Skip to content

Read DuckDB's V1 scans through open - #11664

Draft
gatesn wants to merge 10 commits into
ngates/datafusion-scan-poolfrom
ngates/duckdb-scan-pool
Draft

gatesn wants to merge 10 commits into
ngates/datafusion-scan-poolfrom
ngates/duckdb-scan-pool

Conversation

@gatesn

@gatesn gatesn commented Nov 19, 2026 •

Copy link
Copy Markdown
Contributor

DuckDB iterates files itself: its MultiFileReader opens each file, initializes it once under a file lock, and hands threads to it under a global lock until the file reports no more work, then closes it and never returns to it. The Vortex table function maps that onto the Scan API with one open per file. Initializing a file builds a ScanRequest from DuckDB's pushed-down projection, filter, row range and selection, scans it through LayoutReaderDataSource, and opens it into a ScanStream. Each DuckDB thread that joins the file polls its own clone of that stream with block_on, so DuckDB's threads are the scan's parallelism and the scan starts no threads of its own.

One open per file, rather than one across the glob, keeps DuckDB in charge of what it already owns: file order, globs, the filename and hive columns, file-level skipping from footer statistics, and per-file progress. A file's V1 scan opens into a stream over a queue of its splits. DuckDB runs one thread per core, so its threads are the scan's parallelism, and the request sets split_concurrency to one: each clone claims the next split only once it has finished its last, as each DuckDB thread runs one split at a time. A clone with room for more would claim splits that other threads could have run, so on a file with few splits, such as TPC-H's smaller tables, most threads would find the queue empty. With one split in flight, a clone polls the split itself rather than spawning it on DuckDB's shared runtime, so the thread that wants the split's array decodes it, with no wakeup from another thread, and exports it from the same core. The request is unordered with no limit: DuckDB's parallel scan does not promise row order across threads, and file_row_number is projected per row rather than counted, so it stays correct whichever thread reads a split.

A clone ends only once no split is left to claim and the splits it claimed are done, so DuckDB hands threads to a file until the first of its clones ends. From then on threads move on to the next file while the clones holding the last splits finish them: a clone yields only the splits it claimed, so no thread waits on another's work, and splits of neighbouring files overlap. DuckDB's progress bar reads ScanStream::progress, which this PR adds: the fraction of the scan's work its clones have taken, which a layout scan reports as the fraction of its splits claimed.

DuckDB's private per-file list of V1 split tasks is removed; the stream's split queue is that list behind the Scan API. V2 scans keep their own queue until the V2 source is behind open.

This PR is the stack's caller of open: a morsel-driven host whose threads each pull from a shared stream, the counterpart to DataFusion's fixed partitions through split in #11666. It is part of the Scan API work tracked in #9062.

Local DuckDB benchmarks, as head time over develop time (geomean), with each query interleaved with develop over at least three rounds:

Suite V1 V2
TPC-H 1.011 1.002
ClickBench 1.014 0.985

ClickBench was rotated so the head ran in every position, because run order alone moved a query's time by up to 35%. Develop against itself measured 1.012, and no query was above 1.05 in every rotation.

Stacked on #11666

🤖 Generated with Claude Code

@gatesn gatesn added changelog/performance A performance improvement changelog/chore A trivial change and removed changelog/performance A performance improvement labels Nov 19, 2026
@gatesn gatesn changed the title Read DuckDB scans through scan pools Read DuckDB's V1 scans through scan pools Nov 19, 2026
@gatesn gatesn mentioned this pull request Nov 19, 2026
16 of 46 tasks
@gatesn

gatesn commented Nov 19, 2026

Copy link
Copy Markdown
Contributor Author

DuckDB benchmarks, V1 scan (duckdb-bench --scan-v1 true, --formats vortex), develop (d3688347ff) against this PR (d891c22319) on an M-series laptop. TPC-H is SF1 and ClickBench is the 100-file partitioned set, which also covers multi-file globs. Each target ran two interleaved rounds of 5 iterations, and the tables show the median of 8 timings after dropping each round's first iteration.

Headline: total time is flat: TPC-H 0.99× and ClickBench 1.00×. The geomean is 0.98× for TPC-H and 0.97× for ClickBench.

Regressions over 5% that held in both rounds:

  • TPC-H: q22 (1.43× and 1.39×), q10 (1.23× and 1.16×), q16 (1.09× and 1.12×).
  • ClickBench: q23 (1.12× and 1.21×), q39 (1.12× and 1.13×), q15, and q28 (1.07× and 1.05×).

These are mostly short queries, 9–50 ms. I haven't tuned anything.

Caveat: the machine was not idle. Other sessions were compiling on it, with a load average of about 74 during the follow-up A/B. That A/B ran the flagged queries on develop, this PR as it is (4 splits in flight per thread), and this PR with split_concurrency = 1:

query develop 4 in flight 1 in flight
tpch q02 8.6 13.5 12.7
tpch q10 13.2 20.0 18.4
tpch q16 8.9 12.3 12.2
tpch q22 8.3 16.0 11.8
clickbench q09 190 286 263
clickbench q23 37.7 51.4 54.6
clickbench q28 1917 2699 2621
clickbench q39 23.6 22.5 22.4

The gaps grew with the load. Lowering in-flight splits to 1 closes little of it, except on q22, so the new default of 4 is not the main cause. These numbers need a rerun on a quiet machine before reading anything into single queries.

Full first run (ms, median)

tpch

query develop (ms) pool (ms) ratio
q01 10.7 11.2 1.05
q02 8.7 10.5 1.21
q03 10.5 9.8 0.93
q04 7.0 6.7 0.96
q05 12.4 13.0 1.04
q06 4.2 2.9 0.69
q07 12.5 12.8 1.02
q08 13.9 13.9 1.00
q09 21.8 21.0 0.97
q10 13.7 16.4 1.20 ⚠️
q11 4.6 4.6 1.01
q12 10.2 9.5 0.93
q13 15.5 15.3 0.99
q14 7.0 5.9 0.85
q15 5.5 4.3 0.78
q16 9.4 10.3 1.10
q17 9.6 8.3 0.86
q18 17.8 17.0 0.96
q19 7.9 7.3 0.93
q20 10.2 9.5 0.93
q21 26.6 23.6 0.89
q22 9.2 13.1 1.43 ⚠️
total 249 247 0.993

geomean ratio: 0.976

clickbench

query develop (ms) pool (ms) ratio
q00 1.1 1.1 1.04
q01 9.5 6.8 0.71
q02 12.7 9.9 0.78
q03 36.3 34.1 0.94
q04 107.0 112.8 1.05 ⚠️
q05 107.4 108.0 1.01
q06 1.5 1.8 1.22
q07 11.2 10.4 0.93
q08 139.3 148.5 1.07 ⚠️
q09 207.4 228.1 1.10 ⚠️
q10 41.7 33.4 0.80
q11 45.7 39.9 0.87
q12 111.8 116.0 1.04
q13 234.4 237.9 1.01
q14 132.7 139.5 1.05 ⚠️
q15 147.1 158.7 1.08 ⚠️
q16 308.5 324.8 1.05 ⚠️
q17 301.8 237.5 0.79
q18 611.9 559.9 0.92
q19 18.8 11.9 0.64
q20 85.3 74.1 0.87
q21 183.7 177.8 0.97
q22 299.5 231.2 0.77
q23 40.0 48.5 1.21 ⚠️
q24 12.7 13.4 1.05
q25 36.7 31.1 0.85
q26 10.7 11.2 1.04
q27 112.4 100.3 0.89
q28 1974.2 2129.4 1.08 ⚠️
q29 21.1 16.9 0.80
q30 102.5 95.6 0.93
q31 137.7 128.9 0.94
q32 598.8 620.8 1.04
q33 529.0 519.0 0.98
q34 628.8 639.9 1.02
q35 142.7 151.9 1.06 ⚠️
q36 12.4 11.4 0.92
q37 7.2 7.6 1.05
q38 9.4 8.4 0.90
q39 16.8 18.9 1.13 ⚠️
q40 5.7 6.2 1.09
q41 5.9 6.3 1.08
q42 5.3 6.3 1.19
total 7566 7576 1.001

geomean ratio: 0.966

@gatesn
gatesn changed the base branch from ngates/scan-native-pool to ngates/v2-scan-pool November 20, 2026 00:31
@gatesn
gatesn changed the base branch from ngates/v2-scan-pool to ngates/scan-plan-pool November 20, 2026 00:40
gatesn added a commit that referenced this pull request Nov 20, 2026
Replaces #11664's tree with its port onto #11663's split and open, keeping the
earlier history so the branch moves forward without a force push.

Signed-off-by: Nicholas Gates <nick@nickgates.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@gatesn
gatesn changed the base branch from ngates/scan-plan-pool to ngates/scan-native-pool November 20, 2026 18:25
@gatesn gatesn changed the title Read DuckDB's V1 scans through scan pools Read DuckDB's V1 scans through split and open Nov 20, 2026
@gatesn
gatesn changed the base branch from ngates/scan-native-pool to ngates/datafusion-scan-pool November 20, 2026 19:00
@gatesn
gatesn added this pull request to stack #11706 November 20, 2026 19:00
@gatesn
gatesn force-pushed the ngates/duckdb-scan-pool branch 3 times, most recently from 975aa4f to 5c76944 Compare November 20, 2026 19:38
@gatesn gatesn changed the title Read DuckDB's V1 scans through split and open Read DuckDB's V1 scans through open Nov 20, 2026
@gatesn
gatesn force-pushed the ngates/duckdb-scan-pool branch 2 times, most recently from aafa512 to 0bf4d78 Compare November 20, 2026 19:46
gatesn and others added 2 commits November 20, 2026 14:48
A host reading an opened scan, such as DuckDB showing a progress bar per file, asks its stream how much of the work its clones have taken. A layout scan's stream reports the fraction of its splits claimed; a stream over partitions reports none.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Nicholas Gates <nick@nickgates.com>
Each file DuckDB's MultiFileReader opens is opened once as a V1 layout scan,
and every DuckDB thread that joins the file polls a clone of its stream. A
thread joins a file until the first clone ends, which happens only once no
split is left to claim, so threads move on to the next file without waiting
for the last splits to finish. DuckDB's private per-file split list is removed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Nicholas Gates <nick@nickgates.com>
@gatesn
gatesn force-pushed the ngates/duckdb-scan-pool branch from 0bf4d78 to cdb5fad Compare November 20, 2026 19:48
A ScanStream is Send but not Sync, so the file state that every thread scanning the file shares holds it behind a mutex, and each thread clones it there.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Nicholas Gates <nick@nickgates.com>
gatesn and others added 7 commits November 20, 2026 15:05
DuckDB runs one thread per core, and each polls its own clone of a file's scan stream. A clone that keeps several splits in flight claims splits that other threads could have run, so on a file with few splits, such as TPC-H's smaller tables, most threads find the queue empty. Capping each clone at one split matches DuckDB's own parallelism.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Nicholas Gates <nick@nickgates.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Nicholas Gates <nick@nickgates.com>
…ates/duckdb-stream-3

Signed-off-by: Nicholas Gates <nick@nickgates.com>
…o ngates/duckdb-stream-3

Signed-off-by: Nicholas Gates <nick@nickgates.com>
…o ngates/duckdb-stream-3

Signed-off-by: Nicholas Gates <nick@nickgates.com>
…o ngates/duckdb-stream-3

Signed-off-by: Nicholas Gates <nick@nickgates.com>
…o ngates/duckdb-stream-3

Signed-off-by: Nicholas Gates <nick@nickgates.com>
@github-actions

github-actions Bot commented Nov 20, 2026 •

Copy link
Copy Markdown
Contributor

Polar Signals Profiling Results

Latest Run

Status Commit Job Attempt Link
🟢 Done b20fb3e clickbench-nvme 1 Explore Profiling Data
🟢 Done b20fb3e clickbench-nvme-datafusion-datasource-v1 1 Explore Profiling Data
🟢 Done b20fb3e clickbench-nvme-datafusion-file-v1 1 Explore Profiling Data
🟢 Done b20fb3e clickbench-nvme-duckdb-file-v1 1 Explore Profiling Data
🟢 Done b20fb3e tpch-nvme 1 Explore Profiling Data
🟢 Done b20fb3e tpch-nvme-duckdb-file-v1 1 Explore Profiling Data
🟢 Done b20fb3e tpch-nvme-datafusion-datasource-v1 1 Explore Profiling Data
🟢 Done b20fb3e tpch-nvme-datafusion-file-v1 1 Explore Profiling Data

Powered by Polar Signals Cloud

@github-actions

Copy link
Copy Markdown
Contributor

Benchmarks: TPC-H SF=1 on NVME 📖

Commits: PR b20fb3ee vs base 75910442
Verdict: No clear signal (low confidence)
Attributed Vortex impact: +0.3%
Engines: DataFusion No clear signal (+0.1%, low confidence) · DuckDB No clear signal (+0.7%, low confidence)
Vortex (geomean): hot 1.004x ➖ · cold 1.012x ➖
Parquet (geomean): hot 1.002x ➖ · cold 1.005x ➖
Shifts: Parquet (control) +0.2% · Median polish +0.2%

How to read Verdict and Engines
  • Verdict: Overall PR-level signal after subtracting baseline drift estimated from Parquet control rows. It can be Likely improvement, Likely regression, or No clear signal.
  • Engines: Per-engine attribution. DataFusion is compared against DataFusion/Parquet controls; DuckDB is compared against DuckDB/Parquet controls. This answers whether each engine improved or regressed independently.
  • Confidence: Based on directional consistency, share of rows above the noise floor, and control-run noise.
  • Hot vs cold: Every measurement is run several times. The first run is reported as the cold run, and the median of the runs after it is reported as the hot run. The verdict and significance use hot runs; each target's geomean reports hot and cold beside each other where the individual runs were recorded and hot alone where they were not, the cold column shows first-run cost per row, and hot/cold is how much of each run the warm path saves. Rows whose results predate per-run reporting show only one value, taken from the value the runner reported.
  • Query tables: One row per query and one column per target. Each cell reads develop → PR (PR ÷ develop), so (0.80) is 20% faster than develop. A change that clears this suite's threshold is bold and marked 🔴 slower or 🟢 faster; anything smaller is noise here and is left plain. The geomean row compares each target's geomean over its queries with develop's. Hot runs are shown first and cold runs are folded below them.
  • Other tables: Each cell reads PR / base / %diff, marked the same way.

Hot runs: geomean 1.003x ➖ · 22 queries × 7 targets (0↑ 1↓)

Each cell is develop 75910442 → this PR b20fb3ee (hot runs), with PR ÷ develop in parentheses: below 1.00 is faster. Changes past ±10% are bold. The geomean row compares each target's geomean over its queries.
Vortex columns name the integration and the scan engine: file reads through the engine's file format (DataFusion ListingTable, DuckDB MultiFileReader), datasource through DataFusion's VortexTable over a Vortex DataSource; v1 and v2 are the LayoutReader and plan scans. Each column compares with its own develop times.

query datafusion:vortex [file, v2] datafusion:vortex [file, v1] datafusion:vortex [datasource, v1] datafusion:parquet duckdb:vortex [file, v2] duckdb:vortex [file, v1] duckdb:parquet
geomean 20.1 → 20.1 ms (1.00) 21.2 → 21.2 ms (1.00) 63.9 → 63.7 ms (1.00) 77.7 → 77.6 ms (1.00) 15.5 → 15.6 ms (1.01) 15.5 → 15.7 ms (1.01) 55.8 → 56.1 ms (1.00)
tpch_q01 15.6 → 16.2 ms (1.04) 16.2 → 16.3 ms (1.01) 253 → 254 ms (1.00) 71.2 → 71.4 ms (1.00) 9.53 → 9.76 ms (1.02) 8.97 → 10.0 ms (1.12) 🔴 47.5 → 47.4 ms (1.00)
tpch_q02 17.6 → 17.7 ms (1.00) 17.9 → 17.7 ms (0.99) 37.0 → 37.1 ms (1.00) 76.2 → 76.4 ms (1.00) 15.8 → 15.8 ms (1.00) 14.8 → 14.9 ms (1.01) 48.2 → 48.3 ms (1.00)
tpch_q03 19.6 → 18.9 ms (0.96) 19.5 → 19.2 ms (0.98) 52.3 → 52.2 ms (1.00) 81.9 → 81.1 ms (0.99) 18.6 → 18.7 ms (1.01) 18.9 → 18.8 ms (0.99) 60.3 → 60.2 ms (1.00)
tpch_q04 10.2 → 10.7 ms (1.04) 9.80 → 9.82 ms (1.00) 56.7 → 56.8 ms (1.00) 39.5 → 39.1 ms (0.99) 12.0 → 12.3 ms (1.02) 11.9 → 12.1 ms (1.01) 34.6 → 34.7 ms (1.00)
tpch_q05 37.8 → 37.9 ms (1.00) 37.8 → 37.3 ms (0.99) 161 → 161 ms (1.00) 110 → 110 ms (1.00) 20.2 → 20.5 ms (1.01) 20.0 → 19.9 ms (1.00) 67.0 → 67.1 ms (1.00)
tpch_q06 4.57 → 4.37 ms (0.96) 6.11 → 5.97 ms (0.98) 7.33 → 7.36 ms (1.00) 31.4 → 31.3 ms (0.99) 4.70 → 4.74 ms (1.01) 4.78 → 5.11 ms (1.07) 21.9 → 21.8 ms (0.99)
tpch_q07 52.7 → 53.0 ms (1.01) 51.8 → 51.6 ms (0.99) 194 → 191 ms (0.99) 113 → 113 ms (1.00) 24.3 → 24.1 ms (0.99) 23.9 → 24.0 ms (1.00) 63.9 → 64.1 ms (1.00)
tpch_q08 51.7 → 52.2 ms (1.01) 51.0 → 51.9 ms (1.02) 89.4 → 89.1 ms (1.00) 118 → 117 ms (1.00) 21.4 → 21.3 ms (0.99) 20.7 → 21.0 ms (1.01) 75.6 → 75.8 ms (1.00)
tpch_q09 47.2 → 47.4 ms (1.00) 48.9 → 50.0 ms (1.02) 191 → 190 ms (1.00) 137 → 137 ms (1.00) 22.9 → 23.5 ms (1.03) 21.4 → 21.8 ms (1.02) 107 → 108 ms (1.01)
tpch_q10 18.4 → 17.4 ms (0.95) 18.8 → 18.4 ms (0.98) 85.6 → 85.6 ms (1.00) 104 → 104 ms (1.00) 27.8 → 27.8 ms (1.00) 27.7 → 27.9 ms (1.01) 105 → 106 ms (1.01)
tpch_q11 14.7 → 14.5 ms (0.99) 15.1 → 15.1 ms (1.00) 28.6 → 28.3 ms (0.99) 61.9 → 62.9 ms (1.02) 8.00 → 8.21 ms (1.03) 7.71 → 7.80 ms (1.01) 29.6 → 29.4 ms (1.00)
tpch_q12 18.2 → 18.8 ms (1.03) 22.3 → 22.0 ms (0.99) 47.3 → 47.2 ms (1.00) 64.0 → 63.5 ms (0.99) 14.3 → 14.6 ms (1.02) 15.7 → 15.8 ms (1.00) 37.1 → 37.3 ms (1.01)
tpch_q13 19.6 → 19.8 ms (1.01) 19.3 → 19.3 ms (1.00) 34.9 → 35.4 ms (1.01) 143 → 143 ms (1.00) 19.3 → 19.6 ms (1.02) 19.7 → 19.8 ms (1.01) 175 → 175 ms (1.00)
tpch_q14 14.2 → 14.1 ms (0.99) 14.5 → 14.6 ms (1.01) 23.0 → 23.0 ms (1.00) 47.8 → 47.4 ms (0.99) 11.2 → 11.3 ms (1.01) 11.1 → 11.3 ms (1.02) 42.3 → 42.3 ms (1.00)
tpch_q15 18.9 → 18.7 ms (0.99) 18.8 → 18.8 ms (1.00) 37.1 → 36.5 ms (0.99) 73.8 → 73.7 ms (1.00) 8.18 → 8.43 ms (1.03) 8.50 → 8.44 ms (0.99) 32.7 → 32.7 ms (1.00)
tpch_q16 18.4 → 17.8 ms (0.97) 18.4 → 18.6 ms (1.01) 31.4 → 31.0 ms (0.99) 43.4 → 43.4 ms (1.00) 16.7 → 16.5 ms (0.99) 16.7 → 16.8 ms (1.01) 47.0 → 47.6 ms (1.01)
tpch_q17 29.4 → 29.7 ms (1.01) 28.7 → 29.1 ms (1.01) 214 → 216 ms (1.01) 123 → 123 ms (1.00) 13.0 → 12.9 ms (0.99) 13.6 → 13.3 ms (0.98) 47.9 → 48.0 ms (1.00)
tpch_q18 36.5 → 36.5 ms (1.00) 36.2 → 36.0 ms (0.99) 280 → 278 ms (0.99) 121 → 121 ms (1.00) 23.9 → 24.1 ms (1.01) 25.6 → 25.7 ms (1.00) 73.8 → 74.5 ms (1.01)
tpch_q19 9.76 → 9.55 ms (0.98) 15.0 → 15.0 ms (1.00) 18.3 → 17.7 ms (0.97) 57.6 → 57.0 ms (0.99) 11.1 → 11.4 ms (1.02) 12.0 → 12.7 ms (1.06) 47.5 → 50.0 ms (1.05)
tpch_q20 20.5 → 20.8 ms (1.01) 20.9 → 21.2 ms (1.01) 51.9 → 51.1 ms (0.98) 78.4 → 77.6 ms (0.99) 14.5 → 14.4 ms (1.00) 14.9 → 15.0 ms (1.00) 63.4 → 63.5 ms (1.00)
tpch_q21 34.6 → 35.2 ms (1.02) 37.8 → 38.4 ms (1.02) 227 → 227 ms (1.00) 114 → 114 ms (1.00) 48.9 → 47.7 ms (0.98) 48.1 → 48.1 ms (1.00) 107 → 107 ms (1.00)
tpch_q22 10.7 → 10.9 ms (1.02) 12.2 → 12.5 ms (1.02) 35.0 → 35.2 ms (1.00) 51.4 → 52.0 ms (1.01) 16.6 → 17.0 ms (1.02) 16.7 → 16.9 ms (1.01) 50.4 → 50.7 ms (1.01)
Cold runs (first iteration): geomean 1.010x ➖ · 22 queries × 7 targets (0↑ 1↓)
query datafusion:vortex [file, v2] datafusion:vortex [file, v1] datafusion:vortex [datasource, v1] datafusion:parquet duckdb:vortex [file, v2] duckdb:vortex [file, v1] duckdb:parquet
geomean 46.6 → 47.4 ms (1.02) 46.1 → 46.7 ms (1.01) 91.2 → 91.5 ms (1.00) 100 → 99.9 ms (1.00) 29.0 → 29.4 ms (1.01) 29.0 → 29.4 ms (1.02) 64.4 → 65.2 ms (1.01)
tpch_q01 43.5 → 45.0 ms (1.04) 42.2 → 44.8 ms (1.06) 273 → 274 ms (1.00) 89.0 → 88.2 ms (0.99) 21.1 → 21.2 ms (1.00) 20.3 → 21.1 ms (1.04) 54.0 → 54.4 ms (1.01)
tpch_q02 37.8 → 39.5 ms (1.05) 34.3 → 36.6 ms (1.07) 54.2 → 53.7 ms (0.99) 92.6 → 92.2 ms (1.00) 24.7 → 25.4 ms (1.03) 24.2 → 26.1 ms (1.08) 55.4 → 55.3 ms (1.00)
tpch_q03 46.9 → 45.3 ms (0.97) 46.4 → 44.6 ms (0.96) 78.2 → 79.2 ms (1.01) 104 → 103 ms (0.99) 35.5 → 35.8 ms (1.01) 34.1 → 35.3 ms (1.04) 68.6 → 69.1 ms (1.01)
tpch_q04 34.1 → 34.9 ms (1.02) 32.0 → 32.2 ms (1.01) 74.1 → 74.3 ms (1.00) 58.9 → 58.0 ms (0.99) 25.8 → 25.6 ms (0.99) 23.7 → 23.8 ms (1.00) 42.8 → 43.3 ms (1.01)
tpch_q05 70.3 → 67.2 ms (0.96) 63.5 → 66.1 ms (1.04) 185 → 186 ms (1.01) 136 → 135 ms (0.99) 37.7 → 37.0 ms (0.98) 36.7 → 38.2 ms (1.04) 77.1 → 77.7 ms (1.01)
tpch_q06 24.5 → 26.4 ms (1.08) 27.4 → 27.4 ms (1.00) 23.0 → 24.5 ms (1.07) 48.2 → 48.1 ms (1.00) 14.3 → 14.4 ms (1.00) 15.0 → 15.9 ms (1.06) 26.9 → 27.5 ms (1.02)
tpch_q07 78.5 → 80.9 ms (1.03) 76.5 → 73.9 ms (0.97) 220 → 222 ms (1.01) 141 → 138 ms (0.98) 42.0 → 41.4 ms (0.99) 40.3 → 40.2 ms (1.00) 72.9 → 74.0 ms (1.01)
tpch_q08 83.5 → 83.7 ms (1.00) 78.6 → 82.3 ms (1.05) 113 → 114 ms (1.01) 143 → 144 ms (1.01) 36.8 → 36.8 ms (1.00) 38.6 → 38.7 ms (1.00) 86.7 → 86.9 ms (1.00)
tpch_q09 76.6 → 82.5 ms (1.08) 79.9 → 77.4 ms (0.97) 220 → 218 ms (0.99) 168 → 167 ms (1.00) 37.1 → 38.8 ms (1.05) 39.5 → 40.5 ms (1.03) 118 → 120 ms (1.01)
tpch_q10 49.3 → 49.3 ms (1.00) 48.3 → 48.1 ms (1.00) 117 → 119 ms (1.01) 128 → 127 ms (0.99) 45.2 → 44.1 ms (0.98) 43.0 → 43.4 ms (1.01) 117 → 118 ms (1.01)
tpch_q11 34.7 → 35.9 ms (1.03) 32.0 → 33.7 ms (1.05) 47.6 → 45.8 ms (0.96) 81.3 → 80.3 ms (0.99) 17.2 → 17.2 ms (1.00) 17.2 → 17.8 ms (1.03) 34.9 → 35.2 ms (1.01)
tpch_q12 43.4 → 42.5 ms (0.98) 43.7 → 43.8 ms (1.00) 67.0 → 67.1 ms (1.00) 83.3 → 82.7 ms (0.99) 25.9 → 26.0 ms (1.00) 29.2 → 29.7 ms (1.02) 44.4 → 44.8 ms (1.01)
tpch_q13 37.9 → 38.2 ms (1.01) 41.6 → 41.7 ms (1.00) 51.1 → 51.2 ms (1.00) 163 → 163 ms (1.00) 31.7 → 34.5 ms (1.09) 31.7 → 30.7 ms (0.97) 182 → 184 ms (1.01)
tpch_q14 38.7 → 38.0 ms (0.98) 37.7 → 38.1 ms (1.01) 43.5 → 42.5 ms (0.98) 68.5 → 71.3 ms (1.04) 24.3 → 25.2 ms (1.03) 24.2 → 23.6 ms (0.97) 51.0 → 51.6 ms (1.01)
tpch_q15 47.4 → 48.4 ms (1.02) 44.8 → 43.5 ms (0.97) 54.3 → 54.0 ms (0.99) 93.6 → 95.6 ms (1.02) 21.5 → 21.1 ms (0.98) 21.2 → 21.8 ms (1.03) 41.2 → 41.4 ms (1.00)
tpch_q16 39.3 → 40.2 ms (1.02) 38.5 → 38.8 ms (1.01) 50.4 → 48.3 ms (0.96) 61.6 → 62.5 ms (1.01) 23.4 → 24.0 ms (1.03) 23.3 → 24.1 ms (1.04) 52.2 → 53.0 ms (1.01)
tpch_q17 59.2 → 59.3 ms (1.00) 59.4 → 57.7 ms (0.97) 236 → 239 ms (1.01) 147 → 149 ms (1.01) 27.8 → 29.4 ms (1.06) 27.3 → 27.6 ms (1.01) 57.1 → 58.0 ms (1.02)
tpch_q18 58.3 → 59.2 ms (1.02) 58.3 → 58.5 ms (1.00) 309 → 309 ms (1.00) 143 → 141 ms (0.99) 34.6 → 34.5 ms (1.00) 34.8 → 36.1 ms (1.04) 80.0 → 80.7 ms (1.01)
tpch_q19 35.6 → 38.5 ms (1.08) 40.1 → 42.6 ms (1.06) 38.8 → 40.6 ms (1.05) 81.9 → 82.3 ms (1.00) 27.2 → 27.4 ms (1.01) 26.7 → 26.5 ms (0.99) 54.3 → 60.4 ms (1.11) 🔴
tpch_q20 44.8 → 47.3 ms (1.06) 45.3 → 45.6 ms (1.01) 74.3 → 75.0 ms (1.01) 102 → 99.0 ms (0.97) 29.9 → 30.8 ms (1.03) 30.4 → 29.7 ms (0.98) 74.2 → 74.1 ms (1.00)
tpch_q21 66.6 → 65.7 ms (0.99) 66.9 → 69.2 ms (1.03) 265 → 269 ms (1.02) 139 → 137 ms (0.99) 63.1 → 63.6 ms (1.01) 64.7 → 63.2 ms (0.98) 115 → 116 ms (1.01)
tpch_q22 30.5 → 30.0 ms (0.98) 29.1 → 30.1 ms (1.03) 52.6 → 52.3 ms (0.99) 68.1 → 67.2 ms (0.99) 26.3 → 26.9 ms (1.02) 27.2 → 27.0 ms (0.99) 58.6 → 59.1 ms (1.01)

No file size changes detected.

@github-actions

Copy link
Copy Markdown
Contributor

Benchmarks: Clickbench on NVME 📖

Commits: PR b20fb3ee vs base 75910442
Verdict: No clear signal (low confidence)
Attributed Vortex impact: +1.2%
Engines: DataFusion No clear signal (+0.8%, low confidence) · DuckDB No clear signal (+1.9%, low confidence)
Vortex (geomean): hot 1.008x ➖ · cold 1.024x ➖
Parquet (geomean): hot 0.996x ➖ · cold 1.016x ➖
Shifts: Parquet (control) -0.4% · Median polish -0.1%

How to read Verdict and Engines
  • Verdict: Overall PR-level signal after subtracting baseline drift estimated from Parquet control rows. It can be Likely improvement, Likely regression, or No clear signal.
  • Engines: Per-engine attribution. DataFusion is compared against DataFusion/Parquet controls; DuckDB is compared against DuckDB/Parquet controls. This answers whether each engine improved or regressed independently.
  • Confidence: Based on directional consistency, share of rows above the noise floor, and control-run noise.
  • Hot vs cold: Every measurement is run several times. The first run is reported as the cold run, and the median of the runs after it is reported as the hot run. The verdict and significance use hot runs; each target's geomean reports hot and cold beside each other where the individual runs were recorded and hot alone where they were not, the cold column shows first-run cost per row, and hot/cold is how much of each run the warm path saves. Rows whose results predate per-run reporting show only one value, taken from the value the runner reported.
  • Query tables: One row per query and one column per target. Each cell reads develop → PR (PR ÷ develop), so (0.80) is 20% faster than develop. A change that clears this suite's threshold is bold and marked 🔴 slower or 🟢 faster; anything smaller is noise here and is left plain. The geomean row compares each target's geomean over its queries with develop's. Hot runs are shown first and cold runs are folded below them.
  • Other tables: Each cell reads PR / base / %diff, marked the same way.

Hot runs: geomean 1.004x ➖ · 43 queries × 7 targets (3↑ 10↓)

Each cell is develop 75910442 → this PR b20fb3ee (hot runs), with PR ÷ develop in parentheses: below 1.00 is faster. Changes past ±10% are bold. The geomean row compares each target's geomean over its queries.
Vortex columns name the integration and the scan engine: file reads through the engine's file format (DataFusion ListingTable, DuckDB MultiFileReader), datasource through DataFusion's VortexTable over a Vortex DataSource; v1 and v2 are the LayoutReader and plan scans. Each column compares with its own develop times.

query datafusion:vortex [file, v2] datafusion:vortex [file, v1] datafusion:vortex [datasource, v1] datafusion:parquet duckdb:vortex [file, v2] duckdb:vortex [file, v1] duckdb:parquet
geomean 41.1 → 40.9 ms (0.99) 42.4 → 42.0 ms (0.99) 52.4 → 53.4 ms (1.02) 65.3 → 65.0 ms (0.99) 35.7 → 35.9 ms (1.00) 35.1 → 36.2 ms (1.03) 67.1 → 66.9 ms (1.00)
clickbench_q00 1.25 → 1.27 ms (1.02) 1.25 → 1.24 ms (1.00) 2.47 → 2.37 ms (0.96) 1.27 → 1.26 ms (1.00) 2.22 → 2.23 ms (1.00) 2.26 → 2.21 ms (0.98) 9.78 → 9.04 ms (0.92)
clickbench_q01 8.99 → 9.36 ms (1.04) 9.22 → 8.45 ms (0.92) 7.71 → 7.58 ms (0.98) 15.9 → 16.4 ms (1.03) 7.34 → 7.10 ms (0.97) 7.06 → 9.00 ms (1.27) 🔴 15.7 → 15.0 ms (0.95)
clickbench_q02 13.8 → 13.3 ms (0.97) 12.3 → 12.6 ms (1.02) 9.88 → 10.0 ms (1.01) 18.7 → 18.9 ms (1.01) 9.54 → 9.26 ms (0.97) 8.06 → 10.1 ms (1.26) 🔴 22.7 → 22.3 ms (0.98)
clickbench_q03 14.0 → 14.0 ms (1.00) 12.9 → 12.8 ms (0.99) 9.76 → 10.8 ms (1.10) 🔴 16.6 → 17.2 ms (1.03) 13.9 → 14.5 ms (1.04) 13.7 → 13.9 ms (1.01) 23.0 → 23.2 ms (1.01)
clickbench_q04 58.8 → 59.3 ms (1.01) 56.9 → 56.9 ms (1.00) 73.0 → 71.4 ms (0.98) 65.5 → 63.0 ms (0.96) 56.1 → 55.8 ms (0.99) 54.3 → 54.5 ms (1.00) 62.3 → 61.4 ms (0.99)
clickbench_q05 82.2 → 79.2 ms (0.96) 76.1 → 77.7 ms (1.02) 120 → 116 ms (0.97) 88.9 → 88.4 ms (0.99) 63.7 → 64.9 ms (1.02) 62.4 → 63.0 ms (1.01) 82.2 → 81.9 ms (1.00)
clickbench_q06 1.26 → 1.24 ms (0.98) 1.25 → 1.22 ms (0.97) 8.47 → 8.37 ms (0.99) 1.30 → 1.26 ms (0.97) 2.67 → 2.75 ms (1.03) 2.70 → 2.75 ms (1.02) 20.8 → 20.3 ms (0.98)
clickbench_q07 12.4 → 12.2 ms (0.98) 12.2 → 12.2 ms (1.00) 13.6 → 14.3 ms (1.05) 17.1 → 17.1 ms (1.00) 9.00 → 9.52 ms (1.06) 8.61 → 10.1 ms (1.17) 🔴 17.1 → 17.1 ms (1.00)
clickbench_q08 82.4 → 81.4 ms (0.99) 79.4 → 78.5 ms (0.99) 98.3 → 99.6 ms (1.01) 84.2 → 84.5 ms (1.00) 69.1 → 70.6 ms (1.02) 67.0 → 66.6 ms (0.99) 76.4 → 74.0 ms (0.97)
clickbench_q09 180 → 174 ms (0.97) 169 → 170 ms (1.01) 184 → 186 ms (1.01) 190 → 187 ms (0.99) 92.7 → 93.8 ms (1.01) 90.1 → 86.7 ms (0.96) 102 → 103 ms (1.01)
clickbench_q10 24.6 → 24.5 ms (1.00) 24.5 → 24.7 ms (1.01) 26.2 → 25.4 ms (0.97) 36.8 → 35.8 ms (0.97) 27.3 → 27.8 ms (1.02) 26.8 → 27.0 ms (1.01) 40.8 → 41.4 ms (1.01)
clickbench_q11 25.8 → 26.2 ms (1.02) 27.2 → 27.0 ms (0.99) 29.8 → 29.4 ms (0.99) 42.4 → 42.5 ms (1.00) 30.6 → 30.6 ms (1.00) 28.9 → 29.4 ms (1.02) 46.3 → 47.1 ms (1.02)
clickbench_q12 75.1 → 75.5 ms (1.00) 72.6 → 72.8 ms (1.00) 103 → 101 ms (0.98) 95.3 → 94.9 ms (1.00) 69.6 → 69.6 ms (1.00) 68.0 → 67.8 ms (1.00) 92.5 → 93.6 ms (1.01)
clickbench_q13 122 → 119 ms (0.98) 120 → 122 ms (1.01) 123 → 124 ms (1.01) 146 → 147 ms (1.01) 117 → 118 ms (1.01) 116 → 117 ms (1.01) 149 → 152 ms (1.02)
clickbench_q14 78.4 → 77.8 ms (0.99) 78.7 → 77.0 ms (0.98) 101 → 114 ms (1.12) 🔴 102 → 102 ms (1.00) 72.2 → 75.1 ms (1.04) 73.9 → 71.9 ms (0.97) 99.9 → 101 ms (1.02)
clickbench_q15 65.9 → 66.0 ms (1.00) 65.3 → 63.3 ms (0.97) 81.8 → 87.0 ms (1.06) 73.6 → 72.1 ms (0.98) 75.5 → 76.0 ms (1.01) 75.7 → 75.6 ms (1.00) 85.8 → 85.9 ms (1.00)
clickbench_q16 156 → 157 ms (1.00) 154 → 155 ms (1.01) 204 → 217 ms (1.06) 157 → 156 ms (0.99) 134 → 137 ms (1.02) 134 → 137 ms (1.03) 169 → 168 ms (0.99)
clickbench_q17 158 → 156 ms (0.99) 151 → 153 ms (1.01) 207 → 202 ms (0.98) 157 → 156 ms (0.99) 137 → 145 ms (1.06) 136 → 137 ms (1.01) 163 → 165 ms (1.01)
clickbench_q18 300 → 298 ms (0.99) 296 → 288 ms (0.97) 347 → 342 ms (0.99) 310 → 306 ms (0.99) 256 → 257 ms (1.00) 253 → 250 ms (0.99) 290 → 290 ms (1.00)
clickbench_q19 10.0 → 9.74 ms (0.97) 9.15 → 8.91 ms (0.97) 8.92 → 9.11 ms (1.02) 19.1 → 19.5 ms (1.02) 9.03 → 9.36 ms (1.04) 9.21 → 9.91 ms (1.07) 20.1 → 20.0 ms (0.99)
clickbench_q20 46.2 → 44.9 ms (0.97) 45.3 → 45.5 ms (1.01) 41.6 → 41.0 ms (0.98) 178 → 178 ms (1.00) 51.6 → 52.0 ms (1.01) 48.2 → 50.9 ms (1.06) 180 → 180 ms (1.00)
clickbench_q21 59.4 → 59.2 ms (1.00) 89.0 → 92.6 ms (1.04) 84.0 → 85.1 ms (1.01) 174 → 173 ms (0.99) 56.7 → 56.5 ms (1.00) 88.6 → 88.3 ms (1.00) 221 → 220 ms (0.99)
clickbench_q22 75.4 → 73.3 ms (0.97) 99.3 → 99.6 ms (1.00) 92.1 → 91.5 ms (0.99) 262 → 261 ms (1.00) 89.6 → 91.3 ms (1.02) 137 → 138 ms (1.01) 305 → 305 ms (1.00)
clickbench_q23 96.7 → 89.0 ms (0.92) 147 → 126 ms (0.86) 🟢 150 → 181 ms (1.20) 🔴 1.01 → 1.01 s (1.00) 70.8 → 62.9 ms (0.89) 🟢 34.4 → 39.8 ms (1.16) 🔴 173 → 166 ms (0.96)
clickbench_q24 8.78 → 8.68 ms (0.99) 15.4 → 15.4 ms (1.00) 13.1 → 12.8 ms (0.98) 29.9 → 29.7 ms (1.00) 16.8 → 17.1 ms (1.02) 13.6 → 14.6 ms (1.07) 30.2 → 31.1 ms (1.03)
clickbench_q25 19.8 → 19.9 ms (1.01) 19.7 → 19.2 ms (0.98) 16.6 → 16.9 ms (1.01) 38.2 → 35.8 ms (0.94) 19.1 → 19.5 ms (1.02) 19.0 → 18.9 ms (1.00) 42.8 → 42.3 ms (0.99)
clickbench_q26 14.1 → 14.8 ms (1.05) 16.0 → 15.9 ms (0.99) 17.6 → 17.6 ms (1.00) 30.8 → 30.0 ms (0.98) 13.8 → 13.9 ms (1.01) 11.7 → 12.7 ms (1.09) 30.5 → 30.9 ms (1.02)
clickbench_q27 101 → 104 ms (1.03) 99.3 → 97.0 ms (0.98) 100 → 104 ms (1.04) 193 → 192 ms (1.00) 51.2 → 51.4 ms (1.00) 50.4 → 49.9 ms (0.99) 185 → 191 ms (1.03)
clickbench_q28 559 → 551 ms (0.99) 564 → 553 ms (0.98) 840 → 814 ms (0.97) 622 → 614 ms (0.99) 515 → 520 ms (1.01) 505 → 507 ms (1.00) 956 → 971 ms (1.02)
clickbench_q29 26.5 → 26.5 ms (1.00) 25.2 → 25.3 ms (1.01) 21.8 → 21.7 ms (1.00) 29.4 → 30.1 ms (1.03) 17.0 → 17.1 ms (1.00) 16.4 → 18.2 ms (1.11) 🔴 24.3 → 24.0 ms (0.99)
clickbench_q30 60.3 → 60.2 ms (1.00) 60.1 → 58.2 ms (0.97) 75.2 → 80.7 ms (1.07) 91.1 → 89.7 ms (0.98) 61.7 → 60.1 ms (0.97) 58.6 → 59.1 ms (1.01) 106 → 105 ms (0.99)
clickbench_q31 69.4 → 69.1 ms (1.00) 69.5 → 70.6 ms (1.02) 77.1 → 78.0 ms (1.01) 104 → 103 ms (0.98) 110 → 108 ms (0.98) 109 → 98.7 ms (0.91) 157 → 158 ms (1.00)
clickbench_q32 293 → 292 ms (1.00) 288 → 288 ms (1.00) 290 → 291 ms (1.00) 307 → 305 ms (1.00) 359 → 375 ms (1.04) 371 → 354 ms (0.95) 386 → 385 ms (1.00)
clickbench_q33 325 → 322 ms (0.99) 326 → 330 ms (1.01) 454 → 462 ms (1.02) 381 → 375 ms (0.98) 343 → 348 ms (1.01) 350 → 343 ms (0.98) 478 → 479 ms (1.00)
clickbench_q34 328 → 321 ms (0.98) 326 → 328 ms (1.01) 514 → 457 ms (0.89) 🟢 378 → 375 ms (0.99) 354 → 355 ms (1.00) 357 → 355 ms (0.99) 457 → 474 ms (1.04)
clickbench_q35 61.6 → 62.5 ms (1.01) 60.2 → 60.2 ms (1.00) 85.4 → 86.3 ms (1.01) 64.4 → 63.9 ms (0.99) 57.4 → 57.8 ms (1.01) 56.5 → 57.7 ms (1.02) 60.9 → 61.1 ms (1.00)
clickbench_q36 37.5 → 37.1 ms (0.99) 37.3 → 37.3 ms (1.00) 46.9 → 57.0 ms (1.21) 🔴 73.9 → 73.5 ms (0.99) 11.2 → 11.1 ms (0.99) 11.8 → 11.8 ms (1.00) 30.1 → 30.7 ms (1.02)
clickbench_q37 21.2 → 21.3 ms (1.00) 21.9 → 21.9 ms (1.00) 29.6 → 30.7 ms (1.04) 35.7 → 35.7 ms (1.00) 7.79 → 7.67 ms (0.98) 7.86 → 8.51 ms (1.08) 23.6 → 23.3 ms (0.99)
clickbench_q38 12.9 → 13.4 ms (1.04) 12.0 → 11.8 ms (0.98) 24.9 → 25.3 ms (1.01) 48.1 → 47.9 ms (1.00) 8.36 → 8.10 ms (0.97) 8.53 → 8.91 ms (1.04) 23.1 → 23.5 ms (1.02)
clickbench_q39 78.4 → 77.9 ms (0.99) 76.9 → 76.1 ms (0.99) 124 → 130 ms (1.04) 144 → 143 ms (1.00) 17.9 → 17.5 ms (0.98) 18.5 → 18.7 ms (1.01) 52.5 → 52.6 ms (1.00)
clickbench_q40 10.5 → 10.3 ms (0.98) 11.1 → 11.0 ms (0.99) 18.3 → 18.7 ms (1.02) 20.0 → 20.1 ms (1.00) 6.81 → 6.80 ms (1.00) 6.99 → 7.27 ms (1.04) 14.7 → 14.5 ms (0.98)
clickbench_q41 9.59 → 9.34 ms (0.97) 10.5 → 10.8 ms (1.03) 17.1 → 17.2 ms (1.01) 19.0 → 18.6 ms (0.98) 7.81 → 7.81 ms (1.00) 7.80 → 8.10 ms (1.04) 15.4 → 15.2 ms (0.98)
clickbench_q42 9.76 → 9.77 ms (1.00) 11.1 → 10.8 ms (0.97) 15.5 → 17.7 ms (1.14) 🔴 19.3 → 19.3 ms (1.00) 7.00 → 6.98 ms (1.00) 7.09 → 7.24 ms (1.02) 14.4 → 14.0 ms (0.97)
Cold runs (first iteration): geomean 1.022x ➖ · 43 queries × 7 targets (1↑ 22↓)
query datafusion:vortex [file, v2] datafusion:vortex [file, v1] datafusion:vortex [datasource, v1] datafusion:parquet duckdb:vortex [file, v2] duckdb:vortex [file, v1] duckdb:parquet
geomean 105 → 109 ms (1.03) 103 → 106 ms (1.03) 112 → 113 ms (1.01) 156 → 159 ms (1.02) 82.8 → 84.9 ms (1.03) 80.5 → 82.6 ms (1.03) 82.2 → 83.2 ms (1.01)
clickbench_q00 22.4 → 25.4 ms (1.13) 🔴 22.2 → 24.8 ms (1.12) 🔴 24.9 → 26.1 ms (1.05) 47.6 → 49.8 ms (1.05) 24.7 → 26.0 ms (1.05) 24.0 → 24.2 ms (1.01) 14.1 → 14.3 ms (1.02)
clickbench_q01 36.3 → 34.9 ms (0.96) 37.8 → 38.7 ms (1.02) 31.7 → 30.6 ms (0.97) 67.4 → 73.1 ms (1.08) 28.2 → 27.3 ms (0.97) 25.9 → 27.3 ms (1.05) 21.8 → 20.8 ms (0.96)
clickbench_q02 46.4 → 48.8 ms (1.05) 43.8 → 42.1 ms (0.96) 33.5 → 35.0 ms (1.05) 82.2 → 80.4 ms (0.98) 35.8 → 36.1 ms (1.01) 31.1 → 33.9 ms (1.09) 30.8 → 30.9 ms (1.00)
clickbench_q03 51.6 → 55.6 ms (1.08) 47.9 → 50.7 ms (1.06) 43.2 → 46.2 ms (1.07) 76.6 → 81.4 ms (1.06) 44.1 → 46.3 ms (1.05) 47.6 → 46.5 ms (0.98) 33.3 → 35.0 ms (1.05)
clickbench_q04 106 → 103 ms (0.97) 88.4 → 93.1 ms (1.05) 106 → 111 ms (1.04) 124 → 124 ms (1.00) 86.9 → 89.2 ms (1.03) 86.2 → 87.4 ms (1.01) 69.6 → 75.0 ms (1.08)
clickbench_q05 128 → 126 ms (0.99) 126 → 121 ms (0.96) 158 → 161 ms (1.02) 155 → 158 ms (1.02) 94.0 → 94.0 ms (1.00) 89.5 → 91.9 ms (1.03) 93.2 → 96.6 ms (1.04)
clickbench_q06 22.2 → 24.0 ms (1.08) 22.4 → 24.7 ms (1.10) 🔴 28.9 → 28.2 ms (0.97) 48.5 → 50.7 ms (1.05) 32.2 → 31.4 ms (0.97) 28.9 → 32.0 ms (1.11) 🔴 29.9 → 29.7 ms (0.99)
clickbench_q07 36.6 → 38.0 ms (1.04) 35.4 → 42.9 ms (1.21) 🔴 36.7 → 37.6 ms (1.03) 67.7 → 68.2 ms (1.01) 30.3 → 28.7 ms (0.95) 29.3 → 29.6 ms (1.01) 23.3 → 23.8 ms (1.02)
clickbench_q08 121 → 130 ms (1.07) 124 → 122 ms (0.99) 146 → 144 ms (0.99) 163 → 161 ms (0.99) 96.5 → 101 ms (1.05) 99.4 → 104 ms (1.05) 91.4 → 88.2 ms (0.97)
clickbench_q09 212 → 217 ms (1.02) 211 → 206 ms (0.98) 229 → 229 ms (1.00) 268 → 270 ms (1.01) 123 → 129 ms (1.04) 123 → 112 ms (0.91) 122 → 118 ms (0.97)
clickbench_q10 67.4 → 68.8 ms (1.02) 62.6 → 64.4 ms (1.03) 61.1 → 58.0 ms (0.95) 99.3 → 102 ms (1.03) 63.3 → 68.3 ms (1.08) 61.4 → 61.5 ms (1.00) 53.8 → 55.1 ms (1.02)
clickbench_q11 67.5 → 67.5 ms (1.00) 63.5 → 66.8 ms (1.05) 65.9 → 62.5 ms (0.95) 107 → 109 ms (1.02) 65.3 → 71.5 ms (1.10) 61.0 → 61.3 ms (1.01) 61.5 → 61.6 ms (1.00)
clickbench_q12 124 → 129 ms (1.04) 120 → 124 ms (1.03) 151 → 137 ms (0.91) 159 → 166 ms (1.04) 101 → 99.4 ms (0.98) 94.1 → 95.6 ms (1.02) 108 → 108 ms (1.00)
clickbench_q13 184 → 187 ms (1.01) 179 → 179 ms (1.00) 188 → 185 ms (0.99) 224 → 230 ms (1.03) 158 → 167 ms (1.06) 151 → 154 ms (1.02) 168 → 169 ms (1.00)
clickbench_q14 129 → 139 ms (1.07) 123 → 123 ms (0.99) 171 → 157 ms (0.92) 169 → 174 ms (1.03) 100 → 102 ms (1.02) 99.3 → 101 ms (1.02) 117 → 115 ms (0.99)
clickbench_q15 108 → 106 ms (0.98) 99.4 → 102 ms (1.03) 124 → 125 ms (1.01) 137 → 132 ms (0.96) 106 → 110 ms (1.03) 108 → 110 ms (1.02) 103 → 103 ms (1.01)
clickbench_q16 213 → 225 ms (1.06) 215 → 211 ms (0.98) 277 → 261 ms (0.94) 234 → 236 ms (1.01) 171 → 171 ms (1.00) 172 → 173 ms (1.00) 180 → 179 ms (0.99)
clickbench_q17 224 → 226 ms (1.01) 208 → 216 ms (1.04) 276 → 291 ms (1.05) 237 → 239 ms (1.01) 168 → 170 ms (1.01) 168 → 170 ms (1.01) 181 → 179 ms (0.99)
clickbench_q18 384 → 389 ms (1.01) 379 → 376 ms (0.99) 438 → 446 ms (1.02) 415 → 425 ms (1.02) 307 → 300 ms (0.98) 303 → 304 ms (1.00) 319 → 316 ms (0.99)
clickbench_q19 43.1 → 47.2 ms (1.10) 44.4 → 47.1 ms (1.06) 41.0 → 39.5 ms (0.96) 74.6 → 79.4 ms (1.06) 35.5 → 36.8 ms (1.04) 35.6 → 37.7 ms (1.06) 25.5 → 27.5 ms (1.08)
clickbench_q20 236 → 243 ms (1.03) 226 → 221 ms (0.98) 203 → 204 ms (1.01) 276 → 285 ms (1.03) 156 → 156 ms (1.00) 156 → 158 ms (1.02) 224 → 229 ms (1.02)
clickbench_q21 274 → 279 ms (1.02) 274 → 268 ms (0.98) 262 → 265 ms (1.01) 290 → 301 ms (1.04) 176 → 180 ms (1.02) 249 → 257 ms (1.03) 269 → 259 ms (0.96)
clickbench_q22 385 → 381 ms (0.99) 337 → 344 ms (1.02) 328 → 333 ms (1.01) 416 → 432 ms (1.04) 296 → 303 ms (1.03) 343 → 341 ms (1.00) 354 → 369 ms (1.04)
clickbench_q23 366 → 360 ms (0.98) 470 → 464 ms (0.99) 359 → 479 ms (1.34) 🔴 1.18 → 1.19 s (1.01) 160 → 152 ms (0.95) 84.3 → 84.7 ms (1.00) 247 → 239 ms (0.97)
clickbench_q24 55.5 → 66.2 ms (1.19) 🔴 62.1 → 67.2 ms (1.08) 53.1 → 42.4 ms (0.80) 🟢 87.2 → 89.7 ms (1.03) 44.3 → 47.8 ms (1.08) 35.0 → 33.8 ms (0.96) 45.3 → 44.9 ms (0.99)
clickbench_q25 63.7 → 65.3 ms (1.02) 57.4 → 61.5 ms (1.07) 47.2 → 49.9 ms (1.06) 101 → 103 ms (1.03) 50.2 → 53.8 ms (1.07) 48.5 → 44.7 ms (0.92) 61.2 → 57.4 ms (0.94)
clickbench_q26 63.3 → 70.8 ms (1.12) 🔴 64.1 → 65.9 ms (1.03) 48.0 → 52.4 ms (1.09) 89.2 → 91.2 ms (1.02) 41.6 → 45.9 ms (1.10) 🔴 31.0 → 34.2 ms (1.10) 🔴 39.3 → 51.1 ms (1.30) 🔴
clickbench_q27 258 → 258 ms (1.00) 245 → 239 ms (0.98) 225 → 230 ms (1.02) 280 → 293 ms (1.05) 170 → 175 ms (1.03) 212 → 214 ms (1.01) 219 → 228 ms (1.04)
clickbench_q28 640 → 656 ms (1.02) 622 → 626 ms (1.01) 991 → 949 ms (0.96) 702 → 705 ms (1.00) 550 → 558 ms (1.01) 559 → 559 ms (1.00) 938 → 941 ms (1.00)
clickbench_q29 59.2 → 58.8 ms (0.99) 56.2 → 52.0 ms (0.92) 46.9 → 46.9 ms (1.00) 90.1 → 91.9 ms (1.02) 41.1 → 44.4 ms (1.08) 40.0 → 44.5 ms (1.11) 🔴 32.0 → 31.1 ms (0.97)
clickbench_q30 108 → 107 ms (0.99) 105 → 109 ms (1.04) 129 → 144 ms (1.11) 🔴 167 → 173 ms (1.03) 102 → 107 ms (1.05) 98.6 → 102 ms (1.03) 127 → 128 ms (1.01)
clickbench_q31 154 → 155 ms (1.01) 154 → 155 ms (1.01) 165 → 167 ms (1.01) 191 → 198 ms (1.04) 183 → 177 ms (0.97) 157 → 182 ms (1.16) 🔴 173 → 172 ms (1.00)
clickbench_q32 391 → 390 ms (1.00) 383 → 385 ms (1.01) 415 → 414 ms (1.00) 427 → 424 ms (0.99) 395 → 406 ms (1.03) 404 → 389 ms (0.96) 402 → 420 ms (1.04)
clickbench_q33 437 → 434 ms (0.99) 433 → 444 ms (1.03) 564 → 601 ms (1.07) 468 → 478 ms (1.02) 389 → 411 ms (1.06) 429 → 423 ms (0.99) 440 → 448 ms (1.02)
clickbench_q34 431 → 430 ms (1.00) 436 → 462 ms (1.06) 698 → 631 ms (0.91) 474 → 475 ms (1.00) 415 → 417 ms (1.00) 437 → 438 ms (1.00) 438 → 468 ms (1.07)
clickbench_q35 95.6 → 93.8 ms (0.98) 89.8 → 91.2 ms (1.02) 131 → 135 ms (1.03) 124 → 125 ms (1.00) 85.5 → 87.8 ms (1.03) 89.6 → 86.5 ms (0.96) 75.3 → 73.4 ms (0.97)
clickbench_q36 66.8 → 71.5 ms (1.07) 67.2 → 70.3 ms (1.05) 83.4 → 77.7 ms (0.93) 133 → 136 ms (1.02) 32.9 → 32.7 ms (0.99) 30.6 → 34.5 ms (1.13) 🔴 35.8 → 35.9 ms (1.01)
clickbench_q37 46.6 → 48.4 ms (1.04) 48.6 → 48.2 ms (0.99) 49.6 → 56.2 ms (1.13) 🔴 90.5 → 90.9 ms (1.00) 29.2 → 30.6 ms (1.05) 31.1 → 29.9 ms (0.96) 28.9 → 29.8 ms (1.03)
clickbench_q38 39.9 → 43.1 ms (1.08) 38.9 → 42.5 ms (1.09) 50.4 → 49.6 ms (0.98) 107 → 109 ms (1.02) 29.8 → 30.0 ms (1.00) 26.9 → 30.2 ms (1.12) 🔴 29.8 → 29.4 ms (0.99)
clickbench_q39 111 → 114 ms (1.02) 108 → 114 ms (1.06) 128 → 165 ms (1.29) 🔴 207 → 212 ms (1.03) 42.5 → 42.8 ms (1.01) 41.9 → 44.7 ms (1.07) 59.5 → 60.3 ms (1.01)
clickbench_q40 34.9 → 37.3 ms (1.07) 36.0 → 39.3 ms (1.09) 41.9 → 40.8 ms (0.97) 76.3 → 77.5 ms (1.02) 29.4 → 29.6 ms (1.01) 27.7 → 31.1 ms (1.12) 🔴 20.0 → 20.1 ms (1.00)
clickbench_q41 34.7 → 38.6 ms (1.11) 🔴 35.7 → 38.5 ms (1.08) 39.9 → 38.3 ms (0.96) 76.7 → 75.6 ms (0.99) 28.9 → 30.1 ms (1.04) 26.8 → 29.8 ms (1.11) 🔴 20.2 → 20.4 ms (1.01)
clickbench_q42 35.7 → 36.6 ms (1.03) 36.4 → 38.1 ms (1.05) 39.8 → 36.4 ms (0.91) 74.5 → 74.8 ms (1.00) 26.3 → 29.2 ms (1.11) 🔴 28.5 → 29.6 ms (1.04) 20.3 → 20.7 ms (1.02)

No file size changes detected.

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

changelog/chore A trivial change

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant