Skip to content

perf: replace the routed hash join dynamic filter with one global filter (InList or bloom filter) - #26442

Draft
adriangb wants to merge 3 commits into
apache:mainfrom
adriangb:hash-join-global-bloom-filter
Draft

perf: replace the routed hash join dynamic filter with one global filter (InList or bloom filter)#26442
adriangb wants to merge 3 commits into
apache:mainfrom
adriangb:hash-join-global-bloom-filter

Conversation

@adriangb

@adriangb adriangb commented Oct 17, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Which issue does this PR close?

Depends on apache/arrow-rs#11641 (new Sbbf APIs). Until arrow-rs releases them, this PR patches the workspace with [patch.crates-io] git dependencies to adriangb/arrow-rs:sbbf-batch-union-apis-60: the arrow-rs 60.0.0 tag (the version in Cargo.lock) + apache/arrow-rs#11088 (Sbbf::estimated_fpp) + the changes of apache/arrow-rs#11641 (Sbbf::insert_hashes, Sbbf::check_hashes, Sbbf::union). No other arrow-rs main changes are included. The patch (first commit) must be removed before merge.

Rationale for this change

The dynamic filter of a partitioned hash join routes each probe row to its partition with a CASE on the repartition hash. Profiles show that the CASE machinery costs about 6 / 50 / 150 ns per probe row at 1 / 16 / 128 partitions, and it is most of the query CPU time for join-heavy queries.

Before (snapshot of test_hashjoin_dynamic_filter_pushdown_partitioned, 12 partitions, two key columns):

predicate=DynamicFilter [ CASE hash_repartition % 12 WHEN 5 THEN a@0 >= ab AND a@0 <= ab AND b@1 >= bb AND b@1 <= bb AND struct(a@0, b@1) IN (SET) ([{c0:ab,c1:bb}]) WHEN 8 THEN a@0 >= aa AND a@0 <= aa AND b@1 >= ba AND b@1 <= ba AND struct(a@0, b@1) IN (SET) ([{c0:aa,c1:ba}]) ELSE false END ]

After:

predicate=DynamicFilter [ a@0 >= aa AND a@0 <= ab AND b@1 >= ba AND b@1 <= bb AND struct(a@0, b@1) IN (SET) ([{c0:ab,c1:bb}, {c0:aa,c1:ba}]) ]

With more build keys than the InList limit, the membership test is bloom_filter (was hash_lookup inside each CASE branch):

predicate=DynamicFilter [ k@0 >= 5 AND k@0 <= 390 AND bloom_filter ]

TPC-H SF100, wall time as a ratio to main (lower is better), 3 rounds. The A/A control (main against itself) stayed within 0.99–1.04.

query, partitions, pushdown #26249 this PR (previous version) this PR
Q7, 16, off 0.99 1.11 1.05
Q8, 128, off 1.07 1.18 1.05
Q9, 16, off 1.02 1.25 1.07
Q9, 128, off 0.99 1.29 1.05
Q21, 128, off 1.01 1.12 0.96
Q7, 16, on 0.77 0.74 0.78
Q8, 128, on 0.48 0.45 0.42
Q9, 128, on 0.38 0.36 0.27
Q17, 16, on 0.45 0.39 0.38
Q17, 128, on 0.26 0.23 0.22
Q21, 16, on 0.67 0.63 0.54
Q21, 128, on 0.40 0.40 0.32
  • With pushdown on, this PR is the fastest on every query above except Q7 at 16 partitions.
  • With pushdown off, CPU time matches main (0.98–1.01). The remaining 3–7% wall time is the wait at the build barrier.
  • The previous version built one filter serially at the barrier from sorted per-partition hashes. Per-partition filters merged with Sbbf::union move the fill into the parallel build tasks: fill takes 5–10 ms per partition, union ~0.5 ms, fold <1 ms.
  • Larger filters were slower: on Q9 at 16 partitions, 0.46 with no bloom filter, 0.50 at 8 MiB, 0.57 at 128 MiB. The default hash_join_bloom_filter_max_bytes is now 8 MiB.

Methodology: dedicated GCP c4a-standard-16 VM (Arm, 64 GB), TPC-H SF100 parquet. Each config runs 3 A/B rounds with alternating order, plus an A/A control of main against itself in the same run. Ratios are paired per round, with a bootstrap 95% CI.

What changes are included in this PR?

One global filter for all probe rows, with no routing:

[key IS NULL OR] key >= global_min AND key <= global_max AND membership(key)

membership = key IN (SET) (all build keys)  if total distinct keys <= hash_join_dynamic_filter_inlist_max_distinct_values (20)
           | bloom_filter                   otherwise (one parquet Sbbf of the join key hashes)
           | (none, pruning bitmap only)    if the bloom filter is too big, saturated, or does not fit in memory
  • The filter is true for every build key whatever the partitioning, so it is used for hash partitioned, range partitioned and CollectLeft joins.
  • IS NULL OR is kept for null-aware (NOT IN) and null-equal joins, as before.
  • Fallback to the existing routed CASE filter:
    • when a build partition was canceled (its keys are unknown, so the routed filter keeps every row of that route).
  • Bloom filter (BloomFilterExpr, new):
    • Parquet Sbbf of the join key hashes (create_hashes with HASH_JOIN_SEED, the same hashes the probe side computes). Multi-column keys use one filter of the combined hash. All key types the join supports work, including floats (-0.0 and 0.0 hash the same).
    • Each build partition fills its own filter in its build task with Sbbf::insert_hashes, sized for its distinct key count times the number of partitions (an upper bound for the whole build side), rounded to a power of two.
    • At the barrier the partition filters are ORed with Sbbf::union (a larger filter is folded on the fly), then fold_to_target_fpp(fpp). No OS threads, locks or atomics. The build is eager.
    • The probe checks a whole batch with the branch-free Sbbf::check_hashes.
    • Skipped (bounds and range pruning only) when a partition's filter at fpp would be larger than hash_join_bloom_filter_max_bytes.
    • Not used when estimated_fpp() > hash_join_bloom_filter_max_fpp (saturated; safety net).
    • Memory: each partition filter is charged to the join's build reservation; the merged filter to a HashJoinBloomFilter reservation that the expression owns (released on drop). If a reservation fails, no bloom filter is built and the query does not fail.
    • Serializes to proto as lit(true), like HashTableLookupExpr (it holds runtime build state).
  • Pruning: the KeyRangeBitmaps of all partitions (Prune hash join probe side via a build-side key range bitmap #25602) are merged with KeyRangeBitmap::try_union and attached to the membership expression, so parquet keeps row group and page pruning, also with pushdown_filters = false and when no bloom filter is built (key_range_pruning in EXPLAIN).
  • Merged bounds ignore NULL bounds of partitions that hold only NULL keys.

New options:

Option Default Meaning
datafusion.optimizer.hash_join_dynamic_filter_inlist_max_distinct_values 20 Maximum distinct build keys (all partitions) for the global InList; above it, a bloom filter
datafusion.optimizer.hash_join_bloom_filter_fpp 0.01 Target false positive probability
datafusion.optimizer.hash_join_bloom_filter_max_bytes 8388608 (8 MiB) Maximum filter size; a larger filter is not built
datafusion.optimizer.hash_join_bloom_filter_max_fpp 0.15 Do not use a saturated filter

hash_join_inlist_pushdown_max_distinct_values (default 150) keeps its meaning: the per-partition InList limit. It still decides whether each partition collects an InList, and the global InList needs every partition to have one. So a separate option for the global limit is cleaner than changing the meaning of the existing one.

datafusion-physical-plan now depends on parquet (for Sbbf).

What is the testing strategy for this PR?

  • bloom_filter.rs: build keys are found and others rejected, NULL probe keys, merged filter equals one filter of all keys, partition filters of different sizes, too-big build side has no filter, saturated filter is not used, memory is reserved and released, small memory pool skips the filter, multi-column keys, float keys, proto encodes as lit(true).
  • shared_bounds.rs: global bloom filter shape and results, InList vs bloom at the global limit, range partitioning uses the global filter, float keys use the global bloom filter, canceled partition falls back to the routed filter, small memory pool falls back to bounds, merged pruning bitmap, merged bounds ignore NULL bounds. Existing null-aware / null-equal / canceled tests pass.
  • key_range_bitmap.rs: try_union never drops a key.
  • dynamic_row_group_pruning.slt: new EXPLAIN ANALYZE of a partitioned join with pushdown_filters = false: the global filter, and 2 of 4 row groups pruned by the merged bitmap.
  • Updated snapshots: filter_pushdown.rs, push_down_filter_parquet.slt, range_partitioning.slt, dynamic_row_group_pruning.slt, information_schema.slt.

Are there any user-facing changes?

  • EXPLAIN output of hash join dynamic filters changes (no CASE hash_repartition, bloom_filter or key_range_pruning instead of hash_lookup).
  • CollectLeft dynamic filters are now approximate: a superset of the build keys (bloom filter, ~1% false positives) instead of the exact hash table lookup. Join results do not change.
  • Four new configuration options (see above). No breaking API changes.

Decisions

  • Global InList limit: 20, as a new option. The per-partition option keeps its meaning.
  • Bloom filter build: eager. Each partition fills its own filter; the barrier ORs them with Sbbf::union and folds. No shared mutable filter, OS threads or lazy build.
  • Skip the bloom filter when it cannot reach the target fpp within hash_join_bloom_filter_max_bytes (too big to be useful). The 15% saturation check stays as a safety net.
  • Bloom filter memory (partition filters and merged filter) is charged to the memory pool; on failure, bounds and range pruning only.
  • CollectLeft joins use the bloom filter instead of the exact HashTableLookupExpr.
  • The parquet dependency of datafusion-physical-plan is accepted for now.

Open questions

  • Size partition filters from the memory budget instead of num_distinct × partitions capped by hash_join_bloom_filter_max_bytes.
  • Reuse the join hashes computed for the hash table instead of hashing the keys again for the bloom filter.

🤖 Generated with Claude Code

@github-actions github-actions Bot added documentation Improvements or additions to documentation core Core DataFusion crate sqllogictest SQL Logic Tests (.slt) common Related to common crate physical-plan Changes to the physical-plan crate labels Oct 17, 2026
@adriangb adriangb mentioned this pull request Oct 17, 2026
Draft
@github-actions

github-actions Bot commented Oct 17, 2026 •
edited
Loading

Copy link
Copy Markdown

Thank you for opening this pull request!

Reviewer note: 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 
     Cloning 941146372a8915234682636e124abd3afb540963
    Building datafusion v55.1.0 (current)
error: running cargo-doc on crate 'datafusion' failed with output:
-----
   Compiling proc-macro2 v1.0.107
   Compiling quote v1.0.47
   Compiling unicode-ident v1.0.27
    Checking cfg-if v1.0.5
   Compiling libc v0.2.191
   Compiling shlex v2.0.1
   Compiling autocfg v1.5.1
   Compiling syn v3.0.9
   Compiling syn v2.0.119
   Compiling jobserver v0.1.35
   Compiling find-msvc-tools v0.1.15
   Compiling cc v1.11.0
   Compiling libm v0.2.16
    Checking memchr v2.8.3
   Compiling num-traits v0.2.19
   Compiling zerocopy v0.8.63
    Checking bytes v1.12.1
   Compiling serde_core v1.0.229
   Compiling getrandom v0.3.4
    Checking itoa v1.0.18
    Checking once_cell v1.21.4
    Checking num-integer v0.1.47
   Compiling zmij v1.0.23
   Compiling synstructure v0.14.0
   Compiling zerocopy-derive v0.8.63
   Compiling serde_json v1.0.154
   Compiling serde v1.0.229
   Compiling serde_derive v1.0.229
    Checking foldhash v0.2.0
    Checking equivalent v1.0.3
   Compiling version_check v0.9.5
    Checking allocator-api2 v0.2.21
    Checking siphasher v1.0.4
    Checking iana-time-zone v0.1.65
    Checking phf_shared v0.12.1
    Checking chrono v0.4.45
    Checking hashbrown v0.17.1
   Compiling ahash v0.8.12
   Compiling zerofrom-derive v0.1.8
    Checking num-bigint v0.5.1
   Compiling pkg-config v0.3.34
   Compiling chrono-tz v0.10.4
    Checking zerofrom v0.1.8
    Checking phf v0.12.1
   Compiling yoke-derive v0.8.4
    Checking arrow-schema v60.1.0
    Checking stable_deref_trait v1.2.1
    Checking num-complex v0.4.6
    Checking yoke v0.8.4
   Compiling zerovec-derive v0.11.7
   Compiling displaydoc v0.2.7
   Compiling zstd-sys v2.1.2+zstd.1.5.7
    Checking pin-project-lite v0.2.17
    Checking smallvec v1.16.4
    Checking half v2.7.1
    Checking zerovec v0.11.9
    Checking arrow-buffer v60.1.0
    Checking tinystr v0.8.5
    Checking utf_types v1.0.0
   Compiling object v0.39.1
    Checking arrow-data v60.1.0
    Checking writeable v0.6.4
    Checking lexical-util v1.0.7
   Compiling zstd-safe v8.2.0
    Checking futures-core v0.3.34
    Checking futures-sink v0.3.34
    Checking litemap v0.8.3
    Checking utf8_iter v1.1.0
    Checking icu_locale_core v2.4.0
    Checking potential_utf v0.1.6
    Checking zerotrie v0.2.5
    Checking bitflags v2.13.2
   Compiling icu_properties_data v2.4.0
   Compiling icu_normalizer_data v2.4.0
    Checking icu_provider v2.4.0
    Checking icu_collections v2.4.0
    Checking arrow-array v60.1.0
   Compiling tokio-macros v2.7.2
   Compiling crc32fast v1.5.4
   Compiling semver v1.0.28
    Checking scopeguard v1.2.0
   Compiling rustc_version v0.4.1
    Checking tokio v1.54.1
    Checking arrow-cmp v60.1.0
    Checking arrow-select v60.1.0
   Compiling ar_archive_writer v0.5.3
    Checking futures-channel v0.3.34
    Checking lexical-parse-integer v1.0.6
    Checking lexical-write-integer v1.0.6
   Compiling futures-macro v0.3.34
    Checking slab v0.4.12
   Compiling getrandom v0.4.3
   Compiling parking_lot_core v0.9.12
    Checking adler2 v2.0.1
    Checking rand_core v0.10.1
    Checking futures-io v0.3.34
    Checking simd-adler32 v0.3.10
    Checking futures-task v0.3.34
    Checking lexical-write-float v1.0.6
    Checking futures-util v0.3.34
    Checking miniz_oxide v0.9.1
   Compiling psm v0.1.32
    Checking lexical-parse-float v1.0.6
    Checking icu_properties v2.4.0
    Checking icu_normalizer v2.4.0
   Compiling flatbuffers v25.12.19
    Checking aho-corasick v1.1.5
   Compiling cfg_aliases v0.2.2
    Checking unicode-segmentation v1.13.3
    Checking regex-syntax v0.8.11
    Checking base64 v0.23.1
    Checking ryu v1.0.23
    Checking zlib-rs v0.6.8
    Checking unicode-width v0.2.3
    Checking comfy-table v8.0.1
   Compiling nix v0.31.3
    Checking idna_adapter v1.2.2
    Checking lexical-core v1.0.6
    Checking regex-automata v0.4.18
    Checking flate2 v1.1.10
    Checking arrow-ord v60.1.0
    Checking lock_api v0.4.14
    Checking indexmap v2.14.2
    Checking atoi v3.1.1
   Compiling stacker v0.1.25
    Checking alloc-no-stdlib v3.0.0
    Checking zstd v0.14.3
   Compiling snap v1.1.2
    Checking twox-hash v2.1.5
    Checking percent-encoding v2.3.2
   Compiling thiserror v2.0.21
    Checking lz4_flex v0.14.0
    Checking form_urlencoded v1.2.2
    Checking alloc-stdlib v0.3.0
    Checking arrow-cast v60.1.0
    Checking regex v1.13.1
    Checking idna v1.1.0
    Checking futures-executor v0.3.34
   Compiling thiserror-impl v2.0.21
   Compiling tracing-attributes v0.1.31
    Checking tracing-core v0.1.36
   Compiling ring v0.17.14
    Checking csv-core v0.1.13
    Checking either v1.19.0
    Checking simdutf8 v0.1.5
    Checking same-file v1.0.6
    Checking itertools v0.15.0
    Checking walkdir v2.5.0
    Checking csv v1.4.0
    Checking tracing v0.1.44
    Checking arrow-ipc v60.1.0
    Checking url v2.5.8
    Checking futures v0.3.34
    Checking brotli-decompressor v6.2.0
    Checking parking_lot v0.12.5
   Compiling async-trait v0.1.92
   Compiling recursive-proc-macro-impl v0.1.1
    Checking http v1.5.0
    Checking getrandom v0.2.17
    Checking humantime v2.4.0
    Checking untrusted v0.9.0
    Checking log v0.4.35
    Checking brotli v9.1.1
    Checking recursive v0.1.1
    Checking arrow-csv v60.1.0
    Checking object_store v0.14.2
    Checking arrow-json v60.1.0
    Checking arrow-string v60.1.0
    Checking uuid v1.29.0
    Checking arrow-row v60.1.0
    Checking arrow-arith v60.1.0
   Compiling sqlparser_derive v0.6.0
   Compiling seq-macro v0.3.6
    Checking arrow v60.1.0
    Checking sqlparser v0.63.0
    Checking hex v0.4.3
    Checking lock_api v0.5.0
    Checking parquet v60.1.0
    Checking parking_lot_core v0.10.0
   Compiling pin-project-internal v1.1.13
    Checking parking_lot v0.13.0
    Checking typenum v1.20.1
    Checking pin-project v1.1.13
    Checking datafusion-doc v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/doc)
    Checking cpufeatures v0.3.1
   Compiling rustix v1.1.6
    Checking hybrid-array v0.4.15
   Compiling crossbeam-utils v0.8.23
    Checking cmov v0.5.4
    Checking foldhash v0.1.5
    Checking ctutils v0.4.7
    Checking hashbrown v0.15.5
    Checking crypto-common v0.2.2
    Checking block-buffer v0.12.1
    Checking ppv-lite86 v0.2.21
    Checking rand_core v0.9.5
    Checking const-oid v0.10.2
    Checking linux-raw-sys v0.12.1
    Checking fixedbitset v0.5.7
    Checking petgraph v0.8.3
    Checking rand_chacha v0.9.0
    Checking digest v0.11.3
    Checking tokio-util v0.7.20
    Checking fastrand v2.5.0
    Checking hashbrown v0.14.5
    Checking tempfile v3.27.0
   Compiling datafusion-macros v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/macros)
    Checking rand v0.9.5
    Checking dashmap v6.2.1
   Compiling blake3 v1.8.7
    Checking constant_time_eq v0.4.2
    Checking arrayvec v0.7.8
    Checking sha2 v0.11.1
    Checking blake2 v0.11.0
    Checking md-5 v0.11.0
   Compiling liblzma-sys v0.4.9
    Checking libbz2-rs-sys v0.2.6
    Checking datafusion-common-runtime v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/common-runtime)
    Checking compression-core v0.4.34
    Checking glob v0.3.4
    Checking chacha20 v0.10.2
    Checking bzip2 v0.6.1
   Compiling bigdecimal v0.4.11
    Checking crc-catalog v2.5.0
   Compiling heck v0.5.0
    Checking crc v3.4.0
   Compiling strum_macros v0.28.0
    Checking rand v0.10.3
    Checking num-bigint v0.4.8
    Checking tokio-stream v0.1.19
    Checking liblzma v0.4.9
    Checking compression-codecs v0.4.47
    Checking arrow-avro v60.1.0
    Checking async-compression v0.4.52
    Checking datafusion-common v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/common)
    Checking datafusion-expr-common v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/expr-common)
    Checking datafusion-physical-expr-common v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/physical-expr-common)
    Checking datafusion-functions-window-common v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/functions-window-common)
    Checking datafusion-functions-aggregate-common v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/functions-aggregate-common)
    Checking datafusion-expr v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/expr)
    Checking datafusion-physical-expr v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/physical-expr)
    Checking datafusion-execution v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/execution)
    Checking datafusion-functions v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/functions)
    Checking datafusion-functions-aggregate v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/functions-aggregate)
    Checking datafusion-functions-window v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/functions-window)
    Checking datafusion-optimizer v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/optimizer)
    Checking datafusion-physical-plan v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/physical-plan)
    Checking datafusion-physical-expr-adapter v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/physical-expr-adapter)
    Checking datafusion-functions-nested v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/functions-nested)
    Checking datafusion-sql v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/sql)
error[E0599]: no method named `insert_hashes` found for struct `Sbbf` in the current scope
   --> /home/runner/work/datafusion/datafusion/datafusion/physical-plan/src/joins/hash_join/bloom_filter.rs:121:16
    |
121 |         filter.insert_hashes(hashes);
    |                ^^^^^^^^^^^^^ method not found in `Sbbf`

error[E0599]: no method named `union` found for struct `Sbbf` in the current scope
   --> /home/runner/work/datafusion/datafusion/datafusion/physical-plan/src/joins/hash_join/bloom_filter.rs:160:20
    |
160 |             filter.union(partition).map_err(parquet_err)?;
    |                    ^^^^^ method not found in `Sbbf`

error[E0599]: no method named `check_hashes` found for struct `Sbbf` in the current scope
   --> /home/runner/work/datafusion/datafusion/datafusion/physical-plan/src/joins/hash_join/bloom_filter.rs:313:37
    |
313 |         let matches = filter.filter.check_hashes(&hashes);
    |                                     ^^^^^^^^^^^^ method not found in `Sbbf`

For more information about this error, try `rustc --explain E0599`.
error: could not compile `datafusion-physical-plan` (lib) due to 3 previous errors

-----

error: failed to build rustdoc for crate datafusion v55.1.0
note: this is usually due to a compilation error in the crate,
      and is unlikely to be a bug in cargo-semver-checks
note: the following command can be used to reproduce the error:
      cargo new --lib example &&
          cd example &&
          echo '[workspace]' >> Cargo.toml &&
          cargo add --path /home/runner/work/datafusion/datafusion/datafusion/core --features array_expressions,avro,backtrace,bzip2,compression,crypto_expressions,datafusion-datasource-avro,datafusion-datasource-parquet,datafusion-functions-nested,datafusion-sql,datetime_expressions,default,docs_generation,encoding_expressions,extended_tests,flate2,force_hash_collisions,liblzma,math_expressions,nested_expressions,parquet,parquet_encryption,recursive_protection,regex_expressions,serde,sql,sqlparser,string_expressions,unicode_expressions,zstd &&
          cargo check &&
          cargo doc

    Building datafusion-common v55.1.0 (current)
       Built [  11.955s] (current)
     Parsing datafusion-common v55.1.0 (current)
      Parsed [   0.054s] (current)
    Building datafusion-common v55.1.0 (baseline)
       Built [   0.860s] (baseline)
     Parsing datafusion-common v55.1.0 (baseline)
      Parsed [   0.054s] (baseline)
    Checking datafusion-common v55.1.0 -> v55.1.0 (no change; assume patch)
     Checked [   0.554s] 229 checks: 228 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/v0.51.0/src/lints/constructible_struct_adds_field.ron

Failed in:
  field OptimizerOptions.hash_join_dynamic_filter_inlist_max_distinct_values in /home/runner/work/datafusion/datafusion/datafusion/common/src/config.rs:1612
  field OptimizerOptions.hash_join_bloom_filter_fpp in /home/runner/work/datafusion/datafusion/datafusion/common/src/config.rs:1612
  field OptimizerOptions.hash_join_bloom_filter_max_bytes in /home/runner/work/datafusion/datafusion/datafusion/common/src/config.rs:1612
  field OptimizerOptions.hash_join_bloom_filter_max_fpp in /home/runner/work/datafusion/datafusion/datafusion/common/src/config.rs:1612

     Summary semver requires new major version: 1 major and 0 minor checks failed
    Finished [  14.513s] datafusion-common
    Building datafusion-physical-plan v55.1.0 (current)
       Built [  13.061s] (current)
     Parsing datafusion-physical-plan v55.1.0 (current)
      Parsed [   0.176s] (current)
    Building datafusion-physical-plan v55.1.0 (baseline)
       Built [  10.911s] (baseline)
     Parsing datafusion-physical-plan v55.1.0 (baseline)
      Parsed [   0.167s] (baseline)
    Checking datafusion-physical-plan v55.1.0 -> v55.1.0 (no change; assume patch)
     Checked [   0.569s] 229 checks: 229 pass, 31 skip
     Summary no semver update required
    Finished [  26.056s] datafusion-physical-plan
    Building datafusion-pruning v55.1.0 (current)
error: running cargo-doc on crate 'datafusion-pruning' failed with output:
-----
    Checking getrandom v0.4.3
    Checking arrow v60.1.0
    Checking uuid v1.29.0
    Checking tempfile v3.27.0
    Checking datafusion-common v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/common)
    Checking datafusion-expr-common v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/expr-common)
    Checking datafusion-physical-expr-common v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/physical-expr-common)
    Checking datafusion-functions-window-common v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/functions-window-common)
    Checking datafusion-functions-aggregate-common v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/functions-aggregate-common)
    Checking datafusion-expr v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/expr)
    Checking datafusion-execution v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/execution)
    Checking datafusion-physical-expr v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/physical-expr)
    Checking datafusion-functions v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/functions)
    Checking datafusion-physical-plan v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/physical-plan)
    Checking datafusion-physical-expr-adapter v55.1.0 (/home/runner/work/datafusion/datafusion/datafusion/physical-expr-adapter)
error[E0599]: no method named `insert_hashes` found for struct `Sbbf` in the current scope
   --> /home/runner/work/datafusion/datafusion/datafusion/physical-plan/src/joins/hash_join/bloom_filter.rs:121:16
    |
121 |         filter.insert_hashes(hashes);
    |                ^^^^^^^^^^^^^ method not found in `Sbbf`

error[E0599]: no method named `union` found for struct `Sbbf` in the current scope
   --> /home/runner/work/datafusion/datafusion/datafusion/physical-plan/src/joins/hash_join/bloom_filter.rs:160:20
    |
160 |             filter.union(partition).map_err(parquet_err)?;
    |                    ^^^^^ method not found in `Sbbf`

error[E0599]: no method named `check_hashes` found for struct `Sbbf` in the current scope
   --> /home/runner/work/datafusion/datafusion/datafusion/physical-plan/src/joins/hash_join/bloom_filter.rs:313:37
    |
313 |         let matches = filter.filter.check_hashes(&hashes);
    |                                     ^^^^^^^^^^^^ method not found in `Sbbf`

For more information about this error, try `rustc --explain E0599`.
error: could not compile `datafusion-physical-plan` (lib) due to 3 previous errors

-----

error: failed to build rustdoc for crate datafusion-pruning v55.1.0
note: this is usually due to a compilation error in the crate,
      and is unlikely to be a bug in cargo-semver-checks
note: the following command can be used to reproduce the error:
      cargo new --lib example &&
          cd example &&
          echo '[workspace]' >> Cargo.toml &&
          cargo add --path /home/runner/work/datafusion/datafusion/datafusion/pruning &&
          cargo check &&
          cargo doc

error: aborting due to failure to build rustdoc for crate datafusion v55.1.0

@github-actions github-actions Bot added the auto detected api change Auto detected API change label Oct 17, 2026
@adriangb

adriangb commented Oct 17, 2026

Copy link
Copy Markdown
Contributor Author

run benchmarks

@adriangb

adriangb commented Oct 17, 2026

Copy link
Copy Markdown
Contributor Author

run benchmarks

env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: true

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6235567197-3534-txmrp 6.12.94+ #1 SMP Fri Aug 21 08:00:16 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark clickbench_partitioned

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6235567197-3535-9dpcs 6.12.94+ #1 SMP Fri Aug 21 08:00:16 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark tpcds

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6235567197-3536-kmfgh 6.12.94+ #1 SMP Fri Aug 21 08:00:16 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark tpch

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6235583123-3537-85wzt 6.12.94+ #1 SMP Fri Aug 21 08:00:16 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark clickbench_partitioned
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6235583123-3539-xnfcc 6.12.94+ #1 SMP Fri Aug 21 08:00:16 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark tpch
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6235583123-3538-wz44t 6.12.94+ #1 SMP Fri Aug 21 08:00:16 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark tpcds
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark tpch
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃     HEAD ┃ hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │ 41.07 ms │                      41.47 ms │     no change │
│ QQuery 2  │ 18.60 ms │                      18.35 ms │     no change │
│ QQuery 3  │ 29.04 ms │                      29.45 ms │     no change │
│ QQuery 4  │ 17.40 ms │                      17.64 ms │     no change │
│ QQuery 5  │ 35.59 ms │                      36.33 ms │     no change │
│ QQuery 6  │ 13.82 ms │                      13.76 ms │     no change │
│ QQuery 7  │ 42.94 ms │                      46.64 ms │  1.09x slower │
│ QQuery 8  │ 41.62 ms │                      42.72 ms │     no change │
│ QQuery 9  │ 50.68 ms │                      53.07 ms │     no change │
│ QQuery 10 │ 40.88 ms │                      41.19 ms │     no change │
│ QQuery 11 │ 13.53 ms │                      13.44 ms │     no change │
│ QQuery 12 │ 18.65 ms │                      18.38 ms │     no change │
│ QQuery 13 │ 39.89 ms │                      40.15 ms │     no change │
│ QQuery 14 │ 24.57 ms │                      25.04 ms │     no change │
│ QQuery 15 │ 31.07 ms │                      31.18 ms │     no change │
│ QQuery 16 │ 13.65 ms │                      13.94 ms │     no change │
│ QQuery 17 │ 76.30 ms │                      67.74 ms │ +1.13x faster │
│ QQuery 18 │ 46.12 ms │                      62.82 ms │  1.36x slower │
│ QQuery 19 │ 32.31 ms │                      32.11 ms │     no change │
│ QQuery 20 │ 31.98 ms │                      31.90 ms │     no change │
│ QQuery 21 │ 55.87 ms │                      57.40 ms │     no change │
│ QQuery 22 │ 13.56 ms │                      13.69 ms │     no change │
└───────────┴──────────┴───────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ Benchmark Summary                            ┃          ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 729.15ms │
│ Total Time (hash-join-global-bloom-filter)   │ 748.40ms │
│ Average Time (HEAD)                          │  33.14ms │
│ Average Time (hash-join-global-bloom-filter) │  34.02ms │
│ Queries Faster                               │        1 │
│ Queries Slower                               │        2 │
│ Queries with No Change                       │       19 │
│ Queries with Failure                         │        0 │
└──────────────────────────────────────────────┴──────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                           HEAD ┃  hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │ 41.07 / 42.25 ±1.04 / 43.91 ms │ 41.47 / 42.14 ±1.03 / 44.18 ms │     no change │
│ QQuery 2  │ 18.60 / 18.83 ±0.33 / 19.48 ms │ 18.35 / 18.69 ±0.24 / 19.09 ms │     no change │
│ QQuery 3  │ 29.04 / 29.89 ±0.82 / 31.27 ms │ 29.45 / 29.61 ±0.13 / 29.79 ms │     no change │
│ QQuery 4  │ 17.40 / 17.77 ±0.32 / 18.34 ms │ 17.64 / 17.94 ±0.20 / 18.23 ms │     no change │
│ QQuery 5  │ 35.59 / 35.98 ±0.37 / 36.48 ms │ 36.33 / 36.76 ±0.38 / 37.46 ms │     no change │
│ QQuery 6  │ 13.82 / 14.23 ±0.40 / 14.87 ms │ 13.76 / 14.07 ±0.20 / 14.26 ms │     no change │
│ QQuery 7  │ 42.94 / 43.69 ±0.99 / 45.53 ms │ 46.64 / 49.89 ±3.26 / 55.80 ms │  1.14x slower │
│ QQuery 8  │ 41.62 / 42.01 ±0.45 / 42.63 ms │ 42.72 / 43.10 ±0.30 / 43.48 ms │     no change │
│ QQuery 9  │ 50.68 / 51.93 ±0.90 / 53.39 ms │ 53.07 / 54.80 ±1.14 / 56.53 ms │  1.06x slower │
│ QQuery 10 │ 40.88 / 41.05 ±0.15 / 41.28 ms │ 41.19 / 41.34 ±0.13 / 41.49 ms │     no change │
│ QQuery 11 │ 13.53 / 13.81 ±0.31 / 14.35 ms │ 13.44 / 13.80 ±0.19 / 13.99 ms │     no change │
│ QQuery 12 │ 18.65 / 19.02 ±0.22 / 19.31 ms │ 18.38 / 19.19 ±0.67 / 20.30 ms │     no change │
│ QQuery 13 │ 39.89 / 40.76 ±1.28 / 43.27 ms │ 40.15 / 40.53 ±0.59 / 41.71 ms │     no change │
│ QQuery 14 │ 24.57 / 24.96 ±0.27 / 25.41 ms │ 25.04 / 25.72 ±0.87 / 27.36 ms │     no change │
│ QQuery 15 │ 31.07 / 31.45 ±0.31 / 32.01 ms │ 31.18 / 31.53 ±0.41 / 32.30 ms │     no change │
│ QQuery 16 │ 13.65 / 13.91 ±0.18 / 14.18 ms │ 13.94 / 14.11 ±0.12 / 14.25 ms │     no change │
│ QQuery 17 │ 76.30 / 77.37 ±1.15 / 79.57 ms │ 67.74 / 69.64 ±1.31 / 71.76 ms │ +1.11x faster │
│ QQuery 18 │ 46.12 / 47.18 ±1.08 / 48.59 ms │ 62.82 / 64.26 ±0.95 / 65.70 ms │  1.36x slower │
│ QQuery 19 │ 32.31 / 32.90 ±0.67 / 34.15 ms │ 32.11 / 32.32 ±0.15 / 32.56 ms │     no change │
│ QQuery 20 │ 31.98 / 32.32 ±0.22 / 32.63 ms │ 31.90 / 32.50 ±0.43 / 33.06 ms │     no change │
│ QQuery 21 │ 55.87 / 56.36 ±0.39 / 56.95 ms │ 57.40 / 58.71 ±0.70 / 59.51 ms │     no change │
│ QQuery 22 │ 13.56 / 13.84 ±0.21 / 14.05 ms │ 13.69 / 14.25 ±0.49 / 14.98 ms │     no change │
└───────────┴────────────────────────────────┴────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ Benchmark Summary                            ┃          ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 741.54ms │
│ Total Time (hash-join-global-bloom-filter)   │ 764.89ms │
│ Average Time (HEAD)                          │  33.71ms │
│ Average Time (hash-join-global-bloom-filter) │  34.77ms │
│ Queries Faster                               │        1 │
│ Queries Slower                               │        3 │
│ Queries with No Change                       │       18 │
│ Queries with Failure                         │        0 │
└──────────────────────────────────────────────┴──────────┘

Resource Usage

tpch — base (merge-base)

Metric Value
Wall time 5.0s
Peak memory 1.1 GiB
Avg memory 484.0 MiB
CPU user 20.5s
CPU sys 1.7s
Peak spill 0 B

tpch — branch

Metric Value
Wall time 5.0s
Peak memory 1.3 GiB
Avg memory 502.8 MiB
CPU user 21.2s
CPU sys 1.7s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark tpch
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃     HEAD ┃ hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │ 41.83 ms │                      42.55 ms │     no change │
│ QQuery 2  │ 19.31 ms │                      19.44 ms │     no change │
│ QQuery 3  │ 31.95 ms │                      31.03 ms │     no change │
│ QQuery 4  │ 25.75 ms │                      20.26 ms │ +1.27x faster │
│ QQuery 5  │ 43.89 ms │                      40.51 ms │ +1.08x faster │
│ QQuery 6  │ 22.88 ms │                      18.22 ms │ +1.26x faster │
│ QQuery 7  │ 46.20 ms │                      47.51 ms │     no change │
│ QQuery 8  │ 47.57 ms │                      46.86 ms │     no change │
│ QQuery 9  │ 65.70 ms │                      65.42 ms │     no change │
│ QQuery 10 │ 46.58 ms │                      42.95 ms │ +1.08x faster │
│ QQuery 11 │ 15.77 ms │                      14.73 ms │ +1.07x faster │
│ QQuery 12 │ 40.91 ms │                      25.13 ms │ +1.63x faster │
│ QQuery 13 │ 41.22 ms │                      42.71 ms │     no change │
│ QQuery 14 │ 27.64 ms │                      25.11 ms │ +1.10x faster │
│ QQuery 15 │ 41.56 ms │                      36.21 ms │ +1.15x faster │
│ QQuery 16 │ 16.17 ms │                      15.49 ms │     no change │
│ QQuery 17 │ 74.32 ms │                      36.98 ms │ +2.01x faster │
│ QQuery 18 │ 65.41 ms │                      60.88 ms │ +1.07x faster │
│ QQuery 19 │ 32.88 ms │                      30.12 ms │ +1.09x faster │
│ QQuery 20 │ 31.30 ms │                      28.51 ms │ +1.10x faster │
│ QQuery 21 │ 69.43 ms │                      59.77 ms │ +1.16x faster │
│ QQuery 22 │ 16.76 ms │                      15.11 ms │ +1.11x faster │
└───────────┴──────────┴───────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ Benchmark Summary                            ┃          ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 865.04ms │
│ Total Time (hash-join-global-bloom-filter)   │ 765.52ms │
│ Average Time (HEAD)                          │  39.32ms │
│ Average Time (hash-join-global-bloom-filter) │  34.80ms │
│ Queries Faster                               │       14 │
│ Queries Slower                               │        0 │
│ Queries with No Change                       │        8 │
│ Queries with Failure                         │        0 │
└──────────────────────────────────────────────┴──────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                           HEAD ┃  hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │ 41.83 / 42.49 ±1.13 / 44.75 ms │ 42.55 / 44.03 ±1.47 / 46.12 ms │     no change │
│ QQuery 2  │ 19.31 / 20.13 ±0.64 / 20.96 ms │ 19.44 / 19.70 ±0.21 / 19.97 ms │     no change │
│ QQuery 3  │ 31.95 / 32.27 ±0.41 / 33.05 ms │ 31.03 / 31.33 ±0.55 / 32.43 ms │     no change │
│ QQuery 4  │ 25.75 / 25.83 ±0.08 / 25.97 ms │ 20.26 / 21.05 ±0.59 / 21.78 ms │ +1.23x faster │
│ QQuery 5  │ 43.89 / 44.49 ±0.90 / 46.28 ms │ 40.51 / 41.22 ±1.01 / 43.19 ms │ +1.08x faster │
│ QQuery 6  │ 22.88 / 23.18 ±0.28 / 23.64 ms │ 18.22 / 18.40 ±0.16 / 18.63 ms │ +1.26x faster │
│ QQuery 7  │ 46.20 / 46.93 ±0.53 / 47.56 ms │ 47.51 / 48.68 ±1.22 / 50.57 ms │     no change │
│ QQuery 8  │ 47.57 / 48.26 ±0.98 / 50.15 ms │ 46.86 / 47.45 ±0.32 / 47.79 ms │     no change │
│ QQuery 9  │ 65.70 / 67.26 ±1.24 / 68.85 ms │ 65.42 / 66.02 ±0.44 / 66.75 ms │     no change │
│ QQuery 10 │ 46.58 / 47.55 ±0.72 / 48.39 ms │ 42.95 / 43.32 ±0.26 / 43.59 ms │ +1.10x faster │
│ QQuery 11 │ 15.77 / 16.68 ±1.70 / 20.07 ms │ 14.73 / 15.03 ±0.24 / 15.45 ms │ +1.11x faster │
│ QQuery 12 │ 40.91 / 41.96 ±0.77 / 43.09 ms │ 25.13 / 26.25 ±1.42 / 28.90 ms │ +1.60x faster │
│ QQuery 13 │ 41.22 / 41.31 ±0.08 / 41.44 ms │ 42.71 / 43.44 ±0.55 / 44.42 ms │  1.05x slower │
│ QQuery 14 │ 27.64 / 28.11 ±0.34 / 28.64 ms │ 25.11 / 25.49 ±0.20 / 25.72 ms │ +1.10x faster │
│ QQuery 15 │ 41.56 / 43.54 ±2.07 / 47.54 ms │ 36.21 / 37.26 ±1.17 / 38.98 ms │ +1.17x faster │
│ QQuery 16 │ 16.17 / 16.39 ±0.14 / 16.56 ms │ 15.49 / 15.77 ±0.26 / 16.17 ms │     no change │
│ QQuery 17 │ 74.32 / 76.04 ±1.53 / 78.77 ms │ 36.98 / 38.05 ±1.18 / 40.14 ms │ +2.00x faster │
│ QQuery 18 │ 65.41 / 65.79 ±0.46 / 66.68 ms │ 60.88 / 63.17 ±2.52 / 67.99 ms │     no change │
│ QQuery 19 │ 32.88 / 33.28 ±0.25 / 33.64 ms │ 30.12 / 30.27 ±0.12 / 30.44 ms │ +1.10x faster │
│ QQuery 20 │ 31.30 / 31.64 ±0.25 / 31.97 ms │ 28.51 / 28.83 ±0.21 / 29.05 ms │ +1.10x faster │
│ QQuery 21 │ 69.43 / 70.39 ±0.88 / 71.73 ms │ 59.77 / 60.33 ±0.33 / 60.70 ms │ +1.17x faster │
│ QQuery 22 │ 16.76 / 17.10 ±0.21 / 17.37 ms │ 15.11 / 15.34 ±0.12 / 15.47 ms │ +1.11x faster │
└───────────┴────────────────────────────────┴────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ Benchmark Summary                            ┃          ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 880.59ms │
│ Total Time (hash-join-global-bloom-filter)   │ 780.43ms │
│ Average Time (HEAD)                          │  40.03ms │
│ Average Time (hash-join-global-bloom-filter) │  35.47ms │
│ Queries Faster                               │       13 │
│ Queries Slower                               │        1 │
│ Queries with No Change                       │        8 │
│ Queries with Failure                         │        0 │
└──────────────────────────────────────────────┴──────────┘

Resource Usage

tpch — base (merge-base)

Metric Value
Wall time 5.0s
Peak memory 1.2 GiB
Avg memory 735.3 MiB
CPU user 22.7s
CPU sys 1.8s
Peak spill 0 B

tpch — branch

Metric Value
Wall time 5.0s
Peak memory 1.0 GiB
Avg memory 516.5 MiB
CPU user 20.4s
CPU sys 2.0s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark tpcds
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ hash-join-global-bloom-filter ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │    5.17 ms │                       5.18 ms │    no change │
│ QQuery 2  │   80.55 ms │                      79.19 ms │    no change │
│ QQuery 3  │   27.83 ms │                      28.09 ms │    no change │
│ QQuery 4  │  418.06 ms │                     423.12 ms │    no change │
│ QQuery 5  │   48.55 ms │                      49.11 ms │    no change │
│ QQuery 6  │   34.22 ms │                      34.16 ms │    no change │
│ QQuery 7  │   70.32 ms │                      71.55 ms │    no change │
│ QQuery 8  │   34.99 ms │                      35.20 ms │    no change │
│ QQuery 9  │   51.87 ms │                      50.94 ms │    no change │
│ QQuery 10 │   59.19 ms │                      60.61 ms │    no change │
│ QQuery 11 │  252.56 ms │                     251.26 ms │    no change │
│ QQuery 12 │   28.19 ms │                      27.67 ms │    no change │
│ QQuery 13 │  111.80 ms │                     113.11 ms │    no change │
│ QQuery 14 │  380.83 ms │                     369.35 ms │    no change │
│ QQuery 15 │   51.44 ms │                      51.83 ms │    no change │
│ QQuery 16 │    5.80 ms │                       5.80 ms │    no change │
│ QQuery 17 │   77.29 ms │                      75.93 ms │    no change │
│ QQuery 18 │  102.00 ms │                     102.66 ms │    no change │
│ QQuery 19 │   40.08 ms │                      39.72 ms │    no change │
│ QQuery 20 │   35.09 ms │                      34.91 ms │    no change │
│ QQuery 21 │   16.26 ms │                      16.39 ms │    no change │
│ QQuery 22 │   62.59 ms │                      62.77 ms │    no change │
│ QQuery 23 │  298.77 ms │                     288.03 ms │    no change │
│ QQuery 24 │  186.84 ms │                     186.65 ms │    no change │
│ QQuery 25 │  106.19 ms │                     104.48 ms │    no change │
│ QQuery 26 │   44.09 ms │                      44.87 ms │    no change │
│ QQuery 27 │    5.78 ms │                       5.59 ms │    no change │
│ QQuery 28 │   55.42 ms │                      54.74 ms │    no change │
│ QQuery 29 │   93.50 ms │                      91.75 ms │    no change │
│ QQuery 30 │   30.89 ms │                      31.15 ms │    no change │
│ QQuery 31 │  105.51 ms │                     106.00 ms │    no change │
│ QQuery 32 │   19.27 ms │                      19.36 ms │    no change │
│ QQuery 33 │   35.42 ms │                      36.04 ms │    no change │
│ QQuery 34 │    9.45 ms │                       9.95 ms │ 1.05x slower │
│ QQuery 35 │   69.36 ms │                      70.52 ms │    no change │
│ QQuery 36 │    5.37 ms │                       5.63 ms │    no change │
│ QQuery 37 │    6.37 ms │                       6.60 ms │    no change │
│ QQuery 38 │   59.36 ms │                      59.68 ms │    no change │
│ QQuery 39 │   70.97 ms │                      71.11 ms │    no change │
│ QQuery 40 │   22.15 ms │                      21.97 ms │    no change │
│ QQuery 41 │   11.74 ms │                      11.77 ms │    no change │
│ QQuery 42 │   22.88 ms │                      23.18 ms │    no change │
│ QQuery 43 │    4.59 ms │                       4.57 ms │    no change │
│ QQuery 44 │    8.21 ms │                       8.26 ms │    no change │
│ QQuery 45 │   34.88 ms │                      35.52 ms │    no change │
│ QQuery 46 │   11.18 ms │                      11.29 ms │    no change │
│ QQuery 47 │  212.71 ms │                     214.87 ms │    no change │
│ QQuery 48 │   89.86 ms │                      90.53 ms │    no change │
│ QQuery 49 │   68.00 ms │                      68.72 ms │    no change │
│ QQuery 50 │   56.67 ms │                      56.22 ms │    no change │
│ QQuery 51 │   91.14 ms │                      90.73 ms │    no change │
│ QQuery 52 │   23.15 ms │                      23.27 ms │    no change │
│ QQuery 53 │   28.35 ms │                      28.06 ms │    no change │
│ QQuery 54 │   51.60 ms │                      52.17 ms │    no change │
│ QQuery 55 │   22.88 ms │                      22.75 ms │    no change │
│ QQuery 56 │   36.63 ms │                      36.68 ms │    no change │
│ QQuery 57 │  164.32 ms │                     166.04 ms │    no change │
│ QQuery 58 │  106.09 ms │                     105.76 ms │    no change │
│ QQuery 59 │  114.13 ms │                     114.90 ms │    no change │
│ QQuery 60 │   36.79 ms │                      37.06 ms │    no change │
│ QQuery 61 │   10.11 ms │                      10.20 ms │    no change │
│ QQuery 62 │   42.84 ms │                      42.81 ms │    no change │
│ QQuery 63 │   28.24 ms │                      28.55 ms │    no change │
│ QQuery 64 │  348.99 ms │                     353.90 ms │    no change │
│ QQuery 65 │  132.27 ms │                     129.93 ms │    no change │
│ QQuery 66 │   74.62 ms │                      74.19 ms │    no change │
│ QQuery 67 │  237.62 ms │                     235.76 ms │    no change │
│ QQuery 68 │   11.39 ms │                      11.80 ms │    no change │
│ QQuery 69 │   54.24 ms │                      55.67 ms │    no change │
│ QQuery 70 │  102.00 ms │                     102.11 ms │    no change │
│ QQuery 71 │   34.31 ms │                      34.12 ms │    no change │
│ QQuery 72 │ 1593.25 ms │                    1629.90 ms │    no change │
│ QQuery 73 │   10.07 ms │                       9.70 ms │    no change │
│ QQuery 74 │  149.22 ms │                     149.60 ms │    no change │
│ QQuery 75 │  137.54 ms │                     136.63 ms │    no change │
│ QQuery 76 │   33.28 ms │                      33.38 ms │    no change │
│ QQuery 77 │   57.99 ms │                      58.08 ms │    no change │
│ QQuery 78 │  158.71 ms │                     157.37 ms │    no change │
│ QQuery 79 │   65.31 ms │                      64.78 ms │    no change │
│ QQuery 80 │   93.09 ms │                      93.72 ms │    no change │
│ QQuery 81 │   24.52 ms │                      24.55 ms │    no change │
│ QQuery 82 │   15.82 ms │                      15.77 ms │    no change │
│ QQuery 83 │   31.24 ms │                      31.58 ms │    no change │
│ QQuery 84 │   28.65 ms │                      28.89 ms │    no change │
│ QQuery 85 │   96.97 ms │                      95.84 ms │    no change │
│ QQuery 86 │   24.39 ms │                      24.43 ms │    no change │
│ QQuery 87 │   59.14 ms │                      59.25 ms │    no change │
│ QQuery 88 │   57.01 ms │                      57.02 ms │    no change │
│ QQuery 89 │   34.07 ms │                      34.54 ms │    no change │
│ QQuery 90 │   15.38 ms │                      15.72 ms │    no change │
│ QQuery 91 │   39.62 ms │                      40.07 ms │    no change │
│ QQuery 92 │   27.28 ms │                      28.18 ms │    no change │
│ QQuery 93 │   48.94 ms │                      49.40 ms │    no change │
│ QQuery 94 │   36.49 ms │                      36.89 ms │    no change │
│ QQuery 95 │   77.68 ms │                      77.10 ms │    no change │
│ QQuery 96 │   23.00 ms │                      22.98 ms │    no change │
│ QQuery 97 │   50.54 ms │                      49.96 ms │    no change │
│ QQuery 98 │   42.42 ms │                      41.99 ms │    no change │
│ QQuery 99 │   63.87 ms │                      64.27 ms │    no change │
└───────────┴────────────┴───────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 8587.33ms │
│ Total Time (hash-join-global-bloom-filter)   │ 8615.66ms │
│ Average Time (HEAD)                          │   86.74ms │
│ Average Time (hash-join-global-bloom-filter) │   87.03ms │
│ Queries Faster                               │         0 │
│ Queries Slower                               │         1 │
│ Queries with No Change                       │        98 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃         hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │           5.17 / 5.75 ±1.10 / 7.95 ms │           5.18 / 5.76 ±1.03 / 7.82 ms │     no change │
│ QQuery 2  │        80.55 / 80.91 ±0.38 / 81.65 ms │        79.19 / 79.48 ±0.18 / 79.68 ms │     no change │
│ QQuery 3  │        27.83 / 28.22 ±0.26 / 28.52 ms │        28.09 / 28.44 ±0.27 / 28.75 ms │     no change │
│ QQuery 4  │     418.06 / 423.00 ±4.49 / 429.92 ms │     423.12 / 428.17 ±3.04 / 432.27 ms │     no change │
│ QQuery 5  │        48.55 / 49.09 ±0.36 / 49.61 ms │        49.11 / 49.79 ±0.70 / 51.00 ms │     no change │
│ QQuery 6  │        34.22 / 36.85 ±3.75 / 44.24 ms │        34.16 / 36.72 ±3.35 / 43.26 ms │     no change │
│ QQuery 7  │        70.32 / 71.44 ±0.77 / 72.74 ms │        71.55 / 72.31 ±0.60 / 72.91 ms │     no change │
│ QQuery 8  │        34.99 / 35.23 ±0.21 / 35.53 ms │        35.20 / 35.47 ±0.20 / 35.68 ms │     no change │
│ QQuery 9  │        51.87 / 53.50 ±1.88 / 56.69 ms │        50.94 / 53.27 ±1.87 / 56.23 ms │     no change │
│ QQuery 10 │        59.19 / 60.02 ±0.67 / 61.24 ms │        60.61 / 62.85 ±3.40 / 69.61 ms │     no change │
│ QQuery 11 │     252.56 / 256.88 ±2.63 / 260.19 ms │     251.26 / 257.40 ±5.51 / 267.29 ms │     no change │
│ QQuery 12 │        28.19 / 28.51 ±0.30 / 28.94 ms │        27.67 / 28.14 ±0.30 / 28.60 ms │     no change │
│ QQuery 13 │     111.80 / 113.95 ±3.46 / 120.75 ms │     113.11 / 114.39 ±0.96 / 115.89 ms │     no change │
│ QQuery 14 │     380.83 / 385.35 ±5.00 / 394.96 ms │     369.35 / 372.03 ±2.50 / 375.88 ms │     no change │
│ QQuery 15 │        51.44 / 52.68 ±1.08 / 54.64 ms │        51.83 / 53.37 ±1.12 / 55.23 ms │     no change │
│ QQuery 16 │           5.80 / 5.95 ±0.15 / 6.24 ms │           5.80 / 5.90 ±0.10 / 6.08 ms │     no change │
│ QQuery 17 │        77.29 / 77.80 ±0.35 / 78.35 ms │        75.93 / 76.27 ±0.34 / 76.86 ms │     no change │
│ QQuery 18 │     102.00 / 102.80 ±0.63 / 103.87 ms │     102.66 / 103.19 ±0.67 / 104.47 ms │     no change │
│ QQuery 19 │        40.08 / 40.35 ±0.22 / 40.63 ms │        39.72 / 40.91 ±1.49 / 43.80 ms │     no change │
│ QQuery 20 │        35.09 / 35.52 ±0.43 / 36.22 ms │        34.91 / 35.27 ±0.20 / 35.47 ms │     no change │
│ QQuery 21 │        16.26 / 16.62 ±0.31 / 17.05 ms │        16.39 / 16.71 ±0.20 / 16.94 ms │     no change │
│ QQuery 22 │        62.59 / 63.90 ±0.89 / 65.30 ms │        62.77 / 64.11 ±1.15 / 65.74 ms │     no change │
│ QQuery 23 │     298.77 / 305.78 ±5.22 / 313.77 ms │     288.03 / 290.10 ±1.47 / 292.48 ms │ +1.05x faster │
│ QQuery 24 │     186.84 / 190.28 ±3.33 / 195.90 ms │     186.65 / 189.43 ±3.34 / 195.16 ms │     no change │
│ QQuery 25 │     106.19 / 108.91 ±2.55 / 112.25 ms │     104.48 / 105.45 ±0.86 / 106.80 ms │     no change │
│ QQuery 26 │        44.09 / 46.01 ±2.66 / 51.27 ms │        44.87 / 45.41 ±0.39 / 46.04 ms │     no change │
│ QQuery 27 │           5.78 / 5.90 ±0.18 / 6.25 ms │           5.59 / 5.88 ±0.31 / 6.46 ms │     no change │
│ QQuery 28 │        55.42 / 58.59 ±2.51 / 62.05 ms │        54.74 / 58.29 ±2.26 / 60.21 ms │     no change │
│ QQuery 29 │        93.50 / 95.16 ±1.23 / 96.66 ms │        91.75 / 93.03 ±1.22 / 94.65 ms │     no change │
│ QQuery 30 │        30.89 / 31.45 ±0.55 / 32.23 ms │        31.15 / 31.57 ±0.28 / 31.94 ms │     no change │
│ QQuery 31 │     105.51 / 106.50 ±1.06 / 108.52 ms │     106.00 / 107.62 ±2.14 / 111.79 ms │     no change │
│ QQuery 32 │        19.27 / 19.64 ±0.44 / 20.45 ms │        19.36 / 19.78 ±0.44 / 20.63 ms │     no change │
│ QQuery 33 │        35.42 / 35.98 ±0.37 / 36.48 ms │        36.04 / 36.63 ±0.37 / 37.09 ms │     no change │
│ QQuery 34 │         9.45 / 10.09 ±0.34 / 10.39 ms │         9.95 / 10.52 ±0.52 / 11.43 ms │     no change │
│ QQuery 35 │        69.36 / 71.56 ±2.54 / 76.51 ms │        70.52 / 72.76 ±3.01 / 78.71 ms │     no change │
│ QQuery 36 │           5.37 / 5.48 ±0.16 / 5.79 ms │           5.63 / 5.77 ±0.16 / 6.05 ms │  1.05x slower │
│ QQuery 37 │           6.37 / 6.68 ±0.18 / 6.86 ms │           6.60 / 6.82 ±0.14 / 6.99 ms │     no change │
│ QQuery 38 │        59.36 / 59.92 ±0.50 / 60.66 ms │        59.68 / 60.61 ±0.57 / 61.27 ms │     no change │
│ QQuery 39 │        70.97 / 71.72 ±1.28 / 74.27 ms │        71.11 / 71.46 ±0.45 / 72.30 ms │     no change │
│ QQuery 40 │        22.15 / 23.54 ±1.85 / 27.07 ms │        21.97 / 22.64 ±0.44 / 23.18 ms │     no change │
│ QQuery 41 │        11.74 / 11.92 ±0.17 / 12.21 ms │        11.77 / 11.91 ±0.14 / 12.19 ms │     no change │
│ QQuery 42 │        22.88 / 23.53 ±0.65 / 24.69 ms │        23.18 / 24.81 ±1.94 / 28.50 ms │  1.05x slower │
│ QQuery 43 │           4.59 / 4.68 ±0.13 / 4.93 ms │           4.57 / 4.71 ±0.17 / 5.03 ms │     no change │
│ QQuery 44 │           8.21 / 8.34 ±0.12 / 8.56 ms │           8.26 / 8.36 ±0.10 / 8.55 ms │     no change │
│ QQuery 45 │        34.88 / 35.40 ±0.31 / 35.79 ms │        35.52 / 36.17 ±0.44 / 36.77 ms │     no change │
│ QQuery 46 │        11.18 / 11.38 ±0.19 / 11.63 ms │        11.29 / 11.57 ±0.25 / 11.90 ms │     no change │
│ QQuery 47 │     212.71 / 219.42 ±4.97 / 227.49 ms │     214.87 / 215.72 ±0.82 / 216.83 ms │     no change │
│ QQuery 48 │        89.86 / 91.66 ±1.83 / 94.06 ms │        90.53 / 91.30 ±0.95 / 93.16 ms │     no change │
│ QQuery 49 │        68.00 / 68.84 ±0.99 / 70.74 ms │        68.72 / 71.29 ±4.25 / 79.77 ms │     no change │
│ QQuery 50 │        56.67 / 57.27 ±0.44 / 58.03 ms │        56.22 / 56.80 ±0.44 / 57.27 ms │     no change │
│ QQuery 51 │       91.14 / 95.28 ±3.60 / 100.71 ms │       90.73 / 94.01 ±4.37 / 102.65 ms │     no change │
│ QQuery 52 │        23.15 / 23.71 ±0.30 / 24.03 ms │        23.27 / 23.82 ±0.54 / 24.80 ms │     no change │
│ QQuery 53 │        28.35 / 28.44 ±0.07 / 28.53 ms │        28.06 / 28.39 ±0.20 / 28.64 ms │     no change │
│ QQuery 54 │        51.60 / 51.91 ±0.21 / 52.17 ms │        52.17 / 52.76 ±0.41 / 53.31 ms │     no change │
│ QQuery 55 │        22.88 / 24.08 ±1.80 / 27.65 ms │        22.75 / 22.97 ±0.18 / 23.19 ms │     no change │
│ QQuery 56 │        36.63 / 37.15 ±0.49 / 37.99 ms │        36.68 / 37.49 ±1.24 / 39.94 ms │     no change │
│ QQuery 57 │     164.32 / 168.48 ±3.72 / 175.21 ms │     166.04 / 167.33 ±2.07 / 171.44 ms │     no change │
│ QQuery 58 │     106.09 / 107.78 ±1.24 / 109.29 ms │     105.76 / 108.29 ±2.76 / 113.23 ms │     no change │
│ QQuery 59 │     114.13 / 115.26 ±0.99 / 117.03 ms │     114.90 / 115.32 ±0.65 / 116.62 ms │     no change │
│ QQuery 60 │        36.79 / 38.32 ±1.40 / 40.89 ms │        37.06 / 37.46 ±0.29 / 37.89 ms │     no change │
│ QQuery 61 │        10.11 / 10.29 ±0.19 / 10.55 ms │        10.20 / 11.36 ±2.04 / 15.44 ms │  1.10x slower │
│ QQuery 62 │        42.84 / 43.22 ±0.28 / 43.61 ms │        42.81 / 43.53 ±0.40 / 43.98 ms │     no change │
│ QQuery 63 │        28.24 / 28.67 ±0.24 / 28.98 ms │        28.55 / 29.12 ±0.50 / 30.02 ms │     no change │
│ QQuery 64 │     348.99 / 352.34 ±2.73 / 355.96 ms │     353.90 / 357.13 ±2.95 / 362.09 ms │     no change │
│ QQuery 65 │     132.27 / 135.56 ±2.21 / 138.73 ms │     129.93 / 131.24 ±1.50 / 133.66 ms │     no change │
│ QQuery 66 │        74.62 / 76.57 ±3.12 / 82.78 ms │        74.19 / 76.40 ±2.88 / 82.08 ms │     no change │
│ QQuery 67 │     237.62 / 240.83 ±3.78 / 248.05 ms │     235.76 / 240.10 ±3.95 / 246.45 ms │     no change │
│ QQuery 68 │        11.39 / 11.60 ±0.22 / 12.03 ms │        11.80 / 12.03 ±0.17 / 12.31 ms │     no change │
│ QQuery 69 │        54.24 / 55.74 ±1.64 / 58.90 ms │        55.67 / 57.47 ±2.27 / 61.90 ms │     no change │
│ QQuery 70 │     102.00 / 103.68 ±1.41 / 106.24 ms │     102.11 / 106.88 ±4.86 / 115.39 ms │     no change │
│ QQuery 71 │        34.31 / 34.38 ±0.10 / 34.58 ms │        34.12 / 34.77 ±1.02 / 36.81 ms │     no change │
│ QQuery 72 │ 1593.25 / 1691.26 ±57.54 / 1757.44 ms │ 1629.90 / 1697.45 ±45.98 / 1747.21 ms │     no change │
│ QQuery 73 │        10.07 / 11.44 ±2.45 / 16.34 ms │         9.70 / 10.39 ±0.38 / 10.85 ms │ +1.10x faster │
│ QQuery 74 │     149.22 / 152.59 ±2.72 / 157.30 ms │     149.60 / 151.40 ±1.20 / 152.85 ms │     no change │
│ QQuery 75 │     137.54 / 140.16 ±3.58 / 147.06 ms │     136.63 / 137.77 ±1.00 / 139.40 ms │     no change │
│ QQuery 76 │        33.28 / 33.61 ±0.28 / 34.10 ms │        33.38 / 33.87 ±0.38 / 34.31 ms │     no change │
│ QQuery 77 │        57.99 / 59.88 ±1.85 / 63.31 ms │        58.08 / 58.36 ±0.18 / 58.61 ms │     no change │
│ QQuery 78 │     158.71 / 162.49 ±4.04 / 169.47 ms │     157.37 / 162.19 ±4.89 / 168.29 ms │     no change │
│ QQuery 79 │        65.31 / 65.93 ±0.51 / 66.86 ms │        64.78 / 65.35 ±0.40 / 65.81 ms │     no change │
│ QQuery 80 │        93.09 / 94.92 ±1.15 / 96.49 ms │        93.72 / 95.35 ±1.85 / 98.76 ms │     no change │
│ QQuery 81 │        24.52 / 24.84 ±0.28 / 25.29 ms │        24.55 / 24.96 ±0.32 / 25.39 ms │     no change │
│ QQuery 82 │        15.82 / 16.49 ±1.00 / 18.47 ms │        15.77 / 17.02 ±1.81 / 20.63 ms │     no change │
│ QQuery 83 │        31.24 / 31.58 ±0.31 / 32.10 ms │        31.58 / 32.01 ±0.30 / 32.49 ms │     no change │
│ QQuery 84 │        28.65 / 30.41 ±2.39 / 35.09 ms │        28.89 / 28.98 ±0.07 / 29.11 ms │     no change │
│ QQuery 85 │        96.97 / 97.83 ±0.86 / 99.47 ms │       95.84 / 99.16 ±2.70 / 103.57 ms │     no change │
│ QQuery 86 │        24.39 / 24.70 ±0.35 / 25.34 ms │        24.43 / 24.84 ±0.47 / 25.72 ms │     no change │
│ QQuery 87 │        59.14 / 59.90 ±0.57 / 60.63 ms │        59.25 / 59.87 ±0.47 / 60.43 ms │     no change │
│ QQuery 88 │        57.01 / 58.05 ±1.54 / 61.08 ms │        57.02 / 58.01 ±1.41 / 60.77 ms │     no change │
│ QQuery 89 │        34.07 / 34.63 ±0.44 / 35.08 ms │        34.54 / 35.28 ±0.41 / 35.79 ms │     no change │
│ QQuery 90 │        15.38 / 15.82 ±0.31 / 16.28 ms │        15.72 / 16.07 ±0.28 / 16.46 ms │     no change │
│ QQuery 91 │        39.62 / 40.43 ±1.08 / 42.53 ms │        40.07 / 40.81 ±0.48 / 41.32 ms │     no change │
│ QQuery 92 │        27.28 / 28.25 ±0.95 / 29.94 ms │        28.18 / 29.04 ±1.37 / 31.77 ms │     no change │
│ QQuery 93 │        48.94 / 50.59 ±2.16 / 54.81 ms │        49.40 / 51.14 ±2.28 / 55.60 ms │     no change │
│ QQuery 94 │        36.49 / 37.24 ±0.48 / 37.97 ms │        36.89 / 37.59 ±0.67 / 38.68 ms │     no change │
│ QQuery 95 │        77.68 / 78.57 ±0.65 / 79.55 ms │        77.10 / 78.40 ±0.97 / 80.03 ms │     no change │
│ QQuery 96 │        23.00 / 23.50 ±0.81 / 25.12 ms │        22.98 / 23.17 ±0.23 / 23.57 ms │     no change │
│ QQuery 97 │        50.54 / 52.20 ±1.34 / 54.24 ms │        49.96 / 50.65 ±1.09 / 52.82 ms │     no change │
│ QQuery 98 │        42.42 / 42.81 ±0.35 / 43.36 ms │        41.99 / 42.79 ±0.68 / 43.91 ms │     no change │
│ QQuery 99 │        63.87 / 64.71 ±0.56 / 65.34 ms │        64.27 / 64.85 ±0.32 / 65.23 ms │     no change │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 8818.08ms │
│ Total Time (hash-join-global-bloom-filter)   │ 8800.97ms │
│ Average Time (HEAD)                          │   89.07ms │
│ Average Time (hash-join-global-bloom-filter) │   88.90ms │
│ Queries Faster                               │         2 │
│ Queries Slower                               │         3 │
│ Queries with No Change                       │        94 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Resource Usage

tpcds — base (merge-base)

Metric Value
Wall time 45.0s
Peak memory 2.1 GiB
Avg memory 1.5 GiB
CPU user 188.9s
CPU sys 5.7s
Peak spill 0 B

tpcds — branch

Metric Value
Wall time 45.0s
Peak memory 2.2 GiB
Avg memory 1.5 GiB
CPU user 188.8s
CPU sys 5.3s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark tpcds
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃      HEAD ┃ hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │   5.59 ms │                       5.50 ms │     no change │
│ QQuery 2  │  48.93 ms │                      48.44 ms │     no change │
│ QQuery 3  │  28.34 ms │                      25.78 ms │ +1.10x faster │
│ QQuery 4  │ 396.51 ms │                     383.86 ms │     no change │
│ QQuery 5  │  49.31 ms │                      48.59 ms │     no change │
│ QQuery 6  │  38.85 ms │                      39.15 ms │     no change │
│ QQuery 7  │  88.92 ms │                      60.31 ms │ +1.47x faster │
│ QQuery 8  │  15.96 ms │                      16.24 ms │     no change │
│ QQuery 9  │ 117.95 ms │                      56.33 ms │ +2.09x faster │
│ QQuery 10 │  77.04 ms │                      83.52 ms │  1.08x slower │
│ QQuery 11 │ 243.98 ms │                     241.81 ms │     no change │
│ QQuery 12 │  30.46 ms │                      26.30 ms │ +1.16x faster │
│ QQuery 13 │ 119.02 ms │                     109.09 ms │ +1.09x faster │
│ QQuery 14 │ 449.74 ms │                     444.08 ms │     no change │
│ QQuery 15 │  25.24 ms │                      27.25 ms │  1.08x slower │
│ QQuery 16 │   6.71 ms │                       6.67 ms │     no change │
│ QQuery 17 │ 121.26 ms │                      77.47 ms │ +1.57x faster │
│ QQuery 18 │ 170.22 ms │                      87.77 ms │ +1.94x faster │
│ QQuery 19 │  40.61 ms │                      40.56 ms │     no change │
│ QQuery 20 │  32.97 ms │                      34.04 ms │     no change │
│ QQuery 21 │  20.13 ms │                      19.39 ms │     no change │
│ QQuery 22 │  70.46 ms │                      70.89 ms │     no change │
│ QQuery 23 │ 416.41 ms │                     362.46 ms │ +1.15x faster │
│ QQuery 24 │ 188.99 ms │                     166.82 ms │ +1.13x faster │
│ QQuery 25 │ 120.73 ms │                      83.56 ms │ +1.44x faster │
│ QQuery 26 │  68.49 ms │                      42.68 ms │ +1.60x faster │
│ QQuery 27 │   6.39 ms │                       6.35 ms │     no change │
│ QQuery 28 │  77.42 ms │                      62.32 ms │ +1.24x faster │
│ QQuery 29 │ 154.88 ms │                      96.94 ms │ +1.60x faster │
│ QQuery 30 │  34.46 ms │                      33.02 ms │     no change │
│ QQuery 31 │ 130.73 ms │                     128.17 ms │     no change │
│ QQuery 32 │  22.83 ms │                      21.51 ms │ +1.06x faster │
│ QQuery 33 │  40.53 ms │                      40.69 ms │     no change │
│ QQuery 34 │  11.43 ms │                      11.99 ms │     no change │
│ QQuery 35 │  92.00 ms │                      80.57 ms │ +1.14x faster │
│ QQuery 36 │   6.42 ms │                       6.48 ms │     no change │
│ QQuery 37 │   7.65 ms │                       6.85 ms │ +1.12x faster │
│ QQuery 38 │  83.56 ms │                      85.44 ms │     no change │
│ QQuery 39 │  77.97 ms │                      77.71 ms │     no change │
│ QQuery 40 │  22.07 ms │                      20.61 ms │ +1.07x faster │
│ QQuery 41 │  14.60 ms │                      13.70 ms │ +1.07x faster │
│ QQuery 42 │  25.55 ms │                      24.18 ms │ +1.06x faster │
│ QQuery 43 │   5.21 ms │                       4.98 ms │     no change │
│ QQuery 44 │   9.40 ms │                       8.99 ms │     no change │
│ QQuery 45 │  29.05 ms │                      31.92 ms │  1.10x slower │
│ QQuery 46 │  13.40 ms │                      13.33 ms │     no change │
│ QQuery 47 │ 264.14 ms │                     263.46 ms │     no change │
│ QQuery 48 │ 125.90 ms │                     107.72 ms │ +1.17x faster │
│ QQuery 49 │  92.55 ms │                      87.35 ms │ +1.06x faster │
│ QQuery 50 │ 129.55 ms │                      58.27 ms │ +2.22x faster │
│ QQuery 51 │ 108.34 ms │                     106.88 ms │     no change │
│ QQuery 52 │  25.73 ms │                      25.34 ms │     no change │
│ QQuery 53 │  32.20 ms │                      32.82 ms │     no change │
│ QQuery 54 │  30.83 ms │                      28.88 ms │ +1.07x faster │
│ QQuery 55 │  25.25 ms │                      24.35 ms │     no change │
│ QQuery 56 │  36.10 ms │                      35.58 ms │     no change │
│ QQuery 57 │ 142.72 ms │                     149.20 ms │     no change │
│ QQuery 58 │  77.38 ms │                      66.66 ms │ +1.16x faster │
│ QQuery 59 │  83.66 ms │                      81.73 ms │     no change │
│ QQuery 60 │  40.33 ms │                      38.92 ms │     no change │
│ QQuery 61 │  12.13 ms │                      10.94 ms │ +1.11x faster │
│ QQuery 62 │  45.59 ms │                      43.98 ms │     no change │
│ QQuery 63 │  32.53 ms │                      32.63 ms │     no change │
│ QQuery 64 │ 545.35 ms │                     311.09 ms │ +1.75x faster │
│ QQuery 65 │ 165.03 ms │                     155.61 ms │ +1.06x faster │
│ QQuery 66 │  67.25 ms │                      66.14 ms │     no change │
│ QQuery 67 │ 289.93 ms │                     291.15 ms │     no change │
│ QQuery 68 │  13.26 ms │                      12.87 ms │     no change │
│ QQuery 69 │  75.65 ms │                      75.60 ms │     no change │
│ QQuery 70 │ 125.00 ms │                     122.23 ms │     no change │
│ QQuery 71 │  34.42 ms │                      34.13 ms │     no change │
│ QQuery 72 │ 182.76 ms │                     108.03 ms │ +1.69x faster │
│ QQuery 73 │  10.83 ms │                      11.28 ms │     no change │
│ QQuery 74 │ 160.09 ms │                     161.15 ms │     no change │
│ QQuery 75 │ 189.99 ms │                     127.22 ms │ +1.49x faster │
│ QQuery 76 │  42.67 ms │                      38.64 ms │ +1.10x faster │
│ QQuery 77 │  63.71 ms │                      61.58 ms │     no change │
│ QQuery 78 │ 127.61 ms │                     122.34 ms │     no change │
│ QQuery 79 │  70.91 ms │                      70.32 ms │     no change │
│ QQuery 80 │  78.82 ms │                      77.90 ms │     no change │
│ QQuery 81 │  28.68 ms │                      28.20 ms │     no change │
│ QQuery 82 │  33.09 ms │                      25.77 ms │ +1.28x faster │
│ QQuery 83 │  30.25 ms │                      33.25 ms │  1.10x slower │
│ QQuery 84 │  31.91 ms │                      33.69 ms │  1.06x slower │
│ QQuery 85 │ 168.49 ms │                     121.40 ms │ +1.39x faster │
│ QQuery 86 │  31.74 ms │                      32.93 ms │     no change │
│ QQuery 87 │  83.34 ms │                      86.60 ms │     no change │
│ QQuery 88 │  67.75 ms │                      66.45 ms │     no change │
│ QQuery 89 │  44.79 ms │                      40.30 ms │ +1.11x faster │
│ QQuery 90 │  18.64 ms │                      18.81 ms │     no change │
│ QQuery 91 │  44.62 ms │                      44.45 ms │     no change │
│ QQuery 92 │  31.44 ms │                      30.56 ms │     no change │
│ QQuery 93 │  53.17 ms │                      50.38 ms │ +1.06x faster │
│ QQuery 94 │  37.09 ms │                      38.25 ms │     no change │
│ QQuery 95 │  95.41 ms │                      60.03 ms │ +1.59x faster │
│ QQuery 96 │  28.02 ms │                      25.98 ms │ +1.08x faster │
│ QQuery 97 │  56.13 ms │                      57.08 ms │     no change │
│ QQuery 98 │  39.80 ms │                      37.29 ms │ +1.07x faster │
│ QQuery 99 │  60.87 ms │                      60.59 ms │     no change │
└───────────┴───────────┴───────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 8391.80ms │
│ Total Time (hash-join-global-bloom-filter)   │ 7318.33ms │
│ Average Time (HEAD)                          │   84.77ms │
│ Average Time (hash-join-global-bloom-filter) │   73.92ms │
│ Queries Faster                               │        37 │
│ Queries Slower                               │         5 │
│ Queries with No Change                       │        57 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                               HEAD ┃      hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │        5.59 / 6.12 ±1.00 / 8.13 ms │        5.50 / 6.02 ±0.94 / 7.89 ms │     no change │
│ QQuery 2  │     48.93 / 49.32 ±0.44 / 50.08 ms │     48.44 / 48.88 ±0.39 / 49.58 ms │     no change │
│ QQuery 3  │     28.34 / 28.57 ±0.21 / 28.91 ms │     25.78 / 26.18 ±0.36 / 26.83 ms │ +1.09x faster │
│ QQuery 4  │  396.51 / 402.76 ±4.13 / 409.12 ms │  383.86 / 392.81 ±7.00 / 403.51 ms │     no change │
│ QQuery 5  │     49.31 / 49.70 ±0.23 / 49.98 ms │     48.59 / 48.91 ±0.27 / 49.25 ms │     no change │
│ QQuery 6  │     38.85 / 39.09 ±0.18 / 39.31 ms │     39.15 / 39.64 ±0.33 / 40.08 ms │     no change │
│ QQuery 7  │     88.92 / 92.61 ±2.70 / 97.33 ms │     60.31 / 64.09 ±3.90 / 69.71 ms │ +1.44x faster │
│ QQuery 8  │     15.96 / 16.11 ±0.13 / 16.34 ms │     16.24 / 16.38 ±0.12 / 16.53 ms │     no change │
│ QQuery 9  │  117.95 / 123.21 ±4.27 / 127.94 ms │     56.33 / 59.06 ±3.04 / 64.27 ms │ +2.09x faster │
│ QQuery 10 │     77.04 / 79.07 ±1.95 / 82.53 ms │     83.52 / 85.07 ±2.14 / 89.22 ms │  1.08x slower │
│ QQuery 11 │  243.98 / 249.47 ±4.57 / 257.10 ms │  241.81 / 244.13 ±2.63 / 247.68 ms │     no change │
│ QQuery 12 │     30.46 / 30.94 ±0.49 / 31.83 ms │     26.30 / 26.49 ±0.13 / 26.69 ms │ +1.17x faster │
│ QQuery 13 │  119.02 / 123.65 ±5.91 / 135.29 ms │  109.09 / 109.92 ±0.53 / 110.57 ms │ +1.12x faster │
│ QQuery 14 │  449.74 / 456.99 ±5.23 / 463.96 ms │  444.08 / 447.44 ±3.86 / 454.05 ms │     no change │
│ QQuery 15 │     25.24 / 25.40 ±0.18 / 25.74 ms │     27.25 / 27.91 ±0.36 / 28.27 ms │  1.10x slower │
│ QQuery 16 │        6.71 / 6.80 ±0.07 / 6.91 ms │        6.67 / 6.77 ±0.09 / 6.93 ms │     no change │
│ QQuery 17 │  121.26 / 123.79 ±4.06 / 131.87 ms │     77.47 / 78.35 ±1.00 / 80.22 ms │ +1.58x faster │
│ QQuery 18 │  170.22 / 174.26 ±3.09 / 177.53 ms │     87.77 / 89.75 ±2.72 / 95.14 ms │ +1.94x faster │
│ QQuery 19 │     40.61 / 41.47 ±0.86 / 43.09 ms │     40.56 / 41.25 ±1.22 / 43.68 ms │     no change │
│ QQuery 20 │     32.97 / 34.07 ±0.72 / 34.84 ms │     34.04 / 34.72 ±0.35 / 35.04 ms │     no change │
│ QQuery 21 │     20.13 / 20.34 ±0.15 / 20.57 ms │     19.39 / 19.56 ±0.17 / 19.87 ms │     no change │
│ QQuery 22 │     70.46 / 72.87 ±3.12 / 79.01 ms │     70.89 / 71.57 ±0.60 / 72.49 ms │     no change │
│ QQuery 23 │  416.41 / 426.75 ±7.47 / 435.62 ms │  362.46 / 370.89 ±6.96 / 379.29 ms │ +1.15x faster │
│ QQuery 24 │  188.99 / 197.44 ±9.55 / 215.47 ms │  166.82 / 171.57 ±6.10 / 183.55 ms │ +1.15x faster │
│ QQuery 25 │  120.73 / 122.06 ±2.02 / 126.07 ms │     83.56 / 86.86 ±4.73 / 96.27 ms │ +1.41x faster │
│ QQuery 26 │     68.49 / 70.96 ±4.12 / 79.17 ms │     42.68 / 42.92 ±0.18 / 43.13 ms │ +1.65x faster │
│ QQuery 27 │        6.39 / 6.50 ±0.06 / 6.57 ms │        6.35 / 6.41 ±0.04 / 6.48 ms │     no change │
│ QQuery 28 │     77.42 / 77.94 ±0.48 / 78.58 ms │     62.32 / 63.48 ±1.12 / 65.62 ms │ +1.23x faster │
│ QQuery 29 │  154.88 / 160.09 ±7.46 / 174.76 ms │   96.94 / 101.68 ±7.12 / 115.83 ms │ +1.57x faster │
│ QQuery 30 │     34.46 / 34.97 ±0.44 / 35.53 ms │     33.02 / 33.62 ±0.41 / 34.27 ms │     no change │
│ QQuery 31 │  130.73 / 133.30 ±2.80 / 138.70 ms │  128.17 / 130.45 ±1.67 / 132.72 ms │     no change │
│ QQuery 32 │     22.83 / 23.25 ±0.33 / 23.60 ms │     21.51 / 23.60 ±2.93 / 29.38 ms │     no change │
│ QQuery 33 │     40.53 / 40.85 ±0.26 / 41.18 ms │     40.69 / 43.19 ±3.52 / 50.17 ms │  1.06x slower │
│ QQuery 34 │     11.43 / 11.57 ±0.11 / 11.68 ms │     11.99 / 12.27 ±0.49 / 13.26 ms │  1.06x slower │
│ QQuery 35 │    92.00 / 95.96 ±4.58 / 104.58 ms │     80.57 / 82.42 ±1.03 / 83.65 ms │ +1.16x faster │
│ QQuery 36 │        6.42 / 6.55 ±0.10 / 6.73 ms │        6.48 / 6.64 ±0.15 / 6.87 ms │     no change │
│ QQuery 37 │        7.65 / 7.78 ±0.07 / 7.86 ms │        6.85 / 6.90 ±0.06 / 7.01 ms │ +1.13x faster │
│ QQuery 38 │     83.56 / 84.90 ±1.17 / 86.96 ms │     85.44 / 88.60 ±4.07 / 96.46 ms │     no change │
│ QQuery 39 │     77.97 / 81.76 ±6.27 / 94.17 ms │     77.71 / 77.95 ±0.26 / 78.33 ms │     no change │
│ QQuery 40 │     22.07 / 23.15 ±0.74 / 24.29 ms │     20.61 / 20.73 ±0.15 / 21.02 ms │ +1.12x faster │
│ QQuery 41 │     14.60 / 15.05 ±0.32 / 15.59 ms │     13.70 / 13.85 ±0.13 / 14.08 ms │ +1.09x faster │
│ QQuery 42 │     25.55 / 26.02 ±0.41 / 26.66 ms │     24.18 / 24.52 ±0.24 / 24.91 ms │ +1.06x faster │
│ QQuery 43 │        5.21 / 5.35 ±0.10 / 5.50 ms │        4.98 / 5.07 ±0.08 / 5.17 ms │ +1.05x faster │
│ QQuery 44 │        9.40 / 9.63 ±0.17 / 9.85 ms │      8.99 / 10.67 ±3.07 / 16.80 ms │  1.11x slower │
│ QQuery 45 │     29.05 / 29.32 ±0.15 / 29.46 ms │     31.92 / 33.95 ±2.04 / 37.07 ms │  1.16x slower │
│ QQuery 46 │     13.40 / 13.73 ±0.30 / 14.14 ms │     13.33 / 13.93 ±0.49 / 14.78 ms │     no change │
│ QQuery 47 │  264.14 / 271.17 ±5.81 / 277.61 ms │  263.46 / 272.07 ±6.90 / 280.64 ms │     no change │
│ QQuery 48 │  125.90 / 127.91 ±1.87 / 131.36 ms │  107.72 / 109.99 ±2.75 / 115.36 ms │ +1.16x faster │
│ QQuery 49 │    92.55 / 97.47 ±7.00 / 110.76 ms │     87.35 / 89.85 ±2.34 / 93.33 ms │ +1.08x faster │
│ QQuery 50 │  129.55 / 131.21 ±1.12 / 132.49 ms │     58.27 / 58.59 ±0.34 / 59.05 ms │ +2.24x faster │
│ QQuery 51 │  108.34 / 112.66 ±4.36 / 119.90 ms │  106.88 / 109.03 ±1.95 / 112.70 ms │     no change │
│ QQuery 52 │     25.73 / 25.81 ±0.06 / 25.90 ms │     25.34 / 25.58 ±0.14 / 25.76 ms │     no change │
│ QQuery 53 │     32.20 / 32.43 ±0.15 / 32.63 ms │     32.82 / 33.20 ±0.41 / 33.96 ms │     no change │
│ QQuery 54 │     30.83 / 33.05 ±1.70 / 35.56 ms │     28.88 / 30.87 ±1.74 / 34.08 ms │ +1.07x faster │
│ QQuery 55 │     25.25 / 27.11 ±2.51 / 31.67 ms │     24.35 / 25.60 ±1.40 / 28.21 ms │ +1.06x faster │
│ QQuery 56 │     36.10 / 36.69 ±0.43 / 37.41 ms │     35.58 / 35.96 ±0.24 / 36.22 ms │     no change │
│ QQuery 57 │  142.72 / 146.11 ±2.71 / 149.61 ms │  149.20 / 150.91 ±1.02 / 151.99 ms │     no change │
│ QQuery 58 │     77.38 / 79.66 ±3.99 / 87.62 ms │     66.66 / 69.01 ±2.86 / 74.62 ms │ +1.15x faster │
│ QQuery 59 │     83.66 / 84.36 ±0.44 / 84.84 ms │     81.73 / 82.58 ±0.82 / 83.95 ms │     no change │
│ QQuery 60 │     40.33 / 40.72 ±0.29 / 41.22 ms │     38.92 / 39.42 ±0.62 / 40.51 ms │     no change │
│ QQuery 61 │     12.13 / 13.01 ±1.64 / 16.29 ms │     10.94 / 11.29 ±0.22 / 11.57 ms │ +1.15x faster │
│ QQuery 62 │     45.59 / 48.77 ±5.13 / 59.00 ms │     43.98 / 46.78 ±2.12 / 49.61 ms │     no change │
│ QQuery 63 │     32.53 / 32.94 ±0.51 / 33.87 ms │     32.63 / 33.29 ±0.74 / 34.63 ms │     no change │
│ QQuery 64 │ 545.35 / 555.01 ±12.38 / 579.20 ms │  311.09 / 318.54 ±9.95 / 338.14 ms │ +1.74x faster │
│ QQuery 65 │  165.03 / 169.94 ±4.78 / 178.53 ms │  155.61 / 159.80 ±3.25 / 163.26 ms │ +1.06x faster │
│ QQuery 66 │     67.25 / 69.81 ±4.14 / 78.06 ms │     66.14 / 66.96 ±0.76 / 68.24 ms │     no change │
│ QQuery 67 │  289.93 / 302.48 ±6.71 / 308.68 ms │  291.15 / 300.87 ±8.79 / 313.34 ms │     no change │
│ QQuery 68 │     13.26 / 13.51 ±0.26 / 13.85 ms │     12.87 / 13.15 ±0.16 / 13.33 ms │     no change │
│ QQuery 69 │     75.65 / 81.86 ±6.15 / 92.66 ms │     75.60 / 78.39 ±3.36 / 84.93 ms │     no change │
│ QQuery 70 │  125.00 / 126.16 ±0.61 / 126.68 ms │  122.23 / 125.91 ±4.63 / 134.35 ms │     no change │
│ QQuery 71 │     34.42 / 38.62 ±4.99 / 47.68 ms │     34.13 / 34.53 ±0.46 / 35.43 ms │ +1.12x faster │
│ QQuery 72 │  182.76 / 186.32 ±3.53 / 193.02 ms │ 108.03 / 117.25 ±10.53 / 136.53 ms │ +1.59x faster │
│ QQuery 73 │     10.83 / 11.27 ±0.42 / 11.97 ms │     11.28 / 11.58 ±0.25 / 12.03 ms │     no change │
│ QQuery 74 │  160.09 / 163.03 ±2.34 / 166.69 ms │  161.15 / 165.82 ±6.23 / 178.03 ms │     no change │
│ QQuery 75 │  189.99 / 194.53 ±6.49 / 207.43 ms │  127.22 / 128.33 ±0.93 / 129.76 ms │ +1.52x faster │
│ QQuery 76 │     42.67 / 48.29 ±8.37 / 64.96 ms │     38.64 / 40.85 ±2.40 / 45.29 ms │ +1.18x faster │
│ QQuery 77 │     63.71 / 64.36 ±0.64 / 65.50 ms │     61.58 / 62.98 ±1.14 / 64.81 ms │     no change │
│ QQuery 78 │  127.61 / 130.88 ±2.70 / 135.13 ms │  122.34 / 125.76 ±1.94 / 128.12 ms │     no change │
│ QQuery 79 │     70.91 / 71.65 ±0.45 / 72.11 ms │     70.32 / 73.25 ±4.40 / 82.01 ms │     no change │
│ QQuery 80 │     78.82 / 79.10 ±0.29 / 79.61 ms │     77.90 / 78.94 ±1.06 / 80.65 ms │     no change │
│ QQuery 81 │     28.68 / 30.51 ±2.41 / 35.29 ms │     28.20 / 29.77 ±2.32 / 34.38 ms │     no change │
│ QQuery 82 │     33.09 / 37.02 ±5.88 / 48.71 ms │     25.77 / 27.09 ±1.87 / 30.79 ms │ +1.37x faster │
│ QQuery 83 │     30.25 / 30.97 ±0.49 / 31.75 ms │     33.25 / 34.43 ±1.31 / 36.80 ms │  1.11x slower │
│ QQuery 84 │     31.91 / 32.10 ±0.17 / 32.35 ms │     33.69 / 34.37 ±0.77 / 35.85 ms │  1.07x slower │
│ QQuery 85 │ 168.49 / 177.01 ±14.69 / 206.30 ms │  121.40 / 122.86 ±2.17 / 127.12 ms │ +1.44x faster │
│ QQuery 86 │     31.74 / 32.45 ±0.51 / 33.19 ms │     32.93 / 36.20 ±5.25 / 46.60 ms │  1.12x slower │
│ QQuery 87 │     83.34 / 85.87 ±1.62 / 88.08 ms │     86.60 / 87.66 ±0.87 / 88.65 ms │     no change │
│ QQuery 88 │     67.75 / 75.68 ±9.72 / 93.71 ms │     66.45 / 71.23 ±5.83 / 82.71 ms │ +1.06x faster │
│ QQuery 89 │     44.79 / 44.98 ±0.11 / 45.13 ms │     40.30 / 40.81 ±0.40 / 41.50 ms │ +1.10x faster │
│ QQuery 90 │     18.64 / 19.06 ±0.44 / 19.88 ms │     18.81 / 18.96 ±0.12 / 19.15 ms │     no change │
│ QQuery 91 │     44.62 / 45.47 ±0.72 / 46.50 ms │     44.45 / 45.66 ±0.83 / 46.69 ms │     no change │
│ QQuery 92 │     31.44 / 33.97 ±3.02 / 39.79 ms │     30.56 / 31.03 ±0.34 / 31.46 ms │ +1.09x faster │
│ QQuery 93 │     53.17 / 56.69 ±5.90 / 68.42 ms │     50.38 / 51.80 ±2.01 / 55.78 ms │ +1.09x faster │
│ QQuery 94 │     37.09 / 37.82 ±0.72 / 39.09 ms │     38.25 / 41.56 ±4.42 / 49.92 ms │  1.10x slower │
│ QQuery 95 │    95.41 / 98.94 ±3.74 / 106.14 ms │     60.03 / 61.78 ±1.58 / 64.13 ms │ +1.60x faster │
│ QQuery 96 │     28.02 / 29.31 ±1.56 / 32.08 ms │     25.98 / 26.29 ±0.23 / 26.65 ms │ +1.12x faster │
│ QQuery 97 │     56.13 / 57.66 ±1.60 / 60.64 ms │     57.08 / 57.99 ±0.78 / 59.12 ms │     no change │
│ QQuery 98 │     39.80 / 42.40 ±1.92 / 45.62 ms │     37.29 / 38.01 ±0.99 / 39.92 ms │ +1.12x faster │
│ QQuery 99 │     60.87 / 61.85 ±0.64 / 62.76 ms │     60.59 / 64.93 ±5.96 / 76.70 ms │     no change │
└───────────┴────────────────────────────────────┴────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 8629.15ms │
│ Total Time (hash-join-global-bloom-filter)   │ 7506.46ms │
│ Average Time (HEAD)                          │   87.16ms │
│ Average Time (hash-join-global-bloom-filter) │   75.82ms │
│ Queries Faster                               │        41 │
│ Queries Slower                               │        10 │
│ Queries with No Change                       │        48 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Resource Usage

tpcds — base (merge-base)

Metric Value
Wall time 45.0s
Peak memory 1.9 GiB
Avg memory 1.4 GiB
CPU user 125.3s
CPU sys 6.8s
Peak spill 0 B

tpcds — branch

Metric Value
Wall time 40.0s
Peak memory 1.9 GiB
Avg memory 1.3 GiB
CPU user 104.4s
CPU sys 6.6s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark clickbench_partitioned
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │    1.18 ms │                       1.18 ms │     no change │
│ QQuery 1  │   12.29 ms │                      11.96 ms │     no change │
│ QQuery 2  │   35.75 ms │                      34.59 ms │     no change │
│ QQuery 3  │   31.13 ms │                      30.69 ms │     no change │
│ QQuery 4  │  232.72 ms │                     237.19 ms │     no change │
│ QQuery 5  │  270.84 ms │                     270.89 ms │     no change │
│ QQuery 6  │    1.22 ms │                       1.25 ms │     no change │
│ QQuery 7  │   15.63 ms │                      15.69 ms │     no change │
│ QQuery 8  │  327.92 ms │                     328.95 ms │     no change │
│ QQuery 9  │  453.31 ms │                     460.95 ms │     no change │
│ QQuery 10 │   88.14 ms │                      88.14 ms │     no change │
│ QQuery 11 │   98.53 ms │                      98.65 ms │     no change │
│ QQuery 12 │  295.10 ms │                     280.67 ms │     no change │
│ QQuery 13 │  395.72 ms │                     390.14 ms │     no change │
│ QQuery 14 │  310.41 ms │                     297.62 ms │     no change │
│ QQuery 15 │  282.25 ms │                     281.21 ms │     no change │
│ QQuery 16 │  617.52 ms │                     609.88 ms │     no change │
│ QQuery 17 │  617.86 ms │                     625.67 ms │     no change │
│ QQuery 18 │ 1245.62 ms │                    1227.20 ms │     no change │
│ QQuery 19 │   28.85 ms │                      29.22 ms │     no change │
│ QQuery 20 │  513.06 ms │                     513.83 ms │     no change │
│ QQuery 21 │  555.11 ms │                     537.22 ms │     no change │
│ QQuery 22 │  778.69 ms │                     763.97 ms │     no change │
│ QQuery 23 │  128.01 ms │                     125.77 ms │     no change │
│ QQuery 24 │   43.81 ms │                      40.88 ms │ +1.07x faster │
│ QQuery 25 │  133.79 ms │                     106.48 ms │ +1.26x faster │
│ QQuery 26 │   47.49 ms │                      44.91 ms │ +1.06x faster │
│ QQuery 27 │  549.55 ms │                     538.04 ms │     no change │
│ QQuery 28 │ 2897.48 ms │                    2847.61 ms │     no change │
│ QQuery 29 │   38.69 ms │                      38.67 ms │     no change │
│ QQuery 30 │  308.72 ms │                     294.31 ms │     no change │
│ QQuery 31 │  281.09 ms │                     268.15 ms │     no change │
│ QQuery 32 │  919.42 ms │                     952.16 ms │     no change │
│ QQuery 33 │ 1473.36 ms │                    1438.98 ms │     no change │
│ QQuery 34 │ 1462.73 ms │                    1473.00 ms │     no change │
│ QQuery 35 │  287.86 ms │                     291.90 ms │     no change │
│ QQuery 36 │   67.60 ms │                      65.49 ms │     no change │
│ QQuery 37 │   35.77 ms │                      34.43 ms │     no change │
│ QQuery 38 │   35.10 ms │                      33.11 ms │ +1.06x faster │
│ QQuery 39 │  133.99 ms │                     136.87 ms │     no change │
│ QQuery 40 │   17.37 ms │                      16.93 ms │     no change │
│ QQuery 41 │   15.67 ms │                      15.51 ms │     no change │
│ QQuery 42 │   13.84 ms │                      13.58 ms │     no change │
└───────────┴────────────┴───────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 16100.12ms │
│ Total Time (hash-join-global-bloom-filter)   │ 15913.56ms │
│ Average Time (HEAD)                          │   374.42ms │
│ Average Time (hash-join-global-bloom-filter) │   370.08ms │
│ Queries Faster                               │          4 │
│ Queries Slower                               │          0 │
│ Queries with No Change                       │         39 │
│ Queries with Failure                         │          0 │
└──────────────────────────────────────────────┴────────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃         hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │          1.18 / 4.14 ±5.82 / 15.78 ms │          1.18 / 4.13 ±5.84 / 15.82 ms │     no change │
│ QQuery 1  │        12.29 / 12.51 ±0.23 / 12.79 ms │        11.96 / 12.09 ±0.10 / 12.24 ms │     no change │
│ QQuery 2  │        35.75 / 36.29 ±0.46 / 36.96 ms │        34.59 / 34.98 ±0.33 / 35.58 ms │     no change │
│ QQuery 3  │        31.13 / 31.95 ±0.68 / 33.19 ms │        30.69 / 31.27 ±0.48 / 31.98 ms │     no change │
│ QQuery 4  │     232.72 / 236.80 ±3.18 / 241.17 ms │     237.19 / 238.59 ±1.32 / 240.93 ms │     no change │
│ QQuery 5  │     270.84 / 278.30 ±7.83 / 293.13 ms │     270.89 / 274.48 ±3.86 / 280.17 ms │     no change │
│ QQuery 6  │           1.22 / 1.38 ±0.25 / 1.86 ms │           1.25 / 1.41 ±0.25 / 1.90 ms │     no change │
│ QQuery 7  │        15.63 / 15.92 ±0.17 / 16.12 ms │        15.69 / 15.84 ±0.10 / 16.00 ms │     no change │
│ QQuery 8  │     327.92 / 334.67 ±6.35 / 346.40 ms │     328.95 / 332.28 ±3.27 / 337.82 ms │     no change │
│ QQuery 9  │     453.31 / 466.80 ±7.11 / 473.27 ms │     460.95 / 470.20 ±7.56 / 479.28 ms │     no change │
│ QQuery 10 │       88.14 / 91.56 ±5.53 / 102.59 ms │        88.14 / 89.26 ±1.02 / 90.90 ms │     no change │
│ QQuery 11 │      98.53 / 100.18 ±1.17 / 102.18 ms │       98.65 / 99.47 ±1.39 / 102.23 ms │     no change │
│ QQuery 12 │     295.10 / 298.36 ±2.43 / 302.57 ms │     280.67 / 289.06 ±7.75 / 302.74 ms │     no change │
│ QQuery 13 │    395.72 / 406.66 ±12.08 / 425.56 ms │    390.14 / 408.17 ±12.66 / 427.88 ms │     no change │
│ QQuery 14 │     310.41 / 314.88 ±4.90 / 323.42 ms │     297.62 / 304.28 ±6.88 / 317.21 ms │     no change │
│ QQuery 15 │    282.25 / 294.28 ±13.83 / 321.25 ms │     281.21 / 284.74 ±3.35 / 289.15 ms │     no change │
│ QQuery 16 │     617.52 / 624.05 ±5.36 / 631.05 ms │     609.88 / 627.63 ±9.21 / 635.18 ms │     no change │
│ QQuery 17 │     617.86 / 622.15 ±2.80 / 626.04 ms │     625.67 / 636.54 ±9.41 / 652.29 ms │     no change │
│ QQuery 18 │ 1245.62 / 1270.80 ±19.20 / 1297.06 ms │ 1227.20 / 1242.68 ±21.92 / 1285.82 ms │     no change │
│ QQuery 19 │        28.85 / 30.05 ±0.96 / 31.56 ms │        29.22 / 33.17 ±4.38 / 40.11 ms │  1.10x slower │
│ QQuery 20 │     513.06 / 521.68 ±7.17 / 534.66 ms │     513.83 / 521.28 ±6.62 / 533.03 ms │     no change │
│ QQuery 21 │    555.11 / 565.77 ±13.42 / 591.28 ms │    537.22 / 551.57 ±11.49 / 572.51 ms │     no change │
│ QQuery 22 │     778.69 / 783.09 ±4.99 / 792.34 ms │     763.97 / 770.63 ±3.97 / 775.92 ms │     no change │
│ QQuery 23 │     128.01 / 133.78 ±4.65 / 142.13 ms │    125.77 / 137.75 ±11.50 / 156.89 ms │     no change │
│ QQuery 24 │        43.81 / 46.59 ±3.98 / 54.25 ms │        40.88 / 41.63 ±0.86 / 43.14 ms │ +1.12x faster │
│ QQuery 25 │     133.79 / 138.52 ±5.05 / 145.70 ms │     106.48 / 107.44 ±0.95 / 109.13 ms │ +1.29x faster │
│ QQuery 26 │        47.49 / 49.70 ±2.45 / 54.14 ms │        44.91 / 48.85 ±4.60 / 56.89 ms │     no change │
│ QQuery 27 │     549.55 / 557.38 ±6.16 / 566.24 ms │    538.04 / 560.10 ±13.19 / 572.45 ms │     no change │
│ QQuery 28 │ 2897.48 / 2915.81 ±15.21 / 2940.02 ms │ 2847.61 / 2873.59 ±17.85 / 2901.10 ms │     no change │
│ QQuery 29 │        38.69 / 39.08 ±0.26 / 39.39 ms │        38.67 / 42.48 ±7.25 / 56.98 ms │  1.09x slower │
│ QQuery 30 │    308.72 / 321.02 ±11.99 / 342.51 ms │     294.31 / 304.66 ±6.14 / 311.06 ms │ +1.05x faster │
│ QQuery 31 │    281.09 / 290.77 ±11.97 / 313.99 ms │     268.15 / 278.26 ±9.13 / 295.16 ms │     no change │
│ QQuery 32 │    919.42 / 952.22 ±19.03 / 976.60 ms │    952.16 / 968.87 ±10.72 / 985.36 ms │     no change │
│ QQuery 33 │ 1473.36 / 1487.81 ±20.33 / 1527.52 ms │ 1438.98 / 1483.31 ±31.73 / 1525.93 ms │     no change │
│ QQuery 34 │ 1462.73 / 1505.46 ±53.28 / 1608.99 ms │ 1473.00 / 1524.70 ±34.96 / 1568.17 ms │     no change │
│ QQuery 35 │    287.86 / 310.41 ±19.38 / 342.54 ms │    291.90 / 319.82 ±46.18 / 411.85 ms │     no change │
│ QQuery 36 │        67.60 / 77.97 ±5.95 / 84.97 ms │        65.49 / 73.13 ±6.54 / 82.60 ms │ +1.07x faster │
│ QQuery 37 │        35.77 / 39.91 ±4.45 / 47.53 ms │        34.43 / 39.59 ±4.33 / 44.89 ms │     no change │
│ QQuery 38 │        35.10 / 38.57 ±2.99 / 43.28 ms │        33.11 / 35.97 ±2.22 / 38.05 ms │ +1.07x faster │
│ QQuery 39 │     133.99 / 142.99 ±7.89 / 154.62 ms │    136.87 / 150.04 ±12.45 / 173.18 ms │     no change │
│ QQuery 40 │        17.37 / 17.73 ±0.36 / 18.31 ms │        16.93 / 21.27 ±4.30 / 26.70 ms │  1.20x slower │
│ QQuery 41 │        15.67 / 19.80 ±5.12 / 28.81 ms │        15.51 / 15.97 ±0.37 / 16.47 ms │ +1.24x faster │
│ QQuery 42 │        13.84 / 14.09 ±0.22 / 14.44 ms │        13.58 / 13.72 ±0.13 / 13.93 ms │     no change │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 16441.87ms │
│ Total Time (hash-join-global-bloom-filter)   │ 16314.91ms │
│ Average Time (HEAD)                          │   382.37ms │
│ Average Time (hash-join-global-bloom-filter) │   379.42ms │
│ Queries Faster                               │          6 │
│ Queries Slower                               │          3 │
│ Queries with No Change                       │         34 │
│ Queries with Failure                         │          0 │
└──────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric Value
Wall time 85.0s
Peak memory 11.1 GiB
Avg memory 4.7 GiB
CPU user 838.7s
CPU sys 56.7s
Peak spill 0 B

clickbench_partitioned — branch

Metric Value
Wall time 85.0s
Peak memory 13.0 GiB
Avg memory 5.1 GiB
CPU user 829.1s
CPU sys 56.7s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark clickbench_partitioned
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │    1.20 ms │                       1.20 ms │     no change │
│ QQuery 1  │   11.68 ms │                      11.85 ms │     no change │
│ QQuery 2  │   35.96 ms │                      34.47 ms │     no change │
│ QQuery 3  │   31.32 ms │                      31.73 ms │     no change │
│ QQuery 4  │  235.51 ms │                     234.91 ms │     no change │
│ QQuery 5  │  273.21 ms │                     271.38 ms │     no change │
│ QQuery 6  │    1.26 ms │                       1.27 ms │     no change │
│ QQuery 7  │   12.79 ms │                      12.86 ms │     no change │
│ QQuery 8  │  332.15 ms │                     327.04 ms │     no change │
│ QQuery 9  │  471.82 ms │                     452.93 ms │     no change │
│ QQuery 10 │   64.35 ms │                      65.39 ms │     no change │
│ QQuery 11 │   74.47 ms │                      76.44 ms │     no change │
│ QQuery 12 │  259.08 ms │                     263.29 ms │     no change │
│ QQuery 13 │  357.56 ms │                     356.16 ms │     no change │
│ QQuery 14 │  277.36 ms │                     276.51 ms │     no change │
│ QQuery 15 │  282.94 ms │                     285.25 ms │     no change │
│ QQuery 16 │  618.79 ms │                     639.36 ms │     no change │
│ QQuery 17 │  615.87 ms │                     645.16 ms │     no change │
│ QQuery 18 │ 1281.63 ms │                    1225.69 ms │     no change │
│ QQuery 19 │   28.13 ms │                      28.08 ms │     no change │
│ QQuery 20 │  519.69 ms │                     520.66 ms │     no change │
│ QQuery 21 │  510.10 ms │                     510.79 ms │     no change │
│ QQuery 22 │ 1005.00 ms │                    1026.84 ms │     no change │
│ QQuery 23 │ 3074.46 ms │                    3306.74 ms │  1.08x slower │
│ QQuery 24 │   39.96 ms │                      40.55 ms │     no change │
│ QQuery 25 │  104.48 ms │                     106.93 ms │     no change │
│ QQuery 26 │   40.77 ms │                      41.27 ms │     no change │
│ QQuery 27 │  513.02 ms │                     511.56 ms │     no change │
│ QQuery 28 │ 2858.63 ms │                    2850.50 ms │     no change │
│ QQuery 29 │   40.54 ms │                      38.46 ms │ +1.05x faster │
│ QQuery 30 │  302.02 ms │                     306.06 ms │     no change │
│ QQuery 31 │  288.83 ms │                     285.30 ms │     no change │
│ QQuery 32 │ 1131.39 ms │                    1073.94 ms │ +1.05x faster │
│ QQuery 33 │ 1538.86 ms │                    1558.10 ms │     no change │
│ QQuery 34 │ 1577.41 ms │                    1585.46 ms │     no change │
│ QQuery 35 │  303.87 ms │                     303.97 ms │     no change │
│ QQuery 36 │   66.50 ms │                      65.95 ms │     no change │
│ QQuery 37 │   34.67 ms │                      35.05 ms │     no change │
│ QQuery 38 │   40.75 ms │                      41.25 ms │     no change │
│ QQuery 39 │  147.14 ms │                     134.95 ms │ +1.09x faster │
│ QQuery 40 │   13.91 ms │                      14.00 ms │     no change │
│ QQuery 41 │   13.14 ms │                      13.45 ms │     no change │
│ QQuery 42 │   12.84 ms │                      13.05 ms │     no change │
└───────────┴────────────┴───────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 19445.06ms │
│ Total Time (hash-join-global-bloom-filter)   │ 19625.82ms │
│ Average Time (HEAD)                          │   452.21ms │
│ Average Time (hash-join-global-bloom-filter) │   456.41ms │
│ Queries Faster                               │          3 │
│ Queries Slower                               │          1 │
│ Queries with No Change                       │         39 │
│ Queries with Failure                         │          0 │
└──────────────────────────────────────────────┴────────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃          hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │          1.20 / 4.17 ±5.86 / 15.88 ms │           1.20 / 4.16 ±5.85 / 15.86 ms │     no change │
│ QQuery 1  │        11.68 / 11.83 ±0.15 / 12.11 ms │         11.85 / 12.09 ±0.14 / 12.21 ms │     no change │
│ QQuery 2  │        35.96 / 36.43 ±0.33 / 36.95 ms │         34.47 / 35.27 ±0.89 / 37.00 ms │     no change │
│ QQuery 3  │        31.32 / 32.46 ±0.73 / 33.30 ms │         31.73 / 34.42 ±1.75 / 36.70 ms │  1.06x slower │
│ QQuery 4  │     235.51 / 239.80 ±3.20 / 245.08 ms │      234.91 / 238.79 ±3.25 / 244.21 ms │     no change │
│ QQuery 5  │     273.21 / 277.35 ±3.77 / 282.25 ms │      271.38 / 276.81 ±7.98 / 292.54 ms │     no change │
│ QQuery 6  │           1.26 / 1.44 ±0.25 / 1.93 ms │            1.27 / 1.41 ±0.23 / 1.87 ms │     no change │
│ QQuery 7  │        12.79 / 13.02 ±0.16 / 13.21 ms │         12.86 / 13.16 ±0.23 / 13.50 ms │     no change │
│ QQuery 8  │     332.15 / 338.29 ±4.78 / 343.88 ms │      327.04 / 334.07 ±4.72 / 340.00 ms │     no change │
│ QQuery 9  │     471.82 / 482.83 ±7.06 / 493.96 ms │      452.93 / 462.58 ±5.71 / 468.98 ms │     no change │
│ QQuery 10 │        64.35 / 65.36 ±0.98 / 67.07 ms │         65.39 / 65.66 ±0.28 / 66.05 ms │     no change │
│ QQuery 11 │        74.47 / 76.84 ±2.35 / 81.25 ms │         76.44 / 82.00 ±7.48 / 96.49 ms │  1.07x slower │
│ QQuery 12 │     259.08 / 264.64 ±4.24 / 271.54 ms │      263.29 / 266.88 ±4.23 / 274.40 ms │     no change │
│ QQuery 13 │     357.56 / 365.37 ±9.71 / 384.38 ms │     356.16 / 372.03 ±11.15 / 389.88 ms │     no change │
│ QQuery 14 │     277.36 / 279.07 ±1.81 / 281.68 ms │      276.51 / 282.88 ±6.07 / 291.52 ms │     no change │
│ QQuery 15 │     282.94 / 288.83 ±6.96 / 302.06 ms │      285.25 / 294.55 ±7.75 / 305.14 ms │     no change │
│ QQuery 16 │     618.79 / 624.54 ±5.15 / 632.02 ms │     639.36 / 659.83 ±16.11 / 687.14 ms │  1.06x slower │
│ QQuery 17 │     615.87 / 626.08 ±8.96 / 638.93 ms │     645.16 / 667.40 ±21.95 / 708.03 ms │  1.07x slower │
│ QQuery 18 │ 1281.63 / 1307.99 ±18.48 / 1335.32 ms │  1225.69 / 1260.35 ±33.03 / 1308.26 ms │     no change │
│ QQuery 19 │       28.13 / 34.30 ±11.60 / 57.50 ms │         28.08 / 30.38 ±3.64 / 37.65 ms │ +1.13x faster │
│ QQuery 20 │     519.69 / 528.21 ±9.53 / 542.98 ms │     520.66 / 532.56 ±15.23 / 562.34 ms │     no change │
│ QQuery 21 │     510.10 / 513.84 ±2.68 / 518.14 ms │      510.79 / 524.89 ±8.43 / 533.89 ms │     no change │
│ QQuery 22 │  1005.00 / 1013.72 ±5.17 / 1020.43 ms │  1026.84 / 1041.13 ±13.63 / 1064.04 ms │     no change │
│ QQuery 23 │ 3074.46 / 3201.69 ±85.02 / 3340.71 ms │  3306.74 / 3396.02 ±53.36 / 3459.57 ms │  1.06x slower │
│ QQuery 24 │    39.96 / 175.15 ±204.29 / 567.51 ms │       40.55 / 63.64 ±35.28 / 132.06 ms │ +2.75x faster │
│ QQuery 25 │     104.48 / 106.27 ±1.35 / 108.47 ms │    106.93 / 187.12 ±125.26 / 431.09 ms │  1.76x slower │
│ QQuery 26 │        40.77 / 41.39 ±0.55 / 42.36 ms │         41.27 / 41.62 ±0.30 / 42.05 ms │     no change │
│ QQuery 27 │     513.02 / 518.78 ±4.91 / 525.37 ms │     511.56 / 523.63 ±12.82 / 546.12 ms │     no change │
│ QQuery 28 │ 2858.63 / 2898.44 ±32.80 / 2952.31 ms │  2850.50 / 2881.76 ±25.26 / 2919.84 ms │     no change │
│ QQuery 29 │       40.54 / 63.93 ±16.02 / 84.59 ms │         38.46 / 45.60 ±8.53 / 59.40 ms │ +1.40x faster │
│ QQuery 30 │    302.02 / 317.85 ±15.20 / 344.83 ms │      306.06 / 312.09 ±4.53 / 317.71 ms │     no change │
│ QQuery 31 │    288.83 / 299.78 ±11.55 / 319.79 ms │     285.30 / 299.49 ±13.00 / 322.97 ms │     no change │
│ QQuery 32 │ 1131.39 / 1205.05 ±65.90 / 1317.01 ms │ 1073.94 / 1140.90 ±103.58 / 1345.27 ms │ +1.06x faster │
│ QQuery 33 │ 1538.86 / 1592.81 ±28.52 / 1619.49 ms │  1558.10 / 1593.02 ±34.66 / 1654.83 ms │     no change │
│ QQuery 34 │ 1577.41 / 1611.20 ±30.99 / 1656.28 ms │  1585.46 / 1634.16 ±57.61 / 1742.79 ms │     no change │
│ QQuery 35 │    303.87 / 346.28 ±49.91 / 436.88 ms │     303.97 / 337.31 ±37.28 / 397.76 ms │     no change │
│ QQuery 36 │        66.50 / 73.76 ±7.34 / 87.11 ms │         65.95 / 72.00 ±6.79 / 84.35 ms │     no change │
│ QQuery 37 │        34.67 / 42.43 ±6.01 / 50.71 ms │         35.05 / 38.94 ±5.15 / 48.23 ms │ +1.09x faster │
│ QQuery 38 │        40.75 / 45.61 ±4.61 / 52.29 ms │         41.25 / 50.62 ±5.54 / 58.62 ms │  1.11x slower │
│ QQuery 39 │     147.14 / 154.75 ±8.13 / 169.39 ms │      134.95 / 150.03 ±8.12 / 158.69 ms │     no change │
│ QQuery 40 │        13.91 / 17.24 ±5.86 / 28.94 ms │         14.00 / 16.89 ±5.02 / 26.90 ms │     no change │
│ QQuery 41 │        13.14 / 13.44 ±0.17 / 13.64 ms │         13.45 / 16.04 ±4.71 / 25.45 ms │  1.19x slower │
│ QQuery 42 │        12.84 / 12.97 ±0.07 / 13.04 ms │         13.05 / 15.32 ±4.03 / 23.37 ms │  1.18x slower │
└───────────┴───────────────────────────────────────┴────────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 20165.23ms │
│ Total Time (hash-join-global-bloom-filter)   │ 20309.49ms │
│ Average Time (HEAD)                          │   468.96ms │
│ Average Time (hash-join-global-bloom-filter) │   472.31ms │
│ Queries Faster                               │          5 │
│ Queries Slower                               │          9 │
│ Queries with No Change                       │         29 │
│ Queries with Failure                         │          0 │
└──────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric Value
Wall time 105.0s
Peak memory 18.2 GiB
Avg memory 6.1 GiB
CPU user 995.5s
CPU sys 97.2s
Peak spill 0 B

clickbench_partitioned — branch

Metric Value
Wall time 105.0s
Peak memory 18.3 GiB
Avg memory 6.1 GiB
CPU user 992.6s
CPU sys 102.0s
Peak spill 0 B

File an issue against this benchmark runner

@adriangb

adriangb commented Oct 17, 2026

Copy link
Copy Markdown
Contributor Author

run benchmark tpch10

env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: true

@adriangb

adriangb commented Oct 17, 2026

Copy link
Copy Markdown
Contributor Author

run benchmark tpch10

env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: false

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6235857254-3541-4xbqj 6.12.94+ #1 SMP Fri Aug 21 08:00:16 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "false"

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6235855840-3540-jw8gv 6.12.94+ #1 SMP Fri Aug 21 08:00:16 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "false"
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃      HEAD ┃ hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │ 328.51 ms │                     331.67 ms │     no change │
│ QQuery 2  │  89.40 ms │                      90.19 ms │     no change │
│ QQuery 3  │ 219.03 ms │                     222.53 ms │     no change │
│ QQuery 4  │ 117.73 ms │                     116.47 ms │     no change │
│ QQuery 5  │ 351.20 ms │                     358.92 ms │     no change │
│ QQuery 6  │ 101.24 ms │                     100.13 ms │     no change │
│ QQuery 7  │ 452.04 ms │                     474.96 ms │  1.05x slower │
│ QQuery 8  │ 358.57 ms │                     371.01 ms │     no change │
│ QQuery 9  │ 514.09 ms │                     530.48 ms │     no change │
│ QQuery 10 │ 297.50 ms │                     292.91 ms │     no change │
│ QQuery 11 │  61.09 ms │                      61.49 ms │     no change │
│ QQuery 12 │ 119.64 ms │                     120.81 ms │     no change │
│ QQuery 13 │ 299.25 ms │                     305.08 ms │     no change │
│ QQuery 14 │ 170.95 ms │                     171.79 ms │     no change │
│ QQuery 15 │ 300.03 ms │                     297.95 ms │     no change │
│ QQuery 16 │  64.49 ms │                      64.81 ms │     no change │
│ QQuery 17 │ 581.69 ms │                     582.23 ms │     no change │
│ QQuery 18 │ 728.93 ms │                     430.61 ms │ +1.69x faster │
│ QQuery 19 │ 232.62 ms │                     231.76 ms │     no change │
│ QQuery 20 │ 264.47 ms │                     260.83 ms │     no change │
│ QQuery 21 │ 666.74 ms │                     693.19 ms │     no change │
│ QQuery 22 │  60.25 ms │                      61.33 ms │     no change │
└───────────┴───────────┴───────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 6379.44ms │
│ Total Time (hash-join-global-bloom-filter)   │ 6171.15ms │
│ Average Time (HEAD)                          │  289.97ms │
│ Average Time (hash-join-global-bloom-filter) │  280.51ms │
│ Queries Faster                               │         1 │
│ Queries Slower                               │         1 │
│ Queries with No Change                       │        20 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                               HEAD ┃      hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │  328.51 / 331.83 ±1.99 / 334.36 ms │  331.67 / 333.18 ±0.89 / 334.38 ms │     no change │
│ QQuery 2  │     89.40 / 91.69 ±1.37 / 93.29 ms │     90.19 / 92.32 ±2.86 / 97.83 ms │     no change │
│ QQuery 3  │  219.03 / 220.88 ±1.96 / 224.58 ms │  222.53 / 225.30 ±1.54 / 227.25 ms │     no change │
│ QQuery 4  │  117.73 / 118.61 ±0.65 / 119.52 ms │  116.47 / 119.44 ±2.52 / 123.82 ms │     no change │
│ QQuery 5  │  351.20 / 357.96 ±7.27 / 367.52 ms │  358.92 / 363.78 ±2.92 / 367.96 ms │     no change │
│ QQuery 6  │  101.24 / 102.18 ±1.28 / 104.68 ms │  100.13 / 103.05 ±3.35 / 109.50 ms │     no change │
│ QQuery 7  │  452.04 / 457.47 ±5.09 / 465.10 ms │  474.96 / 484.10 ±8.81 / 497.41 ms │  1.06x slower │
│ QQuery 8  │  358.57 / 363.40 ±4.17 / 370.40 ms │  371.01 / 376.75 ±4.56 / 383.99 ms │     no change │
│ QQuery 9  │ 514.09 / 529.72 ±21.60 / 572.06 ms │ 530.48 / 543.07 ±15.80 / 573.45 ms │     no change │
│ QQuery 10 │  297.50 / 300.36 ±2.07 / 303.77 ms │ 292.91 / 302.77 ±12.25 / 326.34 ms │     no change │
│ QQuery 11 │     61.09 / 62.82 ±1.04 / 64.23 ms │     61.49 / 65.46 ±6.40 / 78.18 ms │     no change │
│ QQuery 12 │  119.64 / 127.66 ±8.49 / 139.66 ms │  120.81 / 124.79 ±5.87 / 136.36 ms │     no change │
│ QQuery 13 │  299.25 / 309.28 ±6.50 / 316.91 ms │ 305.08 / 321.96 ±20.20 / 360.71 ms │     no change │
│ QQuery 14 │  170.95 / 174.15 ±2.88 / 178.76 ms │  171.79 / 175.11 ±4.49 / 183.68 ms │     no change │
│ QQuery 15 │  300.03 / 302.65 ±2.38 / 305.80 ms │  297.95 / 303.43 ±4.31 / 308.77 ms │     no change │
│ QQuery 16 │     64.49 / 65.88 ±1.37 / 68.34 ms │     64.81 / 66.08 ±1.16 / 68.12 ms │     no change │
│ QQuery 17 │  581.69 / 587.75 ±4.83 / 595.93 ms │  582.23 / 588.35 ±4.94 / 595.39 ms │     no change │
│ QQuery 18 │ 728.93 / 760.52 ±35.94 / 822.33 ms │ 430.61 / 452.57 ±17.33 / 477.60 ms │ +1.68x faster │
│ QQuery 19 │ 232.62 / 247.86 ±13.89 / 267.58 ms │ 231.76 / 241.80 ±14.61 / 270.65 ms │     no change │
│ QQuery 20 │ 264.47 / 283.26 ±13.44 / 306.06 ms │  260.83 / 276.19 ±9.92 / 291.43 ms │     no change │
│ QQuery 21 │ 666.74 / 679.35 ±10.50 / 691.77 ms │  693.19 / 696.38 ±3.38 / 701.92 ms │     no change │
│ QQuery 22 │     60.25 / 61.85 ±1.30 / 63.29 ms │     61.33 / 63.02 ±1.65 / 65.41 ms │     no change │
└───────────┴────────────────────────────────────┴────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 6537.12ms │
│ Total Time (hash-join-global-bloom-filter)   │ 6318.89ms │
│ Average Time (HEAD)                          │  297.14ms │
│ Average Time (hash-join-global-bloom-filter) │  287.22ms │
│ Queries Faster                               │         1 │
│ Queries Slower                               │         1 │
│ Queries with No Change                       │        20 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Resource Usage

tpch10 — base (merge-base)

Metric Value
Wall time 35.0s
Peak memory 5.1 GiB
Avg memory 1.5 GiB
CPU user 331.8s
CPU sys 18.4s
Peak spill 0 B

tpch10 — branch

Metric Value
Wall time 35.0s
Peak memory 4.5 GiB
Avg memory 1.4 GiB
CPU user 314.8s
CPU sys 17.6s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Oct 17, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (885155a) to 5f21582 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │  340.58 ms │                     336.28 ms │     no change │
│ QQuery 2  │  107.49 ms │                      88.27 ms │ +1.22x faster │
│ QQuery 3  │  302.00 ms │                     204.50 ms │ +1.48x faster │
│ QQuery 4  │  291.16 ms │                     128.35 ms │ +2.27x faster │
│ QQuery 5  │  434.69 ms │                     247.53 ms │ +1.76x faster │
│ QQuery 6  │  179.28 ms │                     133.62 ms │ +1.34x faster │
│ QQuery 7  │  433.22 ms │                     291.26 ms │ +1.49x faster │
│ QQuery 8  │  400.74 ms │                     323.77 ms │ +1.24x faster │
│ QQuery 9  │  626.36 ms │                     573.09 ms │ +1.09x faster │
│ QQuery 10 │  370.81 ms │                     271.75 ms │ +1.36x faster │
│ QQuery 11 │   67.13 ms │                      67.27 ms │     no change │
│ QQuery 12 │  295.80 ms │                     157.63 ms │ +1.88x faster │
│ QQuery 13 │  393.83 ms │                     323.17 ms │ +1.22x faster │
│ QQuery 14 │  190.98 ms │                     163.42 ms │ +1.17x faster │
│ QQuery 15 │  376.11 ms │                     326.27 ms │ +1.15x faster │
│ QQuery 16 │   88.08 ms │                      63.68 ms │ +1.38x faster │
│ QQuery 17 │  487.32 ms │                     257.91 ms │ +1.89x faster │
│ QQuery 18 │  647.96 ms │                     317.34 ms │ +2.04x faster │
│ QQuery 19 │  241.34 ms │                     206.71 ms │ +1.17x faster │
│ QQuery 20 │  275.62 ms │                     204.91 ms │ +1.35x faster │
│ QQuery 21 │ 1202.20 ms │                     703.25 ms │ +1.71x faster │
│ QQuery 22 │  127.02 ms │                      57.86 ms │ +2.20x faster │
└───────────┴────────────┴───────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 7879.72ms │
│ Total Time (hash-join-global-bloom-filter)   │ 5447.83ms │
│ Average Time (HEAD)                          │  358.17ms │
│ Average Time (hash-join-global-bloom-filter) │  247.63ms │
│ Queries Faster                               │        20 │
│ Queries Slower                               │         0 │
│ Queries with No Change                       │         2 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃      hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │     340.58 / 342.50 ±1.82 / 345.42 ms │  336.28 / 340.87 ±2.61 / 343.43 ms │     no change │
│ QQuery 2  │     107.49 / 109.57 ±1.58 / 112.04 ms │     88.27 / 89.66 ±1.15 / 91.34 ms │ +1.22x faster │
│ QQuery 3  │     302.00 / 304.69 ±1.98 / 307.23 ms │  204.50 / 208.47 ±3.41 / 214.61 ms │ +1.46x faster │
│ QQuery 4  │     291.16 / 293.89 ±3.00 / 299.58 ms │  128.35 / 130.54 ±1.98 / 134.03 ms │ +2.25x faster │
│ QQuery 5  │     434.69 / 436.16 ±1.01 / 437.19 ms │  247.53 / 251.96 ±2.24 / 253.48 ms │ +1.73x faster │
│ QQuery 6  │     179.28 / 180.46 ±1.02 / 182.31 ms │  133.62 / 134.62 ±1.04 / 135.89 ms │ +1.34x faster │
│ QQuery 7  │     433.22 / 433.49 ±0.39 / 434.25 ms │  291.26 / 295.37 ±4.99 / 305.19 ms │ +1.47x faster │
│ QQuery 8  │     400.74 / 403.39 ±1.95 / 406.57 ms │  323.77 / 325.33 ±1.04 / 326.81 ms │ +1.24x faster │
│ QQuery 9  │     626.36 / 633.28 ±4.58 / 639.72 ms │  573.09 / 576.00 ±4.68 / 585.27 ms │ +1.10x faster │
│ QQuery 10 │     370.81 / 372.95 ±2.24 / 375.73 ms │  271.75 / 275.20 ±4.67 / 284.10 ms │ +1.36x faster │
│ QQuery 11 │        67.13 / 70.66 ±2.65 / 73.76 ms │     67.27 / 73.10 ±4.43 / 80.97 ms │     no change │
│ QQuery 12 │     295.80 / 299.92 ±4.74 / 306.37 ms │  157.63 / 160.97 ±4.01 / 168.17 ms │ +1.86x faster │
│ QQuery 13 │     393.83 / 403.48 ±5.47 / 407.91 ms │  323.17 / 334.96 ±8.42 / 345.85 ms │ +1.20x faster │
│ QQuery 14 │     190.98 / 193.30 ±2.93 / 199.10 ms │  163.42 / 167.47 ±3.22 / 171.35 ms │ +1.15x faster │
│ QQuery 15 │     376.11 / 379.84 ±2.45 / 382.26 ms │  326.27 / 332.59 ±3.92 / 337.74 ms │ +1.14x faster │
│ QQuery 16 │        88.08 / 91.16 ±3.54 / 97.28 ms │     63.68 / 65.17 ±1.67 / 68.29 ms │ +1.40x faster │
│ QQuery 17 │     487.32 / 492.93 ±4.59 / 500.15 ms │  257.91 / 265.45 ±6.44 / 274.56 ms │ +1.86x faster │
│ QQuery 18 │     647.96 / 655.53 ±8.42 / 671.24 ms │  317.34 / 330.18 ±9.65 / 342.77 ms │ +1.99x faster │
│ QQuery 19 │     241.34 / 248.05 ±6.67 / 260.57 ms │  206.71 / 213.54 ±9.16 / 230.84 ms │ +1.16x faster │
│ QQuery 20 │     275.62 / 277.49 ±1.76 / 280.44 ms │  204.91 / 208.21 ±2.81 / 211.56 ms │ +1.33x faster │
│ QQuery 21 │ 1202.20 / 1219.83 ±13.31 / 1232.99 ms │ 703.25 / 714.58 ±12.06 / 737.01 ms │ +1.71x faster │
│ QQuery 22 │     127.02 / 139.95 ±9.12 / 150.13 ms │     57.86 / 61.67 ±3.83 / 68.87 ms │ +2.27x faster │
└───────────┴───────────────────────────────────────┴────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 7982.55ms │
│ Total Time (hash-join-global-bloom-filter)   │ 5555.90ms │
│ Average Time (HEAD)                          │  362.84ms │
│ Average Time (hash-join-global-bloom-filter) │  252.54ms │
│ Queries Faster                               │        20 │
│ Queries Slower                               │         0 │
│ Queries with No Change                       │         2 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Resource Usage

tpch10 — base (merge-base)

Metric Value
Wall time 40.0s
Peak memory 2.9 GiB
Avg memory 1.2 GiB
CPU user 417.4s
CPU sys 16.9s
Peak spill 0 B

tpch10 — branch

Metric Value
Wall time 30.0s
Peak memory 3.5 GiB
Avg memory 1.3 GiB
CPU user 278.5s
CPU sys 16.8s
Peak spill 0 B

File an issue against this benchmark runner

@codecov-commenter

codecov-commenter commented Oct 17, 2026 •
edited
Loading

Copy link
Copy Markdown

Codecov Report

❌ Patch coverage is 93.20276% with 59 lines in your changes missing coverage. Please review.
✅ Project coverage is 82.93%. Comparing base (a7c7c08) to head (4b3684c).
⚠️ Report is 132 commits behind head on main.

