.27 / 28.45 ms β 26.17 / 26.86 Β±0.54 / 27.75 ms β no change β
β QQuery 82 β 16.78 / 17.58 Β±1.20 / 19.95 ms β 17.36 / 19.22 Β±3.50 / 26.21 ms β 1.09x slower β
β QQuery 83 β 32.70 / 33.87 Β±0.83 / 35.11 ms β 33.60 / 34.32 Β±0.65 / 35.30 ms β no change β
β QQuery 84 β 29.89 / 30.12 Β±0.30 / 30.67 ms β 30.16 / 31.15 Β±1.08 / 33.13 ms β no change β
β QQuery 85 β 104.11 / 108.81 Β±4.63 / 117.50 ms β 104.52 / 105.35 Β±0.79 / 106.37 ms β no change β
β QQuery 86 β 25.78 / 26.19 Β±0.36 / 26.68 ms β 25.58 / 28.38 Β±3.93 / 36.16 ms β 1.08x slower β
β QQuery 87 β 60.84 / 61.27 Β±0.48 / 62.21 ms β 62.12 / 65.36 Β±2.84 / 69.78 ms β 1.07x slower β
β QQuery 88 β 60.75 / 65.10 Β±6.51 / 78.00 ms β 61.16 / 61.99 Β±0.62 / 62.95 ms β no change β
β QQuery 89 β 35.72 / 36.66 Β±0.73 / 37.66 ms β 35.17 / 35.46 Β±0.20 / 35.68 ms β no change β
β QQuery 90 β 16.49 / 16.73 Β±0.23 / 17.08 ms β 17.52 / 17.64 Β±0.10 / 17.82 ms β 1.05x slower β
β QQuery 91 β 44.49 / 46.01 Β±1.03 / 47.58 ms β 45.05 / 47.14 Β±2.92 / 52.84 ms β no change β
β QQuery 92 β 29.50 / 29.79 Β±0.33 / 30.35 ms β 29.35 / 30.81 Β±1.78 / 34.30 ms β no change β
β QQuery 93 β 51.24 / 52.56 Β±1.75 / 55.95 ms β 48.01 / 49.53 Β±1.14 / 51.41 ms β +1.06x faster β
β QQuery 94 β 39.06 / 40.06 Β±1.29 / 42.42 ms β 38.36 / 39.37 Β±0.63 / 40.16 ms β no change β
β QQuery 95 β 79.35 / 82.58 Β±2.70 / 87.00 ms β 81.59 / 83.88 Β±2.24 / 88.06 ms β no change β
β QQuery 96 β 24.52 / 24.71 Β±0.16 / 24.94 ms β 24.16 / 24.32 Β±0.10 / 24.42 ms β no change β
β QQuery 97 β 50.93 / 52.43 Β±1.09 / 53.70 ms β 51.95 / 52.89 Β±1.53 / 55.94 ms β no change β
β QQuery 98 β 43.58 / 44.51 Β±0.52 / 45.11 ms β 43.90 / 44.34 Β±0.40 / 45.05 ms β no change β
β QQuery 99 β 66.39 / 67.43 Β±0.74 / 68.50 ms β 66.79 / 69.77 Β±4.26 / 78.23 ms β no change β
βββββββββββββ΄ββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 9496.69ms β
β Total Time (hash-join-split-optional-filter) β 9408.88ms β
β Average Time (HEAD) β 95.93ms β
β Average Time (hash-join-split-optional-filter) β 95.04ms β
β Queries Faster β 10 β
β Queries Slower β 8 β
β Queries with No Change β 81 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**tpcds β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 50.0s |
| Peak memory | 2.2 GiB |
| Avg memory | 1.5 GiB |
| CPU user | 204.6s |
| CPU sys | 5.5s |
| Peak spill | 0 B |
**tpcds β branch**
| Metric | Value |
|--------|-------|
| Wall time | 50.0s |
| Peak memory | 2.1 GiB |
| Avg memory | 1.4 GiB |
| CPU user | 204.0s |
| CPU sys | 5.8s |
| Peak spill | 0 B |
</details>
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push hash join dynamic filter bounds and membership as separate optional filters [datafusion]
Posted by "codecov-commenter (via GitHub)" <gi...@apache.org>.
codecov-commenter commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5819852507
## [Codecov](https://app.codecov.io/gh/apache/datafusion/pull/25713?dropdown=coverage&src=pr&el=h1&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) Report
:x: Patch coverage is `93.68917%` with `201 lines` in your changes missing coverage. Please review.
:white_check_mark: Project coverage is 82.59%. Comparing base ([`95bb0a0`](https://app.codecov.io/gh/apache/datafusion/commit/95bb0a0dfa48ca20d017faa55b2fc96ee2f300fe?dropdown=coverage&el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache)) to head ([`612efc2`](https://app.codecov.io/gh/apache/datafusion/commit/612efc26e5605bb646310c5037fe7ea220d06d37?dropdown=coverage&el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache)).
:warning: Report is 12 commits behind head on main.
| [Files with missing lines](https://app.codecov.io/gh/apache/datafusion/pull/25713?dropdown=coverage&src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) | Patch % | Lines |
|---|---|---|
| [datafusion/proto-models/src/generated/pbjson.rs](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&filepath=datafusion%2Fproto-models%2Fsrc%2Fgenerated%2Fpbjson.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9wcm90by1tb2RlbHMvc3JjL2dlbmVyYXRlZC9wYmpzb24ucnM=) | 0.00% | [70 Missing :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [...n/physical-expr/src/expressions/optional\_filter.rs](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&filepath=datafusion%2Fphysical-expr%2Fsrc%2Fexpressions%2Foptional_filter.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1leHByL3NyYy9leHByZXNzaW9ucy9vcHRpb25hbF9maWx0ZXIucnM=) | 79.53% | [17 Missing and 18 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [...physical-plan/src/joins/hash\_join/shared\_bounds.rs](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Fjoins%2Fhash_join%2Fshared_bounds.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9qb2lucy9oYXNoX2pvaW4vc2hhcmVkX2JvdW5kcy5ycw==) | 95.54% | [11 Missing and 17 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [...tafusion/physical-plan/src/joins/hash\_join/exec.rs](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Fjoins%2Fhash_join%2Fexec.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9qb2lucy9oYXNoX2pvaW4vZXhlYy5ycw==) | 90.32% | [8 Missing and 7 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [...tafusion/physical-expr/src/optional\_filter\_gate.rs](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&filepath=datafusion%2Fphysical-expr%2Fsrc%2Foptional_filter_gate.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1leHByL3NyYy9vcHRpb25hbF9maWx0ZXJfZ2F0ZS5ycw==) | 98.09% | [10 Missing and 3 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [datafusion/datasource-parquet/src/row\_filter.rs](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&filepath=datafusion%2Fdatasource-parquet%2Fsrc%2Frow_filter.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9kYXRhc291cmNlLXBhcnF1ZXQvc3JjL3Jvd19maWx0ZXIucnM=) | 98.35% | [4 Missing and 5 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [...hysical-plan/src/joins/hash\_join/inlist\_builder.rs](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Fjoins%2Fhash_join%2Finlist_builder.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9qb2lucy9oYXNoX2pvaW4vaW5saXN0X2J1aWxkZXIucnM=) | 83.33% | [2 Missing and 3 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [datafusion/physical-expr/src/utils/mod.rs](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&filepath=datafusion%2Fphysical-expr%2Fsrc%2Futils%2Fmod.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1leHByL3NyYy91dGlscy9tb2QucnM=) | 96.94% | [1 Missing and 3 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [...fusion/physical-plan/src/joins/nested\_loop\_join.rs](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Fjoins%2Fnested_loop_join.