bench(parquet): add same-projection predicate chain benchmark - #11007
Merged
alamb merged 4 commits intoSep 19, 2026
Merged
Conversation
Add a criterion benchmark for RowFilter chains whose predicates share one projection. It scans an in-memory Snappy Parquet file of 262,144 rows through the async reader and covers int64 and string filter columns in fragmented and clustered layouts, with the filter column either projected (predicate cache) or not, chains of one, two and four predicates, and survivor profiles that keep 99% or 50% per predicate or 1% in the first or last predicate. A selection_boundary group probes run lengths around the row selection policy threshold.
Contributor
|
This looks good except it makes 75 benchmarks cargo bench --bench arrow_reader_predicate_fusion --features="arrow async snap"
...
.00%) high severe
same_projection_filter/string/clustered/cached/1/all99
time: [1.6653 ms 1.6686 ms 1.6722 ms]
thrpt: [156.77 Melem/s 157.11 Melem/s 157.42 Melem/s]
Found 10 outliers among 100 measurements (10.00%)
4 (4.00%) high mild
6 (6.00%) high severe
same_projection_filter/string/clustered/cached/2/all99
time: [2.5275 ms 2.5337 ms 2.5405 ms]
thrpt: [103.18 Melem/s 103.46 Melem/s 103.72 Melem/s]
Found 4 outliers among 100 measurements (4.00%)
3 (3.00%) high mild
1 (1.00%) high severe
same_projection_filter/string/clustered/cached/2/all50
time: [2.8156 ms 2.8271 ms 2.8381 ms]
thrpt: [92.366 Melem/s 92.725 Melem/s 93.104 Melem/s]
Found 17 outliers among 100 measurements (17.00%)
11 (11.00%) low severe
6 (6.00%) high severe
same_projection_filter/string/clustered/cached/2/early1
time: [1.3208 ms 1.3229 ms 1.3253 ms]
thrpt: [197.80 Melem/s 198.16 Melem/s 198.47 Melem/s]
Found 8 outliers among 100 measurements (8.00%)
6 (6.00%) high mild
2 (2.00%) high severe
same_projection_filter/string/clustered/cached/2/late1
time: [2.0176 ms 2.0224 ms 2.0277 ms]
thrpt: [129.28 Melem/s 129.62 Melem/s 129.93 Melem/s]
Found 17 outliers among 100 measurements (17.00%)
1 (1.00%) high mild
16 (16.00%) high severe
same_projection_filter/string/clustered/cached/4/all99
time: [4.3241 ms 4.3359 ms 4.3489 ms]
thrpt: [60.278 Melem/s 60.459 Melem/s 60.624 Melem/s]
Found 8 outliers among 100 measurements (8.00%)
5 (5.00%) high mild
3 (3.00%) high severe
same_projection_filter/string/clustered/cached/4/all50
time: [3.3406 ms 3.3495 ms 3.3579 ms]
thrpt: [78.068 Melem/s 78.263 Melem/s 78.472 Melem/s]
Found 11 outliers among 100 measurements (11.00%)
3 (3.00%) low severe
8 (8.00%) high severe
same_projection_filter/string/clustered/cached/4/early1
time: [1.4078 ms 1.4143 ms 1.4211 ms]
thrpt: [184.46 Melem/s 185.35 Melem/s 186.21 Melem/s]
Found 1 outliers among 100 measurements (1.00%)
1 (1.00%) high mild
same_projection_filter/string/clustered/cached/4/late1
time: [3.8255 ms 3.8473 ms 3.8698 ms]
thrpt: [67.742 Melem/s 68.138 Melem/s 68.526 Melem/s]
Found 4 outliers among 100 measurements (4.00%)
4 (4.00%) high mild
same_projection_filter/selection_boundary/int64/uncached/4/run16
time: [6.3433 ms 6.3583 ms 6.3741 ms]
thrpt: [41.127 Melem/s 41.229 Melem/s 41.326 Melem/s]
Found 2 outliers among 100 measurements (2.00%)
2 (2.00%) high mild
same_projection_filter/selection_boundary/int64/uncached/4/run32
time: [6.0143 ms 6.0268 ms 6.0399 ms]
thrpt: [43.402 Melem/s 43.496 Melem/s 43.587 Melem/s]
Found 2 outliers among 100 measurements (2.00%)
2 (2.00%) high mild
same_projection_filter/selection_boundary/int64/uncached/4/run64
time: [5.7467 ms 5.7677 ms 5.7904 ms]
thrpt: [45.272 Melem/s 45.451 Melem/s 45.617 Melem/s]
Found 3 outliers among 100 measurements (3.00%)
2 (2.00%) high mild
1 (1.00%) high severeDo we really need them all? |
alamb
approved these changes
Sep 18, 2026
Contributor
There was a problem hiding this comment.
This looks good to me -- thank you @haohuaijin . My only potential concern is has 75 combinations -- do we really need all of them?
cargo bench --bench arrow_reader_predicate_fusion --features="arrow async snap"
...
.00%) high severe
same_projection_filter/string/clustered/cached/1/all99
time: [1.6653 ms 1.6686 ms 1.6722 ms]
thrpt: [156.77 Melem/s 157.11 Melem/s 157.42 Melem/s]
Found 10 outliers among 100 measurements (10.00%)
4 (4.00%) high mild
6 (6.00%) high severe
same_projection_filter/string/clustered/cached/2/all99
time: [2.5275 ms 2.5337 ms 2.5405 ms]
thrpt: [103.18 Melem/s 103.46 Melem/s 103.72 Melem/s]
Found 4 outliers among 100 measurements (4.00%)
3 (3.00%) high mild
1 (1.00%) high severe
same_projection_filter/string/clustered/cached/2/all50
time: [2.8156 ms 2.8271 ms 2.8381 ms]
thrpt: [92.366 Melem/s 92.725 Melem/s 93.104 Melem/s]
Found 17 outliers among 100 measurements (17.00%)
11 (11.00%) low severe
6 (6.00%) high severe
same_projection_filter/string/clustered/cached/2/early1
time: [1.3208 ms 1.3229 ms 1.3253 ms]
thrpt: [197.80 Melem/s 198.16 Melem/s 198.47 Melem/s]
Found 8 outliers among 100 measurements (8.00%)
6 (6.00%) high mild
2 (2.00%) high severe
same_projection_filter/string/clustered/cached/2/late1
time: [2.0176 ms 2.0224 ms 2.0277 ms]
thrpt: [129.28 Melem/s 129.62 Melem/s 129.93 Melem/s]
Found 17 outliers among 100 measurements (17.00%)
1 (1.00%) high mild
16 (16.00%) high severe
same_projection_filter/string/clustered/cached/4/all99
time: [4.3241 ms 4.3359 ms 4.3489 ms]
thrpt: [60.278 Melem/s 60.459 Melem/s 60.624 Melem/s]
Found 8 outliers among 100 measurements (8.00%)
5 (5.00%) high mild
3 (3.00%) high severe
same_projection_filter/string/clustered/cached/4/all50
time: [3.3406 ms 3.3495 ms 3.3579 ms]
thrpt: [78.068 Melem/s 78.263 Melem/s 78.472 Melem/s]
Found 11 outliers among 100 measurements (11.00%)
3 (3.00%) low severe
8 (8.00%) high severe
same_projection_filter/string/clustered/cached/4/early1
time: [1.4078 ms 1.4143 ms 1.4211 ms]
thrpt: [184.46 Melem/s 185.35 Melem/s 186.21 Melem/s]
Found 1 outliers among 100 measurements (1.00%)
1 (1.00%) high mild
same_projection_filter/string/clustered/cached/4/late1
time: [3.8255 ms 3.8473 ms 3.8698 ms]
thrpt: [67.742 Melem/s 68.138 Melem/s 68.526 Melem/s]
Found 4 outliers among 100 measurements (4.00%)
4 (4.00%) high mild
same_projection_filter/selection_boundary/int64/uncached/4/run16
time: [6.3433 ms 6.3583 ms 6.3741 ms]
thrpt: [41.127 Melem/s 41.229 Melem/s 41.326 Melem/s]
Found 2 outliers among 100 measurements (2.00%)
2 (2.00%) high mild
same_projection_filter/selection_boundary/int64/uncached/4/run32
time: [6.0143 ms 6.0268 ms 6.0399 ms]
thrpt: [43.402 Melem/s 43.496 Melem/s 43.587 Melem/s]
Found 2 outliers among 100 measurements (2.00%)
2 (2.00%) high mild
same_projection_filter/selection_boundary/int64/uncached/4/run64
time: [5.7467 ms 5.7677 ms 5.7904 ms]
thrpt: [45.272 Melem/s 45.451 Melem/s 45.617 Melem/s]
Found 3 outliers among 100 measurements (3.00%)
2 (2.00%) high mild
1 (1.00%) high severeDo we really need them all?
Contributor
Author
Good point! i added 75 cases because i want to have good cover for benchmark, but it can be reduce. I've reduce the case down from 75 to 30 cases while keeping the main scenarios covered. |
Contributor
|
Thanks @haohuaijin |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Which issue does this PR close?
Related to #10926.
Rationale for this change
RowFilterevaluates eachArrowPredicateseparately, so consecutive predicates on the same projection decode that column, or replay it from the predicate cache, once per predicate. #10859 fuses such chains. Per the contributing guide, the benchmark is submitted separately so it can run on the automated runner and serve as the baseline for that change.What changes are included in this PR?
A criterion benchmark,
parquet/benches/arrow_reader_predicate_fusion.rs, that scans an in-memory Snappy Parquet file of 262,144 rows through the async reader. Case names aretype/layout/cache/predicates/profile:4/all99chains across both column types (int64,string), layouts (fragmented,clustered), and cache modes (cached,uncached): 8 cases1/all99controls,2/all99, and2/all50chains across both types and cache modes: 12 cases2/all50and4/all50chains across both types, retaining cases where fusion has shown regressions: 4 casesint64/fragmented/4withearly1andlate1across both cache modes, covering predicate ordering: 4 casesselection_boundarygroup with run lengths 16 and 64 on either side of the default row selection policy threshold: 2 cases30 cases in total. Each case is validated once outside measurement for the expected row count and predicate cache use. Run-length cases also validate the predicate mask run lengths.
Are these changes tested?
The benchmark builds and all 30 cases pass with
cargo bench -p parquet --bench arrow_reader_predicate_fusion --features "arrow async snap" --locked -- --test.Are there any user-facing changes?
No.