Files with missing lines Patch % Lines
.../physical-plan/src/joins/hash_join/bloom_filter.rs 88.76% 30 Missing and 9 partials ⚠️
...physical-plan/src/joins/hash_join/shared_bounds.rs 95.70% 5 Missing and 12 partials ⚠️
...fusion/physical-plan/src/joins/key_range_bitmap.rs 97.22% 1 Missing and 1 partial ⚠️
...tafusion/physical-plan/src/joins/hash_join/exec.rs 95.00% 0 Missing and 1 partial ⚠️
Additional details and impacted files 
@@           Coverage Diff            @@
##             main   #26442    +/-   ##
========================================
  Coverage   82.92%   82.93%            
========================================
  Files        1154     1155     +1     
  Lines      457872   458727   +855     
  Branches   457872   458727   +855     
========================================
+ Hits       379712   380465   +753     
- Misses      55136    55217    +81     
- Partials    23024    23045    +21

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
  • 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

@adriangb

adriangb commented Oct 17, 2026

Copy link
Copy Markdown
Contributor Author

run benchmark tpch10

baseline:
  env:
    DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: false
changed:
  env:
    DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: true

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6628507148-4442-mzq75 6.12.105+ #1 SMP Sat Sep 19 07:55:25 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpcds

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6628506370-4439-js8r7 6.12.105+ #1 SMP Sat Sep 19 07:55:25 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpcds

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6628506732-4437-vqsbm 6.12.105+ #1 SMP Sat Sep 19 07:55:25 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6628506370-4440-k66pw 6.12.105+ #1 SMP Sat Sep 19 07:55:25 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃     HEAD ┃ hash-join-global-bloom-filter ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │ 40.93 ms │                      40.80 ms │    no change │
│ QQuery 2  │ 18.68 ms │                      18.46 ms │    no change │
│ QQuery 3  │ 28.66 ms │                      29.29 ms │    no change │
│ QQuery 4  │ 17.53 ms │                      17.54 ms │    no change │
│ QQuery 5  │ 35.66 ms │                      37.32 ms │    no change │
│ QQuery 6  │ 13.93 ms │                      14.74 ms │ 1.06x slower │
│ QQuery 7  │ 41.77 ms │                      45.53 ms │ 1.09x slower │
│ QQuery 8  │ 40.77 ms │                      41.91 ms │    no change │
│ QQuery 9  │ 49.93 ms │                      52.60 ms │ 1.05x slower │
│ QQuery 10 │ 40.64 ms │                      41.08 ms │    no change │
│ QQuery 11 │ 13.76 ms │                      13.65 ms │    no change │
│ QQuery 12 │ 17.95 ms │                      18.22 ms │    no change │
│ QQuery 13 │ 40.99 ms │                      41.19 ms │    no change │
│ QQuery 14 │ 24.36 ms │                      24.73 ms │    no change │
│ QQuery 15 │ 31.22 ms │                      31.50 ms │    no change │
│ QQuery 16 │ 13.58 ms │                      13.35 ms │    no change │
│ QQuery 17 │ 71.29 ms │                      70.58 ms │    no change │
│ QQuery 18 │ 76.06 ms │                      81.78 ms │ 1.08x slower │
│ QQuery 19 │ 31.90 ms │                      32.13 ms │    no change │
│ QQuery 20 │ 31.98 ms │                      32.26 ms │    no change │
│ QQuery 21 │ 56.12 ms │                      59.95 ms │ 1.07x slower │
│ QQuery 22 │ 13.93 ms │                      13.91 ms │    no change │
└───────────┴──────────┴───────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ Benchmark Summary                            ┃          ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 751.65ms │
│ Total Time (hash-join-global-bloom-filter)   │ 772.51ms │
│ Average Time (HEAD)                          │  34.17ms │
│ Average Time (hash-join-global-bloom-filter) │  35.11ms │
│ Queries Faster                               │        0 │
│ Queries Slower                               │        5 │
│ Queries with No Change                       │       17 │
│ Queries with Failure                         │        0 │
└──────────────────────────────────────────────┴──────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃                           HEAD ┃  hash-join-global-bloom-filter ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │ 40.93 / 42.29 ±1.42 / 44.18 ms │ 40.80 / 41.96 ±1.09 / 43.84 ms │    no change │
│ QQuery 2  │ 18.68 / 18.83 ±0.15 / 19.06 ms │ 18.46 / 18.69 ±0.18 / 18.98 ms │    no change │
│ QQuery 3  │ 28.66 / 29.05 ±0.38 / 29.77 ms │ 29.29 / 29.72 ±0.42 / 30.43 ms │    no change │
│ QQuery 4  │ 17.53 / 18.07 ±0.56 / 19.00 ms │ 17.54 / 17.61 ±0.05 / 17.66 ms │    no change │
│ QQuery 5  │ 35.66 / 35.97 ±0.36 / 36.66 ms │ 37.32 / 37.54 ±0.25 / 38.00 ms │    no change │
│ QQuery 6  │ 13.93 / 14.48 ±0.60 / 15.34 ms │ 14.74 / 15.42 ±0.39 / 15.88 ms │ 1.07x slower │
│ QQuery 7  │ 41.77 / 43.51 ±1.59 / 46.41 ms │ 45.53 / 47.71 ±2.42 / 52.29 ms │ 1.10x slower │
│ QQuery 8  │ 40.77 / 41.28 ±0.74 / 42.73 ms │ 41.91 / 42.73 ±1.01 / 44.71 ms │    no change │
│ QQuery 9  │ 49.93 / 50.77 ±0.61 / 51.56 ms │ 52.60 / 53.28 ±0.46 / 53.95 ms │    no change │
│ QQuery 10 │ 40.64 / 41.72 ±1.60 / 44.90 ms │ 41.08 / 41.82 ±1.18 / 44.16 ms │    no change │
│ QQuery 11 │ 13.76 / 13.88 ±0.12 / 14.11 ms │ 13.65 / 14.03 ±0.43 / 14.86 ms │    no change │
│ QQuery 12 │ 17.95 / 18.31 ±0.21 / 18.60 ms │ 18.22 / 18.69 ±0.57 / 19.80 ms │    no change │
│ QQuery 13 │ 40.99 / 41.99 ±1.24 / 44.35 ms │ 41.19 / 42.39 ±1.40 / 45.12 ms │    no change │
│ QQuery 14 │ 24.36 / 24.49 ±0.12 / 24.69 ms │ 24.73 / 24.84 ±0.08 / 24.95 ms │    no change │
│ QQuery 15 │ 31.22 / 31.95 ±0.84 / 33.32 ms │ 31.50 / 32.03 ±0.82 / 33.66 ms │    no change │
│ QQuery 16 │ 13.58 / 13.69 ±0.13 / 13.93 ms │ 13.35 / 13.59 ±0.14 / 13.78 ms │    no change │
│ QQuery 17 │ 71.29 / 72.37 ±1.05 / 74.30 ms │ 70.58 / 71.46 ±0.68 / 72.33 ms │    no change │
│ QQuery 18 │ 76.06 / 77.90 ±2.14 / 81.64 ms │ 81.78 / 85.54 ±3.30 / 90.97 ms │ 1.10x slower │
│ QQuery 19 │ 31.90 / 33.00 ±1.22 / 34.83 ms │ 32.13 / 34.00 ±2.47 / 38.32 ms │    no change │
│ QQuery 20 │ 31.98 / 33.32 ±1.99 / 37.27 ms │ 32.26 / 32.85 ±0.69 / 34.20 ms │    no change │
│ QQuery 21 │ 56.12 / 57.20 ±0.79 / 58.12 ms │ 59.95 / 61.77 ±1.46 / 63.25 ms │ 1.08x slower │
│ QQuery 22 │ 13.93 / 14.16 ±0.33 / 14.82 ms │ 13.91 / 14.25 ±0.50 / 15.24 ms │    no change │
└───────────┴────────────────────────────────┴────────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ Benchmark Summary                            ┃          ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 768.22ms │
│ Total Time (hash-join-global-bloom-filter)   │ 791.92ms │
│ Average Time (HEAD)                          │  34.92ms │
│ Average Time (hash-join-global-bloom-filter) │  36.00ms │
│ Queries Faster                               │        0 │
│ Queries Slower                               │        4 │
│ Queries with No Change                       │       18 │
│ Queries with Failure                         │        0 │
└──────────────────────────────────────────────┴──────────┘

Resource Usage

tpch — base (merge-base)

Metric Value
Wall time 5.0s
Peak memory 2.3 GiB
Avg memory 696.4 MiB
CPU user 21.6s
CPU sys 1.9s
Peak spill 0 B

tpch — branch

Metric Value
Wall time 5.0s
Peak memory 2.3 GiB
Avg memory 691.7 MiB
CPU user 22.3s
CPU sys 1.9s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃     HEAD ┃ hash-join-global-bloom-filter ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │ 40.46 ms │                      40.36 ms │    no change │
│ QQuery 2  │ 17.90 ms │                      17.98 ms │    no change │
│ QQuery 3  │ 28.05 ms │                      28.75 ms │    no change │
│ QQuery 4  │ 16.87 ms │                      17.01 ms │    no change │
│ QQuery 5  │ 34.74 ms │                      36.12 ms │    no change │
│ QQuery 6  │ 13.52 ms │                      13.72 ms │    no change │
│ QQuery 7  │ 41.75 ms │                      43.45 ms │    no change │
│ QQuery 8  │ 39.77 ms │                      41.48 ms │    no change │
│ QQuery 9  │ 49.35 ms │                      51.91 ms │ 1.05x slower │
│ QQuery 10 │ 39.95 ms │                      40.10 ms │    no change │
│ QQuery 11 │ 12.92 ms │                      13.03 ms │    no change │
│ QQuery 12 │ 18.10 ms │                      17.80 ms │    no change │
│ QQuery 13 │ 39.25 ms │                      39.43 ms │    no change │
│ QQuery 14 │ 24.13 ms │                      24.20 ms │    no change │
│ QQuery 15 │ 30.66 ms │                      30.69 ms │    no change │
│ QQuery 16 │ 13.05 ms │                      13.35 ms │    no change │
│ QQuery 17 │ 67.41 ms │                      66.70 ms │    no change │
│ QQuery 18 │ 74.05 ms │                      79.45 ms │ 1.07x slower │
│ QQuery 19 │ 31.31 ms │                      31.58 ms │    no change │
│ QQuery 20 │ 31.97 ms │                      31.67 ms │    no change │
│ QQuery 21 │ 55.54 ms │                      58.29 ms │    no change │
│ QQuery 22 │ 13.27 ms │                      13.42 ms │    no change │
└───────────┴──────────┴───────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ Benchmark Summary                            ┃          ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 734.01ms │
│ Total Time (hash-join-global-bloom-filter)   │ 750.50ms │
│ Average Time (HEAD)                          │  33.36ms │
│ Average Time (hash-join-global-bloom-filter) │  34.11ms │
│ Queries Faster                               │        0 │
│ Queries Slower                               │        2 │
│ Queries with No Change                       │       20 │
│ Queries with Failure                         │        0 │
└──────────────────────────────────────────────┴──────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃                           HEAD ┃  hash-join-global-bloom-filter ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │ 40.46 / 42.42 ±2.25 / 46.27 ms │ 40.36 / 41.17 ±1.21 / 43.58 ms │    no change │
│ QQuery 2  │ 17.90 / 18.48 ±0.65 / 19.63 ms │ 17.98 / 18.27 ±0.34 / 18.91 ms │    no change │
│ QQuery 3  │ 28.05 / 28.48 ±0.57 / 29.59 ms │ 28.75 / 29.42 ±1.02 / 31.45 ms │    no change │
│ QQuery 4  │ 16.87 / 17.46 ±0.64 / 18.68 ms │ 17.01 / 17.13 ±0.13 / 17.34 ms │    no change │
│ QQuery 5  │ 34.74 / 34.97 ±0.20 / 35.28 ms │ 36.12 / 36.81 ±1.04 / 38.88 ms │ 1.05x slower │
│ QQuery 6  │ 13.52 / 13.65 ±0.10 / 13.80 ms │ 13.72 / 13.86 ±0.15 / 14.16 ms │    no change │
│ QQuery 7  │ 41.75 / 43.19 ±0.76 / 43.84 ms │ 43.45 / 45.09 ±1.27 / 47.09 ms │    no change │
│ QQuery 8  │ 39.77 / 41.69 ±2.65 / 46.95 ms │ 41.48 / 42.15 ±0.93 / 43.98 ms │    no change │
│ QQuery 9  │ 49.35 / 50.79 ±1.01 / 52.36 ms │ 51.91 / 52.99 ±0.97 / 54.74 ms │    no change │
│ QQuery 10 │ 39.95 / 40.42 ±0.81 / 42.04 ms │ 40.10 / 40.31 ±0.28 / 40.85 ms │    no change │
│ QQuery 11 │ 12.92 / 13.42 ±0.50 / 14.34 ms │ 13.03 / 13.46 ±0.49 / 14.34 ms │    no change │
│ QQuery 12 │ 18.10 / 18.24 ±0.14 / 18.52 ms │ 17.80 / 18.34 ±0.55 / 19.35 ms │    no change │
│ QQuery 13 │ 39.25 / 40.07 ±1.18 / 42.41 ms │ 39.43 / 40.44 ±1.14 / 42.60 ms │    no change │
│ QQuery 14 │ 24.13 / 24.28 ±0.19 / 24.65 ms │ 24.20 / 24.52 ±0.18 / 24.75 ms │    no change │
│ QQuery 15 │ 30.66 / 31.75 ±0.77 / 32.65 ms │ 30.69 / 31.44 ±0.69 / 32.54 ms │    no change │
│ QQuery 16 │ 13.05 / 13.61 ±0.54 / 14.57 ms │ 13.35 / 13.51 ±0.11 / 13.68 ms │    no change │
│ QQuery 17 │ 67.41 / 68.67 ±1.07 / 70.46 ms │ 66.70 / 67.41 ±0.46 / 68.03 ms │    no change │
│ QQuery 18 │ 74.05 / 76.22 ±1.66 / 78.74 ms │ 79.45 / 82.78 ±2.74 / 86.19 ms │ 1.09x slower │
│ QQuery 19 │ 31.31 / 32.43 ±1.20 / 34.54 ms │ 31.58 / 31.68 ±0.06 / 31.76 ms │    no change │
│ QQuery 20 │ 31.97 / 33.41 ±2.42 / 38.24 ms │ 31.67 / 32.84 ±2.01 / 36.85 ms │    no change │
│ QQuery 21 │ 55.54 / 57.38 ±1.04 / 58.60 ms │ 58.29 / 58.74 ±0.38 / 59.36 ms │    no change │
│ QQuery 22 │ 13.27 / 13.39 ±0.16 / 13.71 ms │ 13.42 / 14.05 ±0.60 / 15.07 ms │    no change │
└───────────┴────────────────────────────────┴────────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ Benchmark Summary                            ┃          ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 754.45ms │
│ Total Time (hash-join-global-bloom-filter)   │ 766.41ms │
│ Average Time (HEAD)                          │  34.29ms │
│ Average Time (hash-join-global-bloom-filter) │  34.84ms │
│ Queries Faster                               │        0 │
│ Queries Slower                               │        2 │
│ Queries with No Change                       │       20 │
│ Queries with Failure                         │        0 │
└──────────────────────────────────────────────┴──────────┘

Resource Usage

tpch — base (merge-base)

Metric Value
Wall time 5.0s
Peak memory 2.3 GiB
Avg memory 676.9 MiB
CPU user 21.0s
CPU sys 1.8s
Peak spill 0 B

tpch — branch

Metric Value
Wall time 5.0s
Peak memory 2.4 GiB
Avg memory 705.3 MiB
CPU user 21.6s
CPU sys 1.6s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6628508073-4443-wn7f5 6.12.105+ #1 SMP Sat Sep 19 07:55:25 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch10

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpcds
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │    5.24 ms │                       5.31 ms │     no change │
│ QQuery 2  │   78.68 ms │                      78.07 ms │     no change │
│ QQuery 3  │   28.27 ms │                      28.51 ms │     no change │
│ QQuery 4  │  608.39 ms │                     607.01 ms │     no change │
│ QQuery 5  │   49.15 ms │                      49.53 ms │     no change │
│ QQuery 6  │   34.61 ms │                      34.47 ms │     no change │
│ QQuery 7  │   70.86 ms │                      71.11 ms │     no change │
│ QQuery 8  │   35.69 ms │                      35.77 ms │     no change │
│ QQuery 9  │   50.67 ms │                      52.61 ms │     no change │
│ QQuery 10 │   59.10 ms │                      59.01 ms │     no change │
│ QQuery 11 │  323.10 ms │                     328.95 ms │     no change │
│ QQuery 12 │   28.11 ms │                      28.32 ms │     no change │
│ QQuery 13 │  111.08 ms │                     111.41 ms │     no change │
│ QQuery 14 │  371.09 ms │                     373.74 ms │     no change │
│ QQuery 15 │   52.66 ms │                      52.99 ms │     no change │
│ QQuery 16 │    5.86 ms │                       5.96 ms │     no change │
│ QQuery 17 │   74.40 ms │                      76.05 ms │     no change │
│ QQuery 18 │  102.17 ms │                     102.46 ms │     no change │
│ QQuery 19 │   40.75 ms │                      40.44 ms │     no change │
│ QQuery 20 │   34.62 ms │                      34.89 ms │     no change │
│ QQuery 21 │   16.56 ms │                      16.60 ms │     no change │
│ QQuery 22 │   62.45 ms │                      63.32 ms │     no change │
│ QQuery 23 │  287.22 ms │                     289.90 ms │     no change │
│ QQuery 24 │  187.64 ms │                     193.37 ms │     no change │
│ QQuery 25 │  102.53 ms │                     105.19 ms │     no change │
│ QQuery 26 │   44.48 ms │                      44.38 ms │     no change │
│ QQuery 27 │    5.68 ms │                       5.71 ms │     no change │
│ QQuery 28 │   59.70 ms │                      56.75 ms │     no change │
│ QQuery 29 │   90.86 ms │                      92.53 ms │     no change │
│ QQuery 30 │   32.10 ms │                      31.51 ms │     no change │
│ QQuery 31 │  105.83 ms │                     106.17 ms │     no change │
│ QQuery 32 │   19.55 ms │                      19.30 ms │     no change │
│ QQuery 33 │   36.99 ms │                      37.50 ms │     no change │
│ QQuery 34 │    9.73 ms │                      10.02 ms │     no change │
│ QQuery 35 │   70.90 ms │                      71.83 ms │     no change │
│ QQuery 36 │    5.46 ms │                       5.38 ms │     no change │
│ QQuery 37 │    6.74 ms │                       6.49 ms │     no change │
│ QQuery 38 │   59.65 ms │                      60.52 ms │     no change │
│ QQuery 39 │   70.95 ms │                      70.67 ms │     no change │
│ QQuery 40 │   22.50 ms │                      22.16 ms │     no change │
│ QQuery 41 │   12.07 ms │                      11.66 ms │     no change │
│ QQuery 42 │   23.32 ms │                      23.00 ms │     no change │
│ QQuery 43 │    4.52 ms │                       4.44 ms │     no change │
│ QQuery 44 │    8.30 ms │                       8.18 ms │     no change │
│ QQuery 45 │   35.56 ms │                      35.48 ms │     no change │
│ QQuery 46 │   11.13 ms │                      11.76 ms │  1.06x slower │
│ QQuery 47 │  217.73 ms │                     218.48 ms │     no change │
│ QQuery 48 │   90.20 ms │                      90.40 ms │     no change │
│ QQuery 49 │   69.12 ms │                      68.76 ms │     no change │
│ QQuery 50 │   55.72 ms │                      56.36 ms │     no change │
│ QQuery 51 │   94.35 ms │                      93.40 ms │     no change │
│ QQuery 52 │   23.54 ms │                      23.47 ms │     no change │
│ QQuery 53 │   28.27 ms │                      28.70 ms │     no change │
│ QQuery 54 │   53.08 ms │                      52.00 ms │     no change │
│ QQuery 55 │   23.17 ms │                      23.17 ms │     no change │
│ QQuery 56 │   38.29 ms │                      37.93 ms │     no change │
│ QQuery 57 │  166.89 ms │                     166.28 ms │     no change │
│ QQuery 58 │  103.79 ms │                     104.72 ms │     no change │
│ QQuery 59 │  113.22 ms │                     113.26 ms │     no change │
│ QQuery 60 │   38.67 ms │                      38.99 ms │     no change │
│ QQuery 61 │    9.14 ms │                       9.11 ms │     no change │
│ QQuery 62 │   42.57 ms │                      42.75 ms │     no change │
│ QQuery 63 │   28.93 ms │                      28.97 ms │     no change │
│ QQuery 64 │  347.50 ms │                     349.63 ms │     no change │
│ QQuery 65 │   74.06 ms │                      66.91 ms │ +1.11x faster │
│ QQuery 66 │   71.82 ms │                      72.02 ms │     no change │
│ QQuery 67 │  246.60 ms │                     243.76 ms │     no change │
│ QQuery 68 │   11.30 ms │                      12.14 ms │  1.07x slower │
│ QQuery 69 │   54.64 ms │                      55.08 ms │     no change │
│ QQuery 70 │  101.38 ms │                     101.23 ms │     no change │
│ QQuery 71 │   34.25 ms │                      34.26 ms │     no change │
│ QQuery 72 │ 1665.01 ms │                    1628.47 ms │     no change │
│ QQuery 73 │    9.22 ms │                      10.00 ms │  1.08x slower │
│ QQuery 74 │  185.88 ms │                     186.70 ms │     no change │
│ QQuery 75 │  135.61 ms │                     135.87 ms │     no change │
│ QQuery 76 │   33.45 ms │                      34.28 ms │     no change │
│ QQuery 77 │   58.17 ms │                      58.77 ms │     no change │
│ QQuery 78 │  158.48 ms │                     159.76 ms │     no change │
│ QQuery 79 │   65.19 ms │                      66.00 ms │     no change │
│ QQuery 80 │   91.07 ms │                      92.98 ms │     no change │
│ QQuery 81 │   25.19 ms │                      25.36 ms │     no change │
│ QQuery 82 │   15.96 ms │                      15.78 ms │     no change │
│ QQuery 83 │   31.74 ms │                      31.98 ms │     no change │
│ QQuery 84 │   29.33 ms │                      29.50 ms │     no change │
│ QQuery 85 │   95.71 ms │                      95.74 ms │     no change │
│ QQuery 86 │   24.82 ms │                      24.54 ms │     no change │
│ QQuery 87 │   59.48 ms │                      61.57 ms │     no change │
│ QQuery 88 │   57.73 ms │                      58.02 ms │     no change │
│ QQuery 89 │   34.26 ms │                      34.57 ms │     no change │
│ QQuery 90 │   15.60 ms │                      15.82 ms │     no change │
│ QQuery 91 │   39.70 ms │                      39.89 ms │     no change │
│ QQuery 92 │   28.23 ms │                      28.04 ms │     no change │
│ QQuery 93 │   48.72 ms │                      48.10 ms │     no change │
│ QQuery 94 │   37.42 ms │                      36.84 ms │     no change │
│ QQuery 95 │   78.74 ms │                      76.81 ms │     no change │
│ QQuery 96 │   23.26 ms │                      23.36 ms │     no change │
│ QQuery 97 │   50.63 ms │                      49.52 ms │     no change │
│ QQuery 98 │   41.83 ms │                      42.59 ms │     no change │
│ QQuery 99 │   63.65 ms │                      64.00 ms │     no change │
└───────────┴────────────┴───────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 8899.29ms │
│ Total Time (hash-join-global-bloom-filter)   │ 8887.11ms │
│ Average Time (HEAD)                          │   89.89ms │
│ Average Time (hash-join-global-bloom-filter) │   89.77ms │
│ Queries Faster                               │         1 │
│ Queries Slower                               │         3 │
│ Queries with No Change                       │        95 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃         hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │           5.24 / 5.86 ±1.05 / 7.95 ms │           5.31 / 5.88 ±1.01 / 7.90 ms │     no change │
│ QQuery 2  │        78.68 / 78.88 ±0.18 / 79.22 ms │        78.07 / 78.26 ±0.21 / 78.65 ms │     no change │
│ QQuery 3  │        28.27 / 28.58 ±0.23 / 29.00 ms │        28.51 / 28.69 ±0.12 / 28.88 ms │     no change │
│ QQuery 4  │     608.39 / 613.87 ±4.05 / 618.74 ms │     607.01 / 614.58 ±4.32 / 618.62 ms │     no change │
│ QQuery 5  │        49.15 / 49.63 ±0.26 / 49.89 ms │        49.53 / 49.74 ±0.21 / 50.14 ms │     no change │
│ QQuery 6  │        34.61 / 37.66 ±2.57 / 42.16 ms │        34.47 / 35.41 ±0.56 / 36.11 ms │ +1.06x faster │
│ QQuery 7  │        70.86 / 72.29 ±1.05 / 73.77 ms │        71.11 / 72.45 ±0.74 / 73.06 ms │     no change │
│ QQuery 8  │        35.69 / 36.00 ±0.24 / 36.27 ms │        35.77 / 36.17 ±0.24 / 36.40 ms │     no change │
│ QQuery 9  │        50.67 / 53.61 ±2.53 / 57.11 ms │        52.61 / 53.34 ±0.62 / 54.35 ms │     no change │
│ QQuery 10 │        59.10 / 60.15 ±0.76 / 61.35 ms │        59.01 / 60.90 ±2.82 / 66.47 ms │     no change │
│ QQuery 11 │     323.10 / 326.88 ±4.23 / 334.97 ms │     328.95 / 333.93 ±3.62 / 340.20 ms │     no change │
│ QQuery 12 │        28.11 / 28.68 ±0.40 / 29.14 ms │        28.32 / 28.56 ±0.24 / 28.94 ms │     no change │
│ QQuery 13 │     111.08 / 111.61 ±0.30 / 111.99 ms │     111.41 / 113.16 ±2.37 / 117.79 ms │     no change │
│ QQuery 14 │     371.09 / 375.75 ±5.14 / 385.34 ms │     373.74 / 377.80 ±2.44 / 379.94 ms │     no change │
│ QQuery 15 │        52.66 / 53.17 ±0.28 / 53.48 ms │        52.99 / 54.25 ±0.96 / 55.93 ms │     no change │
│ QQuery 16 │           5.86 / 6.11 ±0.24 / 6.57 ms │           5.96 / 6.09 ±0.21 / 6.50 ms │     no change │
│ QQuery 17 │        74.40 / 75.02 ±0.47 / 75.71 ms │        76.05 / 76.51 ±0.39 / 77.09 ms │     no change │
│ QQuery 18 │     102.17 / 102.77 ±0.43 / 103.38 ms │     102.46 / 103.40 ±0.58 / 104.10 ms │     no change │
│ QQuery 19 │        40.75 / 41.07 ±0.26 / 41.49 ms │        40.44 / 41.29 ±1.19 / 43.62 ms │     no change │
│ QQuery 20 │        34.62 / 34.99 ±0.32 / 35.53 ms │        34.89 / 35.94 ±1.51 / 38.92 ms │     no change │
│ QQuery 21 │        16.56 / 17.04 ±0.40 / 17.73 ms │        16.60 / 16.86 ±0.31 / 17.44 ms │     no change │
│ QQuery 22 │        62.45 / 63.01 ±0.65 / 64.26 ms │        63.32 / 64.60 ±0.91 / 65.95 ms │     no change │
│ QQuery 23 │     287.22 / 288.83 ±1.43 / 290.67 ms │     289.90 / 292.32 ±1.71 / 295.19 ms │     no change │
│ QQuery 24 │     187.64 / 190.91 ±3.14 / 195.96 ms │     193.37 / 196.91 ±4.09 / 204.91 ms │     no change │
│ QQuery 25 │     102.53 / 103.50 ±0.80 / 104.78 ms │     105.19 / 107.39 ±1.84 / 110.56 ms │     no change │
│ QQuery 26 │        44.48 / 46.22 ±2.48 / 51.15 ms │        44.38 / 44.95 ±0.29 / 45.22 ms │     no change │
│ QQuery 27 │           5.68 / 5.82 ±0.15 / 6.10 ms │           5.71 / 5.87 ±0.15 / 6.08 ms │     no change │
│ QQuery 28 │        59.70 / 60.50 ±0.47 / 60.90 ms │        56.75 / 59.72 ±2.38 / 61.70 ms │     no change │
│ QQuery 29 │        90.86 / 93.28 ±3.06 / 99.31 ms │        92.53 / 93.86 ±1.47 / 96.00 ms │     no change │
│ QQuery 30 │        32.10 / 34.70 ±2.65 / 39.67 ms │        31.51 / 31.99 ±0.36 / 32.43 ms │ +1.08x faster │
│ QQuery 31 │     105.83 / 106.74 ±0.82 / 108.11 ms │     106.17 / 107.98 ±1.95 / 111.72 ms │     no change │
│ QQuery 32 │        19.55 / 20.55 ±1.21 / 22.92 ms │        19.30 / 19.51 ±0.15 / 19.73 ms │ +1.05x faster │
│ QQuery 33 │        36.99 / 37.61 ±0.35 / 37.95 ms │        37.50 / 37.58 ±0.09 / 37.73 ms │     no change │
│ QQuery 34 │          9.73 / 9.95 ±0.12 / 10.07 ms │        10.02 / 10.21 ±0.15 / 10.47 ms │     no change │
│ QQuery 35 │        70.90 / 72.02 ±0.82 / 73.10 ms │        71.83 / 72.79 ±0.74 / 73.98 ms │     no change │
│ QQuery 36 │           5.46 / 5.63 ±0.26 / 6.14 ms │           5.38 / 5.54 ±0.24 / 6.02 ms │     no change │
│ QQuery 37 │           6.74 / 6.82 ±0.08 / 6.97 ms │           6.49 / 6.56 ±0.07 / 6.69 ms │     no change │
│ QQuery 38 │        59.65 / 60.28 ±0.47 / 61.01 ms │        60.52 / 61.78 ±1.02 / 63.35 ms │     no change │
│ QQuery 39 │        70.95 / 73.26 ±3.56 / 80.34 ms │        70.67 / 71.34 ±0.44 / 71.98 ms │     no change │
│ QQuery 40 │        22.50 / 23.53 ±0.78 / 24.60 ms │        22.16 / 22.46 ±0.23 / 22.88 ms │     no change │
│ QQuery 41 │        12.07 / 12.21 ±0.17 / 12.54 ms │        11.66 / 13.06 ±2.45 / 17.95 ms │  1.07x slower │
│ QQuery 42 │        23.32 / 23.59 ±0.22 / 23.87 ms │        23.00 / 23.28 ±0.26 / 23.74 ms │     no change │
│ QQuery 43 │           4.52 / 4.65 ±0.19 / 5.04 ms │           4.44 / 4.56 ±0.18 / 4.92 ms │     no change │
│ QQuery 44 │           8.30 / 8.38 ±0.05 / 8.43 ms │           8.18 / 8.27 ±0.11 / 8.48 ms │     no change │
│ QQuery 45 │        35.56 / 36.28 ±0.44 / 36.77 ms │        35.48 / 37.11 ±1.41 / 39.71 ms │     no change │
│ QQuery 46 │        11.13 / 11.70 ±0.65 / 12.89 ms │        11.76 / 12.22 ±0.35 / 12.81 ms │     no change │
│ QQuery 47 │     217.73 / 220.47 ±2.30 / 223.89 ms │     218.48 / 222.97 ±3.66 / 228.18 ms │     no change │
│ QQuery 48 │        90.20 / 92.38 ±2.41 / 96.93 ms │        90.40 / 90.84 ±0.28 / 91.18 ms │     no change │
│ QQuery 49 │        69.12 / 69.73 ±0.47 / 70.50 ms │        68.76 / 70.45 ±1.87 / 74.01 ms │     no change │
│ QQuery 50 │        55.72 / 56.27 ±0.43 / 56.78 ms │        56.36 / 57.20 ±0.43 / 57.50 ms │     no change │
│ QQuery 51 │        94.35 / 95.84 ±1.23 / 97.98 ms │        93.40 / 95.58 ±1.86 / 98.21 ms │     no change │
│ QQuery 52 │        23.54 / 23.68 ±0.08 / 23.79 ms │        23.47 / 23.65 ±0.18 / 23.98 ms │     no change │
│ QQuery 53 │        28.27 / 28.68 ±0.38 / 29.37 ms │        28.70 / 30.13 ±2.51 / 35.14 ms │  1.05x slower │
│ QQuery 54 │        53.08 / 53.50 ±0.59 / 54.67 ms │        52.00 / 52.38 ±0.39 / 52.91 ms │     no change │
│ QQuery 55 │        23.17 / 23.48 ±0.18 / 23.67 ms │        23.17 / 23.55 ±0.50 / 24.54 ms │     no change │
│ QQuery 56 │        38.29 / 40.65 ±3.70 / 47.93 ms │        37.93 / 38.30 ±0.21 / 38.52 ms │ +1.06x faster │
│ QQuery 57 │     166.89 / 168.01 ±1.06 / 169.65 ms │     166.28 / 168.44 ±2.26 / 172.60 ms │     no change │
│ QQuery 58 │     103.79 / 105.23 ±1.30 / 107.38 ms │     104.72 / 106.04 ±1.13 / 108.00 ms │     no change │
│ QQuery 59 │     113.22 / 115.62 ±3.33 / 122.16 ms │     113.26 / 113.81 ±0.35 / 114.23 ms │     no change │
│ QQuery 60 │        38.67 / 39.38 ±0.48 / 40.03 ms │        38.99 / 39.42 ±0.33 / 39.95 ms │     no change │
│ QQuery 61 │           9.14 / 9.35 ±0.23 / 9.80 ms │           9.11 / 9.24 ±0.18 / 9.60 ms │     no change │
│ QQuery 62 │        42.57 / 42.81 ±0.15 / 42.94 ms │        42.75 / 43.98 ±1.30 / 46.39 ms │     no change │
│ QQuery 63 │        28.93 / 29.69 ±0.95 / 31.56 ms │        28.97 / 29.38 ±0.72 / 30.82 ms │     no change │
│ QQuery 64 │     347.50 / 354.33 ±5.56 / 363.92 ms │     349.63 / 353.78 ±2.85 / 356.67 ms │     no change │
│ QQuery 65 │        74.06 / 74.71 ±0.54 / 75.63 ms │        66.91 / 67.52 ±0.35 / 67.96 ms │ +1.11x faster │
│ QQuery 66 │        71.82 / 74.94 ±4.92 / 84.72 ms │        72.02 / 72.63 ±0.41 / 73.12 ms │     no change │
│ QQuery 67 │     246.60 / 251.20 ±5.38 / 258.96 ms │     243.76 / 248.44 ±5.16 / 256.19 ms │     no change │
│ QQuery 68 │        11.30 / 11.57 ±0.25 / 12.01 ms │        12.14 / 12.29 ±0.10 / 12.44 ms │  1.06x slower │
│ QQuery 69 │        54.64 / 55.36 ±0.58 / 56.30 ms │        55.08 / 56.11 ±1.44 / 58.94 ms │     no change │
│ QQuery 70 │     101.38 / 105.46 ±6.90 / 119.21 ms │     101.23 / 104.05 ±4.69 / 113.40 ms │     no change │
│ QQuery 71 │        34.25 / 34.88 ±0.80 / 36.39 ms │        34.26 / 34.92 ±0.50 / 35.65 ms │     no change │
│ QQuery 72 │ 1665.01 / 1682.43 ±10.93 / 1696.38 ms │ 1628.47 / 1671.20 ±32.97 / 1704.79 ms │     no change │
│ QQuery 73 │          9.22 / 9.67 ±0.59 / 10.82 ms │        10.00 / 10.24 ±0.30 / 10.82 ms │  1.06x slower │
│ QQuery 74 │     185.88 / 188.17 ±2.97 / 193.54 ms │     186.70 / 189.30 ±2.26 / 193.31 ms │     no change │
│ QQuery 75 │     135.61 / 138.34 ±3.69 / 145.49 ms │     135.87 / 138.54 ±4.15 / 146.81 ms │     no change │
│ QQuery 76 │        33.45 / 33.99 ±0.51 / 34.93 ms │        34.28 / 34.53 ±0.23 / 34.91 ms │     no change │
│ QQuery 77 │        58.17 / 61.01 ±4.47 / 69.90 ms │        58.77 / 59.04 ±0.21 / 59.31 ms │     no change │
│ QQuery 78 │     158.48 / 163.31 ±4.25 / 171.11 ms │     159.76 / 165.02 ±3.50 / 167.95 ms │     no change │
│ QQuery 79 │        65.19 / 66.25 ±0.70 / 67.19 ms │        66.00 / 66.22 ±0.20 / 66.52 ms │     no change │
│ QQuery 80 │        91.07 / 92.78 ±1.85 / 96.01 ms │       92.98 / 95.53 ±2.90 / 100.74 ms │     no change │
│ QQuery 81 │        25.19 / 25.81 ±1.05 / 27.90 ms │        25.36 / 25.54 ±0.23 / 25.97 ms │     no change │
│ QQuery 82 │        15.96 / 16.30 ±0.50 / 17.28 ms │        15.78 / 15.82 ±0.04 / 15.88 ms │     no change │
│ QQuery 83 │        31.74 / 33.78 ±3.53 / 40.84 ms │        31.98 / 32.34 ±0.34 / 32.79 ms │     no change │
│ QQuery 84 │        29.33 / 29.60 ±0.35 / 30.27 ms │        29.50 / 29.69 ±0.15 / 29.93 ms │     no change │
│ QQuery 85 │        95.71 / 97.10 ±1.13 / 98.67 ms │       95.74 / 99.36 ±4.46 / 107.91 ms │     no change │
│ QQuery 86 │        24.82 / 25.76 ±1.05 / 27.53 ms │        24.54 / 24.76 ±0.21 / 25.13 ms │     no change │
│ QQuery 87 │        59.48 / 61.48 ±2.02 / 65.24 ms │        61.57 / 62.00 ±0.48 / 62.93 ms │     no change │
│ QQuery 88 │        57.73 / 58.55 ±0.59 / 59.25 ms │        58.02 / 59.88 ±2.93 / 65.72 ms │     no change │
│ QQuery 89 │        34.26 / 34.49 ±0.18 / 34.81 ms │        34.57 / 35.03 ±0.37 / 35.65 ms │     no change │
│ QQuery 90 │        15.60 / 15.81 ±0.18 / 16.14 ms │        15.82 / 17.00 ±1.78 / 20.52 ms │  1.08x slower │
│ QQuery 91 │        39.70 / 39.93 ±0.17 / 40.23 ms │        39.89 / 40.41 ±0.57 / 41.46 ms │     no change │
│ QQuery 92 │        28.23 / 28.74 ±0.45 / 29.51 ms │        28.04 / 28.42 ±0.30 / 28.90 ms │     no change │
│ QQuery 93 │        48.72 / 50.50 ±2.27 / 54.89 ms │        48.10 / 48.86 ±0.92 / 50.65 ms │     no change │
│ QQuery 94 │        37.42 / 37.89 ±0.40 / 38.46 ms │        36.84 / 37.33 ±0.75 / 38.80 ms │     no change │
│ QQuery 95 │        78.74 / 79.86 ±1.09 / 81.24 ms │        76.81 / 78.17 ±1.26 / 80.16 ms │     no change │
│ QQuery 96 │        23.26 / 23.46 ±0.16 / 23.70 ms │        23.36 / 23.70 ±0.28 / 24.03 ms │     no change │
│ QQuery 97 │        50.63 / 51.50 ±0.99 / 53.34 ms │        49.52 / 50.64 ±1.22 / 52.20 ms │     no change │
│ QQuery 98 │        41.83 / 43.02 ±1.12 / 44.87 ms │        42.59 / 43.37 ±0.63 / 44.45 ms │     no change │
│ QQuery 99 │        63.65 / 64.47 ±1.03 / 66.46 ms │        64.00 / 64.57 ±0.84 / 66.24 ms │     no change │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 9040.82ms │
│ Total Time (hash-join-global-bloom-filter)   │ 9046.69ms │
│ Average Time (HEAD)                          │   91.32ms │
│ Average Time (hash-join-global-bloom-filter) │   91.38ms │
│ Queries Faster                               │         5 │
│ Queries Slower                               │         5 │
│ Queries with No Change                       │        89 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Resource Usage

tpcds — base (merge-base)

Metric Value
Wall time 50.0s
Peak memory 2.1 GiB
Avg memory 1.3 GiB
CPU user 188.7s
CPU sys 5.3s
Peak spill 0 B

tpcds — branch

Metric Value
Wall time 50.0s
Peak memory 1.9 GiB
Avg memory 1.2 GiB
CPU user 188.4s
CPU sys 5.1s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpcds
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │    5.05 ms │                       5.09 ms │     no change │
│ QQuery 2  │   78.57 ms │                      77.34 ms │     no change │
│ QQuery 3  │   27.77 ms │                      27.86 ms │     no change │
│ QQuery 4  │  603.42 ms │                     592.77 ms │     no change │
│ QQuery 5  │   48.95 ms │                      49.35 ms │     no change │
│ QQuery 6  │   34.16 ms │                      35.01 ms │     no change │
│ QQuery 7  │   71.12 ms │                      71.45 ms │     no change │
│ QQuery 8  │   35.50 ms │                      35.67 ms │     no change │
│ QQuery 9  │   50.52 ms │                      50.09 ms │     no change │
│ QQuery 10 │   59.23 ms │                      59.02 ms │     no change │
│ QQuery 11 │  323.37 ms │                     307.90 ms │     no change │
│ QQuery 12 │   28.48 ms │                      27.57 ms │     no change │
│ QQuery 13 │  111.03 ms │                     110.34 ms │     no change │
│ QQuery 14 │  367.10 ms │                     373.34 ms │     no change │
│ QQuery 15 │   52.30 ms │                      51.47 ms │     no change │
│ QQuery 16 │    5.95 ms │                       5.75 ms │     no change │
│ QQuery 17 │   73.68 ms │                      75.02 ms │     no change │
│ QQuery 18 │  101.16 ms │                     101.76 ms │     no change │
│ QQuery 19 │   40.16 ms │                      39.98 ms │     no change │
│ QQuery 20 │   34.53 ms │                      34.52 ms │     no change │
│ QQuery 21 │   16.45 ms │                      16.17 ms │     no change │
│ QQuery 22 │   61.65 ms │                      63.37 ms │     no change │
│ QQuery 23 │  287.18 ms │                     286.20 ms │     no change │
│ QQuery 24 │  186.11 ms │                     191.15 ms │     no change │
│ QQuery 25 │  103.33 ms │                     105.88 ms │     no change │
│ QQuery 26 │   44.55 ms │                      44.93 ms │     no change │
│ QQuery 27 │    5.83 ms │                       5.71 ms │     no change │
│ QQuery 28 │   55.47 ms │                      55.31 ms │     no change │
│ QQuery 29 │   89.95 ms │                      92.72 ms │     no change │
│ QQuery 30 │   32.62 ms │                      32.10 ms │     no change │
│ QQuery 31 │  104.27 ms │                     105.25 ms │     no change │
│ QQuery 32 │   19.49 ms │                      19.20 ms │     no change │
│ QQuery 33 │   37.19 ms │                      36.97 ms │     no change │
│ QQuery 34 │    9.78 ms │                       9.96 ms │     no change │
│ QQuery 35 │   69.40 ms │                      71.32 ms │     no change │
│ QQuery 36 │    5.44 ms │                       5.27 ms │     no change │
│ QQuery 37 │    6.48 ms │                       6.44 ms │     no change │
│ QQuery 38 │   58.87 ms │                      60.23 ms │     no change │
│ QQuery 39 │   70.91 ms │                      70.91 ms │     no change │
│ QQuery 40 │   22.01 ms │                      21.48 ms │     no change │
│ QQuery 41 │   12.15 ms │                      11.50 ms │ +1.06x faster │
│ QQuery 42 │   22.96 ms │                      22.70 ms │     no change │
│ QQuery 43 │    4.70 ms │                       4.42 ms │ +1.06x faster │
│ QQuery 44 │    8.65 ms │                       8.05 ms │ +1.07x faster │
│ QQuery 45 │   35.87 ms │                      34.98 ms │     no change │
│ QQuery 46 │   11.13 ms │                      11.51 ms │     no change │
│ QQuery 47 │  214.81 ms │                     217.21 ms │     no change │
│ QQuery 48 │   90.69 ms │                      89.46 ms │     no change │
│ QQuery 49 │   68.32 ms │                      68.64 ms │     no change │
│ QQuery 50 │   55.87 ms │                      56.40 ms │     no change │
│ QQuery 51 │   92.28 ms │                      91.56 ms │     no change │
│ QQuery 52 │   23.33 ms │                      23.32 ms │     no change │
│ QQuery 53 │   28.37 ms │                      28.36 ms │     no change │
│ QQuery 54 │   52.40 ms │                      51.64 ms │     no change │
│ QQuery 55 │   23.01 ms │                      23.40 ms │     no change │
│ QQuery 56 │   38.21 ms │                      38.31 ms │     no change │
│ QQuery 57 │  163.93 ms │                     166.19 ms │     no change │
│ QQuery 58 │  103.49 ms │                     103.90 ms │     no change │
│ QQuery 59 │  112.89 ms │                     112.86 ms │     no change │
│ QQuery 60 │   38.54 ms │                      38.85 ms │     no change │
│ QQuery 61 │    9.00 ms │                       8.95 ms │     no change │
│ QQuery 62 │   42.20 ms │                      42.75 ms │     no change │
│ QQuery 63 │   28.38 ms │                      28.39 ms │     no change │
│ QQuery 64 │  343.93 ms │                     346.59 ms │     no change │
│ QQuery 65 │   72.08 ms │                      66.55 ms │ +1.08x faster │
│ QQuery 66 │   72.21 ms │                      72.43 ms │     no change │
│ QQuery 67 │  242.51 ms │                     244.30 ms │     no change │
│ QQuery 68 │   11.33 ms │                      12.20 ms │  1.08x slower │
│ QQuery 69 │   54.88 ms │                      54.86 ms │     no change │
│ QQuery 70 │  102.15 ms │                     100.85 ms │     no change │
│ QQuery 71 │   33.91 ms │                      33.72 ms │     no change │
│ QQuery 72 │ 1546.76 ms │                    1677.08 ms │  1.08x slower │
│ QQuery 73 │    9.45 ms │                       9.57 ms │     no change │
│ QQuery 74 │  181.14 ms │                     180.25 ms │     no change │
│ QQuery 75 │  135.02 ms │                     134.45 ms │     no change │
│ QQuery 76 │   33.48 ms │                      33.94 ms │     no change │
│ QQuery 77 │   57.77 ms │                      58.11 ms │     no change │
│ QQuery 78 │  158.42 ms │                     159.13 ms │     no change │
│ QQuery 79 │   65.00 ms │                      65.48 ms │     no change │
│ QQuery 80 │   90.50 ms │                      91.10 ms │     no change │
│ QQuery 81 │   24.92 ms │                      25.46 ms │     no change │
│ QQuery 82 │   15.81 ms │                      15.65 ms │     no change │
│ QQuery 83 │   32.48 ms │                      31.59 ms │     no change │
│ QQuery 84 │   29.09 ms │                      28.88 ms │     no change │
│ QQuery 85 │   95.70 ms │                      95.92 ms │     no change │
│ QQuery 86 │   25.18 ms │                      24.52 ms │     no change │
│ QQuery 87 │   59.27 ms │                      61.29 ms │     no change │
│ QQuery 88 │   57.97 ms │                      57.80 ms │     no change │
│ QQuery 89 │   34.22 ms │                      33.98 ms │     no change │
│ QQuery 90 │   15.48 ms │                      15.54 ms │     no change │
│ QQuery 91 │   38.92 ms │                      39.43 ms │     no change │
│ QQuery 92 │   28.27 ms │                      27.51 ms │     no change │
│ QQuery 93 │   47.16 ms │                      47.66 ms │     no change │
│ QQuery 94 │   36.83 ms │                      35.99 ms │     no change │
│ QQuery 95 │   77.26 ms │                      76.50 ms │     no change │
│ QQuery 96 │   23.39 ms │                      23.39 ms │     no change │
│ QQuery 97 │   49.71 ms │                      49.16 ms │     no change │
│ QQuery 98 │   41.48 ms │                      41.23 ms │     no change │
│ QQuery 99 │   63.10 ms │                      64.02 ms │     no change │
└───────────┴────────────┴───────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 8726.99ms │
│ Total Time (hash-join-global-bloom-filter)   │ 8851.35ms │
│ Average Time (HEAD)                          │   88.15ms │
│ Average Time (hash-join-global-bloom-filter) │   89.41ms │
│ Queries Faster                               │         4 │
│ Queries Slower                               │         2 │
│ Queries with No Change                       │        93 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpcds_sf1.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃         hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │           5.05 / 5.70 ±1.08 / 7.84 ms │           5.09 / 5.68 ±0.98 / 7.64 ms │     no change │
│ QQuery 2  │        78.57 / 78.78 ±0.14 / 78.96 ms │        77.34 / 78.20 ±0.60 / 79.21 ms │     no change │
│ QQuery 3  │        27.77 / 28.06 ±0.16 / 28.25 ms │        27.86 / 28.21 ±0.19 / 28.38 ms │     no change │
│ QQuery 4  │    603.42 / 613.55 ±14.67 / 642.43 ms │    592.77 / 610.68 ±14.62 / 637.39 ms │     no change │
│ QQuery 5  │        48.95 / 49.71 ±0.58 / 50.69 ms │        49.35 / 49.54 ±0.24 / 50.00 ms │     no change │
│ QQuery 6  │        34.16 / 34.76 ±0.36 / 35.16 ms │        35.01 / 36.77 ±2.67 / 42.02 ms │  1.06x slower │
│ QQuery 7  │        71.12 / 71.62 ±0.51 / 72.56 ms │        71.45 / 72.16 ±0.69 / 73.47 ms │     no change │
│ QQuery 8  │        35.50 / 35.78 ±0.28 / 36.31 ms │        35.67 / 36.04 ±0.22 / 36.35 ms │     no change │
│ QQuery 9  │        50.52 / 52.22 ±0.90 / 52.95 ms │        50.09 / 51.77 ±1.37 / 53.67 ms │     no change │
│ QQuery 10 │        59.23 / 60.89 ±1.84 / 64.25 ms │        59.02 / 60.47 ±2.28 / 65.00 ms │     no change │
│ QQuery 11 │     323.37 / 330.97 ±7.60 / 344.02 ms │     307.90 / 315.46 ±6.19 / 325.99 ms │     no change │
│ QQuery 12 │        28.48 / 28.66 ±0.20 / 28.98 ms │        27.57 / 27.99 ±0.32 / 28.55 ms │     no change │
│ QQuery 13 │     111.03 / 111.82 ±0.98 / 113.43 ms │     110.34 / 112.13 ±2.43 / 116.94 ms │     no change │
│ QQuery 14 │     367.10 / 372.51 ±5.46 / 382.54 ms │     373.34 / 378.29 ±4.38 / 386.23 ms │     no change │
│ QQuery 15 │        52.30 / 53.76 ±1.05 / 55.05 ms │        51.47 / 52.41 ±0.94 / 53.65 ms │     no change │
│ QQuery 16 │           5.95 / 6.11 ±0.21 / 6.52 ms │           5.75 / 5.96 ±0.18 / 6.29 ms │     no change │
│ QQuery 17 │        73.68 / 75.04 ±2.24 / 79.52 ms │        75.02 / 75.79 ±0.61 / 76.87 ms │     no change │
│ QQuery 18 │     101.16 / 102.79 ±1.05 / 104.46 ms │     101.76 / 103.99 ±2.80 / 109.44 ms │     no change │
│ QQuery 19 │        40.16 / 40.59 ±0.39 / 41.32 ms │        39.98 / 40.50 ±0.45 / 41.07 ms │     no change │
│ QQuery 20 │        34.53 / 34.98 ±0.31 / 35.25 ms │        34.52 / 34.93 ±0.41 / 35.67 ms │     no change │
│ QQuery 21 │        16.45 / 16.71 ±0.36 / 17.40 ms │        16.17 / 16.57 ±0.34 / 17.16 ms │     no change │
│ QQuery 22 │        61.65 / 63.08 ±1.57 / 65.76 ms │        63.37 / 64.03 ±0.45 / 64.57 ms │     no change │
│ QQuery 23 │     287.18 / 288.61 ±1.81 / 292.18 ms │     286.20 / 290.36 ±5.22 / 300.63 ms │     no change │
│ QQuery 24 │     186.11 / 187.46 ±1.29 / 189.87 ms │     191.15 / 194.33 ±4.20 / 202.59 ms │     no change │
│ QQuery 25 │     103.33 / 105.05 ±2.53 / 110.05 ms │     105.88 / 107.37 ±1.27 / 108.99 ms │     no change │
│ QQuery 26 │        44.55 / 46.30 ±2.08 / 50.37 ms │        44.93 / 45.34 ±0.28 / 45.81 ms │     no change │
│ QQuery 27 │           5.83 / 5.96 ±0.14 / 6.22 ms │           5.71 / 5.83 ±0.17 / 6.17 ms │     no change │
│ QQuery 28 │        55.47 / 59.51 ±2.08 / 61.36 ms │        55.31 / 59.69 ±2.27 / 61.28 ms │     no change │
│ QQuery 29 │        89.95 / 91.89 ±1.66 / 94.25 ms │        92.72 / 94.60 ±1.60 / 97.01 ms │     no change │
│ QQuery 30 │        32.62 / 34.25 ±1.68 / 37.34 ms │        32.10 / 32.68 ±0.65 / 33.85 ms │     no change │
│ QQuery 31 │     104.27 / 105.43 ±0.93 / 107.10 ms │     105.25 / 106.00 ±0.61 / 106.99 ms │     no change │
│ QQuery 32 │        19.49 / 19.81 ±0.33 / 20.32 ms │        19.20 / 20.39 ±1.77 / 23.91 ms │     no change │
│ QQuery 33 │        37.19 / 37.86 ±0.59 / 38.82 ms │        36.97 / 38.79 ±2.38 / 43.46 ms │     no change │
│ QQuery 34 │         9.78 / 10.94 ±2.02 / 14.96 ms │         9.96 / 10.81 ±0.80 / 12.09 ms │     no change │
│ QQuery 35 │        69.40 / 70.74 ±1.15 / 72.61 ms │        71.32 / 72.43 ±0.97 / 74.13 ms │     no change │
│ QQuery 36 │           5.44 / 5.58 ±0.20 / 5.97 ms │           5.27 / 5.48 ±0.25 / 5.95 ms │     no change │
│ QQuery 37 │           6.48 / 6.57 ±0.07 / 6.68 ms │           6.44 / 6.57 ±0.09 / 6.70 ms │     no change │
│ QQuery 38 │        58.87 / 59.24 ±0.41 / 59.76 ms │        60.23 / 60.85 ±0.77 / 62.38 ms │     no change │
│ QQuery 39 │        70.91 / 73.13 ±2.29 / 77.28 ms │        70.91 / 73.30 ±3.58 / 80.41 ms │     no change │
│ QQuery 40 │        22.01 / 22.62 ±0.33 / 23.01 ms │        21.48 / 22.21 ±0.40 / 22.68 ms │     no change │
│ QQuery 41 │        12.15 / 12.34 ±0.20 / 12.72 ms │        11.50 / 11.67 ±0.16 / 11.97 ms │ +1.06x faster │
│ QQuery 42 │        22.96 / 23.36 ±0.45 / 24.20 ms │        22.70 / 23.03 ±0.24 / 23.39 ms │     no change │
│ QQuery 43 │           4.70 / 4.82 ±0.17 / 5.16 ms │           4.42 / 4.56 ±0.23 / 5.02 ms │ +1.06x faster │
│ QQuery 44 │           8.65 / 8.79 ±0.14 / 9.03 ms │           8.05 / 8.20 ±0.10 / 8.37 ms │ +1.07x faster │
│ QQuery 45 │        35.87 / 36.43 ±0.61 / 37.58 ms │        34.98 / 35.73 ±0.66 / 36.91 ms │     no change │
│ QQuery 46 │        11.13 / 11.41 ±0.43 / 12.25 ms │        11.51 / 11.73 ±0.15 / 11.91 ms │     no change │
│ QQuery 47 │     214.81 / 219.56 ±2.46 / 221.97 ms │     217.21 / 222.54 ±2.99 / 225.64 ms │     no change │
│ QQuery 48 │        90.69 / 91.29 ±0.45 / 91.95 ms │        89.46 / 90.39 ±0.67 / 91.32 ms │     no change │
│ QQuery 49 │        68.32 / 69.21 ±0.70 / 70.13 ms │        68.64 / 69.05 ±0.46 / 69.83 ms │     no change │
│ QQuery 50 │        55.87 / 57.81 ±2.48 / 62.64 ms │        56.40 / 57.12 ±0.97 / 59.00 ms │     no change │
│ QQuery 51 │        92.28 / 94.32 ±1.33 / 95.86 ms │        91.56 / 93.98 ±2.31 / 97.67 ms │     no change │
│ QQuery 52 │        23.33 / 23.74 ±0.22 / 23.92 ms │        23.32 / 23.47 ±0.10 / 23.64 ms │     no change │
│ QQuery 53 │        28.37 / 28.95 ±0.38 / 29.49 ms │        28.36 / 28.64 ±0.20 / 28.86 ms │     no change │
│ QQuery 54 │        52.40 / 53.95 ±2.60 / 59.11 ms │        51.64 / 52.32 ±1.03 / 54.36 ms │     no change │
│ QQuery 55 │        23.01 / 23.19 ±0.23 / 23.64 ms │        23.40 / 24.73 ±1.89 / 28.45 ms │  1.07x slower │
│ QQuery 56 │        38.21 / 38.90 ±0.54 / 39.74 ms │        38.31 / 38.91 ±0.77 / 40.39 ms │     no change │
│ QQuery 57 │     163.93 / 166.00 ±1.68 / 168.09 ms │     166.19 / 168.22 ±1.56 / 170.71 ms │     no change │
│ QQuery 58 │     103.49 / 104.75 ±1.02 / 106.12 ms │     103.90 / 106.06 ±1.68 / 108.31 ms │     no change │
│ QQuery 59 │     112.89 / 113.45 ±0.34 / 113.86 ms │     112.86 / 113.55 ±0.61 / 114.52 ms │     no change │
│ QQuery 60 │        38.54 / 38.99 ±0.36 / 39.45 ms │        38.85 / 39.13 ±0.32 / 39.73 ms │     no change │
│ QQuery 61 │           9.00 / 9.15 ±0.23 / 9.61 ms │           8.95 / 9.11 ±0.22 / 9.54 ms │     no change │
│ QQuery 62 │        42.20 / 42.53 ±0.22 / 42.80 ms │        42.75 / 42.87 ±0.13 / 43.11 ms │     no change │
│ QQuery 63 │        28.38 / 28.66 ±0.30 / 29.23 ms │        28.39 / 28.53 ±0.11 / 28.72 ms │     no change │
│ QQuery 64 │     343.93 / 348.52 ±2.61 / 351.68 ms │     346.59 / 352.26 ±3.27 / 356.28 ms │     no change │
│ QQuery 65 │        72.08 / 73.75 ±2.05 / 77.59 ms │        66.55 / 67.83 ±1.56 / 70.89 ms │ +1.09x faster │
│ QQuery 66 │        72.21 / 74.94 ±4.70 / 84.33 ms │        72.43 / 73.83 ±1.73 / 77.16 ms │     no change │
│ QQuery 67 │     242.51 / 250.07 ±4.68 / 257.24 ms │     244.30 / 248.06 ±4.11 / 254.82 ms │     no change │
│ QQuery 68 │        11.33 / 11.52 ±0.15 / 11.78 ms │        12.20 / 12.42 ±0.21 / 12.77 ms │  1.08x slower │
│ QQuery 69 │        54.88 / 55.34 ±0.41 / 56.03 ms │        54.86 / 55.22 ±0.28 / 55.54 ms │     no change │
│ QQuery 70 │     102.15 / 106.53 ±4.10 / 113.04 ms │     100.85 / 104.73 ±4.12 / 112.38 ms │     no change │
│ QQuery 71 │        33.91 / 34.53 ±0.71 / 35.83 ms │        33.72 / 34.20 ±0.44 / 34.78 ms │     no change │
│ QQuery 72 │ 1546.76 / 1724.22 ±99.14 / 1851.40 ms │ 1677.08 / 1733.27 ±34.42 / 1775.99 ms │     no change │
│ QQuery 73 │         9.45 / 11.37 ±1.87 / 14.77 ms │         9.57 / 11.17 ±1.08 / 12.51 ms │     no change │
│ QQuery 74 │     181.14 / 185.58 ±3.28 / 190.67 ms │     180.25 / 184.47 ±2.26 / 186.69 ms │     no change │
│ QQuery 75 │     135.02 / 135.68 ±0.54 / 136.64 ms │     134.45 / 135.24 ±0.69 / 136.51 ms │     no change │
│ QQuery 76 │        33.48 / 34.61 ±1.38 / 37.13 ms │        33.94 / 34.93 ±1.57 / 38.04 ms │     no change │
│ QQuery 77 │        57.77 / 58.27 ±0.65 / 59.52 ms │        58.11 / 58.90 ±0.89 / 60.57 ms │     no change │
│ QQuery 78 │     158.42 / 161.27 ±2.83 / 166.49 ms │     159.13 / 162.67 ±2.26 / 165.67 ms │     no change │
│ QQuery 79 │        65.00 / 66.53 ±2.38 / 71.27 ms │        65.48 / 66.06 ±0.42 / 66.72 ms │     no change │
│ QQuery 80 │        90.50 / 92.26 ±0.99 / 93.20 ms │        91.10 / 92.37 ±0.76 / 93.34 ms │     no change │
│ QQuery 81 │        24.92 / 26.07 ±1.82 / 29.68 ms │        25.46 / 25.75 ±0.19 / 26.04 ms │     no change │
│ QQuery 82 │        15.81 / 16.06 ±0.31 / 16.66 ms │        15.65 / 18.78 ±4.07 / 26.74 ms │  1.17x slower │
│ QQuery 83 │        32.48 / 34.37 ±3.21 / 40.79 ms │        31.59 / 31.94 ±0.48 / 32.86 ms │ +1.08x faster │
│ QQuery 84 │        29.09 / 29.66 ±0.79 / 31.22 ms │        28.88 / 29.41 ±0.50 / 30.16 ms │     no change │
│ QQuery 85 │        95.70 / 97.02 ±0.96 / 98.42 ms │        95.92 / 96.55 ±0.47 / 97.30 ms │     no change │
│ QQuery 86 │        25.18 / 25.53 ±0.25 / 25.90 ms │        24.52 / 27.45 ±4.73 / 36.88 ms │  1.08x slower │
│ QQuery 87 │        59.27 / 61.07 ±1.65 / 64.05 ms │        61.29 / 63.25 ±1.26 / 64.54 ms │     no change │
│ QQuery 88 │        57.97 / 58.48 ±0.30 / 58.87 ms │        57.80 / 58.07 ±0.24 / 58.34 ms │     no change │
│ QQuery 89 │        34.22 / 34.55 ±0.21 / 34.77 ms │        33.98 / 34.37 ±0.62 / 35.60 ms │     no change │
│ QQuery 90 │        15.48 / 15.69 ±0.31 / 16.29 ms │        15.54 / 16.33 ±0.95 / 18.21 ms │     no change │
│ QQuery 91 │        38.92 / 39.22 ±0.17 / 39.41 ms │        39.43 / 42.00 ±3.60 / 49.09 ms │  1.07x slower │
│ QQuery 92 │        28.27 / 30.07 ±1.93 / 33.81 ms │        27.51 / 28.22 ±0.55 / 29.15 ms │ +1.07x faster │
│ QQuery 93 │        47.16 / 47.94 ±0.79 / 49.13 ms │        47.66 / 48.73 ±0.98 / 50.57 ms │     no change │
│ QQuery 94 │        36.83 / 37.63 ±0.57 / 38.50 ms │        35.99 / 36.31 ±0.28 / 36.75 ms │     no change │
│ QQuery 95 │        77.26 / 78.96 ±1.34 / 80.52 ms │        76.50 / 78.54 ±1.83 / 81.85 ms │     no change │
│ QQuery 96 │        23.39 / 23.71 ±0.31 / 24.25 ms │        23.39 / 23.56 ±0.13 / 23.76 ms │     no change │
│ QQuery 97 │        49.71 / 51.86 ±2.35 / 56.40 ms │        49.16 / 50.66 ±1.12 / 52.32 ms │     no change │
│ QQuery 98 │        41.48 / 42.18 ±0.45 / 42.82 ms │        41.23 / 42.13 ±0.49 / 42.62 ms │     no change │
│ QQuery 99 │        63.10 / 63.92 ±0.91 / 65.69 ms │        64.02 / 64.60 ±0.72 / 66.02 ms │     no change │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 9039.64ms │
│ Total Time (hash-join-global-bloom-filter)   │ 9052.44ms │
│ Average Time (HEAD)                          │   91.31ms │
│ Average Time (hash-join-global-bloom-filter) │   91.44ms │
│ Queries Faster                               │         6 │
│ Queries Slower                               │         6 │
│ Queries with No Change                       │        87 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Resource Usage

tpcds — base (merge-base)

Metric Value
Wall time 50.0s
Peak memory 2.0 GiB
Avg memory 1.3 GiB
CPU user 184.4s
CPU sys 5.1s
Peak spill 0 B

tpcds — branch

Metric Value
Wall time 50.0s
Peak memory 2.0 GiB
Avg memory 1.3 GiB
CPU user 188.4s
CPU sys 5.0s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark clickbench_partitioned
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ hash-join-global-bloom-filter ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 0  │    1.24 ms │                       1.21 ms │    no change │
│ QQuery 1  │   11.67 ms │                      11.52 ms │    no change │
│ QQuery 2  │   34.58 ms │                      34.64 ms │    no change │
│ QQuery 3  │   30.26 ms │                      30.35 ms │    no change │
│ QQuery 4  │  236.39 ms │                     233.76 ms │    no change │
│ QQuery 5  │  272.02 ms │                     271.03 ms │    no change │
│ QQuery 6  │    1.26 ms │                       1.24 ms │    no change │
│ QQuery 7  │   12.81 ms │                      12.82 ms │    no change │
│ QQuery 8  │  328.55 ms │                     325.59 ms │    no change │
│ QQuery 9  │  448.33 ms │                     462.64 ms │    no change │
│ QQuery 10 │   63.77 ms │                      64.15 ms │    no change │
│ QQuery 11 │   74.48 ms │                      75.26 ms │    no change │
│ QQuery 12 │  257.47 ms │                     261.49 ms │    no change │
│ QQuery 13 │  357.75 ms │                     356.73 ms │    no change │
│ QQuery 14 │  281.69 ms │                     279.28 ms │    no change │
│ QQuery 15 │  280.16 ms │                     282.08 ms │    no change │
│ QQuery 16 │  613.51 ms │                     608.14 ms │    no change │
│ QQuery 17 │  626.04 ms │                     628.52 ms │    no change │
│ QQuery 18 │ 1234.23 ms │                    1224.23 ms │    no change │
│ QQuery 19 │   27.27 ms │                      27.27 ms │    no change │
│ QQuery 20 │  515.51 ms │                     515.48 ms │    no change │
│ QQuery 21 │  504.08 ms │                     508.95 ms │    no change │
│ QQuery 22 │  976.38 ms │                     980.87 ms │    no change │
│ QQuery 23 │ 3039.00 ms │                    3083.94 ms │    no change │
│ QQuery 24 │   40.01 ms │                      40.72 ms │    no change │
│ QQuery 25 │  104.63 ms │                     104.88 ms │    no change │
│ QQuery 26 │   40.63 ms │                      41.06 ms │    no change │
│ QQuery 27 │  505.77 ms │                     511.16 ms │    no change │
│ QQuery 28 │ 2856.82 ms │                    2878.23 ms │    no change │
│ QQuery 29 │   37.62 ms │                      38.01 ms │    no change │
│ QQuery 30 │  301.70 ms │                     302.90 ms │    no change │
│ QQuery 31 │  272.68 ms │                     280.99 ms │    no change │
│ QQuery 32 │  946.57 ms │                     947.81 ms │    no change │
│ QQuery 33 │ 1438.53 ms │                    1458.68 ms │    no change │
│ QQuery 34 │ 1465.36 ms │                    1480.27 ms │    no change │
│ QQuery 35 │  291.44 ms │                     294.88 ms │    no change │
│ QQuery 36 │   64.79 ms │                      66.41 ms │    no change │
│ QQuery 37 │   34.61 ms │                      36.53 ms │ 1.06x slower │
│ QQuery 38 │   40.06 ms │                      40.55 ms │    no change │
│ QQuery 39 │  136.10 ms │                     145.32 ms │ 1.07x slower │
│ QQuery 40 │   13.75 ms │                      14.24 ms │    no change │
│ QQuery 41 │   13.16 ms │                      13.21 ms │    no change │
│ QQuery 42 │   12.90 ms │                      12.87 ms │    no change │
└───────────┴────────────┴───────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 18845.58ms │
│ Total Time (hash-join-global-bloom-filter)   │ 18989.95ms │
│ Average Time (HEAD)                          │   438.27ms │
│ Average Time (hash-join-global-bloom-filter) │   441.63ms │
│ Queries Faster                               │          0 │
│ Queries Slower                               │          2 │
│ Queries with No Change                       │         41 │
│ Queries with Failure                         │          0 │
└──────────────────────────────────────────────┴────────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃         hash-join-global-bloom-filter ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 0  │           1.24 / 1.85 ±1.10 / 4.04 ms │           1.21 / 1.81 ±1.08 / 3.96 ms │    no change │
│ QQuery 1  │        11.67 / 11.78 ±0.08 / 11.87 ms │        11.52 / 11.57 ±0.04 / 11.62 ms │    no change │
│ QQuery 2  │        34.58 / 34.83 ±0.21 / 35.16 ms │        34.64 / 35.18 ±0.46 / 36.00 ms │    no change │
│ QQuery 3  │        30.26 / 30.87 ±0.68 / 32.18 ms │        30.35 / 30.72 ±0.48 / 31.65 ms │    no change │
│ QQuery 4  │     236.39 / 239.21 ±1.73 / 241.27 ms │     233.76 / 241.21 ±4.54 / 246.04 ms │    no change │
│ QQuery 5  │     272.02 / 277.88 ±3.66 / 282.09 ms │     271.03 / 274.63 ±3.19 / 280.55 ms │    no change │
│ QQuery 6  │           1.26 / 1.42 ±0.24 / 1.89 ms │           1.24 / 1.40 ±0.25 / 1.90 ms │    no change │
│ QQuery 7  │        12.81 / 13.00 ±0.17 / 13.21 ms │        12.82 / 12.94 ±0.09 / 13.07 ms │    no change │
│ QQuery 8  │     328.55 / 331.32 ±2.22 / 335.35 ms │     325.59 / 333.37 ±5.76 / 340.58 ms │    no change │
│ QQuery 9  │    448.33 / 470.28 ±14.61 / 491.33 ms │    462.64 / 475.11 ±18.38 / 511.66 ms │    no change │
│ QQuery 10 │        63.77 / 66.85 ±4.76 / 76.34 ms │        64.15 / 65.14 ±0.59 / 65.86 ms │    no change │
│ QQuery 11 │        74.48 / 77.84 ±5.61 / 89.04 ms │        75.26 / 79.32 ±4.89 / 88.64 ms │    no change │
│ QQuery 12 │     257.47 / 262.84 ±3.00 / 265.88 ms │     261.49 / 267.38 ±5.32 / 275.60 ms │    no change │
│ QQuery 13 │     357.75 / 365.59 ±5.78 / 373.31 ms │     356.73 / 367.40 ±8.25 / 377.64 ms │    no change │
│ QQuery 14 │     281.69 / 286.27 ±5.27 / 296.46 ms │     279.28 / 281.34 ±2.05 / 285.12 ms │    no change │
│ QQuery 15 │     280.16 / 292.57 ±9.93 / 304.82 ms │     282.08 / 288.47 ±6.33 / 299.95 ms │    no change │
│ QQuery 16 │     613.51 / 621.72 ±4.40 / 625.12 ms │     608.14 / 620.19 ±6.98 / 626.76 ms │    no change │
│ QQuery 17 │     626.04 / 631.23 ±3.85 / 635.53 ms │     628.52 / 638.51 ±6.35 / 645.05 ms │    no change │
│ QQuery 18 │  1234.23 / 1244.64 ±7.82 / 1253.69 ms │  1224.23 / 1234.64 ±8.72 / 1246.13 ms │    no change │
│ QQuery 19 │        27.27 / 27.78 ±0.66 / 29.05 ms │        27.27 / 29.63 ±4.17 / 37.96 ms │ 1.07x slower │
│ QQuery 20 │     515.51 / 524.62 ±9.53 / 542.95 ms │     515.48 / 525.60 ±6.43 / 535.25 ms │    no change │
│ QQuery 21 │     504.08 / 514.11 ±6.09 / 521.73 ms │     508.95 / 514.84 ±6.16 / 525.22 ms │    no change │
│ QQuery 22 │   976.38 / 987.73 ±14.26 / 1014.99 ms │   980.87 / 997.07 ±14.83 / 1016.63 ms │    no change │
│ QQuery 23 │ 3039.00 / 3097.38 ±33.24 / 3135.43 ms │ 3083.94 / 3122.51 ±22.62 / 3144.37 ms │    no change │
│ QQuery 24 │       40.01 / 47.72 ±14.67 / 77.05 ms │        40.72 / 48.54 ±9.16 / 62.43 ms │    no change │
│ QQuery 25 │     104.63 / 107.96 ±5.41 / 118.71 ms │     104.88 / 109.40 ±5.74 / 119.64 ms │    no change │
│ QQuery 26 │        40.63 / 40.94 ±0.24 / 41.35 ms │        41.06 / 44.99 ±6.30 / 57.55 ms │ 1.10x slower │
│ QQuery 27 │     505.77 / 511.46 ±4.42 / 518.75 ms │     511.16 / 519.16 ±5.73 / 528.68 ms │    no change │
│ QQuery 28 │ 2856.82 / 2894.83 ±29.23 / 2932.81 ms │ 2878.23 / 2903.93 ±16.27 / 2922.92 ms │    no change │
│ QQuery 29 │        37.62 / 42.67 ±7.51 / 57.51 ms │       38.01 / 47.88 ±13.00 / 71.04 ms │ 1.12x slower │
│ QQuery 30 │     301.70 / 309.26 ±5.42 / 318.08 ms │    302.90 / 313.15 ±10.48 / 332.82 ms │    no change │
│ QQuery 31 │    272.68 / 283.26 ±11.16 / 302.94 ms │     280.99 / 288.82 ±7.00 / 300.82 ms │    no change │
│ QQuery 32 │   946.57 / 978.12 ±22.96 / 1011.59 ms │    947.81 / 982.11 ±17.41 / 994.69 ms │    no change │
│ QQuery 33 │ 1438.53 / 1463.16 ±19.91 / 1485.00 ms │ 1458.68 / 1488.58 ±39.47 / 1561.98 ms │    no change │
│ QQuery 34 │ 1465.36 / 1504.83 ±42.81 / 1582.95 ms │ 1480.27 / 1500.57 ±18.37 / 1532.52 ms │    no change │
│ QQuery 35 │    291.44 / 301.93 ±12.63 / 326.20 ms │    294.88 / 331.51 ±40.35 / 382.55 ms │ 1.10x slower │
│ QQuery 36 │        64.79 / 69.01 ±4.42 / 77.54 ms │        66.41 / 74.04 ±6.96 / 85.88 ms │ 1.07x slower │
│ QQuery 37 │        34.61 / 37.73 ±3.33 / 43.01 ms │        36.53 / 40.72 ±3.86 / 47.02 ms │ 1.08x slower │
│ QQuery 38 │        40.06 / 43.46 ±1.81 / 45.31 ms │        40.55 / 43.58 ±3.34 / 49.71 ms │    no change │
│ QQuery 39 │    136.10 / 151.67 ±11.95 / 169.72 ms │     145.32 / 154.65 ±6.98 / 163.19 ms │    no change │
│ QQuery 40 │        13.75 / 15.74 ±2.46 / 20.56 ms │        14.24 / 16.37 ±3.97 / 24.29 ms │    no change │
│ QQuery 41 │        13.16 / 13.35 ±0.14 / 13.53 ms │        13.21 / 15.03 ±2.10 / 18.81 ms │ 1.13x slower │
│ QQuery 42 │        12.90 / 13.24 ±0.56 / 14.37 ms │        12.87 / 12.99 ±0.15 / 13.27 ms │    no change │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 19243.93ms │
│ Total Time (hash-join-global-bloom-filter)   │ 19385.97ms │
│ Average Time (HEAD)                          │   447.53ms │
│ Average Time (hash-join-global-bloom-filter) │   450.84ms │
│ Queries Faster                               │          0 │
│ Queries Slower                               │          7 │
│ Queries with No Change                       │         36 │
│ Queries with Failure                         │          0 │
└──────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric Value
Wall time 100.0s
Peak memory 11.3 GiB
Avg memory 4.4 GiB
CPU user 987.5s
CPU sys 68.7s
Peak spill 0 B

clickbench_partitioned — branch

Metric Value
Wall time 100.0s
Peak memory 11.5 GiB
Avg memory 4.6 GiB
CPU user 987.8s
CPU sys 71.7s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark clickbench_partitioned
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │    1.20 ms │                       1.21 ms │     no change │
│ QQuery 1  │   11.56 ms │                      11.41 ms │     no change │
│ QQuery 2  │   34.53 ms │                      34.61 ms │     no change │
│ QQuery 3  │   30.80 ms │                      30.29 ms │     no change │
│ QQuery 4  │  232.73 ms │                     232.07 ms │     no change │
│ QQuery 5  │  269.58 ms │                     268.59 ms │     no change │
│ QQuery 6  │    1.27 ms │                       1.26 ms │     no change │
│ QQuery 7  │   12.64 ms │                      12.59 ms │     no change │
│ QQuery 8  │  322.28 ms │                     324.25 ms │     no change │
│ QQuery 9  │  447.75 ms │                     459.60 ms │     no change │
│ QQuery 10 │   62.54 ms │                      63.68 ms │     no change │
│ QQuery 11 │   74.58 ms │                      74.70 ms │     no change │
│ QQuery 12 │  257.58 ms │                     257.24 ms │     no change │
│ QQuery 13 │  356.89 ms │                     360.57 ms │     no change │
│ QQuery 14 │  274.28 ms │                     277.19 ms │     no change │
│ QQuery 15 │  273.77 ms │                     284.41 ms │     no change │
│ QQuery 16 │  604.51 ms │                     606.37 ms │     no change │
│ QQuery 17 │  612.43 ms │                     621.24 ms │     no change │
│ QQuery 18 │ 1219.14 ms │                    1204.28 ms │     no change │
│ QQuery 19 │   27.16 ms │                      27.21 ms │     no change │
│ QQuery 20 │  514.46 ms │                     516.34 ms │     no change │
│ QQuery 21 │  505.92 ms │                     508.89 ms │     no change │
│ QQuery 22 │  971.88 ms │                     969.96 ms │     no change │
│ QQuery 23 │ 3002.43 ms │                    3048.10 ms │     no change │
│ QQuery 24 │   39.73 ms │                      39.83 ms │     no change │
│ QQuery 25 │  105.17 ms │                     104.05 ms │     no change │
│ QQuery 26 │   40.44 ms │                      40.30 ms │     no change │
│ QQuery 27 │  510.39 ms │                     510.40 ms │     no change │
│ QQuery 28 │ 2873.24 ms │                    2844.15 ms │     no change │
│ QQuery 29 │   37.41 ms │                      37.13 ms │     no change │
│ QQuery 30 │  298.90 ms │                     302.45 ms │     no change │
│ QQuery 31 │  269.70 ms │                     273.75 ms │     no change │
│ QQuery 32 │  945.33 ms │                     926.54 ms │     no change │
│ QQuery 33 │ 1410.52 ms │                    1420.90 ms │     no change │
│ QQuery 34 │ 1432.68 ms │                    1468.45 ms │     no change │
│ QQuery 35 │  280.70 ms │                     290.09 ms │     no change │
│ QQuery 36 │   69.92 ms │                      68.61 ms │     no change │
│ QQuery 37 │   35.58 ms │                      34.53 ms │     no change │
│ QQuery 38 │   41.34 ms │                      41.36 ms │     no change │
│ QQuery 39 │  151.77 ms │                     139.53 ms │ +1.09x faster │
│ QQuery 40 │   13.91 ms │                      13.62 ms │     no change │
│ QQuery 41 │   13.10 ms │                      13.27 ms │     no change │
│ QQuery 42 │   12.81 ms │                      12.91 ms │     no change │
└───────────┴────────────┴───────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 18704.56ms │
│ Total Time (hash-join-global-bloom-filter)   │ 18777.96ms │
│ Average Time (HEAD)                          │   434.99ms │
│ Average Time (hash-join-global-bloom-filter) │   436.70ms │
│ Queries Faster                               │          1 │
│ Queries Slower                               │          0 │
│ Queries with No Change                       │         42 │
│ Queries with Failure                         │          0 │
└──────────────────────────────────────────────┴────────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark clickbench_partitioned.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃         hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 0  │           1.20 / 1.80 ±1.09 / 3.98 ms │           1.21 / 1.81 ±1.11 / 4.03 ms │     no change │
│ QQuery 1  │        11.56 / 11.64 ±0.06 / 11.71 ms │        11.41 / 11.53 ±0.09 / 11.64 ms │     no change │
│ QQuery 2  │        34.53 / 34.83 ±0.21 / 35.15 ms │        34.61 / 34.79 ±0.16 / 35.04 ms │     no change │
│ QQuery 3  │        30.80 / 31.43 ±0.58 / 32.46 ms │        30.29 / 30.76 ±0.34 / 31.27 ms │     no change │
│ QQuery 4  │     232.73 / 236.40 ±3.93 / 242.50 ms │     232.07 / 238.15 ±4.25 / 245.18 ms │     no change │
│ QQuery 5  │     269.58 / 272.60 ±3.19 / 277.00 ms │     268.59 / 271.54 ±2.28 / 275.05 ms │     no change │
│ QQuery 6  │           1.27 / 1.42 ±0.24 / 1.90 ms │           1.26 / 1.41 ±0.23 / 1.86 ms │     no change │
│ QQuery 7  │        12.64 / 12.87 ±0.16 / 13.04 ms │        12.59 / 12.75 ±0.11 / 12.91 ms │     no change │
│ QQuery 8  │     322.28 / 328.71 ±4.13 / 334.04 ms │     324.25 / 327.06 ±2.36 / 330.17 ms │     no change │
│ QQuery 9  │     447.75 / 461.27 ±7.16 / 467.27 ms │     459.60 / 469.52 ±6.75 / 480.38 ms │     no change │
│ QQuery 10 │        62.54 / 63.56 ±0.67 / 64.28 ms │        63.68 / 65.34 ±2.54 / 70.39 ms │     no change │
│ QQuery 11 │        74.58 / 78.31 ±4.21 / 84.14 ms │        74.70 / 77.48 ±3.77 / 84.84 ms │     no change │
│ QQuery 12 │     257.58 / 263.34 ±4.83 / 269.85 ms │     257.24 / 262.88 ±3.50 / 267.92 ms │     no change │
│ QQuery 13 │    356.89 / 369.09 ±11.20 / 382.12 ms │    360.57 / 378.36 ±16.90 / 404.09 ms │     no change │
│ QQuery 14 │     274.28 / 282.38 ±7.78 / 293.37 ms │     277.19 / 284.60 ±8.50 / 300.23 ms │     no change │
│ QQuery 15 │     273.77 / 283.06 ±6.41 / 291.45 ms │     284.41 / 293.15 ±7.69 / 305.87 ms │     no change │
│ QQuery 16 │     604.51 / 614.18 ±7.55 / 624.42 ms │     606.37 / 617.96 ±8.25 / 629.78 ms │     no change │
│ QQuery 17 │     612.43 / 618.62 ±7.02 / 629.50 ms │     621.24 / 632.30 ±9.13 / 644.94 ms │     no change │
│ QQuery 18 │ 1219.14 / 1242.45 ±14.73 / 1263.35 ms │ 1204.28 / 1229.98 ±18.11 / 1254.76 ms │     no change │
│ QQuery 19 │        27.16 / 30.51 ±5.41 / 41.27 ms │        27.21 / 30.78 ±6.33 / 43.40 ms │     no change │
│ QQuery 20 │    514.46 / 525.48 ±12.22 / 546.92 ms │     516.34 / 518.50 ±1.57 / 520.40 ms │     no change │
│ QQuery 21 │     505.92 / 510.82 ±4.06 / 517.98 ms │     508.89 / 516.44 ±9.21 / 533.77 ms │     no change │
│ QQuery 22 │     971.88 / 984.35 ±7.94 / 991.48 ms │     969.96 / 977.67 ±5.07 / 984.92 ms │     no change │
│ QQuery 23 │ 3002.43 / 3073.26 ±49.18 / 3138.06 ms │ 3048.10 / 3097.46 ±27.39 / 3125.59 ms │     no change │
│ QQuery 24 │        39.73 / 44.33 ±4.58 / 50.44 ms │        39.83 / 41.50 ±1.78 / 44.05 ms │ +1.07x faster │
│ QQuery 25 │     105.17 / 111.87 ±7.99 / 125.76 ms │     104.05 / 105.85 ±1.40 / 107.36 ms │ +1.06x faster │
│ QQuery 26 │        40.44 / 40.60 ±0.20 / 40.98 ms │        40.30 / 40.85 ±0.73 / 42.27 ms │     no change │
│ QQuery 27 │    510.39 / 522.67 ±15.76 / 553.58 ms │     510.40 / 514.18 ±3.72 / 519.19 ms │     no change │
│ QQuery 28 │ 2873.24 / 2889.96 ±18.52 / 2915.36 ms │ 2844.15 / 2863.43 ±19.26 / 2900.14 ms │     no change │
│ QQuery 29 │      37.41 / 58.33 ±40.56 / 139.45 ms │       37.13 / 44.98 ±10.14 / 63.02 ms │ +1.30x faster │
│ QQuery 30 │     298.90 / 307.40 ±6.73 / 316.32 ms │    302.45 / 316.01 ±14.95 / 341.60 ms │     no change │
│ QQuery 31 │    269.70 / 282.24 ±10.61 / 299.64 ms │     273.75 / 281.79 ±5.29 / 290.04 ms │     no change │
│ QQuery 32 │    945.33 / 964.03 ±12.91 / 977.36 ms │   926.54 / 964.90 ±34.26 / 1013.31 ms │     no change │
│ QQuery 33 │ 1410.52 / 1463.64 ±37.30 / 1504.72 ms │ 1420.90 / 1468.82 ±36.95 / 1528.58 ms │     no change │
│ QQuery 34 │ 1432.68 / 1468.30 ±28.74 / 1506.34 ms │ 1468.45 / 1494.10 ±23.69 / 1532.96 ms │     no change │
│ QQuery 35 │    280.70 / 314.95 ±50.74 / 414.43 ms │    290.09 / 320.25 ±39.71 / 397.87 ms │     no change │
│ QQuery 36 │        69.92 / 77.93 ±6.63 / 87.21 ms │        68.61 / 74.71 ±6.57 / 87.29 ms │     no change │
│ QQuery 37 │        35.58 / 38.19 ±3.85 / 45.85 ms │        34.53 / 35.77 ±1.75 / 39.24 ms │ +1.07x faster │
│ QQuery 38 │        41.34 / 41.82 ±0.37 / 42.27 ms │        41.36 / 42.70 ±1.14 / 44.48 ms │     no change │
│ QQuery 39 │     151.77 / 159.65 ±6.48 / 167.67 ms │     139.53 / 152.46 ±7.41 / 162.15 ms │     no change │
│ QQuery 40 │        13.91 / 14.16 ±0.26 / 14.64 ms │        13.62 / 13.93 ±0.28 / 14.46 ms │     no change │
│ QQuery 41 │        13.10 / 13.39 ±0.16 / 13.53 ms │        13.27 / 13.39 ±0.11 / 13.56 ms │     no change │
│ QQuery 42 │        12.81 / 12.98 ±0.17 / 13.29 ms │        12.91 / 15.48 ±4.98 / 25.43 ms │  1.19x slower │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 19158.80ms │
│ Total Time (hash-join-global-bloom-filter)   │ 19187.32ms │
│ Average Time (HEAD)                          │   445.55ms │
│ Average Time (hash-join-global-bloom-filter) │   446.22ms │
│ Queries Faster                               │          4 │
│ Queries Slower                               │          1 │
│ Queries with No Change                       │         38 │
│ Queries with Failure                         │          0 │
└──────────────────────────────────────────────┴────────────┘

Resource Usage

clickbench_partitioned — base (merge-base)

Metric Value
Wall time 100.0s
Peak memory 12.2 GiB
Avg memory 4.6 GiB
CPU user 984.4s
CPU sys 67.8s
Peak spill 0 B

clickbench_partitioned — branch

Metric Value
Wall time 100.0s
Peak memory 12.7 GiB
Avg memory 4.7 GiB
CPU user 981.8s
CPU sys 69.0s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch10
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃      HEAD ┃ hash-join-global-bloom-filter ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │ 329.60 ms │                     330.30 ms │    no change │
│ QQuery 2  │  88.57 ms │                      89.48 ms │    no change │
│ QQuery 3  │ 215.85 ms │                     216.33 ms │    no change │
│ QQuery 4  │ 114.35 ms │                     115.35 ms │    no change │
│ QQuery 5  │ 342.52 ms │                     342.76 ms │    no change │
│ QQuery 6  │ 100.29 ms │                     101.33 ms │    no change │
│ QQuery 7  │ 442.65 ms │                     460.53 ms │    no change │
│ QQuery 8  │ 357.15 ms │                     368.95 ms │    no change │
│ QQuery 9  │ 511.71 ms │                     519.32 ms │    no change │
│ QQuery 10 │ 284.86 ms │                     294.97 ms │    no change │
│ QQuery 11 │  61.83 ms │                      62.74 ms │    no change │
│ QQuery 12 │ 118.11 ms │                     120.34 ms │    no change │
│ QQuery 13 │ 292.49 ms │                     307.03 ms │    no change │
│ QQuery 14 │ 169.11 ms │                     169.94 ms │    no change │
│ QQuery 15 │ 298.93 ms │                     298.22 ms │    no change │
│ QQuery 16 │  65.31 ms │                      63.49 ms │    no change │
│ QQuery 17 │ 575.72 ms │                     575.25 ms │    no change │
│ QQuery 18 │ 916.93 ms │                    1004.81 ms │ 1.10x slower │
│ QQuery 19 │ 231.47 ms │                     232.46 ms │    no change │
│ QQuery 20 │ 270.97 ms │                     262.70 ms │    no change │
│ QQuery 21 │ 667.34 ms │                     671.12 ms │    no change │
│ QQuery 22 │  60.02 ms │                      59.61 ms │    no change │
└───────────┴───────────┴───────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 6515.77ms │
│ Total Time (hash-join-global-bloom-filter)   │ 6667.02ms │
│ Average Time (HEAD)                          │  296.17ms │
│ Average Time (hash-join-global-bloom-filter) │  303.05ms │
│ Queries Faster                               │         0 │
│ Queries Slower                               │         1 │
│ Queries with No Change                       │        21 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃                               HEAD ┃         hash-join-global-bloom-filter ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │  329.60 / 332.06 ±2.22 / 335.00 ms │     330.30 / 332.18 ±1.41 / 334.67 ms │    no change │
│ QQuery 2  │     88.57 / 89.71 ±0.92 / 91.12 ms │        89.48 / 90.83 ±0.88 / 92.18 ms │    no change │
│ QQuery 3  │  215.85 / 218.94 ±1.64 / 220.50 ms │     216.33 / 221.89 ±4.36 / 227.20 ms │    no change │
│ QQuery 4  │  114.35 / 116.65 ±2.57 / 121.09 ms │     115.35 / 116.39 ±0.70 / 117.12 ms │    no change │
│ QQuery 5  │  342.52 / 351.68 ±7.25 / 360.38 ms │     342.76 / 354.39 ±8.29 / 366.23 ms │    no change │
│ QQuery 6  │  100.29 / 103.18 ±3.13 / 109.21 ms │     101.33 / 103.66 ±2.45 / 108.39 ms │    no change │
│ QQuery 7  │ 442.65 / 463.19 ±16.66 / 493.27 ms │     460.53 / 473.04 ±9.13 / 486.79 ms │    no change │
│ QQuery 8  │  357.15 / 361.73 ±6.05 / 373.69 ms │     368.95 / 378.00 ±9.26 / 393.15 ms │    no change │
│ QQuery 9  │  511.71 / 520.22 ±6.45 / 527.05 ms │    519.32 / 538.65 ±22.88 / 583.50 ms │    no change │
│ QQuery 10 │ 284.86 / 301.38 ±13.85 / 319.11 ms │     294.97 / 302.44 ±6.18 / 310.71 ms │    no change │
│ QQuery 11 │     61.83 / 63.41 ±1.04 / 65.05 ms │        62.74 / 63.49 ±0.71 / 64.74 ms │    no change │
│ QQuery 12 │ 118.11 / 126.31 ±13.28 / 152.81 ms │     120.34 / 125.78 ±6.25 / 136.79 ms │    no change │
│ QQuery 13 │ 292.49 / 307.83 ±10.53 / 324.95 ms │     307.03 / 317.47 ±8.09 / 328.92 ms │    no change │
│ QQuery 14 │  169.11 / 174.11 ±8.77 / 191.63 ms │     169.94 / 173.09 ±3.71 / 179.60 ms │    no change │
│ QQuery 15 │  298.93 / 303.53 ±5.29 / 313.38 ms │     298.22 / 303.78 ±6.71 / 316.93 ms │    no change │
│ QQuery 16 │     65.31 / 66.19 ±0.95 / 67.97 ms │        63.49 / 65.35 ±2.78 / 70.89 ms │    no change │
│ QQuery 17 │ 575.72 / 589.67 ±10.92 / 601.22 ms │    575.25 / 588.02 ±11.80 / 602.81 ms │    no change │
│ QQuery 18 │ 916.93 / 954.12 ±29.86 / 990.13 ms │ 1004.81 / 1032.90 ±28.61 / 1086.20 ms │ 1.08x slower │
│ QQuery 19 │ 231.47 / 248.65 ±16.56 / 277.68 ms │    232.46 / 257.18 ±20.49 / 289.88 ms │    no change │
│ QQuery 20 │  270.97 / 278.04 ±3.86 / 281.43 ms │    262.70 / 274.53 ±11.30 / 294.46 ms │    no change │
│ QQuery 21 │  667.34 / 673.89 ±6.32 / 685.20 ms │    671.12 / 686.28 ±10.43 / 703.00 ms │    no change │
│ QQuery 22 │     60.02 / 61.88 ±0.99 / 62.86 ms │        59.61 / 65.92 ±4.03 / 69.56 ms │ 1.07x slower │
└───────────┴────────────────────────────────────┴───────────────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 6706.37ms │
│ Total Time (hash-join-global-bloom-filter)   │ 6865.24ms │
│ Average Time (HEAD)                          │  304.84ms │
│ Average Time (hash-join-global-bloom-filter) │  312.06ms │
│ Queries Faster                               │         0 │
│ Queries Slower                               │         2 │
│ Queries with No Change                       │        20 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Resource Usage

tpch10 — base (merge-base)

Metric Value
Wall time 35.0s
Peak memory 14.5 GiB
Avg memory 2.9 GiB
CPU user 339.7s
CPU sys 19.9s
Peak spill 0 B

tpch10 — branch

Metric Value
Wall time 35.0s
Peak memory 12.7 GiB
Avg memory 2.9 GiB
CPU user 338.3s
CPU sys 20.2s
Peak spill 0 B

File an issue against this benchmark runner

@adriangb

adriangb commented Nov 4, 2026

Copy link
Copy Markdown
Contributor Author

run benchmark tpch10

env:
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: 4
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: false

@adriangb

adriangb commented Nov 4, 2026

Copy link
Copy Markdown
Contributor Author

run benchmark tpch10

env:
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: 4
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: true

@adriangb

adriangb commented Nov 4, 2026

Copy link
Copy Markdown
Contributor Author

run benchmark tpch10

env:
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: 16
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: false

@adriangb

adriangb commented Nov 4, 2026

Copy link
Copy Markdown
Contributor Author

run benchmark tpch10

env:
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: 16
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: true

@adriangb

adriangb commented Nov 4, 2026

Copy link
Copy Markdown
Contributor Author

run benchmark tpch10

env:
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: 64
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: false

@adriangb

adriangb commented Nov 4, 2026

Copy link
Copy Markdown
Contributor Author

run benchmark tpch10

env:
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: 64
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: true

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6628799277-4444-n8qs2 6.12.105+ #1 SMP Sat Sep 19 07:55:25 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "false"
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: "4"

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6628801196-4447-pt26r 6.12.105+ #1 SMP Sat Sep 19 07:55:25 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "false"
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: "64"

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6628799719-4446-gj29c 6.12.105+ #1 SMP Sat Sep 19 07:55:25 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: "4"

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6628800662-4448-27vxr 6.12.105+ #1 SMP Sat Sep 19 07:55:25 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: "16"

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6628800186-4445-b48bf 6.12.105+ #1 SMP Sat Sep 19 07:55:25 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "false"
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: "16"

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "false"
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: "64"
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃      HEAD ┃ hash-join-global-bloom-filter ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │ 336.86 ms │                     334.71 ms │    no change │
│ QQuery 2  │ 110.38 ms │                     110.94 ms │    no change │
│ QQuery 3  │ 258.97 ms │                     264.99 ms │    no change │
│ QQuery 4  │ 134.42 ms │                     137.69 ms │    no change │
│ QQuery 5  │ 382.80 ms │                     419.08 ms │ 1.09x slower │
│ QQuery 6  │ 100.74 ms │                      99.68 ms │    no change │
│ QQuery 7  │ 448.79 ms │                     456.53 ms │    no change │
│ QQuery 8  │ 377.37 ms │                     423.77 ms │ 1.12x slower │
│ QQuery 9  │ 516.73 ms │                     513.85 ms │    no change │
│ QQuery 10 │ 307.15 ms │                     315.70 ms │    no change │
│ QQuery 11 │  73.12 ms │                      73.99 ms │    no change │
│ QQuery 12 │ 126.07 ms │                     133.47 ms │ 1.06x slower │
│ QQuery 13 │ 244.34 ms │                     257.90 ms │ 1.06x slower │
│ QQuery 14 │ 178.60 ms │                     184.09 ms │    no change │
│ QQuery 15 │ 299.88 ms │                     307.54 ms │    no change │
│ QQuery 16 │  79.18 ms │                      83.30 ms │ 1.05x slower │
│ QQuery 17 │ 468.77 ms │                     463.44 ms │    no change │
│ QQuery 18 │ 853.79 ms │                     870.50 ms │    no change │
│ QQuery 19 │ 237.05 ms │                     241.15 ms │    no change │
│ QQuery 20 │ 260.79 ms │                     259.35 ms │    no change │
│ QQuery 21 │ 780.61 ms │                     786.99 ms │    no change │
│ QQuery 22 │  76.81 ms │                      79.50 ms │    no change │
└───────────┴───────────┴───────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 6653.22ms │
│ Total Time (hash-join-global-bloom-filter)   │ 6818.15ms │
│ Average Time (HEAD)                          │  302.42ms │
│ Average Time (hash-join-global-bloom-filter) │  309.92ms │
│ Queries Faster                               │         0 │
│ Queries Slower                               │         5 │
│ Queries with No Change                       │        17 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃                               HEAD ┃      hash-join-global-bloom-filter ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │  336.86 / 340.30 ±1.94 / 342.65 ms │  334.71 / 341.43 ±4.10 / 347.33 ms │    no change │
│ QQuery 2  │  110.38 / 111.68 ±0.85 / 113.03 ms │  110.94 / 111.73 ±0.66 / 112.63 ms │    no change │
│ QQuery 3  │  258.97 / 262.30 ±3.29 / 266.36 ms │  264.99 / 269.81 ±5.25 / 279.97 ms │    no change │
│ QQuery 4  │  134.42 / 136.43 ±1.82 / 139.24 ms │  137.69 / 139.45 ±2.20 / 143.61 ms │    no change │
│ QQuery 5  │  382.80 / 390.38 ±4.34 / 396.03 ms │  419.08 / 424.08 ±3.33 / 428.88 ms │ 1.09x slower │
│ QQuery 6  │  100.74 / 102.08 ±1.08 / 103.75 ms │   99.68 / 100.98 ±1.11 / 102.79 ms │    no change │
│ QQuery 7  │  448.79 / 461.73 ±7.06 / 469.21 ms │  456.53 / 471.05 ±7.98 / 480.31 ms │    no change │
│ QQuery 8  │  377.37 / 385.53 ±5.05 / 392.14 ms │ 423.77 / 437.92 ±11.75 / 452.63 ms │ 1.14x slower │
│ QQuery 9  │ 516.73 / 529.45 ±10.89 / 542.96 ms │ 513.85 / 530.79 ±16.38 / 550.47 ms │    no change │
│ QQuery 10 │  307.15 / 316.09 ±7.33 / 326.95 ms │ 315.70 / 328.23 ±18.41 / 364.76 ms │    no change │
│ QQuery 11 │     73.12 / 74.12 ±1.10 / 76.25 ms │     73.99 / 76.48 ±2.80 / 81.49 ms │    no change │
│ QQuery 12 │  126.07 / 133.66 ±6.95 / 145.51 ms │ 133.47 / 139.28 ±10.34 / 159.95 ms │    no change │
│ QQuery 13 │  244.34 / 251.36 ±5.77 / 259.25 ms │  257.90 / 269.60 ±7.47 / 277.26 ms │ 1.07x slower │
│ QQuery 14 │  178.60 / 182.70 ±3.72 / 188.95 ms │  184.09 / 193.98 ±9.52 / 211.85 ms │ 1.06x slower │
│ QQuery 15 │  299.88 / 305.85 ±4.52 / 312.89 ms │  307.54 / 312.11 ±3.02 / 315.23 ms │    no change │
│ QQuery 16 │     79.18 / 82.40 ±2.95 / 87.91 ms │     83.30 / 84.97 ±1.99 / 88.85 ms │    no change │
│ QQuery 17 │  468.77 / 475.51 ±5.35 / 483.19 ms │ 463.44 / 475.25 ±11.62 / 495.87 ms │    no change │
│ QQuery 18 │ 853.79 / 890.75 ±23.71 / 925.67 ms │ 870.50 / 889.67 ±15.00 / 909.30 ms │    no change │
│ QQuery 19 │ 237.05 / 248.58 ±10.60 / 266.02 ms │ 241.15 / 250.90 ±15.13 / 280.86 ms │    no change │
│ QQuery 20 │  260.79 / 263.10 ±2.23 / 266.83 ms │  259.35 / 266.85 ±5.74 / 277.10 ms │    no change │
│ QQuery 21 │  780.61 / 790.11 ±8.81 / 802.28 ms │  786.99 / 794.59 ±5.85 / 804.90 ms │    no change │
│ QQuery 22 │     76.81 / 82.42 ±5.68 / 92.12 ms │     79.50 / 84.05 ±2.46 / 86.63 ms │    no change │
└───────────┴────────────────────────────────────┴────────────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 6816.53ms │
│ Total Time (hash-join-global-bloom-filter)   │ 6993.23ms │
│ Average Time (HEAD)                          │  309.84ms │
│ Average Time (hash-join-global-bloom-filter) │  317.87ms │
│ Queries Faster                               │         0 │
│ Queries Slower                               │         4 │
│ Queries with No Change                       │        18 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Resource Usage

tpch10 — base (merge-base)

Metric Value
Wall time 35.0s
Peak memory 15.8 GiB
Avg memory 4.3 GiB
CPU user 344.2s
CPU sys 21.1s
Peak spill 0 B

tpch10 — branch

Metric Value
Wall time 40.0s
Peak memory 15.9 GiB
Avg memory 3.8 GiB
CPU user 348.4s
CPU sys 22.0s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "false"
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: "16"
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃      HEAD ┃ hash-join-global-bloom-filter ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │ 324.67 ms │                     326.22 ms │    no change │
│ QQuery 2  │  91.57 ms │                      92.23 ms │    no change │
│ QQuery 3  │ 220.75 ms │                     224.60 ms │    no change │
│ QQuery 4  │ 106.55 ms │                     110.67 ms │    no change │
│ QQuery 5  │ 336.70 ms │                     347.62 ms │    no change │
│ QQuery 6  │  96.19 ms │                      96.16 ms │    no change │
│ QQuery 7  │ 443.05 ms │                     451.03 ms │    no change │
│ QQuery 8  │ 356.53 ms │                     375.90 ms │ 1.05x slower │
│ QQuery 9  │ 554.55 ms │                     561.59 ms │    no change │
│ QQuery 10 │ 289.90 ms │                     293.55 ms │    no change │
│ QQuery 11 │  60.13 ms │                      61.11 ms │    no change │
│ QQuery 12 │ 109.51 ms │                     110.98 ms │    no change │
│ QQuery 13 │ 286.39 ms │                     294.07 ms │    no change │
│ QQuery 14 │ 163.48 ms │                     164.87 ms │    no change │
│ QQuery 15 │ 288.58 ms │                     290.44 ms │    no change │
│ QQuery 16 │  66.38 ms │                      65.71 ms │    no change │
│ QQuery 17 │ 579.62 ms │                     571.20 ms │    no change │
│ QQuery 18 │ 977.32 ms │                    1010.92 ms │    no change │
│ QQuery 19 │ 227.50 ms │                     225.88 ms │    no change │
│ QQuery 20 │ 269.59 ms │                     268.02 ms │    no change │
│ QQuery 21 │ 687.13 ms │                     686.39 ms │    no change │
│ QQuery 22 │  68.42 ms │                      68.06 ms │    no change │
└───────────┴───────────┴───────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 6604.50ms │
│ Total Time (hash-join-global-bloom-filter)   │ 6697.23ms │
│ Average Time (HEAD)                          │  300.20ms │
│ Average Time (hash-join-global-bloom-filter) │  304.42ms │
│ Queries Faster                               │         0 │
│ Queries Slower                               │         1 │
│ Queries with No Change                       │        21 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃                                 HEAD ┃         hash-join-global-bloom-filter ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │    324.67 / 327.87 ±1.88 / 330.12 ms │     326.22 / 329.48 ±2.36 / 332.53 ms │    no change │
│ QQuery 2  │       91.57 / 94.02 ±2.83 / 99.26 ms │        92.23 / 93.83 ±1.94 / 97.43 ms │    no change │
│ QQuery 3  │    220.75 / 224.07 ±2.27 / 227.39 ms │     224.60 / 226.83 ±1.80 / 229.34 ms │    no change │
│ QQuery 4  │    106.55 / 111.94 ±3.59 / 116.01 ms │     110.67 / 112.58 ±1.24 / 114.52 ms │    no change │
│ QQuery 5  │    336.70 / 339.79 ±2.69 / 344.25 ms │     347.62 / 353.85 ±5.51 / 362.43 ms │    no change │
│ QQuery 6  │       96.19 / 96.63 ±0.58 / 97.72 ms │       96.16 / 97.97 ±1.39 / 100.19 ms │    no change │
│ QQuery 7  │   443.05 / 460.97 ±11.33 / 472.79 ms │     451.03 / 463.01 ±7.59 / 472.43 ms │    no change │
│ QQuery 8  │    356.53 / 361.83 ±3.70 / 366.32 ms │     375.90 / 380.02 ±2.66 / 383.10 ms │ 1.05x slower │
│ QQuery 9  │    554.55 / 558.05 ±2.09 / 560.29 ms │     561.59 / 569.32 ±4.17 / 572.45 ms │    no change │
│ QQuery 10 │   289.90 / 303.24 ±11.29 / 317.99 ms │     293.55 / 305.45 ±8.04 / 317.55 ms │    no change │
│ QQuery 11 │       60.13 / 60.93 ±0.85 / 61.98 ms │        61.11 / 62.14 ±1.14 / 64.18 ms │    no change │
│ QQuery 12 │   109.51 / 118.55 ±12.10 / 142.11 ms │     110.98 / 119.46 ±6.49 / 130.55 ms │    no change │
│ QQuery 13 │   286.39 / 303.03 ±11.51 / 321.83 ms │     294.07 / 301.74 ±5.76 / 310.18 ms │    no change │
│ QQuery 14 │    163.48 / 167.28 ±2.95 / 171.93 ms │     164.87 / 169.90 ±4.55 / 176.37 ms │    no change │
│ QQuery 15 │    288.58 / 292.15 ±3.06 / 296.29 ms │     290.44 / 293.93 ±3.07 / 299.36 ms │    no change │
│ QQuery 16 │       66.38 / 68.56 ±2.02 / 71.08 ms │        65.71 / 69.42 ±3.07 / 72.67 ms │    no change │
│ QQuery 17 │    579.62 / 582.58 ±2.80 / 587.72 ms │     571.20 / 577.82 ±5.17 / 587.08 ms │    no change │
│ QQuery 18 │ 977.32 / 1006.98 ±20.03 / 1028.44 ms │ 1010.92 / 1052.40 ±29.34 / 1098.58 ms │    no change │
│ QQuery 19 │   227.50 / 246.05 ±19.65 / 282.06 ms │    225.88 / 257.31 ±34.40 / 308.42 ms │    no change │
│ QQuery 20 │    269.59 / 278.45 ±5.06 / 283.39 ms │    268.02 / 281.82 ±15.65 / 307.68 ms │    no change │
│ QQuery 21 │    687.13 / 691.97 ±6.71 / 705.10 ms │    686.39 / 701.44 ±11.37 / 715.55 ms │    no change │
│ QQuery 22 │       68.42 / 70.30 ±1.32 / 72.46 ms │        68.06 / 70.08 ±1.11 / 71.29 ms │    no change │
└───────────┴──────────────────────────────────────┴───────────────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 6765.25ms │
│ Total Time (hash-join-global-bloom-filter)   │ 6889.79ms │
│ Average Time (HEAD)                          │  307.51ms │
│ Average Time (hash-join-global-bloom-filter) │  313.17ms │
│ Queries Faster                               │         0 │
│ Queries Slower                               │         1 │
│ Queries with No Change                       │        21 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Resource Usage

tpch10 — base (merge-base)

Metric Value
Wall time 35.0s
Peak memory 15.3 GiB
Avg memory 3.2 GiB
CPU user 339.2s
CPU sys 20.3s
Peak spill 0 B

tpch10 — branch

Metric Value
Wall time 35.0s
Peak memory 12.5 GiB
Avg memory 3.0 GiB
CPU user 337.2s
CPU sys 19.9s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: "16"
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │  335.21 ms │                     332.89 ms │     no change │
│ QQuery 2  │   86.42 ms │                      85.85 ms │     no change │
│ QQuery 3  │  209.19 ms │                     196.16 ms │ +1.07x faster │
│ QQuery 4  │  150.92 ms │                     123.34 ms │ +1.22x faster │
│ QQuery 5  │  290.35 ms │                     231.85 ms │ +1.25x faster │
│ QQuery 6  │  124.66 ms │                     125.83 ms │     no change │
│ QQuery 7  │  293.77 ms │                     279.05 ms │ +1.05x faster │
│ QQuery 8  │  352.30 ms │                     305.01 ms │ +1.16x faster │
│ QQuery 9  │  594.28 ms │                     514.15 ms │ +1.16x faster │
│ QQuery 10 │  278.63 ms │                     266.68 ms │     no change │
│ QQuery 11 │   63.63 ms │                      68.23 ms │  1.07x slower │
│ QQuery 12 │  159.53 ms │                     152.54 ms │     no change │
│ QQuery 13 │  336.86 ms │                     331.55 ms │     no change │
│ QQuery 14 │  165.00 ms │                     162.69 ms │     no change │
│ QQuery 15 │  320.23 ms │                     324.17 ms │     no change │
│ QQuery 16 │   70.11 ms │                      64.07 ms │ +1.09x faster │
│ QQuery 17 │  267.58 ms │                     240.15 ms │ +1.11x faster │
│ QQuery 18 │ 1405.80 ms │                    1081.92 ms │ +1.30x faster │
│ QQuery 19 │  213.02 ms │                     212.88 ms │     no change │
│ QQuery 20 │  204.63 ms │                     198.50 ms │     no change │
│ QQuery 21 │  822.93 ms │                     712.87 ms │ +1.15x faster │
│ QQuery 22 │   73.90 ms │                      59.23 ms │ +1.25x faster │
└───────────┴────────────┴───────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 6818.97ms │
│ Total Time (hash-join-global-bloom-filter)   │ 6069.62ms │
│ Average Time (HEAD)                          │  309.95ms │
│ Average Time (hash-join-global-bloom-filter) │  275.89ms │
│ Queries Faster                               │        11 │
│ Queries Slower                               │         1 │
│ Queries with No Change                       │        10 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                  HEAD ┃         hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │     335.21 / 337.20 ±2.04 / 341.06 ms │     332.89 / 337.74 ±3.11 / 341.16 ms │     no change │
│ QQuery 2  │        86.42 / 88.27 ±1.68 / 91.04 ms │        85.85 / 86.86 ±1.00 / 88.78 ms │     no change │
│ QQuery 3  │     209.19 / 222.25 ±8.60 / 234.26 ms │     196.16 / 200.78 ±3.23 / 204.59 ms │ +1.11x faster │
│ QQuery 4  │     150.92 / 165.50 ±7.84 / 172.37 ms │     123.34 / 136.26 ±7.41 / 143.05 ms │ +1.21x faster │
│ QQuery 5  │    290.35 / 328.27 ±24.74 / 350.60 ms │     231.85 / 242.67 ±9.40 / 258.04 ms │ +1.35x faster │
│ QQuery 6  │     124.66 / 128.08 ±2.62 / 132.77 ms │     125.83 / 128.29 ±1.91 / 131.14 ms │     no change │
│ QQuery 7  │     293.77 / 295.36 ±1.26 / 297.18 ms │     279.05 / 281.78 ±2.81 / 286.73 ms │     no change │
│ QQuery 8  │     352.30 / 356.01 ±3.00 / 360.91 ms │     305.01 / 307.80 ±1.78 / 309.90 ms │ +1.16x faster │
│ QQuery 9  │    594.28 / 636.28 ±22.56 / 660.82 ms │    514.15 / 537.39 ±22.33 / 568.92 ms │ +1.18x faster │
│ QQuery 10 │    278.63 / 303.44 ±19.43 / 337.14 ms │    266.68 / 289.15 ±15.01 / 310.56 ms │     no change │
│ QQuery 11 │        63.63 / 64.60 ±0.50 / 65.01 ms │        68.23 / 72.92 ±6.06 / 84.47 ms │  1.13x slower │
│ QQuery 12 │     159.53 / 170.98 ±7.29 / 182.19 ms │    152.54 / 167.41 ±15.09 / 196.46 ms │     no change │
│ QQuery 13 │    336.86 / 349.58 ±11.71 / 368.77 ms │     331.55 / 338.95 ±7.06 / 349.46 ms │     no change │
│ QQuery 14 │     165.00 / 169.15 ±4.80 / 178.46 ms │     162.69 / 168.15 ±6.71 / 181.32 ms │     no change │
│ QQuery 15 │     320.23 / 323.10 ±1.56 / 324.69 ms │     324.17 / 326.45 ±2.05 / 330.19 ms │     no change │
│ QQuery 16 │        70.11 / 74.50 ±5.46 / 84.55 ms │        64.07 / 65.33 ±1.38 / 67.87 ms │ +1.14x faster │
│ QQuery 17 │     267.58 / 273.84 ±8.39 / 290.19 ms │     240.15 / 246.49 ±8.44 / 263.02 ms │ +1.11x faster │
│ QQuery 18 │ 1405.80 / 1429.23 ±23.85 / 1475.01 ms │ 1081.92 / 1128.06 ±34.82 / 1186.82 ms │ +1.27x faster │
│ QQuery 19 │    213.02 / 234.79 ±21.18 / 269.78 ms │    212.88 / 252.61 ±69.98 / 392.48 ms │  1.08x slower │
│ QQuery 20 │     204.63 / 206.37 ±1.59 / 208.42 ms │     198.50 / 202.97 ±2.86 / 207.51 ms │     no change │
│ QQuery 21 │    822.93 / 857.28 ±21.22 / 884.23 ms │    712.87 / 750.84 ±25.93 / 784.06 ms │ +1.14x faster │
│ QQuery 22 │        73.90 / 76.23 ±1.69 / 78.21 ms │        59.23 / 60.67 ±1.43 / 63.18 ms │ +1.26x faster │
└───────────┴───────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 7090.31ms │
│ Total Time (hash-join-global-bloom-filter)   │ 6329.57ms │
│ Average Time (HEAD)                          │  322.29ms │
│ Average Time (hash-join-global-bloom-filter) │  287.71ms │
│ Queries Faster                               │        10 │
│ Queries Slower                               │         2 │
│ Queries with No Change                       │        10 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Resource Usage

tpch10 — base (merge-base)

Metric Value
Wall time 40.0s
Peak memory 13.1 GiB
Avg memory 3.0 GiB
CPU user 354.1s
CPU sys 19.8s
Peak spill 0 B

tpch10 — branch

Metric Value
Wall time 35.0s
Peak memory 12.8 GiB
Avg memory 3.0 GiB
CPU user 311.0s
CPU sys 20.6s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c6628801687-4449-vkmf2 6.12.105+ #1 SMP Sat Sep 19 07:55:25 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: "64"

Results will be posted here when complete

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "false"
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: "4"
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ hash-join-global-bloom-filter ┃       Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━┩
│ QQuery 1  │  952.50 ms │                     947.11 ms │    no change │
│ QQuery 2  │  206.63 ms │                     207.34 ms │    no change │
│ QQuery 3  │  574.61 ms │                     575.01 ms │    no change │
│ QQuery 4  │  275.18 ms │                     276.68 ms │    no change │
│ QQuery 5  │  766.20 ms │                     780.38 ms │    no change │
│ QQuery 6  │  285.59 ms │                     287.09 ms │    no change │
│ QQuery 7  │  887.90 ms │                     913.62 ms │    no change │
│ QQuery 8  │ 1008.45 ms │                    1012.62 ms │    no change │
│ QQuery 9  │ 1303.77 ms │                    1328.04 ms │    no change │
│ QQuery 10 │  661.02 ms │                     661.31 ms │    no change │
│ QQuery 11 │  154.25 ms │                     156.14 ms │    no change │
│ QQuery 12 │  290.02 ms │                     289.14 ms │    no change │
│ QQuery 13 │  628.11 ms │                     644.90 ms │    no change │
│ QQuery 14 │  472.81 ms │                     474.85 ms │    no change │
│ QQuery 15 │  823.50 ms │                     828.09 ms │    no change │
│ QQuery 16 │  104.01 ms │                     101.81 ms │    no change │
│ QQuery 17 │ 1375.62 ms │                    1466.20 ms │ 1.07x slower │
│ QQuery 18 │ 1992.45 ms │                    2121.79 ms │ 1.06x slower │
│ QQuery 19 │  647.44 ms │                     650.77 ms │    no change │
│ QQuery 20 │  594.54 ms │                     590.87 ms │    no change │
│ QQuery 21 │ 1612.34 ms │                    1605.54 ms │    no change │
│ QQuery 22 │  142.91 ms │                     144.15 ms │    no change │
└───────────┴────────────┴───────────────────────────────┴──────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 15759.87ms │
│ Total Time (hash-join-global-bloom-filter)   │ 16063.45ms │
│ Average Time (HEAD)                          │   716.36ms │
│ Average Time (hash-join-global-bloom-filter) │   730.16ms │
│ Queries Faster                               │          0 │
│ Queries Slower                               │          2 │
│ Queries with No Change                       │         20 │
│ Queries with Failure                         │          0 │
└──────────────────────────────────────────────┴────────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Query     ┃                                   HEAD ┃          hash-join-global-bloom-filter ┃    Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ QQuery 1  │      952.50 / 956.81 ±3.72 / 963.58 ms │      947.11 / 951.82 ±4.26 / 956.83 ms │ no change │
│ QQuery 2  │      206.63 / 209.45 ±1.99 / 212.76 ms │      207.34 / 209.81 ±1.96 / 212.27 ms │ no change │
│ QQuery 3  │      574.61 / 581.37 ±6.87 / 594.00 ms │      575.01 / 581.19 ±3.57 / 584.59 ms │ no change │
│ QQuery 4  │      275.18 / 279.65 ±3.07 / 283.22 ms │      276.68 / 281.05 ±3.22 / 285.93 ms │ no change │
│ QQuery 5  │      766.20 / 771.63 ±4.41 / 779.36 ms │      780.38 / 783.83 ±2.35 / 786.65 ms │ no change │
│ QQuery 6  │      285.59 / 288.35 ±3.27 / 294.70 ms │      287.09 / 288.45 ±1.57 / 291.42 ms │ no change │
│ QQuery 7  │      887.90 / 900.14 ±7.22 / 907.60 ms │      913.62 / 918.99 ±3.62 / 924.16 ms │ no change │
│ QQuery 8  │   1008.45 / 1013.70 ±3.37 / 1017.78 ms │   1012.62 / 1015.36 ±3.34 / 1021.81 ms │ no change │
│ QQuery 9  │  1303.77 / 1330.07 ±14.11 / 1344.72 ms │   1328.04 / 1337.61 ±7.76 / 1351.64 ms │ no change │
│ QQuery 10 │     661.02 / 682.81 ±15.15 / 705.99 ms │      661.31 / 674.79 ±8.47 / 687.82 ms │ no change │
│ QQuery 11 │      154.25 / 155.57 ±1.37 / 158.13 ms │      156.14 / 156.57 ±0.34 / 157.09 ms │ no change │
│ QQuery 12 │      290.02 / 296.33 ±9.79 / 315.81 ms │     289.14 / 297.73 ±12.95 / 323.44 ms │ no change │
│ QQuery 13 │     628.11 / 643.86 ±10.65 / 657.60 ms │      644.90 / 654.25 ±8.73 / 667.58 ms │ no change │
│ QQuery 14 │      472.81 / 476.05 ±3.84 / 483.34 ms │      474.85 / 479.67 ±4.67 / 487.96 ms │ no change │
│ QQuery 15 │      823.50 / 829.31 ±4.70 / 836.07 ms │      828.09 / 833.99 ±4.25 / 841.14 ms │ no change │
│ QQuery 16 │      104.01 / 106.60 ±3.00 / 112.08 ms │      101.81 / 106.36 ±3.21 / 110.25 ms │ no change │
│ QQuery 17 │  1375.62 / 1436.21 ±36.25 / 1484.07 ms │  1466.20 / 1484.80 ±25.60 / 1535.31 ms │ no change │
│ QQuery 18 │ 1992.45 / 2269.59 ±161.78 / 2432.41 ms │ 2121.79 / 2349.39 ±146.52 / 2583.11 ms │ no change │
│ QQuery 19 │     647.44 / 681.45 ±44.62 / 765.31 ms │     650.77 / 683.85 ±39.44 / 734.89 ms │ no change │
│ QQuery 20 │      594.54 / 601.18 ±6.27 / 613.10 ms │      590.87 / 600.58 ±8.96 / 617.07 ms │ no change │
│ QQuery 21 │   1612.34 / 1616.64 ±2.82 / 1620.92 ms │  1605.54 / 1625.23 ±14.69 / 1643.75 ms │ no change │
│ QQuery 22 │      142.91 / 146.28 ±1.93 / 148.08 ms │      144.15 / 146.79 ±2.30 / 149.69 ms │ no change │
└───────────┴────────────────────────────────────────┴────────────────────────────────────────┴───────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 16273.06ms │
│ Total Time (hash-join-global-bloom-filter)   │ 16462.09ms │
│ Average Time (HEAD)                          │   739.68ms │
│ Average Time (hash-join-global-bloom-filter) │   748.28ms │
│ Queries Faster                               │          0 │
│ Queries Slower                               │          0 │
│ Queries with No Change                       │         22 │
│ Queries with Failure                         │          0 │
└──────────────────────────────────────────────┴────────────┘

Resource Usage

tpch10 — base (merge-base)

Metric Value
Wall time 85.0s
Peak memory 11.5 GiB
Avg memory 1.9 GiB
CPU user 326.8s
CPU sys 25.8s
Peak spill 0 B

tpch10 — branch

Metric Value
Wall time 85.0s
Peak memory 11.3 GiB
Avg memory 1.8 GiB
CPU user 328.5s
CPU sys 25.5s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: "4"
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │  992.46 ms │                     988.17 ms │     no change │
│ QQuery 2  │  216.23 ms │                     211.09 ms │     no change │
│ QQuery 3  │  595.62 ms │                     551.98 ms │ +1.08x faster │
│ QQuery 4  │  410.08 ms │                     340.41 ms │ +1.20x faster │
│ QQuery 5  │  827.71 ms │                     662.87 ms │ +1.25x faster │
│ QQuery 6  │  381.22 ms │                     381.54 ms │     no change │
│ QQuery 7  │  834.43 ms │                     788.26 ms │ +1.06x faster │
│ QQuery 8  │ 1035.98 ms │                     879.58 ms │ +1.18x faster │
│ QQuery 9  │ 1554.01 ms │                    1298.38 ms │ +1.20x faster │
│ QQuery 10 │  707.50 ms │                     675.03 ms │     no change │
│ QQuery 11 │  167.35 ms │                     180.41 ms │  1.08x slower │
│ QQuery 12 │  455.38 ms │                     436.55 ms │     no change │
│ QQuery 13 │  771.55 ms │                     695.98 ms │ +1.11x faster │
│ QQuery 14 │  457.25 ms │                     460.76 ms │     no change │
│ QQuery 15 │  916.11 ms │                     926.47 ms │     no change │
│ QQuery 16 │  119.70 ms │                     102.93 ms │ +1.16x faster │
│ QQuery 17 │  782.42 ms │                     713.19 ms │ +1.10x faster │
│ QQuery 18 │ 2894.93 ms │                    2376.18 ms │ +1.22x faster │
│ QQuery 19 │  590.98 ms │                     594.91 ms │     no change │
│ QQuery 20 │  591.35 ms │                     571.30 ms │     no change │
│ QQuery 21 │ 2054.38 ms │                    1773.92 ms │ +1.16x faster │
│ QQuery 22 │  184.76 ms │                     139.82 ms │ +1.32x faster │
└───────────┴────────────┴───────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 17541.38ms │
│ Total Time (hash-join-global-bloom-filter)   │ 15749.73ms │
│ Average Time (HEAD)                          │   797.34ms │
│ Average Time (hash-join-global-bloom-filter) │   715.90ms │
│ Queries Faster                               │         12 │
│ Queries Slower                               │          1 │
│ Queries with No Change                       │          9 │
│ Queries with Failure                         │          0 │
└──────────────────────────────────────────────┴────────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                   HEAD ┃         hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │     992.46 / 999.03 ±5.53 / 1005.60 ms │    988.17 / 996.35 ±5.52 / 1005.54 ms │     no change │
│ QQuery 2  │      216.23 / 219.26 ±3.67 / 226.15 ms │     211.09 / 214.00 ±2.35 / 217.83 ms │     no change │
│ QQuery 3  │      595.62 / 599.53 ±2.93 / 602.88 ms │     551.98 / 553.79 ±1.25 / 555.23 ms │ +1.08x faster │
│ QQuery 4  │      410.08 / 412.31 ±2.74 / 417.34 ms │     340.41 / 342.54 ±1.96 / 346.23 ms │ +1.20x faster │
│ QQuery 5  │      827.71 / 833.59 ±4.46 / 838.25 ms │    662.87 / 669.18 ±10.09 / 689.13 ms │ +1.25x faster │
│ QQuery 6  │      381.22 / 384.65 ±4.05 / 391.81 ms │     381.54 / 383.19 ±1.55 / 385.93 ms │     no change │
│ QQuery 7  │      834.43 / 842.79 ±5.30 / 850.81 ms │     788.26 / 797.28 ±6.88 / 807.43 ms │ +1.06x faster │
│ QQuery 8  │   1035.98 / 1040.39 ±3.32 / 1044.37 ms │     879.58 / 884.13 ±2.91 / 887.60 ms │ +1.18x faster │
│ QQuery 9  │  1554.01 / 1570.74 ±16.14 / 1595.11 ms │ 1298.38 / 1323.35 ±16.09 / 1344.94 ms │ +1.19x faster │
│ QQuery 10 │     707.50 / 724.42 ±19.62 / 760.24 ms │    675.03 / 694.38 ±14.86 / 712.82 ms │     no change │
│ QQuery 11 │      167.35 / 168.95 ±1.20 / 170.36 ms │     180.41 / 188.10 ±9.02 / 199.16 ms │  1.11x slower │
│ QQuery 12 │     455.38 / 465.89 ±10.10 / 481.87 ms │     436.55 / 438.58 ±1.66 / 440.53 ms │ +1.06x faster │
│ QQuery 13 │     771.55 / 779.86 ±12.72 / 804.56 ms │    695.98 / 715.01 ±15.47 / 741.63 ms │ +1.09x faster │
│ QQuery 14 │      457.25 / 465.36 ±7.86 / 479.62 ms │     460.76 / 466.94 ±3.61 / 472.03 ms │     no change │
│ QQuery 15 │     916.11 / 930.43 ±12.78 / 950.13 ms │     926.47 / 932.60 ±4.37 / 938.12 ms │     no change │
│ QQuery 16 │      119.70 / 123.11 ±3.19 / 128.61 ms │     102.93 / 106.06 ±3.43 / 112.07 ms │ +1.16x faster │
│ QQuery 17 │      782.42 / 788.61 ±9.86 / 808.27 ms │    713.19 / 720.56 ±10.25 / 740.78 ms │ +1.09x faster │
│ QQuery 18 │ 2894.93 / 3073.63 ±119.81 / 3234.57 ms │ 2376.18 / 2551.08 ±93.47 / 2622.42 ms │ +1.20x faster │
│ QQuery 19 │     590.98 / 627.17 ±44.37 / 706.85 ms │    594.91 / 630.94 ±42.02 / 688.21 ms │     no change │
│ QQuery 20 │      591.35 / 594.05 ±2.33 / 597.83 ms │     571.30 / 574.62 ±2.57 / 577.58 ms │     no change │
│ QQuery 21 │  2054.38 / 2094.37 ±28.39 / 2140.75 ms │ 1773.92 / 1799.65 ±22.00 / 1837.98 ms │ +1.16x faster │
│ QQuery 22 │      184.76 / 188.52 ±2.69 / 191.50 ms │     139.82 / 142.27 ±1.45 / 144.24 ms │ +1.33x faster │
└───────────┴────────────────────────────────────────┴───────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃            ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 17926.66ms │
│ Total Time (hash-join-global-bloom-filter)   │ 16124.61ms │
│ Average Time (HEAD)                          │   814.85ms │
│ Average Time (hash-join-global-bloom-filter) │   732.94ms │
│ Queries Faster                               │         13 │
│ Queries Slower                               │          1 │
│ Queries with No Change                       │          8 │
│ Queries with Failure                         │          0 │
└──────────────────────────────────────────────┴────────────┘

Resource Usage

tpch10 — base (merge-base)

Metric Value
Wall time 90.0s
Peak memory 10.0 GiB
Avg memory 2.0 GiB
CPU user 346.0s
CPU sys 25.6s
Peak spill 0 B

tpch10 — branch

Metric Value
Wall time 85.0s
Peak memory 11.4 GiB
Avg memory 2.0 GiB
CPU user 306.4s
CPU sys 25.7s
Peak spill 0 B

File an issue against this benchmark runner

@adriangbot

adriangbot commented Nov 4, 2026

Copy link
Copy Markdown

🤖 Benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

Comparing hash-join-global-bloom-filter (a57d1f4) to 9411463 (merge-base) diff

Run configuration 
run benchmark tpch10
env:
  DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true"
  DATAFUSION_EXECUTION_TARGET_PARTITIONS: "64"
CPU Details (lscpu) 
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected
Details 

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃       HEAD ┃ hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │  340.61 ms │                     336.95 ms │     no change │
│ QQuery 2  │  102.26 ms │                     101.27 ms │     no change │
│ QQuery 3  │  217.35 ms │                     205.97 ms │ +1.06x faster │
│ QQuery 4  │  154.87 ms │                     134.41 ms │ +1.15x faster │
│ QQuery 5  │  305.15 ms │                     257.08 ms │ +1.19x faster │
│ QQuery 6  │  127.27 ms │                     129.23 ms │     no change │
│ QQuery 7  │  316.16 ms │                     308.42 ms │     no change │
│ QQuery 8  │  379.79 ms │                     337.39 ms │ +1.13x faster │
│ QQuery 9  │  575.71 ms │                     491.31 ms │ +1.17x faster │
│ QQuery 10 │  285.33 ms │                     282.37 ms │     no change │
│ QQuery 11 │   72.15 ms │                      75.48 ms │     no change │
│ QQuery 12 │  174.05 ms │                     165.16 ms │ +1.05x faster │
│ QQuery 13 │  269.22 ms │                     263.55 ms │     no change │
│ QQuery 14 │  165.61 ms │                     172.06 ms │     no change │
│ QQuery 15 │  322.36 ms │                     324.07 ms │     no change │
│ QQuery 16 │   81.33 ms │                      76.74 ms │ +1.06x faster │
│ QQuery 17 │  280.81 ms │                     273.08 ms │     no change │
│ QQuery 18 │ 1202.64 ms │                     876.14 ms │ +1.37x faster │
│ QQuery 19 │  212.04 ms │                     209.63 ms │     no change │
│ QQuery 20 │  216.95 ms │                     210.30 ms │     no change │
│ QQuery 21 │  796.53 ms │                     736.78 ms │ +1.08x faster │
│ QQuery 22 │   85.19 ms │                      69.11 ms │ +1.23x faster │
└───────────┴────────────┴───────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 6683.37ms │
│ Total Time (hash-join-global-bloom-filter)   │ 6036.49ms │
│ Average Time (HEAD)                          │  303.79ms │
│ Average Time (hash-join-global-bloom-filter) │  274.39ms │
│ Queries Faster                               │        10 │
│ Queries Slower                               │         0 │
│ Queries with No Change                       │        12 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Distribution per query (min / mean ±stddev / max):

Comparing HEAD and hash-join-global-bloom-filter
--------------------
Benchmark tpch_sf10.json
--------------------
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Query     ┃                                 HEAD ┃      hash-join-global-bloom-filter ┃        Change ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ QQuery 1  │    340.61 / 344.97 ±4.62 / 353.64 ms │  336.95 / 341.14 ±2.86 / 344.72 ms │     no change │
│ QQuery 2  │    102.26 / 103.56 ±0.86 / 104.63 ms │  101.27 / 103.09 ±1.56 / 105.94 ms │     no change │
│ QQuery 3  │    217.35 / 218.85 ±1.11 / 220.51 ms │  205.97 / 209.39 ±2.34 / 213.24 ms │     no change │
│ QQuery 4  │    154.87 / 156.42 ±1.18 / 157.94 ms │  134.41 / 136.35 ±1.13 / 137.76 ms │ +1.15x faster │
│ QQuery 5  │    305.15 / 307.86 ±2.26 / 310.86 ms │  257.08 / 262.01 ±3.81 / 267.93 ms │ +1.18x faster │
│ QQuery 6  │    127.27 / 129.57 ±2.20 / 133.48 ms │  129.23 / 131.99 ±2.68 / 136.05 ms │     no change │
│ QQuery 7  │    316.16 / 320.73 ±4.98 / 328.93 ms │  308.42 / 310.49 ±1.77 / 313.40 ms │     no change │
│ QQuery 8  │    379.79 / 383.89 ±4.82 / 393.08 ms │  337.39 / 339.28 ±2.94 / 345.12 ms │ +1.13x faster │
│ QQuery 9  │    575.71 / 581.06 ±3.29 / 584.63 ms │  491.31 / 494.75 ±2.79 / 498.80 ms │ +1.17x faster │
│ QQuery 10 │    285.33 / 291.71 ±5.29 / 299.01 ms │  282.37 / 287.87 ±5.30 / 297.42 ms │     no change │
│ QQuery 11 │       72.15 / 73.58 ±1.24 / 75.78 ms │     75.48 / 78.28 ±1.71 / 80.43 ms │  1.06x slower │
│ QQuery 12 │    174.05 / 175.71 ±2.18 / 179.86 ms │  165.16 / 172.61 ±8.15 / 187.09 ms │     no change │
│ QQuery 13 │    269.22 / 282.36 ±9.80 / 294.97 ms │  263.55 / 268.75 ±5.28 / 277.92 ms │     no change │
│ QQuery 14 │    165.61 / 169.98 ±3.26 / 174.20 ms │  172.06 / 174.32 ±2.76 / 179.48 ms │     no change │
│ QQuery 15 │    322.36 / 327.45 ±4.29 / 334.27 ms │  324.07 / 329.98 ±3.93 / 336.27 ms │     no change │
│ QQuery 16 │       81.33 / 82.09 ±0.40 / 82.47 ms │     76.74 / 78.25 ±1.07 / 79.75 ms │     no change │
│ QQuery 17 │    280.81 / 284.83 ±3.58 / 290.06 ms │  273.08 / 274.59 ±1.57 / 277.07 ms │     no change │
│ QQuery 18 │ 1202.64 / 1216.68 ±9.32 / 1228.36 ms │ 876.14 / 903.68 ±17.21 / 927.73 ms │ +1.35x faster │
│ QQuery 19 │    212.04 / 219.68 ±9.06 / 235.95 ms │ 209.63 / 220.80 ±14.13 / 246.71 ms │     no change │
│ QQuery 20 │    216.95 / 218.34 ±0.98 / 219.49 ms │  210.30 / 212.90 ±2.09 / 215.82 ms │     no change │
│ QQuery 21 │    796.53 / 804.21 ±5.66 / 810.81 ms │  736.78 / 743.44 ±5.43 / 751.48 ms │ +1.08x faster │
│ QQuery 22 │       85.19 / 88.00 ±1.82 / 90.00 ms │     69.11 / 72.00 ±3.28 / 77.19 ms │ +1.22x faster │
└───────────┴──────────────────────────────────────┴────────────────────────────────────┴───────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━┓
┃ Benchmark Summary                            ┃           ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━┩
│ Total Time (HEAD)                            │ 6781.53ms │
│ Total Time (hash-join-global-bloom-filter)   │ 6145.97ms │
│ Average Time (HEAD)                          │  308.25ms │
│ Average Time (hash-join-global-bloom-filter) │  279.36ms │
│ Queries Faster                               │         7 │
│ Queries Slower                               │         1 │
│ Queries with No Change                       │        14 │
│ Queries with Failure                         │         0 │
└──────────────────────────────────────────────┴───────────┘

Resource Usage

tpch10 — base (merge-base)

Metric Value
Wall time 35.0s
Peak memory 14.9 GiB
Avg memory 3.9 GiB
CPU user 351.7s
CPU sys 18.8s
Peak spill 0 B

tpch10 — branch

Metric Value
Wall time 35.0s
Peak memory 15.8 GiB
Avg memory 3.3 GiB
CPU user 310.6s
CPU sys 19.4s
Peak spill 0 B

File an issue against this benchmark runner

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

auto detected api change Auto detected API change common Related to common crate core Core DataFusion crate documentation Improvements or additions to documentation physical-plan Changes to the physical-plan crate proto Related to proto crate sqllogictest SQL Logic Tests (.slt)

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

Improve Performance (Polars 2.0 Call to Action)

3 participants

@adriangb @adriangbot @codecov-commenter