perf(compressor): Improve string dictionary selection - #9762
Draft
lwwmanning wants to merge 2 commits into
Draft
Conversation
Signed-off-by: Will Manning <will@willmanning.io>
lwwmanning
added
changelog/performance
github-actions
Bot
removed
the
action/bench-compress
Merging this PR will regress 3 benchmarks
|
| Mode | Benchmark | BASE |
HEAD |
Efficiency | |
|---|---|---|---|---|---|
| ❌ | WallTime | mul_u64_nonnull_neon |
15.3 µs | 21.1 µs | -27.31% |
| ❌ | WallTime | multiply_shapes_neon[(16384, PerRowPerRow)] |
17.2 µs | 20.2 µs | -14.58% |
| ❌ | WallTime | mul_i64_nonnull_neon |
17.2 µs | 20.1 µs | -14.28% |
| ⚡ | WallTime | arrow_checked_add_u32_neon[16384] |
20.5 µs | 12.7 µs | +60.66% |
| ⚡ | WallTime | mul_u32_nonnull_avx2 |
6.9 µs | 5.9 µs | +16.83% |
| ⚡ | WallTime | mul_u32_nonnull_avx512 |
6.3 µs | 5.6 µs | +11.8% |
| ⚡ | Simulation | allocate_drop_bytes[0] |
520.2 ns | 466 ns | +11.62% |
Tip
Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.
Comparing wm/string-dictionary-selection (6ed06bd) with develop (265b705)
Footnotes
-
206 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩
Contributor
Polar Signals Profiling ResultsLatest Run
Powered by Polar Signals Cloud |
Contributor
Benchmarks: Compression 📖Commits: PR vortex / vortex-file-compressed / ns (1.016x ➖, 0↑ 2↓)
vortex / vortex-file-compressed / bytes (1.000x ➖, 0↑ 0↓)
vortex / vortex-file-compressed / ratio (1.008x ➖, 0↑ 2↓)
vortex / parquet / ns (1.004x ➖, 0↑ 0↓)
vortex / parquet / bytes (1.000x ➖, 0↑ 0↓)
vortex / arrow-ipc / ns (1.012x ➖, 1↑ 1↓)
vortex / arrow-ipc / bytes (1.000x ➖, 0↑ 0↓)
|
Signed-off-by: Will Manning <will@willmanning.io>
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
1 participant
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.
String Dictionary selection now probes a small distributed sample before it builds an exact candidate. The old estimator missed profitable dictionaries when columns contained thousands of repeated strings. Samples above 75% distinct values reject Dictionary. An incomplete probe uses the existing estimator. An admitted candidate has explicit limits for values, value bytes, code bytes, and input bytes. The compressor compares the candidate's actual size and reuses the candidate if it wins.
The probe reads at most 8,192 rows and 1 MiB of values. String samples now compact their payload before size estimates, so unrelated retained bytes do not distort the score.
On seven local runs of the standard compression benchmark, Euro2016 stayed effectively neutral: write +0.01%, full materialization +0.46%, and size -0.14%. HashTags wrote 4.4% faster and used 2.1% fewer bytes, but full materialization was 10.9% slower. The benchmark expands Dictionary into flat strings. Engines that preserve Dictionary can retain the encoded form.
🤖 Generated with Codex