Optimize slice for dict and minor changes in other arrays#8321
Merged
Conversation
Signed-off-by: Adam Gutglick <adam@spiraldb.com>
Signed-off-by: Adam Gutglick <adam@spiraldb.com>
Merging this PR will improve performance by 27.21%
|
| Mode | Benchmark | BASE |
HEAD |
Efficiency | |
|---|---|---|---|---|---|
| ❌ | Simulation | bitwise_not_vortex_buffer_mut[128] |
216.9 ns | 275.3 ns | -21.19% |
| ❌ | Simulation | bitwise_not_vortex_buffer_mut[1024] |
278.6 ns | 336.9 ns | -17.31% |
| ❌ | Simulation | bitwise_not_vortex_buffer_mut[2048] |
342.2 ns | 400.6 ns | -14.56% |
| ⚡ | Simulation | compare[63] |
360.5 µs | 245 µs | +47.16% |
| ⚡ | Simulation | compare[56] |
327.6 µs | 225.1 µs | +45.55% |
| ⚡ | Simulation | compare[62] |
363.9 µs | 250.2 µs | +45.44% |
| ⚡ | Simulation | compare[63] |
371.2 µs | 255.8 µs | +45.14% |
| ⚡ | Simulation | compare[60] |
353.8 µs | 243.9 µs | +45.05% |
| ⚡ | Simulation | compare[56] |
332.4 µs | 229.8 µs | +44.65% |
| ⚡ | Simulation | compare[62] |
368.7 µs | 255 µs | +44.61% |
| ⚡ | Simulation | compare[61] |
362.7 µs | 250.9 µs | +44.54% |
| ⚡ | Simulation | compare[60] |
358.6 µs | 248.7 µs | +44.22% |
| ⚡ | Simulation | compare[58] |
347.3 µs | 241.1 µs | +44.09% |
| ⚡ | Simulation | compare[59] |
354.5 µs | 246.5 µs | +43.84% |
| ⚡ | Simulation | compare[61] |
367.5 µs | 255.7 µs | +43.75% |
| ⚡ | Simulation | compare[58] |
352.2 µs | 245.9 µs | +43.22% |
| ⚡ | Simulation | compare[57] |
345.9 µs | 241.6 µs | +43.17% |
| ⚡ | Simulation | compare[59] |
359.3 µs | 251.2 µs | +43.03% |
| ⚡ | Simulation | compare[54] |
330.4 µs | 231.7 µs | +42.62% |
| ⚡ | Simulation | compare[55] |
337.2 µs | 236.7 µs | +42.49% |
| ... | ... | ... | ... | ... | ... |
ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.
Tip
Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.
Comparing adamg/dict-slice-thing (915c696) with develop (47d2041)
Signed-off-by: Adam Gutglick <adam@spiraldb.com>
Signed-off-by: Adam Gutglick <adam@spiraldb.com>
github-actions
Bot
removed
the
action/benchmark
Contributor
Polar Signals Profiling ResultsLatest Run
Previous Runs (1)
Powered by Polar Signals Cloud |
Contributor
Benchmarks: PolarSignals ProfilingVortex (geomean): 0.936x ➖ How to read Verdict and Engines
datafusion / vortex-file-compressed (0.936x ➖, 2↑ 0↓)
No file size changes detected. |
Contributor
Benchmarks: FineWeb NVMeVerdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.964x ➖, 1↑ 0↓)
datafusion / vortex-compact (0.984x ➖, 1↑ 0↓)
datafusion / parquet (0.997x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (1.003x ➖, 0↑ 1↓)
duckdb / vortex-compact (1.000x ➖, 0↑ 0↓)
duckdb / parquet (1.008x ➖, 0↑ 1↓)
File Size Changes (1 files changed, -0.0% overall, 0↑ 1↓)
Totals:
|
Contributor
Benchmarks: TPC-H SF=1 on NVMEVerdict: No clear signal (environment too noisy confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.959x ➖, 0↑ 0↓)
datafusion / vortex-compact (0.899x ✅, 9↑ 0↓)
