Run SQL benchmarks with plan v2 (full stack vs develop) - #9342
Draft
joseph-isaacs wants to merge 9 commits into
Draft
Run SQL benchmarks with plan v2 (full stack vs develop)#9342joseph-isaacs wants to merge 9 commits into
joseph-isaacs wants to merge 9 commits into
Conversation
joseph-isaacs
added
the
action/bench-all
Claude
github-actions
Bot
removed
the
action/bench-all
joseph-isaacs
added
the
changelog/skip
Claude
Contributor
Polar Signals Profiling ResultsLatest Run
Previous Runs (60)
Powered by Polar Signals Cloud |
Contributor
Benchmarks: PolarSignals Profiling 📖Vortex (geomean): 1.024x ➖ datafusion / vortex-file-compressed / ns (1.024x ➖, 0↑ 1↓)
No file size changes detected. |
Contributor
Benchmarks: String Encoding 📖vortex / vortex-file-compressed / ms (0.999x ➖, 0↑ 0↓)
vortex / vortex-file-compressed / % (1.000x ➖, 0↑ 0↓)
|
Contributor
Benchmarks: TPC-H SF=1 on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (1.073x ➖, 0↑ 6↓)
datafusion / vortex-compact / ns (1.043x ➖, 1↑ 6↓)
datafusion / parquet / ns (0.991x ➖, 2↑ 2↓)
duckdb / vortex-file-compressed / ns (0.972x ➖, 1↑ 0↓)
duckdb / vortex-compact / ns (0.984x ➖, 0↑ 1↓)
duckdb / parquet / ns (0.998x ➖, 0↑ 0↓)
No file size changes detected. |
Contributor
Benchmarks: FineWeb NVMe 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (1.280x ❌, 0↑ 5↓)
datafusion / vortex-compact / ns (1.017x ➖, 1↑ 1↓)
datafusion / parquet / ns (0.998x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed / ns (1.087x ➖, 0↑ 4↓)
duckdb / vortex-compact / ns (0.969x ➖, 0↑ 0↓)
duckdb / parquet / ns (1.002x ➖, 0↑ 0↓)
No file size changes detected. |
Contributor
Benchmarks: Clickbench Sorted on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (1.043x ➖, 1↑ 4↓)
datafusion / vortex-compact / ns (1.201x ❌, 0↑ 5↓)
datafusion / parquet / ns (0.992x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed / ns (1.026x ➖, 0↑ 2↓)
duckdb / vortex-compact / ns (1.026x ➖, 0↑ 1↓)
duckdb / parquet / ns (0.997x ➖, 1↑ 0↓)
File Size Changes (200 files changed, +0.0% overall, 99↑ 101↓)
Totals:
|
Contributor
Benchmarks: TPC-DS SF=1 on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (1.060x ➖, 1↑ 21↓)
datafusion / vortex-compact / ns (1.015x ➖, 5↑ 7↓)
datafusion / parquet / ns (0.995x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed / ns (0.996x ➖, 5↑ 4↓)
duckdb / vortex-compact / ns (0.979x ➖, 5↑ 1↓)
duckdb / parquet / ns (1.005x ➖, 3↑ 6↓)
No file size changes detected. |
Contributor
Benchmarks: Statistical and Population Genetics 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
duckdb / vortex-file-compressed / ns (0.973x ➖, 3↑ 3↓)
duckdb / vortex-compact / ns (1.038x ➖, 0↑ 1↓)
duckdb / parquet / ns (1.004x ➖, 0↑ 0↓)
No file size changes detected. |
Contributor
Benchmarks: Clickbench on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (1.060x ➖, 3↑ 7↓)
datafusion / vortex-compact / ns (1.117x ❌, 1↑ 13↓)
datafusion / parquet / ns (0.999x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed / ns (0.999x ➖, 2↑ 3↓)
duckdb / vortex-compact / ns (1.018x ➖, 2↑ 5↓)
duckdb / parquet / ns (0.989x ➖, 1↑ 0↓)
No file size changes detected. |
Contributor
Benchmarks: Appian on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-compact / ns (1.022x ➖, 0↑ 0↓)
datafusion / parquet / ns (1.005x ➖, 0↑ 0↓)
duckdb / vortex-compact / ns (1.005x ➖, 0↑ 0↓)
duckdb / parquet / ns (0.996x ➖, 0↑ 0↓)
File Size Changes (10 files changed, -63.8% overall, 0↑ 10↓)
Totals:
|
Contributor
Benchmarks: TPC-H SF=1 on S3 📖Verdict: No clear signal (environment too noisy confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (0.990x ➖, 7↑ 8↓)
datafusion / vortex-compact / ns (1.115x ➖, 1↑ 7↓)
datafusion / parquet / ns (0.968x ➖, 1↑ 0↓)
duckdb / vortex-file-compressed / ns (0.992x ➖, 1↑ 0↓)
duckdb / vortex-compact / ns (1.044x ➖, 0↑ 3↓)
duckdb / parquet / ns (0.995x ➖, 0↑ 0↓)
|
Contributor
Benchmarks: TPC-H SF=10 on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (1.186x ❌, 0↑ 16↓)
datafusion / vortex-compact / ns (1.128x ❌, 1↑ 14↓)
datafusion / parquet / ns (1.009x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed / ns (0.994x ➖, 0↑ 0↓)
duckdb / vortex-compact / ns (0.992x ➖, 0↑ 0↓)
duckdb / parquet / ns (1.008x ➖, 0↑ 1↓)
No file size changes detected. |
Contributor
Benchmarks: Random Access 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
random-access / vortex-file-compressed / ns (1.000x ➖, 0↑ 0↓)
random-access / parquet / ns (0.996x ➖, 0↑ 0↓)
random-access / lance / ns (1.006x ➖, 0↑ 0↓)
|
Contributor
Benchmarks: Compression 📖vortex / vortex-file-compressed / ns (0.993x ➖, 1↑ 0↓)
vortex / vortex-file-compressed / bytes (1.001x ➖, 0↑ 0↓)
vortex / vortex-file-compressed / ratio (0.992x ➖, 1↑ 0↓)
vortex / parquet / ns (1.005x ➖, 0↑ 0↓)
vortex / parquet / bytes (1.000x ➖, 0↑ 0↓)
|
Contributor
Benchmarks: FineWeb S3 📖Verdict: No clear signal (environment too noisy confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (0.716x ➖, 5↑ 0↓)
datafusion / vortex-compact / ns (0.775x ➖, 4↑ 0↓)
datafusion / parquet / ns (0.910x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed / ns (0.996x ➖, 0↑ 0↓)
duckdb / vortex-compact / ns (1.008x ➖, 0↑ 0↓)
duckdb / parquet / ns (1.159x ➖, 1↑ 2↓)
|
Contributor
Benchmarks: TPC-H SF=10 on S3 📖Verdict: No clear signal (environment too noisy confidence) How to read Verdict and Engines
datafusion / vortex-compact / ns (1.182x ➖, 1↑ 7↓)
datafusion / parquet / ns (0.897x ➖, 4↑ 4↓)
duckdb / vortex-compact / ns (0.953x ➖, 0↑ 0↓)
duckdb / parquet / ns (0.973x ➖, 0↑ 0↓)
|
joseph-isaacs
added
the
action/bench-all
Claude
github-actions
Bot
removed
the
action/bench-all
joseph-isaacs
added
the
do not merge
Merging this PR will improve performance by 12.71%
Performance Changes
Tip Curious why this is faster? Comment Comparing 
Footnotes
|
joseph-isaacs
force-pushed
the
claude/benchmark-ci-draft-pr-1xvebp
branch
from
August 11, 2026 20:56
f2c5c38 to
daa0791
Compare
joseph-isaacs
added
the
action/bench-all
ChatGPT Codex Connector
github-actions
Bot
removed
the
action/bench-all
joseph-isaacs
force-pushed
the
claude/benchmark-ci-draft-pr-1xvebp
branch
2 times, most recently
from
August 12, 2026 09:57
19bdd2c to
e9aefff
Compare
joseph-isaacs
added
the
action/bench-all
ChatGPT Codex Connector
github-actions
Bot
removed
the
action/bench-all
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
Translate each partition's byte range to the row range whose natural splits it owns, mirroring the LayoutReader path, so plan-v2 scans no longer require repartitioning to be disabled. Benchmarks can now run against the same DataFusion configuration as the stored baseline. Log the plan-v2 scan path once per process so benchmark and CI logs record which scan path produced their timings. Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk> Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MjXw22CLhx6NgA8mxjJCmQ
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk> Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MjXw22CLhx6NgA8mxjJCmQ
joseph-isaacs
force-pushed
the
claude/benchmark-ci-draft-pr-1xvebp
branch
from
August 12, 2026 11:12
e9aefff to
b472e97
Compare
joseph-isaacs
added
the
action/bench-all
ChatGPT Codex Connector
github-actions
Bot
removed
the
action/bench-all
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Rationale for this change
#9341 targets
ji/vx-plan-perf, so it sits mid-stack and the PR benchmark dispatch—which listens forpull_requestlabel events on PRs intodevelop—does not run for it. This draft carries the complete remaining Plan V2 stack directly on currentdevelopso CI can benchmark it against the current baseline.This is the benchmarking-only sibling of #9341.
What is included?
The branch contains 9 commits:
VORTEX_USE_PLAN_V2=1DataFusion path and CI benchmark settingVORTEX_PLAN_V2_FILTER_MODEThe physical plan model is already in
developvia #9142. The former standalone pushdown PR #9166 is folded into #9196.COUNT(*)reduction belongs to #9196, childless-Packsplit handling belongs to #9240, and the zero-read end-to-end regression belongs to #9244.Rebase verification
developata0a0b207d8range-diffmatches the zoned, execution, pruning, performance, CI, DataFusion repartitioning, and tracing commitsUser-facing behavior
No stable API changes.
LayoutReaderremains the default whenVORTEX_USE_PLAN_V2is unset; Plan V2 remains opt-in and experimental.Checks
cargo +nightly fmt --all -- --checkyamllint --strict -c .yamllint.yaml .github/workflows/sql-bench-matrix.ymlcargo test -p vortex-layout -p vortex-scan-v2 -p vortex-datafusioncargo clippy -p vortex-layout -p vortex-scan-v2 -p vortex-datafusion --all-targets --all-features -- -D warningsgit diff --check origin/develop...HEAD