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9qb2lucy9uZXN0ZWRfbG9vcF9qb2luLnJz) | 86.66% | [2 Missing and 2 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [datafusion/datasource-parquet/src/metrics.rs](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&filepath=datafusion%2Fdatasource-parquet%2Fsrc%2Fmetrics.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9kYXRhc291cmNlLXBhcnF1ZXQvc3JjL21ldHJpY3MucnM=) | 94.23% | [1 Missing and 2 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| ... and [9 more](https://app.codecov.io/gh/apache/datafusion/pull/25713?src=pr&el=tree-more&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) | |
<details><summary>Additional details and impacted files</summary>
```diff
@@ Coverage Diff @@
## main #25713 +/- ##
==========================================
+ Coverage 82.48% 82.59% +0.10%
==========================================
Files 1140 1145 +5
Lines 437614 441545 +3931
Branches 437614 441545 +3931
==========================================
+ Hits 360986 364679 +3693
- Misses 54834 54950 +116
- Partials 21794 21916 +122
```
</details>
[:umbrella: View full report in Codecov by Harness](https://app.codecov.io/gh/apache/datafusion/pull/25713?dropdown=coverage&src=pr&el=continue&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache).
:loudspeaker: Have feedback on the report? [Share it here](https://about.codecov.io/codecov-pr-comment-feedback/?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache).
<details><summary> :rocket: New features to boost your workflow: </summary>
- :snowflake: [Test Analytics](https://docs.codecov.com/docs/test-analytics): Detect flaky tests, report on failures, and find test suite problems.
- :package: [JS Bundle Analysis](https://docs.codecov.com/docs/javascript-bundle-analysis): Save yourself from yourself by tracking and limiting bundle sizes in JS merges.
</details>
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push partitioned hash join dynamic filter bounds and membership as separate filters [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5821688886
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/25713#issuecomment-5821635949)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c5821635949-2703-mxfp8 6.12.94+ #1 SMP Wed Aug 19 07:47:20 UTC 2026 aarch64 GNU/Linux`
<details><summary>CPU Details (lscpu)</summary>
```
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 hash-join-split-optional-filter (83d19600b4a207db882058b35ec5126b1a031f61) to 95bb0a0 (merge-base) [diff](https://github.com/apache/datafusion/compare/95bb0a0dfa48ca20d017faa55b2fc96ee2f300fe..83d19600b4a207db882058b35ec5126b1a031f61)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpcds
```
</details>
Results will be posted here when complete
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push partitioned hash join dynamic filter bounds and membership as separate filters [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5821673688
π€ Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/25713#issuecomment-5821635949)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c5821635949-2702-nzgmn 6.12.94+ #1 SMP Wed Aug 19 07:47:20 UTC 2026 aarch64 GNU/Linux`
<details><summary>CPU Details (lscpu)</summary>
```
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 hash-join-split-optional-filter (83d19600b4a207db882058b35ec5126b1a031f61) to 95bb0a0 (merge-base) [diff](https://github.com/apache/datafusion/compare/95bb0a0dfa48ca20d017faa55b2fc96ee2f300fe..83d19600b4a207db882058b35ec5126b1a031f61)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpch
```
</details>
Results will be posted here when complete
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push partitioned hash join dynamic filter bounds and membership as separate filters [datafusion]
Posted by "adriangb (via GitHub)" <gi...@apache.org>.
adriangb commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5821636144
run benchmark tpch tpch10 tpcds
env:
DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] perf: push partitioned hash join dynamic filter bounds and membership as separate filters [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #25713:
URL: https://github.com/apache/datafusion/pull/25713#issuecomment-5821783995
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/25713#issuecomment-5821636144)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
Comparing hash-join-split-optional-filter (83d19600b4a207db882058b35ec5126b1a031f61) to 95bb0a0 (merge-base) [diff](https://github.com/apache/datafusion/compare/95bb0a0dfa48ca20d017faa55b2fc96ee2f300fe..83d19600b4a207db882058b35ec5126b1a031f61)
<details><summary>Run configuration</summary>
```yaml
run benchmark tpch10
env:
DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
```
</details>
<details><summary>CPU Details (lscpu)</summary>
```
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>
<details><summary>Details</summary>
<p>
```
Comparing HEAD and hash-join-split-optional-filter
--------------------
Benchmark tpch_sf10.json
--------------------
βββββββββββββ³βββββββββββββ³ββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β hash-join-split-optional-filter β Change β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 337.47 ms β 335.49 ms β no change β
β QQuery 2 β 106.34 ms β 105.29 ms β no change β
β QQuery 3 β 310.61 ms β 323.80 ms β no change β
β QQuery 4 β 296.72 ms β 321.48 ms β 1.08x slower β
β QQuery 5 β 435.40 ms β 433.89 ms β no change β
β QQuery 6 β 179.87 ms β 180.79 ms β no change β
β QQuery 7 β 441.45 ms β 453.68 ms β no change β
β QQuery 8 β 406.74 ms β 413.81 ms β no change β
β QQuery 9 β 642.75 ms β 665.95 ms β no change β
β QQuery 10 β 369.43 ms β 388.33 ms β 1.05x slower β
β QQuery 11 β 73.58 ms β 68.87 ms β +1.07x faster β
β QQuery 12 β 298.80 ms β 297.29 ms β no change β
β QQuery 13 β 405.25 ms β 399.44 ms β no change β
β QQuery 14 β 191.69 ms β 192.79 ms β no change β
β QQuery 15 β 383.55 ms β 384.76 ms β no change β
β QQuery 16 β 88.60 ms β 86.69 ms β no change β
β QQuery 17 β 497.07 ms β 479.52 ms β no change β
β QQuery 18 β 666.71 ms β 431.64 ms β +1.54x faster β
β QQuery 19 β 239.29 ms β 249.65 ms β no change β
β QQuery 20 β 280.13 ms β 290.95 ms β no change β
β QQuery 21 β 1231.40 ms β 1267.07 ms β no change β
β QQuery 22 β 129.55 ms β 132.30 ms β no change β
βββββββββββββ΄βββββββββββββ΄ββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 8012.39ms β
β Total Time (hash-join-split-optional-filter) β 7903.47ms β
β Average Time (HEAD) β 364.20ms β
β Average Time (hash-join-split-optional-filter) β 359.25ms β
β Queries Faster β 2 β
β Queries Slower β 2 β
β Queries with No Change β 18 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββ
Distribution per query (min / mean Β±stddev / max):
Comparing HEAD and hash-join-split-optional-filter
--------------------
Benchmark tpch_sf10.json
--------------------
βββββββββββββ³βββββββββββββββββββββββββββββββββββββββ³βββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ
β Query β HEAD β hash-join-split-optional-filter β Change β
β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β QQuery 1 β 337.47 / 340.05 Β±1.68 / 342.10 ms β 335.49 / 337.81 Β±1.47 / 340.13 ms β no change β
β QQuery 2 β 106.34 / 107.22 Β±0.78 / 108.66 ms β 105.29 / 106.69 Β±1.17 / 108.76 ms β no change β
β QQuery 3 β 310.61 / 312.06 Β±1.39 / 314.41 ms β 323.80 / 324.91 Β±0.69 / 325.70 ms β no change β
β QQuery 4 β 296.72 / 299.25 Β±2.97 / 304.70 ms β 321.48 / 327.03 Β±3.95 / 332.97 ms β 1.09x slower β
β QQuery 5 β 435.40 / 438.77 Β±1.89 / 440.67 ms β 433.89 / 435.45 Β±0.92 / 436.79 ms β no change β
β QQuery 6 β 179.87 / 181.95 Β±2.17 / 186.15 ms β 180.79 / 183.45 Β±2.03 / 186.99 ms β no change β
β QQuery 7 β 441.45 / 443.40 Β±1.91 / 445.92 ms β 453.68 / 456.49 Β±2.62 / 460.95 ms β no change β
β QQuery 8 β 406.74 / 409.05 Β±1.79 / 411.85 ms β 413.81 / 416.64 Β±1.46 / 417.98 ms β no change β
β QQuery 9 β 642.75 / 649.30 Β±5.86 / 657.75 ms β 665.95 / 672.46 Β±6.99 / 685.25 ms β no change β
β QQuery 10 β 369.43 / 372.14 Β±2.51 / 376.29 ms β 388.33 / 394.74 Β±5.02 / 400.65 ms β 1.06x slower β
β QQuery 11 β 73.58 / 78.38 Β±4.22 / 83.65 ms β 68.87 / 75.96 Β±7.19 / 89.76 ms β no change β
β QQuery 12 β 298.80 / 302.86 Β±4.03 / 310.49 ms β 297.29 / 303.29 Β±4.47 / 310.29 ms β no change β
β QQuery 13 β 405.25 / 424.12 Β±9.91 / 433.43 ms β 399.44 / 419.64 Β±13.53 / 440.15 ms β no change β
β QQuery 14 β 191.69 / 193.84 Β±1.92 / 196.84 ms β 192.79 / 194.36 Β±0.92 / 195.45 ms β no change β
β QQuery 15 β 383.55 / 388.33 Β±5.60 / 399.24 ms β 384.76 / 386.50 Β±1.48 / 388.66 ms β no change β
β QQuery 16 β 88.60 / 92.38 Β±4.17 / 100.45 ms β 86.69 / 90.04 Β±4.54 / 98.86 ms β no change β
β QQuery 17 β 497.07 / 501.15 Β±6.54 / 514.16 ms β 479.52 / 483.62 Β±3.92 / 490.79 ms β no change β
β QQuery 18 β 666.71 / 670.78 Β±3.45 / 675.92 ms β 431.64 / 437.92 Β±4.48 / 444.25 ms β +1.53x faster β
β QQuery 19 β 239.29 / 243.66 Β±4.51 / 251.40 ms β 249.65 / 254.46 Β±8.10 / 270.63 ms β no change β
β QQuery 20 β 280.13 / 281.50 Β±1.48 / 284.21 ms β 290.95 / 293.21 Β±1.42 / 294.72 ms β no change β
β QQuery 21 β 1231.40 / 1242.59 Β±8.24 / 1252.87 ms β 1267.07 / 1276.26 Β±5.93 / 1284.84 ms β no change β
β QQuery 22 β 129.55 / 151.41 Β±13.75 / 166.71 ms β 132.30 / 139.01 Β±6.22 / 150.01 ms β +1.09x faster β
βββββββββββββ΄βββββββββββββββββββββββββββββββββββββββ΄βββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ
ββββββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββ
β Benchmark Summary β β
β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ©
β Total Time (HEAD) β 8124.18ms β
β Total Time (hash-join-split-optional-filter) β 8009.93ms β
β Average Time (HEAD) β 369.28ms β
β Average Time (hash-join-split-optional-filter) β 364.09ms β
β Queries Faster β 2 β
β Queries Slower β 2 β
β Queries with No Change β 18 β
β Queries with Failure β 0 β
ββββββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**tpch10 β base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 45.0s |
| Peak memory | 2.8 GiB |
| Avg memory | 1.2 GiB |
| CPU user | 432.0s |
| CPU sys | 17.5s |
| Peak spill | 0 B |
**tpch10 β branch**
| Metric | Value |
|--------|-------|
| Wall time | 45.0s |
| Peak memory | 3.6 GiB |
| Avg memory | 1.2 GiB |
| CPU user | 426.0s |
| CPU sys | 17.5s |
| Peak spill | 0 B |
</details>
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org