Open
Conversation
AdamGS
force-pushed
the
adamg/aggregate-fn-array-uncompressed
branch
from
1c091b1 to
c8699a6
Compare
Merging this PR will degrade performance by 37.77%
|
| Mode | Benchmark | BASE |
HEAD |
Efficiency | |
|---|---|---|---|---|---|
| ❌ | Simulation | chunked_bool_canonical_into[(1000, 10)] |
31.3 µs | 50.2 µs | -37.77% |
| ⚡ | Simulation | take_search[(0.005, 0.05)] |
168 µs | 131.4 µs | +27.88% |
| ⚡ | Simulation | take_search[(0.005, 0.1)] |
320.1 µs | 247.1 µs | +29.56% |
| ⚡ | Simulation | take_search[(0.005, 0.5)] |
1.5 ms | 1.2 ms | +31.04% |
| ⚡ | Simulation | take_search[(0.005, 1.0)] |
3.1 ms | 2.3 ms | +31.26% |
| ⚡ | Simulation | take_search[(0.01, 0.05)] |
179.1 µs | 142.5 µs | +25.71% |
| ⚡ | Simulation | take_search[(0.01, 0.1)] |
341 µs | 267.9 µs | +27.26% |
| ⚡ | Simulation | take_search[(0.01, 0.5)] |
1.6 ms | 1.3 ms | +28.58% |
| ⚡ | Simulation | take_search[(0.01, 1.0)] |
3.3 ms | 2.5 ms | +28.78% |
| ⚡ | Simulation | take_search[(0.1, 0.05)] |
249 µs | 212.4 µs | +17.25% |
| ⚡ | Simulation | take_search[(0.1, 0.1)] |
459.3 µs | 386.2 µs | +18.91% |
| ⚡ | Simulation | take_search[(0.1, 0.5)] |
2.2 ms | 1.8 ms | +20.4% |
| ⚡ | Simulation | take_search[(0.1, 1.0)] |
4.3 ms | 3.5 ms | +20.62% |
| ⚡ | Simulation | take_search_chunked[(0.005, 0.05)] |
193.4 µs | 162.1 µs | +19.28% |
| ⚡ | Simulation | take_search_chunked[(0.005, 0.1)] |
369.2 µs | 307 µs | +20.26% |
| ⚡ | Simulation | take_search_chunked[(0.005, 0.5)] |
1.8 ms | 1.5 ms | +21.07% |
| ⚡ | Simulation | take_search_chunked[(0.005, 1.0)] |
3.5 ms | 2.9 ms | +21.2% |
| ⚡ | Simulation | take_search_chunked[(0.01, 0.05)] |
205.9 µs | 174.6 µs | +17.9% |
| ⚡ | Simulation | take_search_chunked[(0.01, 0.1)] |
394 µs | 331.8 µs | +18.75% |
| ⚡ | Simulation | take_search_chunked[(0.01, 0.5)] |
1.9 ms | 1.6 ms | +19.5% |
| ... | ... | ... | ... | ... | ... |
ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.
Comparing adamg/aggregate-fn-array-uncompressed (b7fb3d0) with develop (6331959)
AdamGS
added
changelog/feature
github-actions
Bot
removed
the
action/benchmark
Contributor
Polar Signals Profiling ResultsLatest Run
Previous Runs (2)
Powered by Polar Signals Cloud |
Contributor
Benchmarks: PolarSignals ProfilingVortex (geomean): 1.033x ➖ datafusion / vortex-file-compressed (1.033x ➖, 0↑ 1↓)
|
Contributor
File Sizes: PolarSignals ProfilingNo file size changes detected. |
Contributor
Benchmarks: FineWeb NVMeVerdict: No clear signal (low confidence) datafusion / vortex-file-compressed (1.015x ➖, 0↑ 1↓)
datafusion / vortex-compact (0.991x ➖, 0↑ 1↓)
datafusion / parquet (0.954x ➖, 1↑ 0↓)
duckdb / vortex-file-compressed (0.944x ➖, 1↑ 0↓)
duckdb / vortex-compact (0.974x ➖, 1↑ 1↓)
duckdb / parquet (0.974x ➖, 0↑ 0↓)
Full attributed analysis
|
Contributor
File Sizes: FineWeb NVMeNo file size changes detected. |
Contributor
Benchmarks: TPC-H SF=1 on NVMEVerdict: No clear signal (environment too noisy confidence) datafusion / vortex-file-compressed (0.897x ✅, 10↑ 0↓)
datafusion / vortex-compact (0.904x ➖, 9↑ 0↓)
datafusion / parquet (0.933x ➖, 6↑ 0↓)
datafusion / arrow (0.883x ✅, 12↑ 0↓)
duckdb / vortex-file-compressed (0.907x ➖, 7↑ 0↓)
duckdb / vortex-compact (0.921x ➖, 7↑ 0↓)
duckdb / parquet (0.953x ➖, 3↑ 1↓)
duckdb / duckdb (0.923x ➖, 4↑ 0↓)
Full attributed analysis
|
Contributor
File Sizes: TPC-H SF=1 on NVMENo file size changes detected. |
Contributor
Benchmarks: TPC-DS SF=1 on NVMEVerdict: No clear signal (low confidence) datafusion / vortex-file-compressed (1.109x ❌, 0↑ 56↓)
datafusion / vortex-compact (1.095x ➖, 0↑ 47↓)
datafusion / parquet (1.089x ➖, 0↑ 39↓)
duckdb / vortex-file-compressed (1.092x ➖, 1↑ 44↓)
duckdb / vortex-compact (1.077x ➖, 0↑ 28↓)
duckdb / parquet (1.064x ➖, 0↑ 22↓)
duckdb / duckdb (1.090x ➖, 0↑ 38↓)
Full attributed analysis
|
Contributor
File Sizes: TPC-DS SF=1 on NVMENo file size changes detected. |
Contributor
Benchmarks: FineWeb S3Verdict: No clear signal (environment too noisy confidence) datafusion / vortex-file-compressed (0.995x ➖, 0↑ 1↓)
datafusion / vortex-compact (1.073x ➖, 1↑ 2↓)
datafusion / parquet (1.086x ➖, 0↑ 1↓)
duckdb / vortex-file-compressed (1.070x ➖, 0↑ 1↓)
duckdb / vortex-compact (1.053x ➖, 0↑ 1↓)
duckdb / parquet (1.006x ➖, 0↑ 0↓)
Full attributed analysis
|
Contributor
Benchmarks: Random AccessVortex (geomean): 0.967x ➖ unknown / unknown (0.990x ➖, 3↑ 0↓)
|
Contributor
Benchmarks: Statistical and Population GeneticsVerdict: No clear signal (low confidence) duckdb / vortex-file-compressed (1.010x ➖, 0↑ 0↓)
duckdb / vortex-compact (1.004x ➖, 0↑ 0↓)
duckdb / parquet (1.010x ➖, 0↑ 0↓)
Full attributed analysis
|
Contributor
File Sizes: Statistical and Population GeneticsNo file size changes detected. |
Contributor
Benchmarks: Clickbench on NVMEVerdict: No clear signal (low confidence) datafusion / vortex-file-compressed (0.997x ➖, 1↑ 0↓)
datafusion / parquet (1.002x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (1.014x ➖, 1↑ 2↓)
duckdb / parquet (0.990x ➖, 1↑ 0↓)
duckdb / duckdb (0.985x ➖, 1↑ 0↓)
Full attributed analysis
|
Contributor
File Sizes: Clickbench on NVMEFile Size Changes (1 files changed, -0.0% overall, 0↑ 1↓)
Totals:
|
Contributor
Benchmarks: TPC-H SF=1 on S3Verdict: No clear signal (environment too noisy confidence) datafusion / vortex-file-compressed (1.089x ➖, 0↑ 4↓)
datafusion / vortex-compact (1.057x ➖, 0↑ 3↓)
datafusion / parquet (1.115x ➖, 0↑ 5↓)
duckdb / vortex-file-compressed (1.029x ➖, 0↑ 2↓)
duckdb / vortex-compact (1.103x ➖, 0↑ 2↓)
duckdb / parquet (1.024x ➖, 0↑ 0↓)
Full attributed analysis
|
Contributor
Benchmarks: CompressionVortex (geomean): 1.002x ➖ unknown / unknown (0.982x ➖, 10↑ 3↓)
|
Contributor
Benchmarks: TPC-H SF=10 on S3Verdict: No clear signal (environment too noisy confidence) datafusion / vortex-file-compressed (1.165x ➖, 1↑ 6↓)
datafusion / vortex-compact (1.162x ➖, 0↑ 5↓)
datafusion / parquet (1.033x ➖, 1↑ 1↓)
duckdb / vortex-file-compressed (1.127x ➖, 0↑ 0↓)
duckdb / vortex-compact (1.026x ➖, 0↑ 0↓)
duckdb / parquet (1.073x ➖, 0↑ 0↓)
Full attributed analysis
|
Contributor
Benchmarks: TPC-H SF=10 on NVMEVerdict: No clear signal (low confidence) datafusion / vortex-file-compressed (1.011x ➖, 0↑ 0↓)
datafusion / vortex-compact (1.009x ➖, 0↑ 0↓)
datafusion / parquet (1.018x ➖, 0↑ 0↓)
datafusion / arrow (1.011x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (1.001x ➖, 0↑ 0↓)
duckdb / vortex-compact (1.010x ➖, 0↑ 0↓)
duckdb / parquet (1.003x ➖, 0↑ 0↓)
duckdb / duckdb (1.001x ➖, 0↑ 0↓)
Full attributed analysis
|
Contributor
File Sizes: TPC-H SF=10 on NVMENo file size changes detected. |
robert3005
reviewed
Apr 29, 2026
| /// Applies to all types and returns a non-null `u64`. Encoding kernels can return this aggregate | ||
| /// directly from metadata to avoid decoding arrays whose uncompressed size is known. | ||
| #[derive(Clone, Debug)] | ||
| pub struct UncompressedSizeInBytes; |
Contributor
There was a problem hiding this comment.
So this is only buffers of the array and all children.
Are we sure we should not include metadata?
Maybe just justify why?
Contributor
Author
There was a problem hiding this comment.
I don't think we do now, and it should really be a rounding error right?
It also seems to be more of an implementation details and not the number execution engines care about.
Contributor
There was a problem hiding this comment.
You need a more precise definition of this value. I guess its sum of buffers in the fully canonical variant?
Contributor
There was a problem hiding this comment.
Let's define it as byte size of buffers and children in canonical representation. Can you add a note?
AdamGS
force-pushed
the
adamg/aggregate-fn-array-uncompressed
branch
from
0e19932 to
2ac1bd9
Compare
robert3005
reviewed
May 7, 2026
robert3005
reviewed
May 7, 2026
robert3005
requested changes
May 7, 2026
Contributor
robert3005
left a comment
robert3005
left a comment
There was a problem hiding this comment.
Constant impl needs changing, this reads a bit too much like LLM generated code
AdamGS
force-pushed
the
adamg/aggregate-fn-array-uncompressed
branch
from
2ac1bd9 to
7cb7c12
Compare
Signed-off-by: Adam Gutglick <[email protected]>
Signed-off-by: Adam Gutglick <[email protected]>
Signed-off-by: Adam Gutglick <[email protected]>
Signed-off-by: Adam Gutglick <[email protected]>
AdamGS
force-pushed
the
adamg/aggregate-fn-array-uncompressed
branch
from
7cb7c12 to
673cf82
Compare
Signed-off-by: Adam Gutglick <[email protected]>
github-actions
Bot
removed
the
action/benchmark-sql
robert3005
reviewed
May 7, 2026
Comment on lines
+45
to
+46
| #[cfg(test)] | ||
| use crate::builders::builder_with_capacity; |
Contributor
There was a problem hiding this comment.
can you move this to the tests module?
robert3005
reviewed
May 7, 2026
Comment on lines
+311
to
+318
| #[cfg(test)] | ||
| fn materialized_uncompressed_size_in_bytes(array: &ArrayRef) -> u64 { | ||
| let mut builder = builder_with_capacity(array.dtype(), array.len()); | ||
| unsafe { | ||
| builder.extend_from_array_unchecked(array); | ||
| } | ||
| builder.finish().nbytes() | ||
| } |
robert3005
approved these changes
May 7, 2026
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
This stat turns out to be very useful for helping execution engines plan joins better, but its currently requires a full-decompression. This PR includes the basic structure for a new AggregateFn, including implementation for all canonical types.