datafusion / parquet (0.948x ➖, 3↑ 0↓)
datafusion / arrow (0.785x ✅, 21↑ 0↓)
duckdb / vortex-file-compressed (0.939x ➖, 5↑ 0↓)
duckdb / vortex-compact (0.940x ➖, 7↑ 1↓)
duckdb / parquet (0.944x ➖, 4↑ 1↓)
duckdb / duckdb (0.968x ➖, 0↑ 0↓)
File Size Changes (9 files changed, +0.1% overall, 7↑ 2↓)
Totals:
|
Contributor
Benchmarks: TPC-DS SF=1 on NVMEVerdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.996x ➖, 0↑ 2↓)
datafusion / vortex-compact (1.000x ➖, 1↑ 3↓)
datafusion / parquet (0.987x ➖, 4↑ 1↓)
duckdb / vortex-file-compressed (0.989x ➖, 1↑ 0↓)
duckdb / vortex-compact (0.998x ➖, 1↑ 1↓)
duckdb / parquet (1.007x ➖, 0↑ 1↓)
duckdb / duckdb (1.008x ➖, 1↑ 6↓)
File Size Changes (7 files changed, -0.0% overall, 4↑ 3↓)
Totals:
|
Contributor
Benchmarks: FineWeb S3Verdict: No clear signal (environment too noisy confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.908x ➖, 0↑ 0↓)
datafusion / vortex-compact (0.719x ➖, 3↑ 0↓)
datafusion / parquet (0.741x ➖, 3↑ 0↓)
duckdb / vortex-file-compressed (0.787x ➖, 3↑ 0↓)
duckdb / vortex-compact (0.874x ➖, 1↑ 0↓)
duckdb / parquet (0.948x ➖, 0↑ 0↓)
|
Contributor
Benchmarks: Statistical and Population GeneticsVerdict: No clear signal (low confidence) How to read Verdict and Engines
duckdb / vortex-file-compressed (1.047x ➖, 0↑ 1↓)
duckdb / vortex-compact (1.031x ➖, 0↑ 0↓)
duckdb / parquet (1.046x ➖, 0↑ 0↓)
File Size Changes (1 files changed, -0.0% overall, 0↑ 1↓)
Totals:
|
Contributor
Benchmarks: Random AccessVortex (geomean): 0.871x ✅ How to read Verdict and Engines
unknown / unknown (0.961x ➖, 12↑ 2↓)
|
Contributor
Benchmarks: TPC-H SF=10 on NVMEVerdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.985x ➖, 1↑ 0↓)
datafusion / vortex-compact (0.986x ➖, 0↑ 0↓)
datafusion / parquet (0.990x ➖, 0↑ 0↓)
datafusion / arrow (0.948x ➖, 5↑ 0↓)
duckdb / vortex-file-compressed (0.999x ➖, 0↑ 0↓)
duckdb / vortex-compact (0.998x ➖, 0↑ 0↓)
duckdb / parquet (0.995x ➖, 0↑ 0↓)
duckdb / duckdb (0.994x ➖, 0↑ 0↓)
File Size Changes (26 files changed, -0.0% overall, 11↑ 15↓)
Totals:
|
Contributor
Benchmarks: Clickbench on NVMEVerdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.987x ➖, 1↑ 3↓)
datafusion / parquet (0.956x ➖, 8↑ 1↓)
duckdb / vortex-file-compressed (1.014x ➖, 2↑ 3↓)
duckdb / parquet (1.003x ➖, 0↑ 0↓)
duckdb / duckdb (1.003x ➖, 0↑ 0↓)
File Size Changes (106 files changed, +0.0% overall, 66↑ 40↓)
Totals:
|
Contributor
Benchmarks: TPC-H SF=1 on S3Verdict: No clear signal (environment too noisy confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.729x ➖, 11↑ 0↓)
datafusion / vortex-compact (0.801x ➖, 8↑ 0↓)
datafusion / parquet (0.952x ➖, 1↑ 2↓)
duckdb / vortex-file-compressed (1.008x ➖, 0↑ 1↓)
duckdb / vortex-compact (1.017x ➖, 0↑ 1↓)
duckdb / parquet (0.928x ➖, 0↑ 0↓)
|
Contributor
Benchmarks: Appian on NVMEVerdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.023x ➖, 0↑ 0↓)
datafusion / parquet (1.035x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (1.024x ➖, 0↑ 0↓)
duckdb / parquet (1.017x ➖, 0↑ 0↓)
duckdb / duckdb (1.021x ➖, 0↑ 0↓)
File Size Changes (4 files changed, -0.1% overall, 0↑ 4↓)
Totals:
|
Contributor
Benchmarks: CompressionVortex (geomean): 0.994x ➖ How to read Verdict and Engines
unknown / unknown (1.011x ➖, 0↑ 3↓)
|
AdamGS
requested review from
connortsui20 and
joseph-isaacs
and removed request for
connortsui20
AdamGS
changed the title
Contributor
Benchmarks: TPC-H SF=10 on S3Verdict: No clear signal (environment too noisy confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.833x ➖, 2↑ 0↓)
datafusion / vortex-compact (0.963x ➖, 2↑ 3↓)
datafusion / parquet (0.784x ➖, 4↑ 0↓)
duckdb / vortex-file-compressed (1.096x ➖, 0↑ 1↓)
duckdb / vortex-compact (1.156x ➖, 0↑ 1↓)
duckdb / parquet (1.030x ➖, 0↑ 0↓)
|
connortsui20
reviewed
Jun 9, 2026
Comment on lines
+242
to
+245
| ) && value | ||
| .as_ref() | ||
| .as_exact() | ||
| .is_some_and(|v| matches!(v, ScalarValue::Bool(true))) |
connortsui20
reviewed
Jun 9, 2026
connortsui20
reviewed
Jun 9, 2026
Comment on lines
40
to
+41
| if let Some(code) = sliced_code.as_opt::<Constant>() { | ||
| let code = code.scalar().as_primitive().as_::<usize>(); | ||
| return if let Some(code) = code { | ||
| let values = array.values().slice(code..code + 1)?; | ||
| Ok(Some( | ||
| DictArray::new( | ||
| ConstantArray::new(0u8, sliced_code.len()).into_array(), | ||
| values, | ||
| ) | ||
| .into_array(), | ||
| )) | ||
| } else { | ||
| Ok(Some( | ||
| ConstantArray::new(Scalar::null(array.dtype().clone()), sliced_code.len()) | ||
| .into_array(), | ||
| )) | ||
| }; | ||
| return slice_constant_code(array, code.scalar(), sliced_code.len()); |
Contributor
There was a problem hiding this comment.
question (not review): do we have the optimization where if we slice something it can BECOME a constant array? The only cases I can think of this happening would be in runend where we happen to slice into a run
Contributor
Author
There was a problem hiding this comment.
I don't think so, but maybe I was just looking in the wrong place.
connortsui20
reviewed
Jun 9, 2026
connortsui20
approved these changes
Jun 9, 2026
connortsui20
left a comment
connortsui20
left a comment
Contributor
There was a problem hiding this comment.
lgtm, though I am still wondering about those constant codes optimization, it feels like we should have optimized those away before we ever call slice? And then if slicing causes codes to be constant then thats an additional optimization. That can be in a separate PR
AdamGS
force-pushed
the
adamg/dict-slice-thing
branch
from
e0c2f5f to
d2f0bbc
Compare
Signed-off-by: Adam Gutglick <adam@spiraldb.com>
AdamGS
force-pushed
the
adamg/dict-slice-thing
branch
from
d2f0bbc to
915c696
Compare
github-actions
Bot
removed
the
action/benchmark-sql
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
3 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.
Summary
Includes a few optimizations around array slicing: