Skip to content

Add take benchmarks for FSL(Chunked) and Chunked(FSL) arrays - #8887

Merged
robert3005 merged 1 commit into
developfrom
rk/chunkedfsltakebench
Jul 21, 2026
Merged

Add take benchmarks for FSL(Chunked) and Chunked(FSL) arrays#8887
robert3005 merged 1 commit into
developfrom
rk/chunkedfsltakebench

Conversation

@robert3005

Copy link
Copy Markdown
Contributor

Benchmarks take over a FixedSizeListArray with chunked elements and over
a chunked array of FixedSizeList chunks, with random and sorted indices.

Benchmarks take over a FixedSizeListArray with chunked elements and over
a chunked array of FixedSizeList chunks, with random and sorted indices.
Sizes are chosen so the slowest current take path stays under 1ms per
iteration under codspeed simulation.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: Robert Kruszewski <github@robertk.io>
@robert3005 robert3005 added the changelog/chore A trivial change label Jul 21, 2026
@github-actions

github-actions Bot commented Jul 21, 2026

Copy link
Copy Markdown
Contributor

Polar Signals Profiling Results

Latest Run

Status Commit Job Attempt Link
🟢 Done bb43db6 1 Explore Profiling Data

Powered by Polar Signals Cloud

@github-actions

Copy link
Copy Markdown
Contributor

Benchmarks: Vortex queries 📖

Verdict: No clear signal (low confidence)
Attributed Vortex impact: +3.5%
Engines: DataFusion No clear signal (+5.1%, low confidence) · DuckDB No clear signal (+1.8%, low confidence)
Vortex (geomean): 1.020x ➖
Parquet (geomean): 1.003x ➖
Shifts: Parquet (control) +0.3% · Median polish +1.9%

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.

datafusion / vortex-file-compressed (1.046x ➖, 0↑ 0↓)
name PR bb43db6 (ns) base b7ee873 (ns) ratio (PR/base)
vortex_q00/datafusion:vortex-file-compressed 9977406 9472638 1.05
vortex_q01/datafusion:vortex-file-compressed 6175482 5947344 1.04
datafusion / parquet (0.995x ➖, 0↑ 0↓)
name PR bb43db6 (ns) base b7ee873 (ns) ratio (PR/base)
vortex_q00/datafusion:parquet 19980064 20526611 0.97
vortex_q01/datafusion:parquet 4676326 4597130 1.02
duckdb / vortex-file-compressed (1.029x ➖, 0↑ 0↓)
name PR bb43db6 (ns) base b7ee873 (ns) ratio (PR/base)
vortex_q00/duckdb:vortex-file-compressed 10238920 10002024 1.02
vortex_q01/duckdb:vortex-file-compressed 6030228 5833329 1.03
duckdb / parquet (1.010x ➖, 0↑ 0↓)
name PR bb43db6 (ns) base b7ee873 (ns) ratio (PR/base)
vortex_q00/duckdb:parquet 23414720 23225002 1.01
vortex_q01/duckdb:parquet 9430748 9318548 1.01

No file size changes detected.

@robert3005
robert3005 enabled auto-merge (squash) July 21, 2026 18:07
@robert3005
robert3005 merged commit 69f5bc5 into develop Jul 21, 2026
80 of 81 checks passed
@robert3005
robert3005 deleted the rk/chunkedfsltakebench branch July 21, 2026 21:03
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.

2 participants