Skip to content

DO NOT MERGE: TPC-DS SF=10 benchmark run for vortex-data/vortex#11215 - #11216

Closed
joseph-isaacs wants to merge 8 commits into
developfrom
ji/sleepy-clarke-5utr8u-tpcds-sf10
Closed

joseph-isaacs wants to merge 8 commits into
developfrom
ji/sleepy-clarke-5utr8u-tpcds-sf10

Conversation

@joseph-isaacs

Copy link
Copy Markdown
Contributor

Summary

Benchmark-only PR. Do not merge.

It contains #11215 plus one commit that adds TPC-DS SF=10 on NVMe to the pr benchmark preset. SF=10 is otherwise only in the nightly preset. The action/bench-sql comment here therefore includes TPC-DS SF=10 against develop, which is where #11073 reported its gains.

Changes

  • bench-orchestrator/bench_orchestrator/ci_matrix/catalog.py: tpcds-nvme-10 also runs in the pr preset.

Close this once the numbers are in.


Generated by Claude Code

joseph-isaacs and others added 8 commits November 4, 2026 10:14
Inline roots shared one session, and the run clears a root's session when the root retires. Once the first root finished, the remaining roots waited for reads that were never issued.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MZ3d6HFyGGeUGy8brEr6BM
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
Evaluate TopK and hash join dynamic filters in the V2 file scan, so zones and rows are dropped before the engine sees them. The producing operator keeps enforcing them. V1 scans keep them out.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MZ3d6HFyGGeUGy8brEr6BM
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
A join's hash table lookup has no Vortex equivalent, so one unconvertible conjunct turned the whole filter into a constant true and its bounds were lost. Convert each conjunct alone and keep the ones that convert.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MZ3d6HFyGGeUGy8brEr6BM
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
A join's dynamic filter bounds its probe keys with key >= min AND key <= max. Each
comparison decoded the key column again, and the scan's between rewrite only ran on
static filters, so the bounds now merge into one between when the filter is read.

TPC-DS SF10, 10 join queries against the previous commit: geomean 0.972, Q69 and Q10
0.93, none slower than 1.007.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MZ3d6HFyGGeUGy8brEr6BM
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
A scan of at most one filtered chunk evaluates its conjuncts on that chunk next, so
pruning its zones first saves little and delays the chunk, and pruning a dynamic conjunct
proves its current predicate over the zones again. DataFusion splits small join probes into
many such scans. Larger scans keep pruning their dynamic conjuncts: skipping it for every
scan made TPC-DS SF10 Q82 and Q37 3.4-3.7x slower.

TPC-DS, all 99 queries against the previous commit: SF1 geomean 0.97, SF10 0.99 on best
runs and 1.01 on medians; the SF10 outliers were flat when rerun.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MZ3d6HFyGGeUGy8brEr6BM
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
…es change

A dynamic conjunct whose bound never prunes still costs a zone-pruning pass
before each scan and an evaluation per chunk. With
`ScanBuilder::with_skip_unselective_dynamic`, a conjunct containing dynamic
expressions that kept at least 99% of its input rows on four consecutive
chunks at one dynamic version is switched off: its zone pruning and its
evaluation are skipped until its dynamic values change.

The scan can then return rows the conjunct would reject, so this is opt-in.
Only the DataFusion persistent opener enables it, because DataFusion's TopK
and hash-join dynamic filters are pushed down inexactly and the producing
operator applies them again. DuckDB and other V2 scans are unchanged. Cached
filter plans are reused only by scans with the same setting.

`find_between` only merges comparisons against literals, so a switched-off
conjunct never carries a static restriction with it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UeniBBZYz5ae2yLro3TPmH
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
A chunk announced the reads of every conjunct before evaluating any, so a
switched-off conjunct still had its column fetched. The scan test now checks
that a switched-off scan reads no more than an unfiltered one, and tightens
the bound enough that the conjunct stays switched on.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UeniBBZYz5ae2yLro3TPmH
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
Temporary, so this branch's benchmark run covers TPC-DS SF=10.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UeniBBZYz5ae2yLro3TPmH
Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
@joseph-isaacs joseph-isaacs added action/bench-sql Run SQL benchmarks without Vortex Compact on this PR and removed action/bench-sql Run SQL benchmarks without Vortex Compact on this PR labels Nov 4, 2026 — with Claude
@robert3005
robert3005 deleted the ji/sleepy-clarke-5utr8u-tpcds-sf10 branch November 10, 2026 11:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

action/bench-sql Run SQL benchmarks without Vortex Compact on this PR

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant