duckdb: don't optimize_recursive() before exporting chunks - #9795
Conversation
Polar Signals Profiling ResultsLatest Run
Previous Runs (1)
Powered by Polar Signals Cloud |
Merging this PR will degrade performance by 5.52%
|
| Mode | Benchmark | BASE |
HEAD |
Efficiency | |
|---|---|---|---|---|---|
| ❌ | Simulation | random_i8[0.5] |
67.9 µs | 91.4 µs | -25.73% |
| ❌ | Simulation | decompress[u64, (4000, 1024)] |
70.8 µs | 86 µs | -17.66% |
| ❌ | WallTime | words_gather_scalar_avx2[65536] |
8.2 µs | 9.4 µs | -12.15% |
| ⚡ | Simulation | random_i16[0.95] |
95.7 µs | 77.5 µs | +23.5% |
| ⚡ | Simulation | allocate_drop_arrow[0] |
456.9 ns | 402.7 ns | +13.45% |
Tip
Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.
Comparing myrrc/duckdb-no-optimize-recursive (d651f5e) with develop (5c92e4e)2
Footnotes
-
218 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. ↩
-
No successful run was found on
develop(ae8fddc) during the generation of this report, so 5c92e4e was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩
Benchmarks: PolarSignals Profiling 📖Commits: PR datafusion / vortex-file-compressed / ns (0.990x ➖, 0↑ 0↓)
No file size changes detected. |
Benchmarks: TPC-H SF=1 on NVME 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (1.000x ➖, 0↑ 0↓)
datafusion / parquet / ns (0.998x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed / ns (0.977x ➖, 1↑ 0↓)
duckdb / parquet / ns (0.997x ➖, 0↑ 0↓)
No file size changes detected. |
Benchmarks: FineWeb NVMe 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-compact / ns (0.992x ➖, 2↑ 1↓)
datafusion / parquet / ns (0.993x ➖, 0↑ 0↓)
duckdb / vortex-compact / ns (1.027x ➖, 1↑ 2↓)
duckdb / parquet / ns (1.002x ➖, 0↑ 0↓)
No file size changes detected. |
Benchmarks: FineWeb NVMe 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (0.989x ➖, 0↑ 0↓)
datafusion / parquet / ns (1.001x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed / ns (1.017x ➖, 0↑ 0↓)
duckdb / parquet / ns (1.000x ➖, 0↑ 0↓)
No file size changes detected. |
Benchmarks: TPC-DS SF=1 on NVME 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (0.996x ➖, 2↑ 0↓)
datafusion / parquet / ns (0.998x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed / ns (0.998x ➖, 3↑ 4↓)
duckdb / parquet / ns (1.004x ➖, 2↑ 5↓)
No file size changes detected. |
Benchmarks: Clickbench on NVME 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (1.004x ➖, 0↑ 0↓)
datafusion / parquet / ns (1.000x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed / ns (0.979x ➖, 2↑ 0↓)
duckdb / parquet / ns (0.985x ➖, 3↑ 1↓)
No file size changes detected. |
Benchmarks: TPC-H SF=10 on NVME 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (0.994x ➖, 0↑ 0↓)
datafusion / parquet / ns (0.997x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed / ns (0.994x ➖, 0↑ 0↓)
duckdb / parquet / ns (0.985x ➖, 1↑ 0↓)
No file size changes detected. |
Benchmarks: FineWeb S3 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (1.145x ➖, 0↑ 2↓)
datafusion / parquet / ns (1.090x ➖, 0↑ 2↓)
duckdb / vortex-file-compressed / ns (0.993x ➖, 0↑ 0↓)
duckdb / parquet / ns (1.030x ➖, 0↑ 0↓)
|
Benchmarks: Statistical and Population Genetics 📖Commits: PR How to read Verdict and Engines
duckdb / vortex-file-compressed / ns (1.001x ➖, 0↑ 0↓)
duckdb / parquet / ns (0.990x ➖, 0↑ 0↓)
No file size changes detected. |
Benchmarks: FineWeb S3 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-compact / ns (1.092x ➖, 0↑ 3↓)
datafusion / parquet / ns (1.101x ➖, 0↑ 2↓)
duckdb / vortex-compact / ns (1.082x ➖, 0↑ 1↓)
duckdb / parquet / ns (1.042x ➖, 0↑ 0↓)
|
Benchmarks: TPC-DS SF=1 on NVME 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-compact / ns (0.998x ➖, 0↑ 1↓)
datafusion / parquet / ns (0.999x ➖, 0↑ 0↓)
duckdb / vortex-compact / ns (0.996x ➖, 4↑ 0↓)
duckdb / parquet / ns (1.004x ➖, 2↑ 4↓)
No file size changes detected. |
Benchmarks: Clickbench on NVME 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-compact / ns (0.995x ➖, 0↑ 0↓)
datafusion / parquet / ns (1.000x ➖, 0↑ 0↓)
duckdb / vortex-compact / ns (0.991x ➖, 1↑ 1↓)
duckdb / parquet / ns (1.002x ➖, 2↑ 1↓)
No file size changes detected. |
Benchmarks: TPC-H SF=1 on S3 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (1.120x ➖, 0↑ 4↓)
datafusion / parquet / ns (0.980x ➖, 1↑ 0↓)
duckdb / vortex-file-compressed / ns (1.117x ➖, 0↑ 3↓)
duckdb / parquet / ns (1.049x ➖, 0↑ 0↓)
|
Benchmarks: TPC-H SF=1 on S3 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-compact / ns (1.130x ➖, 0↑ 7↓)
datafusion / parquet / ns (0.999x ➖, 0↑ 1↓)
duckdb / vortex-compact / ns (1.000x ➖, 0↑ 0↓)
duckdb / parquet / ns (1.050x ➖, 0↑ 0↓)
|
Benchmarks: TPC-H SF=1 on NVME 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-compact / ns (0.996x ➖, 0↑ 0↓)
datafusion / parquet / ns (0.998x ➖, 0↑ 0↓)
duckdb / vortex-compact / ns (0.998x ➖, 0↑ 0↓)
duckdb / parquet / ns (0.996x ➖, 0↑ 0↓)
No file size changes detected. |
Benchmarks: Clickbench Sorted on NVME 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-compact / ns (0.996x ➖, 0↑ 0↓)
datafusion / parquet / ns (1.006x ➖, 0↑ 0↓)
duckdb / vortex-compact / ns (1.012x ➖, 0↑ 1↓)
duckdb / parquet / ns (1.000x ➖, 0↑ 1↓)
File Size Changes (100 files changed, -0.0% overall, 58↑ 42↓)
Totals:
|
Benchmarks: TPC-H SF=10 on NVME 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-compact / ns (1.002x ➖, 0↑ 0↓)
datafusion / parquet / ns (0.998x ➖, 0↑ 0↓)
duckdb / vortex-compact / ns (0.987x ➖, 0↑ 0↓)
duckdb / parquet / ns (0.985x ➖, 1↑ 0↓)
No file size changes detected. |
Benchmarks: Statistical and Population Genetics 📖Commits: PR How to read Verdict and Engines
duckdb / vortex-compact / ns (1.002x ➖, 0↑ 0↓)
duckdb / parquet / ns (0.997x ➖, 0↑ 0↓)
No file size changes detected. |
Benchmarks: Appian on NVME 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-compact / ns (0.998x ➖, 0↑ 0↓)
datafusion / parquet / ns (0.999x ➖, 0↑ 0↓)
duckdb / vortex-compact / ns (1.012x ➖, 0↑ 0↓)
duckdb / parquet / ns (1.001x ➖, 0↑ 0↓)
No file size changes detected. |
Benchmarks: Clickbench Sorted on NVME 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-file-compressed / ns (1.014x ➖, 0↑ 1↓)
datafusion / parquet / ns (1.015x ➖, 0↑ 1↓)
duckdb / vortex-file-compressed / ns (0.964x ➖, 1↑ 0↓)
duckdb / parquet / ns (0.978x ➖, 0↑ 0↓)
File Size Changes (100 files changed, -0.0% overall, 39↑ 61↓)
Totals:
|
Benchmarks: TPC-H SF=10 on S3 📖Commits: PR How to read Verdict and Engines
datafusion / vortex-compact / ns (1.039x ➖, 2↑ 4↓)
datafusion / parquet / ns (1.044x ➖, 0↑ 1↓)
duckdb / vortex-compact / ns (1.088x ➖, 0↑ 1↓)
duckdb / parquet / ns (1.099x ➖, 0↑ 2↓)
|
Signed-off-by: Mikhail Kot <mikhail@spiraldb.com>
6e90205 to
d651f5e
Compare
| // In very rare cases we may get a DType::Struct which is not a Struct | ||
| // array | ||
| array.execute::<Canonical>(ctx)?.into_struct() |
There was a problem hiding this comment.
this is not rare
| // In very rare cases we may get a DType::Struct which is not a Struct | |
| // array | |
| array.execute::<Canonical>(ctx)?.into_struct() | |
| array.execute::<Canonical>(ctx)?.into_struct() |
When we wrap an Array with another Array, we call optimize() on a child.
This means, by the time we get to the exporter, the ArrayRef of a split
is fully optimized. Remove this call.